Model Report by GCMPC

Transportation Model Update Report

o Measure economic effectiveness and economic impacts of project decision making; o Prepare for possible new air quality conformity standards; o Design an easier tool to use and increase staff training on the model.

Model Update Activities

Purpose of Model Update The purpose of the transportation demand model improvement and update is to support regional transportation planning activities leading to the development of the 2045 Long Range Transportation Plan (LRTP). The purpose of a truly multi‐modal transportation plan is to establish physical and cultural environments that support and encourage safe, comfortable and convenient travel by a variety of modes. The technical modeling and performance measurements procedures to be used in the transportation planning process should be designed to meet these needs. The Genesee County Travel Demand Model is a representation of travel patterns of the major roads in the county. The model uses population

and employment projections to project where people may be traveling on the road network in the future. The model was developed through a cooperative effort consisting of GCMPC staff, Michigan Department of Transportation (MDOT), the Model Development Committee and the modeling consultants from The Corradino Group. The model is used to determine the road network capacity deficiencies, to develop the list of capacity improvement projects for the LRTP, to test alternative projects to alleviate congestion and for air quality conformity analysis.

What were the goals of the update? o Update to a new 2014 Base Year and develop model scenarios out to 2045; o Improve the model capabilities;

The approved model years for this update included a new calibrated base year of 2014 with future years of 2020, 2025, 2035 and 2045. The Genesee County 2045 Population Projections were used to update the population, households, and persons per household attribute fields in the traffic analysis zones. The 2014 base highway and transit network geographical database were updated, and both use a master network which keeps all future network scenarios in one file. The speed and capacity calculators were also updated to the most recent methodologies of Highway Capacity Manual 2010 (HCM2010) which is used to assess the traffic and environmental effects of highway projects. The Genesee travel demand model uses a four‐ step modeling process with a travel time feedback loop. The four steps are trip generation, trip distribution, mode choice, and traffic assignment. The trip generation model was updated and the “MI Travel Counts” data was used for calibration. A gravity model with friction factors was used for trip distribution. A nested logit model was used for the mode choice model. The model divides the person trips into trips of five modes: car driver alone, car share ride, transit (bus), and nonmotorized (walk/bike). For traffic assignment,

the Genesee County model uses a time‐of‐day modeling procedure in which auto and transit assignments are made for each of the four periods which are summed to produce daily assignments. The time of day model includes an AM Peak, Midday, PM Peak and Night time periods and the MI Travel Counts dataset for all Transportation Management Area (TMA) trips was used to calculate a frequency distribution by departure hour for each trip purpose. Key Elements of the GCMPC Network: Number of Links: 4,310 Number of Nodes: 2,930 Number of Centroid Connectors: 1,452 Number of Traffic Analysis Zones: 639 Number of External Zones: 37

Model Page 1

Transportation Model Update Report

Model Validation / Calibration

The Genesee County travel demand model was validated to replicate the observed 2014 traffic count data. Model and network parameters were adjusted so that final root‐mean‐square error (RMSE) and volume/count summaries met MDOT model calibration guidelines. Trip lengths and trip distribution were adjusted during the trip distribution model calibration to ensure that the modeled trip lengths for each trip purpose were comparable to the observed trip lengths from household surveys. Mode choice constants were also adjusted to match observed highway, transit and non‐motorized shares. This model is an improvement on the previous model and appropriate for a medium‐sized MPO such as Genesee County. The updated travel demand model shows appropriate sensitivity to changes in transportation supply and travel demand (Table 1). The 2014 base year validation scenario, and the alternative setups for 2020, 2025, 2035, and 2045 were used as the test cases. Model validation is based on daily comparisons of model volumes to observed counts. The daily model RMSE is at 31.57%. The idea behind the RMSE guidelines is to ensure that model‐estimated volumes are within one lane of actual roadway need. For the daily model, all validation results fall within the MDOT target ranges. The Genesee County travel demand model exceeds MDOT and Federal Highway Administration standards for calibration and is ready for use in the development of the 2045 LRTP.

Key Validation/Calibration Statistic

Table 1: Model Sensitivity to Changes in Transportation Supply and Demand

Measure

INPUTS Population Households Employment Road Miles (non‐centroid) Lane Miles (Thru) MODEL RESULTS internal +IE/EI person trips VMT Transit trips (linked) Estimated delay (daily hrs) CALCULATED VALUES Persons/Household Jobs/Household Trips/Person Trips/Household Trips/Employee Average Link Load VMT/Person

2014

2020

YEAR 2025

2035

2045

RMSE BY VOLUME GROUP Count Range

Count

<1,000

12,661

13,134

405,550 166,254 209,887 1,064 2,645

402,263 166,929 214,685 1,064 2,645

402,688 169,340 217,610 1,064 2,645

407,869 172,223 221,643 1,064 2,645

1,605,707 10,753,449 12,637 16,174

1,600,905 9,863,788 12,406 16,225

1,599,018 9,933,049 12,193 16,736

1,607,773 10,022,876 11,928 17,311

1,633,779 10,174,328 11,918 18,236

2.50 1.16 3.89 9.72 8.39 10,126 26.04

2.44 1.26 3.95 9.63 7.63 9,270 24.32

2.41 1.29 3.98 9.58 7.45 9,336 24.69

2.38 1.29 3.99 9.49 7.39 9,420 24.89

2.37 1.29 4.01 9.49 7.37 9,562 24.95

Flow Count Count Calculated % Target % #Links Ratio % Diff RMSE RMSE

1,000 ‐ 2,500

125,297

2,500 ‐ 5,000

3.733

72.21%

< 100%

151,048

1.206 20.552

67.06%

< 100%

463,531

479,794

1.035

3.508

47.00%

< 100%

124

5,000 ‐ 10,000

1,361,541

1,368,117

1.005

0.483

35.72%

< 45%

190

10,000 ‐ 25,000

3,478,245

3,203,835

0.921

‐7.889

27.19%

< 30%

224

25,000 ‐ 50,000

1,439,041

1,424,762

0.99

‐0.992

17.98%

< 25%

101,500

99,473

0.98

‐1.997

3.55%

< 20%

> 50,000

1.037

Count & VMT Comparision by Functional Class (NFC) Functional Class (NFC) Interstate (1)

412,899 165,199 191,484 1,062 2,643

%RSME 31.46

Count

Flow Count Ratio

1,844,640 1,931,316

Oth Fwy Exp (2)

228,876

Count Flow VMT VMT Ratio VMT % Target(%) #Links VMT Diff.

1.047 2,194,208

2,291,911

1.045

4.50%

+/‐6%

258,944

1.131

385,919

441,133

1.143

14.30%

+/‐6%

Oth Princ Arterial (3) 2,012,736 1,904,678

0.946

698,079

691,055

0.99

‐1.00%

+/‐7%

130

Minor arterial (4)

0.919 1,214,527

1,110,908

0.915

‐8.50%

+/‐10%

278

2,277,548 2,092,643

Major Collector (5)

268,255

221,941

0.827

240,229

207,839

0.865 ‐13.50%

+/‐20%

Minor Collector (6)

334,094

314,644

0.942

178,702

184,172

1.031

3.10%

+/‐20%

0.965 4,911,664

4,927,018

1.003

‐0.30%

+/‐5%

669

TOTAL

6,966,149 6,724,165

Count & VMT Comparision by Area Type Area Type CBD (1)

Count 158,603

Flow Count Ratio 130,402 0.822

Count Flow VMT VMT VMT % Target(%) #Links VMT Ratio Diff. 26,437 27,780 1.051 5.10% +/‐10% 12

Urban (2)

3,167,768 2,974,326

0.939 1,337,767

1,267,990 0.948

‐5.20%

+/‐10%

231

Suburban (3)

2,534,011 2,508,100

0.99 2,440,731

2,507,695 1.027

2.70%

+/‐10%

242

Fringe (4)

400,541

370,073

0.924

339,076

311,990

0.92

‐8.00%

+/‐10%

Rural (5)

720,893

757,261

1.05

787,690

832,782 1.057

5.70%

+/‐10%

119

0.965 4,931,701

4,948,237 1.003

‐0.30%

+/‐5%

671

TOTAL

6,981,816 6,740,162

Model Page 2

G ENESEE CO UNTY TRAVEL DEM AND M O DEL

F I NAL REPORT

AUGUS T2017

GENESEE COUNTY TRAVEL DEMAND MODEL

Final Report

Table of Contents CHAPTER 1: MODEL REVIEW AND IMPROVEMENT RECOMMENDATIONS Introduction ............................................................................................................................................... 1‐1 The Model as Foundation for Performance‐based Planning .................................................................... 1‐1 Factors Considered During Development of Draft Performance Measures ............................................. 1‐2 LRTP Vision, Goals, and Objectives ............................................................................................................ 1‐2 Relationship to FAST Act Emphasis Areas ................................................................................................. 1‐2 Relationship to Federal Livability Principals .............................................................................................. 1‐2 Recommendations for Performance Measures ........................................................................................ 1‐2 Sensitivity to Active Travel Modes ............................................................................................................ 1‐6 Urban Design Score ................................................................................................................................... 1‐7 Diversity – Jobs/Household Ratio .............................................................................................................. 1‐8 Design ........................................................................................................................................................ 1‐8 Destinations ............................................................................................................................................... 1‐8 Distance to Transit ..................................................................................................................................... 1‐8 Economic Analysis ..................................................................................................................................... 1‐8 Roadway Segment Measures .................................................................................................................... 1‐8 BASIC MODEL UPDATE ACTIVITIES ............................................................................................................ 1‐9 Trip Generation ......................................................................................................................................... 1‐9 Household Stratification Curves ................................................................................................................ 1‐9 Destination Choice .................................................................................................................................. 1‐10 Mode Choice ............................................................................................................................................ 1‐10 Truck Model ............................................................................................................................................. 1‐11 Traffic Assignment ................................................................................................................................... 1‐11

CHAPTER 2: 2014 NETWORK DEVELOPMENT 2014 Network Development ..................................................................................................................... 2‐1 Highway Network ...................................................................................................................................... 2‐1 TIP Road Network Changes from 2005 to 2017 ........................................................................................ 2‐4 Traffic Signals ............................................................................................................................................. 2‐5 Bike/Walk Network ................................................................................................................................... 2‐7 Transit Network Updates .......................................................................................................................... 2‐8 Speed and Capacity Calculators .............................................................................................................. 2‐14 Delays on Interrupted Facilities ............................................................................................................... 2‐17 Observed Traffic Data .............................................................................................................................. 2‐18

GENESEE COUNTY TRAVEL DEMAND MODEL

Final Report

Table of Contents (continued) CHAPTER 3: TRAFFIC ANALYSIS ZONE DEVELOPMENT TAZ Development ...................................................................................................................................... 3‐1 Household Model ...................................................................................................................................... 3‐5 External Stations ........................................................................................................................................ 3‐9 Land Use Allocation Model ........................................................................................................................ 3‐9 5D Smart‐Growth Planning Tool ................................................................................................................ 3‐9 5D Variables ........................................................................................................................................ 3‐9 Urban Design Score ........................................................................................................................... 3‐11

CHAPTER 4: EXTERNAL TRIPS Introduction and Overview ....................................................................................................................... 4‐1 Base Year External Station Summary ........................................................................................................ 4‐1 Base Year External Trip Estimation ............................................................................................................ 4‐4 Michigan Statewide Travel Demand Model .............................................................................................. 4‐4 Subarea Analysis ........................................................................................................................................ 4‐4 Trip Table Adjustment ............................................................................................................................... 4‐4 External–Internal and Internal‐External (EI‐IE) Trips by Trip Purpose ...................................................... 4‐5 Future Year External Trip Estimation ...................................................................................................... 4‐10 Growth Rate Calculation ......................................................................................................................... 4‐10 2045 External Trip Estimation ................................................................................................................. 4‐13

CHAPTER 5: TRIP GENERATION Introduction ............................................................................................................................................... 5‐1 Overview of Trip Generation Model.......................................................................................................... 5‐1 MI Travel Counts ....................................................................................................................................... 5‐1 Trip Purpose Taxonomy ............................................................................................................................. 5‐2 Household Stratifications .......................................................................................................................... 5‐2 Production Rates ....................................................................................................................................... 5‐3 Attraction Rates ......................................................................................................................................... 5‐3 External‐Internal Trips ............................................................................................................................... 5‐4

CHAPTER 6: TRIP DISTRIBUTION Introduction and Overview ....................................................................................................................... 6‐1 Trip Distribution Model ............................................................................................................................. 6‐1 Input Data .................................................................................................................................................. 6‐2 Friction Factors .......................................................................................................................................... 6‐3 Feedback Loop ......................................................................................................................................... 6‐16 External‐External Trip Distribution .......................................................................................................... 6‐16 Time of Day Model .................................................................................................................................. 6‐16

GENESEE COUNTY TRAVEL DEMAND MODEL

Final Report

Table of Contents (continued) CHAPTER 7: MODE CHOICE MODEL Introduction and Overview ....................................................................................................................... 7‐1 Calibration, Coefficients and Constants .................................................................................................... 7‐2 Input and Output Files ............................................................................................................................... 7‐4

CHAPTER 8: TIME‐OF‐DAY MODEL Introduction and Overview ....................................................................................................................... 8‐1

CHAPTER 9: TRUCK MODEL Introduction and Overview ....................................................................................................................... 9‐1 Trip Generation ......................................................................................................................................... 9‐1 Origin‐Destination Matrix Estimation (ODME) .......................................................................................... 9‐2 Trip Distribution......................................................................................................................................... 9‐2 Time of Day Choice .................................................................................................................................... 9‐5 Model Validation Check and Adjustment .................................................................................................. 9‐5

CHAPTER 10: TRAFFIC ASSIGNMENT Introduction and Overview ..................................................................................................................... 10‐1 Trip Assignment Procedures .................................................................................................................... 10‐2 MMA Assignments .................................................................................................................................. 10‐3 Traffic Assignment Procedure ................................................................................................................. 10‐3 Feedback Loop ......................................................................................................................................... 10‐3 Trip Assignment Data Inputs ................................................................................................................... 10‐4 Transit Assignment .................................................................................................................................. 10‐4 Time of Day .............................................................................................................................................. 10‐5 Definition of TOD Periods ........................................................................................................................ 10‐6

CHAPTER 11: MODEL CALIBRATION/VALIDATION Introduction ............................................................................................................................................. 11‐1 Trip Distribution Summary ...................................................................................................................... 11‐1 Mode Choice Outputs.............................................................................................................................. 11‐3 Highway Assignment ............................................................................................................................... 11‐9 Screen line/Cutline Analysis .................................................................................................................. 11‐10 Level of Service Analysis ........................................................................................................................ 11‐14 Transit Ridership .................................................................................................................................... 11‐15 Trip Generation ..................................................................................................................................... 11‐16 Trip Production Rate Calibration ..................................................................................................... 11‐16 Cell Compression ............................................................................................................................. 11‐17

iii

GENESEE COUNTY TRAVEL DEMAND MODEL

Final Report

Table of Contents (continued) Home Based Work Income Stratification ........................................................................................ 11‐17 Number of Observations ................................................................................................................. 11‐17 Statistical Difference in Trip Rates .................................................................................................. 11‐17 Recommendation ............................................................................................................................ 11‐19 Combined Trip Rates ....................................................................................................................... 11‐21 Trip Attraction Rate Calibration ............................................................................................................ 11‐21 District Aggregation ........................................................................................................................ 11‐21 Model Estimation ............................................................................................................................ 11‐24 Correlation Analysis ........................................................................................................................ 11‐24 Regression Analysis ......................................................................................................................... 11‐25 Income Distribution of HBW Attractions ......................................................................................... 11‐27 External Models ..................................................................................................................................... 11‐28 Distribution of Trip Purposes ........................................................................................................... 11‐29 Average Trip Length and Distribution of Trip Ends ......................................................................... 11‐30 Model Sensitivity ................................................................................................................................... 11‐33

APPENDIX A – CAPACITY CALCULATION METHODOLOGY APPENDIX B – 2045 POPULATION PROJECTIONS METHODOLOGY REPORT APPENDIX C – 2045 EMPLOYMENT PROJECTIONS METHDOLOGY REPORT

GENESEE COUNTY TRAVEL DEMAND MODEL

Final Report

List of Figures Figure 1‐1: Livability Principals Relationship to Performance Measure Categories ................................. 1‐2 Figure 1‐2: Model Components' Contribution to System Measures ........................................................ 1‐3 Figure 1‐3: Infrastructure Coding for Non‐Motorized .............................................................................. 1‐6 Figure 1‐4: Suitability Index ...................................................................................................................... 1‐7 Figure 1‐5: Urban Design Score Elements ................................................................................................ 1‐7 Figure 1‐6: Economic Analysis Process ..................................................................................................... 1‐8 Figure 2‐1: Genesee County Highway Network........................................................................................ 2‐3 Figure 2‐2: Highway Network Attributes Used in the Suitability Index Calculations ............................... 2‐8 Figure 2‐3: Mass Transportation Authority Primary Bus Routes .............................................................. 2‐9 Figure 2‐4: Ride to Groceries .................................................................................................................. 2‐11 Figure 2‐5: Ride to Wellness ................................................................................................................... 2‐12 Figure 2‐6: 2014–2045 Model Representation of Flint MTA Bus Routes ............................................... 2‐13 Figure 2‐7: Lookup Table for Speed Calculation (mph) .......................................................................... 2‐15 Figure 2‐8: Lookup Table for Capacity Calculation (veh/ln/hr) .............................................................. 2‐15 Figure 2‐9: Auxiliary Lanes in the Genesee County Network ................................................................. 2‐16 Figure 2‐10: FHWA Vehicle Classification Scheme ................................................................................. 2‐19 Figure 3‐1: Genesee County Travel Model Study Area ............................................................................ 3‐2 Figure 3‐2: Genesee County TAZs ............................................................................................................. 3‐3 Figure 3‐3: TAZ Area Types in Genesee County ........................................................................................ 3‐4 Figure 3‐4: Household Stratification Models ............................................................................................ 3‐8 Figure 4‐1: External Station Locations and IDs ......................................................................................... 4‐2 Figure 6‐1: Logarithmic Graph of Friction Factors .................................................................................... 6‐5 Figure 6‐2: Trip Length Frequency Distribution (Travel Time) for AM Peak........................................... 6‐14 Figure 6‐3: Trip Length Frequency Distribution (Travel Time) for Midday ............................................. 6‐14 Figure 6‐4: Trip Length Frequency Distribution (Travel Time) for PM Peak ........................................... 6‐15 Figure 6‐5: Trip Length Frequency Distribution (Travel Time) for Off‐Peak ........................................... 6‐15 Figure 7‐1: Structure of Mode Choice Model ........................................................................................... 7‐1 Figure 8‐1: Time of Day Distribution of Trips ........................................................................................... 8‐1 Figure 9‐1: Truck Trip Length Frequency Distribution .............................................................................. 9‐4 Figure 9‐2: Results of Daily Truck Trip Assignment .................................................................................. 9‐6

GENESEE COUNTY TRAVEL DEMAND MODEL

Final Report

List of Figures (continued) Figure 11‐1: Model Report File ............................................................................................................... 11‐2 Figure 11‐2: Screen line Locations in Genesee County Travel Demand Model .................................... 11‐11 Figure 11‐3: Cutline Locations in Genesee County Travel Demand Model .......................................... 11‐12 Figure 11‐4: District Map ...................................................................................................................... 11‐22 Figure 11‐5: Total Observed vs. Modeled Attractions by Purpose ....................................................... 11‐26 Figure 11‐6: Distribution of Employees by Household Income ............................................................ 11‐27 Figure 11‐7: External Regions to Genesee County ............................................................................... 11‐29 Figure 11‐8: Location of Work Trip Ends .............................................................................................. 11‐31 Figure 11‐10: Population Changes 2014‐2045 ..................................................................................... 11‐34

GENESEE COUNTY TRAVEL DEMAND MODEL

Final Report

List of Tables Table 1‐1: Performance by Planning Category ......................................................................................... 1‐4 Table 1‐2: Potential Expansion of Trip Generation to Add Market Segmentation .................................. 1‐9 Table 2‐1: Road Network Changes 2014 to 2020 ..................................................................................... 2‐4 Table 2‐2: Signal Changes Made in the Node Network ............................................................................ 2‐6 Table 2‐3: MTA Fixed Route System Changes ........................................................................................ 2‐10 Table 2‐4: Link Type Descriptions Used for Speed and Capacity Calculations ..................................... 2‐14 Table 2‐5: Average Percent Increase in Capacity by Adding an Auxiliary Lane .................................... 2‐16 Table 2‐6: Capacity Calculation by Time Period ..................................................................................... 2‐17 Table 3‐1: Genesee County TAZ Numbering Scheme ............................................................................... 3‐5 Table 3‐2: Household Size Model Coefficients and Constants ................................................................. 3‐6 Table 3‐3: Workers Model Coefficients and Constants ............................................................................ 3‐6 Table 3‐4: Vehicles Model Coefficients and Constants ............................................................................ 3‐6 Table 3‐5: Income Model Coefficients and Constants .............................................................................. 3‐6 Table 3‐6: Comparison of Market Shares between the Old and New Model .......................................... 3‐7 Table 3‐7: Example of Growth Factors ..................................................................................................... 3‐9 Table 3‐8: New Urban Design Score Fields in the TAZ File ..................................................................... 3‐12 Table 4‐1: External Station Summary ....................................................................................................... 4‐3 Table 4‐2: 2014 External Auto Trip Estimation ......................................................................................... 4‐6 Table 4‐3: 2014 External Truck Trip Estimation ........................................................................................ 4‐7 Table 4‐4: 2014 External‐to‐External Auto Trips ...................................................................................... 4‐8 Table 4‐5: 2014 External‐to‐External Truck Trips ..................................................................................... 4‐9 Table 4‐6: 2014 EI‐IE Trips by Purpose ................................................................................................... 4‐11 Table 4‐7: Annual Growth Rate of Auto and Truck Trips ........................................................................ 4‐12 Table 4‐8: 2045 External Trips ................................................................................................................ 4‐14 Table 4‐9: 2045 External‐to‐External Auto Trips .................................................................................... 4‐15 Table 4‐10: 2045 External‐to‐External Truck Trips ................................................................................. 4‐16 Table 5‐1: Cross Classification Variables by Trip Purpose ........................................................................ 5‐3 Table 6‐1: Friction Factors by Trip Purpose .............................................................................................. 6‐4 Table 6‐2: Average Daily Travel Time by Purpose (Minutes) ................................................................... 6‐5 Table 6‐3: Trip Length Frequency (Average Travel Time in minutes) for AM Peak .................................. 6‐6 Table 6‐4: Trip Length Frequency (Average Travel Time in minutes) for Midday .................................... 6‐8 Table 6‐5: Trip Length Frequency (Average Travel Time in minutes) for PM Peak ................................ 6‐10 Table 6‐6: Trip Length Frequency (Average Travel Time in Minutes) for Off‐Peak ................................ 6‐12

vii

GENESEE COUNTY TRAVEL DEMAND MODEL

Final Report

List of Tables (continued) Table 7‐1: Nested Logit and Travel Utility Parameters ............................................................................. 7‐3 Table 7‐2: Modeled Mode Shares ............................................................................................................ 7‐4 Table 8‐1: Hourly Distribution of Trips ..................................................................................................... 8‐2 Table 8‐2: TOD Directional Factors by Trip Purpose (PAOD_ToD.bin) ..................................................... 8‐2 Table 9‐1: Daily Trip Generation Rates (QRFM II) .................................................................................... 9‐2 Table 9‐2: Summary of 2014 Trip Generation .......................................................................................... 9‐2 Table 9‐3: Average Travel Time by Trip Type ........................................................................................... 9‐3 Table 9‐4: Time of Day Factors ................................................................................................................. 9‐5 Table 9‐5: Model Volumes vs. Truck Traffic Counts ................................................................................. 9‐5 Table 10‐1: Default Volume Delay Function Parameters by Roadway Class .......................................... 10‐4 Table 11‐1: Average Trips Length Survey Versus Model (using updated times) .................................... 11‐2 Table 11‐2: Trip Distribution Summary Report (Time and Distance) ..................................................... 11‐3 Table 11‐3: Target and Modeled Mode Shares ...................................................................................... 11‐3 Table 11‐4: Mode Choice Summary – AM Peak ..................................................................................... 11‐4 Table 11‐5: Mode Choice Summary – Midday Peak ............................................................................... 11‐5 Table 11‐6: Mode Choice Summary – PM Peak ..................................................................................... 11‐6 Table 11‐7: Mode Choice Summary – Night ........................................................................................... 11‐7 Table 11‐8: Mode Choice Summary – Daily ............................................................................................ 11‐8 Table 11‐9: Count and VMT Comparison by Area Type ......................................................................... 11‐9 Table 11‐10: Count and VMT Comparison by Functional Class (NFC) .................................................... 11‐9 Table 11‐11: Percent RMSE by Volume Group ..................................................................................... 11‐10 Table 11‐12: Summary of the Screenline Counts vs. Volumes ............................................................. 11‐13 Table 11‐13: Volume and Count Comparison for the Cutline Corridors .............................................. 11‐13 Table 11‐14: Model Feedback Report by Time‐of‐Day ......................................................................... 11‐13 Table 11‐15: Level of Service Related Fields Calculated Using the Model Post Processors ................. 11‐14 Table 11‐16: Transit Ridership Comparisons (Model vs. on Board Surveys) for the Base Year ........... 11‐15 Table 11‐17: Work Cross Classification Scheme ................................................................................... 11‐16 Table 11‐18: Non‐Work Cross Classification Scheme ........................................................................... 11‐16 Table 11‐19: Number of Sampled Households (Work Related) ........................................................... 11‐18 Table 11‐20: HBW Low Income Trip Production Rates (Work Related) ............................................... 11‐19 Table 11‐21: HBW High Income Trip Production Rates (Work Related) .............................................. 11‐19 Table 11‐22: NHBW Trip Production Rates (Work Related) ................................................................. 11‐19 Table 11‐23: HBO Trip Production Rates (Non‐work Related) ............................................................. 11‐20 Table 11‐24: HBSH Trip Production Rates (Non‐work Related) ........................................................... 11‐20 Table 11‐25: NHBO Trip Production Rates (Non‐work Related) ........................................................... 11‐20

viii

GENESEE COUNTY TRAVEL DEMAND MODEL

Final Report

List of Tables (continued) Table 11‐26: HBSC (K‐12) Trip Production Rates (School Trips) ........................................................... 11‐20 Table 11‐27: HBSC (College ‐ Univ.) Trip Production Rates (School Trips) ........................................... 11‐20 Table 11‐28: Aggregated Average Daily Production Rate .................................................................... 11‐21 Table 11‐29: District Description .......................................................................................................... 11‐23 Table 11‐30: Correlation Analysis of Observed Trip Ends .................................................................... 11‐25 Table 11‐31: Trip Attraction Step‐Wise Regression Results ................................................................. 11‐26 Table 11‐32: Observed vs. Modeled Attractions by Purpose ............................................................... 11‐26 Table 11‐33: Percent Distribution of Employment Type ...................................................................... 11‐28 Table 11‐34: MI Travel Counts Two Day External Counts .................................................................... 11‐30 Table 11‐35: Average Trip Length in Minutes ...................................................................................... 11‐30 Table 11‐36: Model Sensitivity to Changes in Changes in Transportation Supply and Demand ......... 11‐33

Chapter 1

Model Review and Improvement Recommendations

Chapter 1: Model Review and Improvement Recommendations Introduction o Enhanced ability to conduct scenario testing;

The primary purpose of this project is to update and improve the Genesee County Travel Demand Model, to support regional transportation planning activities, leading to the development of the 2045 LRTP. The model update is a key component in the GCMPC work plan to complete the 2045 LRTP by January 2019. It is therefore important that the model improvements be oriented around the upcoming plan development needs. The project will be done in partnership with the members of the Model Development Committee (MDC) who will oversee the development of the model.

o Strong emphasis on performance based planning; and, o Model sensitivity to the types of projects being planned. • Measure economic effectiveness and economic impacts of project decision making; • Be prepared for new air quality conformity standards that may affect Genesee County; and, • Design an easier tool to use while growing GCMPC staff understanding of the model and its potential uses.

The purpose of a truly multi-modal transportation plan is to establish physical and cultural environments that support and encourage safe, comfortable and convenient travel by a variety of modes. The technical modeling and performance measurement procedures to be used in the transportation planning process should be designed to meet these needs. As such, this model update will have the following goals:

The Model as Foundation for Performancebased Planning The plan development process applies performancebased planning principles, supported by scenario planning and technical modeling. The results from this analysis will feed into the planning, decision making, project selection, and overall action plan to be documented in the remainder of the 2045 LRTP document. This report begins with a discussion of the anticipated set of performance measures needed to evaluate existing conditions, the impact of future travel demand on transportation facilities, and identify future system performance along with specific corridor deficiencies. The need to generate these performance outputs is the driver for the model improvement recommendations.

• Update to a new 2014 Base Year and develop model scenarios in increments out to 2045; • Improve the model capabilities: o Use best practices to guide what should be included in the model and how it is validated; o Add capability for the next LRTP development process;

1-1

Chapter 1

Model Review and Improvement Recommendations

Factors Considered During Development of Draft Performance Measures

Relationship to Federal Livability Principals The MTP Vision, Goals, and Objectives development also account for the livability principals that resulted from a 2009 collaboration of the U.S. Department of Transportation, U.S. Environmental Protection Agency, and the U.S. Department of Housing and Urban Development.

Long-range planning is driven by a number of factors: local growth and land use changes; federal emphasis on performance-based planning; the need to maintain the major highway system; the local city and county roads and multimodal alternatives; available funding; and, the planning process of the Genesee County Metropolitan Planning Commission which integrates these considerations.

1. 2. 3. 4. 5.

Provide more transportation choices; Promote equitable, affordable housing; Enhance economic competitiveness; Support existing communities; Coordinate policies and leverage investment; and, 6. Value communities and neighborhoods.

LRTP Vision, Goals, and Objectives To be consistent with a transparent, performancebased planning approach for the 2045 LRTP, a “performance framework” should be designed to convey the linkage to long-range transportation vision, goals, and objectives, to ensure decisions are made with the desired end-state in mind.

Environmental justice was included as an additional performance category.

Recommendations for Performance Measures

Relationship to FAST Act Emphasis Areas

Generally, performance measures are used to evaluate investment options and monitor progress towards achieving goals and objectives. The specific measures used here are listed in Table 1-1. Criteria

Fixing America’s Surface Transportation (FAST) Act includes an overarching performance-based framework, within which the long-range planning process is embedded. FAST Act Planning Categories are shown in Figure 1-1.

Figure 1-1: Livability Principals Relationship to Performance Measure Categories Livability Principles

Performance Measure Categories

Provide more transportation choices

Promote equitable, affordable housing

Enhance economic competitiveness

Support existing communities

Coordinate policies and leverage investment 

Value communities and neighborhoods

Travel Demand



System Efficiency



System Condition



Safety



Environmental Concerns



System Investment & Economics Environmental Justice (EJ)

 

1-2



Chapter 1

Model Review and Improvement Recommendations

for selecting measures included: feasibility of calculating, policy sensitivity, ease of understanding, and usefulness in decision-making. They are intended to help address questions such as:

From a long-range planning perspective, performance measures are normally used for tracking progress towards overall system-wide or corridor goals and objectives, rather than tracking the performance of individual projects. However, the technical process also generates link specific performance outputs for the auto/truck traffic using each road segment.

â&#x20AC;˘ Are the proposed investment strategies helping to achieve longer-term transportation goals? â&#x20AC;˘ Is the planning process identifying and evaluating appropriate transportation strategies?

Each of the performance measures is computed from a component of the travel demand model as shown in Figure 1-2.

â&#x20AC;˘ Is the region investing in transportation as efficiently and effectively as possible?

Figure 1-2: Model Components' Contribution to System Measures

1-3

Chapter 1

Model Review and Improvement Recommendations

Table 1-1: Performance by Planning Category Travel Demand Vehicle Miles (VMT)

VMT is calculated by multiplying the amount of daily traffic on a roadway segment by the length of the segment, then summing all the segments’ VMT to give a total for the geographical area.

Vehicle Hours (VHT)

VHT is calculated by multiplying the amount of daily traffic on a roadway segment by the travel time of the segment, then summing all the segments’ VHT to give a total for the geographical area.

Work Trip - Vehicle Occupancy

Average number of auto occupants for work trips

Person Trips

Total trip generation output for a given scenario. These represent all trips, regardless of mode.

Transit Share

Percentage of all trips choosing the transit modes under a given network scenario

Daily Ridership

Daily transit ridership under a given network scenario (Note: a trip involving a transfer has two boardings)

Transit Trips

The number of person trips

Transit Person Miles

Aggregate number of miles traveled by all transit riders in one day (one person travelling one mile equals one transit person mile)

Transit Person Hours

Aggregate number of hours traveled by all transit riders in one day (one person travelling one hour equals one transit person hour)

Non-Motorized Share

The percentage of person trips using non-motorized modes

Non-Motorized Trips

The aggregate number of person trips using non-motorized modes

Non-Motorized Person Miles

The aggregate number of miles travelled by all non-motorized travelers in one day

Non-Motorized Person Hours

The aggregate number of hours travelled by all non-motorized travelers in one day

Truck Vehicle Miles of Travel (VMT)

Truck VMT is calculated by multiplying the number of daily trucks on a roadway segment by the length of the segment, then summing all the segments’ VMT to give a total for the geographical area.

Truck Vehicle Hours (VHT) Truck trips Average Truck Speed

Truck VHT is calculated by multiplying the number of daily trucks on a roadway segment by the travel time of the segment, then summing all the segments’ VHT to give a total for the geographical area. This measure will be calculated as the number of trucks estimated by the truck model. Average truck speed will be calculated as truck VMT divided by truck VHT.

System Efficiency Vehicle Hours Under Delayed Conditions

Delay per Vehicle multiplied by the peak hour volume on that roadway segment; expressed in vehicle-hours; this performance measure facilitates the comparison of corridors by considering the number of vehicles impacted by congestion

Avg. PM Peak Speed

The average speed during peak periods (PM) on all non-freeway facilities

Avg. Auto Trip Length

The average trip length of all modeled auto trips for a given scenario

Lane Miles at LOS E or worse

Number of lane miles of non-freeway links where the level of service is E or worse during at least one of the four modeled time periods or overcapacity for the 24-hour period

1-4

Chapter 1

Model Review and Improvement Recommendations

Table 1-1: Performance by Planning Category (continued) System Efficiency (continued) Transit Accessibility (Residents within half-mile)

Number of households within a half-mile of transit service

Transit Accessibility (Jobs within 20 minutes by transit)

Number of jobs within 20 minutes travel by transit

Weighted average wait time for transit

The average wait time (directly related to headways) for transit riders

Urban Design Score

See text

Non-Motorized Accessibility Index

Number of jobs within 20 minutes travel by non-motorized modes

Safety Fatal Accidents

Number of fatal accidents predicted by current crash rates by road type

Injury Accidents

Number of injury accidents predicted by current crash rates by road type

Property Damage Accidents

Number of property damage only accidents predicted by current crash rates by road type

Bike Suitability

See text

Environment Vehicle Emissions (Tons CO2)

Daily tons of carbon dioxide emitted by on-road vehicles under a given scenario. This measure is dependent upon operating speeds and miles traveled. The emission rates were derived from the MOVES model.

Vehicle Emissions (Tons NOX)

Daily tons of NOX emitted by on-road vehicles under a given scenario. This measure is dependent upon operating speeds and miles traveled. The emission rates were derived from the MOVES model.

Vehicle Emissions (Tons HC)

Daily tons of hydrocarbons emitted by on-road vehicles under a given scenario. This measure is dependent upon operating speeds and miles traveled. The emission rates were derived from the MOVES model.

Economic Roadway User Costs

These costs account for both time and operating costs for autos and trucks. Vehicle operating costs include fuel and non-fuel costs.

Capacity Added to Meet Standards (Road Lane Miles)

This is an estimated number of lane miles needed to improve all model links to level of service D or better.

Total Cost of Capacity Improvements

This is an estimated cost to improve all model links to level of service D or better. This is based on a rough estimate of the remaining needs (see measure of lane miles at LOS E or worse), and costs per lane mile to add roadway capacity.

Prosperity Index

Housing costs plus transportation costs for a given zone, in relation to the regional average

Regional GDP Change

Estimated GDP based on MIBEST method to estimate the effect of a scenario compared to baseline conditions

Regional Personal Income Change

Estimated Income based on MIBEST method to estimate the effect of a scenario compared to baseline conditions

1-5

Chapter 1

Model Review and Improvement Recommendations

Table 1-1: Performance by Planning Category (continued) Environmental Justice (EJ) Average travel time for work purpose by auto and transit

Average travel times for work by auto and transit will be calculated and compared among all HH vs. low income HH and minorities

Percentage of population/households within a specified travel time (e.g. 25 minutes AM peak period) from a college

Average travel times by auto and transit between a TAZ with an existing college campus and other TAZs will be calculated and percentage of population/household within a certain travel time will be compared among all HH vs. low income HH and minorities

Percentage of population/households beyond a specific travel time (e.g. 10 minutes) from a healthcare facility

Average travel times by auto between a TAZ with available healthcare facility and other TAZs will be calculated and percentage of population/household beyond a certain travel time will be compared among all HH vs. low income HH and minorities

Average travel time to work by transit for households with fewer cars than workers

Average travel times by transit for households with fewer number of cars than workers for work purposes will be calculated and compared among all HH vs. low income HH and minorities

Sensitivity to Active Travel Modes model uses the combination of these factors to determine a suitability index. It is used as one component of an auto ownership model, and used directly for predicting routes taken by non-motorized travelers via a composite impedance combining travel time and suitability. The composite impedance is used for pathfinding during skims and assignment. The skims are used in the mode choice step, so improving system suitability can result in an increase in non-motorized travel and a reduction in auto travel.

Bike and walk suitability assessment will be added as part of the roadway/trail network for inclusion in multiple steps of the model. The suitability based on how many vehicles will pass a rider/walker while traversing a link, and trucks count disproportionately. The speed of vehicles and how closely they are passing the cyclist/walker is considered. The space available for the non-motorized traveler is considered, and this is based on shoulder width, or presence of bike lanes, sidewalks, sidepaths, etc. The Figure 1-3: Infrastructure Coding for Non-Motorized

1-6

Chapter 1

Model Review and Improvement Recommendations

Figure 1-4: Suitability Index

Urban Design Score

Density

An urban design score methodology is being incorporated into the model using a synthesis of research related to land development types and their impacts on travel behavior (less VMT and vehicle trip making for smart growth). The Design Score will be made up of five elements:

• Household Density – Households per residential acre. • Employment Density – Employment per TAZ Area (jobs per acre).

Figure 1-5: Urban Design Score Elements

1-7

Chapter 1

Model Review and Improvement Recommendations

Diversity – Jobs/Household Ratio

monetized for time savings, operating cost savings, safety cost savings, and air quality impact cost savings. The user benefit calculations are independent of transport mode, and are in the correct format for use in later FHWA TIGER grant applications.

• Number of Jobs in one-mile radius vs. • Number of Households in one-mile radius. Design • Suitability Index – Average suitability score for the TAZ. • Average Block Length – Average centerline miles of road (non-freeway) per link.

Roadway Segment Measures The scenario modeling process will also generate link specific performance outputs for the auto/truck traffic using each road segment and includes measures such as:

Destinations • Commercial establishments within ten-minute walk.

• • • •

Distance to Transit • Distance to nearest transit service. • Access to destinations via transit.

Level of Service; Peak Speeds; Hours of Delay; and, Daily Traffic.

These are similar to those available in the existing Genesee County model, but will be revised and expanded.

Economic Analysis A post-processing tool will be developed and used to extract model data for input into a user benefit analysis system. Under this system, user benefits are

Figure 1-6: Economic Analysis Process

1-8

Chapter 1

Model Review and Improvement Recommendations

Basic Model Update Activities Trip Generation TAZ-level household characteristics expressed in terms of zonal averages and disaggregates them for use in the trip generation process. New stratification curves will be developed using a probabilistic method in place of the traditional linear regression equations that are commonly used.

Corradino intends to use the model developed earlier for the previous LRTP for this study, but will review the trip rates in light of new studies and the traffic assignments resulting from subsequent steps. This model uses a cross-classification process to estimate production trip rates for each market segment (a joint distribution of the number of persons in each household and the number of autos owned).

Trip attractions are calculated using trip rates equations based on the number of retail and non-retail employees in each TAZ. Productions and attractions are calculated for home-based-work, home-basedother, and non-home-based trips.

Household Stratification Curves In a traditional modeling framework, some of the inherent aggregation bias in the trip generation process can be mitigated via use of household stratification curves. Stratification curves are used to take

Table 1-2: Potential Expansion of Trip Generation to Add Market Segmentation

Market Segment

Local Households Zero Autos Autos Less Than Workers – Lowest Income Quartile Autos Less Than Workers – Other Income Quartiles Autos Equal to Workers – Lowest Income Quartile Autos Equal to Workers – Other Income Quartiles University Students On-campus (Group Quarters) Off-campus Businesses Job-related Travel Truck Travel Visitors Hotel Guests Household Guests External External to External Internal to External Commuter Internal to External Other External to Internal Commuter External to Internal Other

Work Commute

Trips Produced Person Trips by Purpose Education K-12 Education Social Shop/Eat School University Recreational

    

 

    

Other

    

 

  

 

Trucks Short Long 4-Tire Distance Distance Commercial Truck Truck

 



 



 



1-9

Chapter 1

Model Review and Improvement Recommendations

Corradino will review the current list of special generators with the GCMPC staff, and will make revisions as necessary. Corradino will use ITE Trip Generation 9th Edition (2012) and other data to estimate appropriate trip rates for special generators.

• Travel distance; and, • Travel time logsums, including travel by transit modes. As in the gravity model, the destination choice model would be calibrated using the 2005 household O-D survey. Separate models would be developed for each trip purpose. In a destination choice model, shadow pricing is used to replicate the observed trip length frequency distribution instead of gravity model friction factors.

The trip generation model also contains a truck submodel, which will be explained further. Trip generation model development will involve balancing the productions and attractions as appropriate. The traditional procedure is to hold home-based productions constant, and balance attractions to their sums. It is best to exempt special generators from the balancing process.

Mode Choice The work will begin with the existing model, make certain improvements, and recalibrate the model to match existing 2014 transit ridership. This work also will include an update of the park-walk model. Note that work on the mode choice model will depend on the transit network having been completed.

After developing the trips rates, Corradino will update the GISDK code as necessary and test the model to produce the list of productions and attractions for each TAZ and trip purpose. The number of trips will be summarized and validated against expected values. Later, after the entire model is complete, Corradino may adjust trip rates, if needed, after comparing assigned trips and traffic counts on the network.

There will be enhancements to the mode choice model. These enhancements will include changing the GISDK structure to use Caliper’s latest nested logit model procedures, adding procedures for allowing external commuter trips to use the transit system, and accounting for transit station microcoding as required in FTA New Starts projects. Corradino also will add a nest to the logit model for the BRT system so that “unincluded attributes” such as reliability, branding, schedule-free service, span of service, station and stop passenger amenities, and dynamic schedule information (when the next bus arrives) that differentiate BRT from ordinary buses can be accounted for correctly.

Destination Choice Many models now use a destination choice model in place of the gravity model (in fact, a gravity model is a type of destination choice model). The advantage of a destination choice model is that other variables in addition to travel times, productions, and attractions, can be used in the trip distribution process. Furthermore, use of travel time logsums provides internal consistency with the transit network. Corradino proposes to implement a destination choice model. While this will be a model based on aggregate data, it would be another “precursor” step toward an activity-based model. Variables to be included in the model would be: • • • •

After all changes are made to the model and data, and the transit network is complete, Corradino will recalibrate the model to replicate the transit ridership counts as provided by GCMPC. In general, recalibration will require modifying the mode-specific constants in the utility expressions.

Trip productions; Trip attractions (as a size term); Other destination zone characteristics; Travel costs;

1-10

Chapter 1

Model Review and Improvement Recommendations

Truck Model

Traffic Assignment

The existing truck model was based on the QRS Freight procedures. In this task, trip rates and trip distribution parameters will be updated to the Quick Response Freight Manual II (2007) values. The model will be validated to available roadway classification counts. More advanced, commodity flow based, truck models will be reviewed under Task 1 and may be implemented where the QRFM covers non-commodity flow trips (more local deliveries), and the commodity flow based trucks (generally long distance trucks) are covered via utilization of MDOT Transearch data.

Corradino recommends the retention of the time-ofday periods used in the existing GCMPC model (AMpeak, midday, PM-peak, and night time periods). The sum of these periods will be the 24-hour daily trips and flows. Auto and transit assignments would continue to be made for each of the four periods, and the periods would be summed to produce daily assignments. Finally, as in the existing model, resulting travel times would then be fed back to trips distribution, assuring consistent travel times in all model steps. After the travel time feedback convergence is reached (doesn’t change between iterations), the travel model highway and traffic assignments will be reported by time periods and for the entire day. To decrease running time (the GCMPC model will run quickly on a modern microcomputer – running time will not be an issue), Corradino will use a method we have developed for other areas whereby the latest travel time skims are saved between runs, greatly reducing the need to apply multiple feedback iterations – the process is started at a point much nearer to the convergent solution than would be the case starting with free-flow travel time skims.

In its experience, Corradino has seen the difficulties in creating meaningful truck traffic assignments in urban travel demand models because the typical household and employment data are not particularly good indicators of truck activities (generation and travel patterns). This is especially true for warehouse activities and truck terminals. Thus, Corradino recommends using TransCAD ODME matrix estimator to develop base year truck trip tables from truck counts. Corradino will assess whether sufficient truck classification counts are available to support this method. If sufficient data are available, Corradino will implement a process whereby the base matrix is “fratared” by the ratio of future QRFMII trip ends/base QRFMII truck trips to produce future year truck trip tables.

The transit assignment will allow trips from external park-and-ride lots to be diverted to the transit mode. These features will have components in the transit network, mode choice, and assignment models. Recent research has shown that historically most highway assignment models were not run to a sufficient level of convergence. Thus, Corradino will use Caliper’s latest highway assignment with warm-start algorithms and will iterate the capacity restrained assignment to a relative gap of less than 0.00001.

1-11

Chapter 2

2014 Network Development

Chapter 2: 2014 Network Development 2014 Network Development There are over 4,000 links serving the 676 zones in the Genesee County travel model. Over 1,400 centroid connectors are used to link the centroids to the greater network. There are 37 external stations in the network. Here is an inventory of these key elements of the GCMPC network:

The purpose of this task was to develop the 2014 base highway network geographical database. The previous Genesee County Travel Demand Model (GCTDM) was developed for a 2005 base year. Future scenario years were 2011, 2018, 2025, and 2035. However, for the new model update effort, the Model Development Committee (MDC) and GCMPC’s policy committee, the Genesee County Metropolitan Alliance, approved 2014 as the new base year with future years of 2020, 2025, 2035, and 2045.

• • • • •

Highway Network

Number of links: 4,310; Number of nodes: 2,930; Number of Centroids: 676; Number of External Centroids: 37; and, Number of Centroid Connectors: 1,452.

The network and TAZ system were developed in a parallel process, with the number and placement of centroids, centroid connectors, and zone boundaries fitted to the important local characteristics of Genesee County. With the framework in place, the next step was to update the characteristics on the nodes, links and centroid connectors that will provide the capability to assign trips.

The updated GCMPC network, developed using TransCAD software, includes the following fundamental elements of travel model networks: • Nodes are elements that describe the position of intersections, junctions or switches in roadway or railway networks. Centroids are nodes that lie at the center of a Traffic Analysis Zone (TAZ).

After receiving a draft 2014 network in TransCAD format from the GCMPC, Corradino started the network development process for the base year of 2014. As a first step during this process, Corradino conducted a comprehensive review of the network for completeness and accuracy. The values for several network attributes were checked and corrected, if wrong, based on aerial photos. These included attributes such as number of lanes, lane direction, functional classes, area types, on-street parking, and posted speeds. In addition to making these checks, Corradino was also asked by the GCMPC to add a set of attributes to the network. The main reason to add

• Links are network model elements that connect the nodes and have attributes including direction, speed, capacity, functional classification, and observed traffic. They represent the street grid. • Centroid connectors link the zones to the network. They represent the distance to be covered between a zone’s center of gravity and the highway nodes or transit stops in the region.

2-1

Chapter 2

2014 Network Development

these fields to the network was that GCMPC is planning to use a TransCAD air quality add-in which converts the assignment outputs into inputs of the MDOT MOVES (Motor Vehicle Emission Simulator) process. Therefore, for this purpose, the following variables were coded into the travel demand model network:

Ramps need to be distinctly identified in the network for the air quality analysis. The values for this attribute can be “null” (links which do not represent ramps), U1_ramp, U2_ramp, R1_ramp, and R2_ramp. 5) MOVES Road Types Each link in the travel demand model network needs to be assigned a unique value which represents what road type it belongs to. Road type classifications are:

1) NFC = National Functional Class In 2010, the Highway Performance Management System (HPMS) changed the coding of the National Functional Classification (NFC). The distinction between urban and rural roads was no longer a factor in this classification, and all roads were classified with codes 1 through 7 with the following definitions. Note that centroid connectors were coded as a separate class: 1. 2. 3. 4. 5. 6. 7. 99.

1. 2. 3. 4. 5.

Off Network; Rural Restricted; Rural Unrestricted; Urban Restricted; and, Urban Unrestricted.

Also, two other variables were added to the network. ACUBL and ACUBR represent the Left and Right Adjusted Census Urban Boundaries, respectively. All of the above attributes were either brought over from Michigan Framework or were coded in the network using lookup tables from a combination of the existing attributes.

Interstate; Other freeways and expressways; Major arterials; Minor arterials; Major collectors; Minor collectors; Local; and, Centroid connectors.

The base year model will be validated to observed traffic count data. Therefore, as another step in this task, observed traffic counts acquired from MDOT and Genesee county agencies were coded in the network. Corradino was provided with traffic counts by vehicle classification, time of day, and direction of travel. Specifically, the truck counts will later be used during an Origin-Destination Matrix Estimation (ODME) process to adjust the base year truck trip table. More details on traffic counts are discussed later in this chapter.

2) AQ_Urban_Rural Since NFC does not distinguish between rural and urban roads, this attribute was added to the network to identify whether a road falls under rural or urban categories. Therefore, rural roads were coded as “R” and urban roads as “U.” 3) AQ_NFC_Urban_Rural

Figure 2-1 shows the Genesee County Travel Model network. Tables of the network attributes can be found in the Model Users Guide. The tables will only include the attributes that are used in the modeling process for different scenarios.

This field combines NFC 2010 variable with AQ_Urban_Rural. Thus, the values for this attribute range from R0, R1 through R7 and U0, U1 through U7. 4) Ramp_AQ_NFC

2-2

Chapter 2

2014 Network Development

Figure 2-1: Genesee County Highway Network

2-3

Chapter 2

2014 Network Development

TIP Road Network Changes from 2005 to 2017

with two in the thru lanes. Also projects that increase capacity by adding additional lanes were identified.

This task was completed before the beginning of this project by GCMPC staff. Data for the road network changes was obtained by reviewing the 2011-2014 Transportation Improvement Program (TIP), 2014-17 TIP and by requesting data from the Michigan Department of Transportation (MDOT), the Genesee County Road Commission and local units of government. This data was used to update the street attributes of the Genesee County model network. Attribute data for the following fields for each of the affected modeling scenarios was updated: NumLanes_xx, Thru_lanes_xx, Trfc_Op_CD_xx, and Dir_xx.

Projects which changed the direction of travel, i.e. one-way to two-way were also included. The “Flint Downtown Traffic and Parking Study” recommended several roadway and network improvements which included converting several downtown streets from one-way to two-way streets and changed the number of lanes on several other roads in the downtown Flint area. These changes occurred in 2010 and were updated in the 2014 model scenario year. Lastly, any project which increased capacity by adding new roads or reconfiguring existing roads was added. An example would be the Dort Highway extension which is modeled in the 2020 scenario year. New road segments were added along with all the corresponding attribute data from 2020 through the 2045 model years. An example of a reconfiguration would be the Fox street realignment or the Bristol Rd at I75 roundabout which are both modeled in the 2020 scenario. See Table 2-1 for a complete list of the road network changes from 2005 thru 2017.

Any projects that change capacity were added to a table to be updated in the model. An example of a capacity change would be a “road diet” which is a reconfiguration or restriping which reduces the number of traffic lanes on a roadway, typically from four lanes down to three. In the model, the network would be coded with a three in the number of lanes

Table 2-1: Road Network Changes 2014 to 2020 LOCATION

MODEL YEAR

TERMINI

LANES

DIRECTION

Beach St Beach St

2014

I-69 to 4th

1 way

2014

4th to Kearsley

1 way

Church St

2014

4th to Kearsley

1 way

1st St

2014

SB Ceasar Chaves to Grand Traverse

3 total 2 thru

2 way

Grand Traverse

2014

Kearsley to 9th

2 way

Harrison

2014

2nd to 4th

1 way

Harrison

2014

4th to 5th

1 way

Kearsley/Glenwood

2014

Harrison to Grand Traverse

3 total 2 thru

2 way

M21/Court/5th

2014

Ann Arbor to NB Chavez

1 way

Saginaw

2014

4th to Union

3 total 2 thru

2 way

Saginaw

2014

7th to 4th

2 way

2nd St

2014

Grand Traverse to NB Cesar Chavez

3 total 2 thru

2 way

Stevens

2014

5th to 4th St

4 total

2 way

Stevens

2014

4th to 1st

3 total 2 thru

2 way

2-4

Chapter 2

2014 Network Development

Table 2-1: Road Network Changes 2014 to 2020 (continued) LOCATION

MODEL YEAR

TERMINI

LANES

DIRECTION

3rd St

2014

NB Chavez to Grand Traverse

3 total 2 thru

2 way

4th St

2014

Beach to Saginaw

3 total 2 thru

2 way

Grand Blanc Rd

2014

W. City Limits to Saginaw

3 total 2 thru

2 way

Chevrolet

2014

University Ave to Glenwood

3 total 2 thru

2 way

Linden Rd

2014

Maple Ave to Bristol

2 way

Hill Rd

2014

Center to Genesee

3 total 2 thru

2 way

Baldwin

2014

Widen bridge to Holly Rd

2 to 5

2 way

Morrish

2014

I-69 overpass to Bristol

3 total 2 thru

2 way

Morrish

2014

Maple Ave to Miller

3 total 2 thru

2 way

Morrish

2014

Miller to I69

3 total 2 thru

2 way

Corunna

2014

Court to Ballenger

3 total 2 thru

2 way

5th Ave

2014

James P. Cole to Saginaw

5 total 4 thru

2 way

5th Ave

2014

MLK to Saginaw

4 total 4 thru

2 way

5th Ave

2014

Prospect to MLK

3 total 2 thru

2 way

Carpenter

2014

Clio to Fleming - resurface

4 total

2 way

Elms

2014

Potter to Flushing

3 total 2 thru

2 way

Carpenter

2014

Fleming to Dupont - resurface

5 total 4 thru

2 way

M15

2014

Mill Point to S. Hegel Mill & HMA overlay

3 total 2 thru

2 way

Carpenter

2020

Dupont to Saginaw

3 total 2 thru

2 way

Irish

2020

Irish @ Potter add center left turn lane

3 total 2 thru

2 way

Flushing

2020

Mill to Eldorado

3 total 2 thru

2 way

Davison (E. Flint)

2020

M15 to E. City Limits resurface

3 total 2 thru

2 way

Fox St

2020

Realignment Court to Glenwood

3 total 2 thru

2 way

Fenton Rd

2020

N Fenton City Limits to Butcher

2 to 3

2 way

Dort

2020

I75 @ M54 interchange new route

4 total

2 way

I75

2020

I75 @ Holly loop ramp to NB I75

1 way

I75

2020

Bristol @ I75 NB roundabout

2 way

M15

2020

Davison Rd to N. City Limit - add left turn lane

5 total 4 thru

2 way

M15

2020

Colonial to Potter - center left turn lane

3 total 2 thru

2 way

Traffic Signals

fields were updated: Traffic Signal, At_Grd_int, Signal TSID, Signal Type, Signal Cycle, Signal Timing, Intersection, Road_1, and Road_2.

This task was also completed before the beginning of this project by GCMPC staff. Traffic signal data was obtained from the Genesee County Road Commission and the City of Flint. This data was used to update the node attributes of the Genesee County model network. Attribute data for the following

The new signal data was compared to the 2005 node network to identify any changes. A total of 61 changes were made to the node layer. There were five nodes changed from flashers to signals while two

2-5

Chapter 2

2014 Network Development

rectly located in the 2005 base year. Table 2-2 identifies the changes made to the signals in the node layer.

were changed from signals to flashers. There were three signals removed and 44 signals were added. There were six signals relocated due to being incor-

Table 2-2: Signal Changes Made in the Node Network SIGNAL ID

ROAD 1

ROAD 2

TYPE

CHANGE

OWNERSHIP

T369

M-15 STATE

GREEN

Added to 2020 - 2045

State

T719

M-15 STATE

HEGEL ( ERIE )

EPAC

Added to 2020 - 2045

State

T808

M-21 CORUNNA

TA MANSOUR

EPAC

Added to 2014 - 2045

State

T884

US-23 SB

SILVERLAKE

EPAC

Added to 2020 - 2045

State

T880

US-23 NB

THOMPSON

EPAC

Added to 2020 - 2045

State

T729

US-23 SB

THOMPSON

EPAC

Flasher to Signal

State

T879

M-54 SAGINAW

WILSON

Added to 2020 - 2045

State

T881

M-57 VIENNA

BELSAY

Added to 2020 - 2045

State

T876

M-57 VIENNA

LINDEN

EPAC

Added to 2014 - 2045

State

T736

I-69 EB

MORRISH

EPAC

Changed to Flasher

State

T886

I-69 WB

MORRISH

EPAC

Added to 2020 - 2045

State

I-75 SB

MT. MORRIS

EPAC

Added to 2014 - 2045

State

T331

BELSAY

MAPLE (E. LEG)

Added to 2014 - 2045

Burton

T259

CENTER

COURTLAND MALL

EPAC

Relocated in model

Burton

T030

MILL

SMITH

EPAC

Added to 2020 - 2045

Clio

T056

OWEN

SILVRKPRKWY

EPAC

Relocated in Model

Fenton

T883

OWEN

TARGET

EPAC

Added to 2014 - 2045

Fenton

T077

SHIAWASEE

ROUNDS

EPAC

Added to 2020 - 2045

Fenton

T882

SILVERLAKE

POPLAR

EPAC

Added to 2020 - 2045

Fenton

T057

SILVRPRKWY

SILVRLKE VILL.

EPIC

Relocated in Model

Fenton

T055

TORREY

S LONG LAKE

EPAC

Flasher to Signal

Fenton

T063

SAGINAW

BELLA VISTA

ECONO

Relocated

Grand Blanc

T018

SAGINAW

HOLLY

ECONO

Relocated

Grand Blanc

T069

SAGINAW

REID

ECONO

Added to 2014 - 2045

Grand Blanc

T025

MILLER

FAIRCHILD

EPAC

Added to 2020 - 2045

Swartz Creek

T039

MILLER

WINSTON

changed to T025

Swartz Creek

T292

MILLER

CARRIAGE PLAZA

EF-140

Removed 2020 -2045

Swartz Creek

COURT ST

FOX/MILLER

EPAC

Added to 2014 - 2045

City of Flint

PIERSON

LONGFELLOW

EPAC

Removed 2020 -2045

City of Flint

T126

BEECHER

CALKINS

EPAC

Added to 2014 - 2045

County

T349

BEECHER

MORRISH

Added to 2014 - 2045

County

T360

BRISTOL

BISHOP EAST

MARC

Added to 2014 - 2045

County

T359

BRISTOL

BISHOP MAIN

EPAC

Added to 2014 - 2045

County

T358

BRISTOL

BISHOP WEST

EPAC

Added to 2014 - 2045

County

2-6

Chapter 2

2014 Network Development

Table 2-2: Signal Changes Made in the Node Network (continued) SIGNAL ID

ROAD 1

ROAD 2

TYPE

CHANGE

OWNERSHIP

T114

CENTER

RICHFIELD

Flasher to Signal 2014-45

County

T222

CLIO

CLIO COURT

EPAC

Removed 2014 - 2045

County

T111

COLDWATER

HORTON

Signal to Flasher 2014 - 45

County

T356

DAVISON

GALE

EPAC

Added to 2014 - 2045

County

T346

ELMS

HILL

Added to 2014 - 2045

County

T339

HOLLY

COOK (EAST LEG)

Added to 2014 - 2045

County

T297

HOLLY

COOK (WEST LEG)

EPAC

Flasher to Signal 2014-45

County

T340

HOLLY

McCANDLISH

EPAC

Added to 2014 - 2045

County

T354

LAHRING

TORREY

Added to 2014 - 2045

County

T303

LAPEER

VASSAR

EPAC

Flasher to Signal 2014-45

County

T353

LENNON

MORRISH

Added to 2014 - 2045

County

T357

LENNON

SEYMOUR

Added to 2014 - 2045

County

T336

LEWIS

WILSON

Added to 2014 - 2045

County

T372

LINDEN

MENARDS

Added to 2014 - 2045

County

T352

LINDEN

THOMPSON

Added to 2014 - 2045

County

T327

LINDEN H.S.

SILVERLKE RD.

EPAC

Relocated

County

T302

MILLER

SEYMOUR

EPIC

Added to 2014 - 2045

County

T370

MT MORRIS

JENNINGS

Added to 2014 - 2045

County

T342

PERRY

BELSAY

EPAC

Added to 2014 - 2045

County

T341

PERRY

GENESEE

EPAC

Added to 2014 - 2045

County

T355

PERRY

PERRY M SCHOOL

EPAC

Added to 2014 - 2045

County

T350

SAGINAW

HERITAGE PARK

EPAC

Added to 2014 - 2045

County

T335

SAGINAW

McCANDLISH

EPAC

Added to 2014 - 2045

County

T362

SEYMOUR

BALDWIN

Added to 2014 - 2045

County

T207

STANLEY

CURVE,W OF GEN.

Added to 2014 - 2045

County

T366

TUSCOLA

LAKE

Added to 2014 - 2045

County

T363

WILSON RD

HENDERSON

Added to 2014 - 2045

County

Bike/Walk Network

travel times for the modes and certain zonal characteristics. Separate travel times will be calculated for walk and bike travel. Travel times will be based on a starting walk speed of 3 miles per hour (mph) and bike speed of 12 mph. The travel time assumption for a given network link will be the time at the starting speed divided by a suitability index, unique for biking and walking, which ranges between 0 and 1. For instance, if the bike suitability index is 1, the speed will be 12 mph. If the suitability index is 0.1, the speed

The highway network is the starting point for the bike/walk network. It is possible, but not likely, that some bike- or pedestrian-only facilities will be added during the calibration phase. The bike/walk network will allow the calculation of bike and walk travel times for use in the mode choice model. Then, like other modes, trips will be allowed to bike and walk modes as a function of the relative

2-7

Chapter 2

2014 Network Development

Transit Network Updates

will be 1.2 mph. If the suitability index is very low, the link will be effectively eliminated from the network.

The transit network section documents the updates made to the transit route system in the model. These updates were made prior to the beginning of this project by GCMPC staff. The Flint MTA operates 14 distinct routes during a typical weekday (Figure 2-3). Note that in the TransCAD transit route file, there are 31 different routes representing the 14 Flint bus route system by different directions (e.g. northbound/southbound). Route service information collected from the Flint MTA in January of 2014 was used to update the transit route system and the bus stop table in the model. The transit route system uses a master network, which keeps all future network scenarios in one file.

The calculation of the suitability index will be based on two publications, “The Bicycle Compatibility Index: A Level of Service Concept, Implementation Manual,” FHWA-RD-98-095, and “Modeling the Roadside Walking Environment: A Pedestrian Level of Service,” TRB Paper No. 01-0511 (Bruce Landis, et al.). Constants developed in these studies will be used in the indices, but might be adjusted during model calibration. The suitability indices will be sensitive to characteristics of the non-motorized travel way and competing motorized travel. The highway network attributes shown in Figure 2 will be used in the calculations.

A review of the 2014 transit routes revealed that no new routes needed to be created and none of the existing routes were eliminated for the model base year (2014). However, some routes were modified to match minor transit route realignments. Some changes of note include the addition of a roundabout in 2020 on Bristol road which required realignment thru the roundabout on the Fenton Road routes. Several minor changes were identified in the downtown area that realigned the approach to the MTA downtown service center. Table 2-3 details the changes made to the 18 routes. Also, two new routes were added to the 14 fixed routes during year 2015. “Rides to Wellness” and “Rides to Groceries” are specialty routes that MTA added. These were created after the closure of all the grocery stores in the City of Flint. These give people access to groceries and health facilities. Since MTA is planning on keeping these routes, we also added them to the TransCAD transit route file. Figures 2-4 and 2-5 show these two routes and some additional information on them. Note that these routes are only used for future year analysis in the model, not base year, since they were added in 2015. Figure 2-6 shows the results of the route realignments in TransCAD for the 2014 thru 2045 modeling scenarios.

Figure 2-2: Highway Network Attributes Used in the Suitability Index Calculations Description (Attribute) Bike Walk Presence and type of Bike Lanes (Bk_lns) X X X Presence and type of sidewalks (sidewalk) X Highway shoulder width (SHOULDER_WIDTH) X X Total number of lanes (NUM_LANES) X X X Number of through lanes (THRU_LANES) X X Presence of on-street parking (PARKING) X Posted speed limit (POSTED_SPEED) X X Area Type (AREA_TYPE) X X X X Assigned vehicular volume (vol) X Assigned truck volume (trucks)

Note that the attributes used in the walk and pedestrian indices are very similar, but their importance in each index is different. Note also that assigned volumes are used in the calculation. In the first iteration of the model, an assumed volume will be used, but the model-assigned volumes will be used in subsequent iterations.

2-8

Chapter 2

2014 Network Development

Figure 2-3: Mass Transportation Authority Primary Bus Routes

2-9

Chapter 2

2014 Network Development

Table 2-3: MTA Fixed Route System Changes

NAME 1NB 1SB 2NB 2SB 3NB 3SB 4NB

ROUTE ID 79 80 52 53 58 59 75

ROUTE_DIR Outbound Inbound Outbound Inbound Outbound Inbound Outbound

ROUTE DESCRIPTION North Saginaw North Saginaw ML King Avenue ML King Avenue Miller-Linden Miller-Linden Civic Park

4SB

Inbound

Civic Park

5NB

Outbound

Dupont

5SB 6NB

54 47

Inbound Outbound

Dupont Lewis-Selby

6SB

Inbound

Lewis-Selby

7NB 7SB 8LNB 8LSB 8SNB 8SSB 9NB 9NB_2

73 72 87 88 67 66 16 49

Outbound Inbound Inbound Outbound Inbound Outbound Inbound Inbound

Franklin Franklin South Saginaw South Saginaw South Saginaw South Saginaw Lapeer Road Lapeer Road

9SB

Outbound

Lapeer Road

10NB 10SB

71 70

Outbound Inbound

Richfield Road Richfield Road

11NB

Inbound

Fenton Road

11SB 11SBfuture 12EB 12WB 13EB 13WB 14

61 89 65 60 57 56 74

Outbound Outbound Inbound Outbound Inbound Outbound Loop

Fenton Road Fenton Road Beecher-Corunna Beecher-Corunna Crosstown North Crosstown North Downtown-Campus

2-10

CHANGE No Change Beach to 3rd to Church to 2nd No Change Beach to 3rd to Church to 2nd 2nd to Wahllenberg to Court, adjust stops No Change No Change 3rd to Beach, removed stops on Mason and 2nd Ave Clio to Hallwood Plaza, remove stops on Stedron and Cloverlawn Beach to 3rd to Church to 2nd No Change Chavez to Robert T. Longway to Kearsley to Chavez, added stops along new route No Change No Change No Change No Change No Change No Change No Change No Change 2nd to Cesar Chavez to 5th, remove stops on Stevens add on Cesar Chavez No Change No Change Delete Bristol to Airport remove stops on segment, Realign thru roundabout on Bristol Realign thru roundabout on Bristol Roundabout on Bristol No Change Kearsley to Church to 3rd St, adjust stops No Change No Change Chavez to 1st to Stevens

Chapter 2

2014 Network Development

Figure 2-4 1: Ride to Groceries 0F

http://mtaflint.org/docs/r2g/NEWR2GFlyer.pdf, July 2016 1

2-11

Chapter 2

2014 Network Development

Figure 2-5 2: Ride to Wellness 1F

http://mtaflint.org/docs/Ride_to_Wellness_flyer_3_16.pdf, July 2016 2

2-12

Chapter 2

2014 Network Development

Figure 2-6: 2014â&#x20AC;&#x201C;2045 Model Representation of Flint MTA Bus Routes

2-13

Chapter 2

2014 Network Development

As another phase of this project, Corradino reviewed the transit network for any potential modifications to represent existing 2014 services. The latest information from the Flint Mass Transit Authority (MTA) suggested that only the model’s fare structure needed to be updated, based on new rates from MTA’s website. 3 The new rates are increased by half a dollar to $1.75. Scenario years were also modified for the model to 2014, 2020, 2025, 2035 and 2045 as approved by GCMPC for the new model update. The transit route system in the previous model was developed concurrently with the development of the roadway network and TAZs to accommodate any special considerations needed for transit modeling in the design of the new TAZ structure and/or road network. Route service information was collected from MTA. It’s important to note that the Flint MTA system operates both fixed routes and curb-to-curb “Your Ride” service. However, only the fixed route portion was represented in the TransCAD model. This fixed-route bus system structure is a classic hub and spoke system centered on the downtown Flint transit center. Route alignments and headways vary by time of day. A complete table of route attributes showing headway, seat capacity and so on can be found in the Model Users Guide.

methodologies used in their small urban travel demand models. A third advantage to this decision was that Corradino had already coded the new methodologies in GISDK for the Tri-County TDM. Only minor changes to the existing code were required to make it work with the GCTDM.

Both speed and capacity calculation methodologies use lookup tables based on area types and link types. Area types (determined by the MPO staff) are similar to the existing “Area_Type” attribute in the network: 1. 2. 3. 4. 5.

Link types, on the other hand, are based on new definitions rather than any existing attribute in the network. However, the values for this attribute could be easily calculated based on lookup tables from a combination of existing attributes. The link type is coded as “SpdCap_Link_Type” in the network and can have any of the values in Table 2-4. Table 2-4: Link Type Descriptions Used for Speed and Capacity Calculations

Speed and Capacity Calculators

SpdCap_Link_Type 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

The previous GCTDM used speed and capacity calculators based on HCM2000 methodologies. However, Corradino was asked by GCMPC to update these calculators to the most recent methodologies of HCM2010. After internal discussions among Corradino team members, it was decided that the TriCounty travel demand model calculators (developed for Lansing, MI) would work well for GCTDM. This decision was recommended to GCMPC PM and staff for approval. GCMPC also concurred on this decision and, therefore, Corradino updated the previous calculators based on the new methodologies. The main benefits of adopting the Tri-County speed and capacity calculators are that these procedures are based on HCM 2010 and in accordance with Michigan DOT 3

Central Business District; Urban; Suburban; Fringe; and, Rural,

http://mtaflint.org/travel/fares.php, April 2016.

2-14

DESCRIPTION Freeway High-speed Ramps On-ramps Off-ramps Principal Arterial w/TWLTL Principal Arterial One-way Minor Arterial Minor Arterial w/TWLTL Minor Arterial One-way Collector Collector w/TWLTL Collector Local Roadway w/TWLTL Local Road Centroid Connectors Trunk Principal Arterial w/TWLTL Trunk Principal Arterial

Chapter 2

2014 Network Development

free flow speeds were the starting values to calculate the free flow speeds for the MDOTâ&#x20AC;&#x2122;s Urban Model Improvement Program (UMIP). These speeds were modified and calibrated for Tri-county Travel Demand Model. During the course of this project, if necessary, these values will be adjusted for Genesee county.

After coding this new attribute into the network, the same lookup tables as for Tri-County model were used for the GCTDM as well. During the course of the project, these values will be validated and modified, if necessary. Figures 2-7 and 2-8 show the lookup table used for speed and capacity calculations (HDUC stands for High Density Urban Clusters). Georgia DOT Figure 2-7: Lookup Table for Speed Calculation (mph)

Figure 2-8: Lookup Table for Capacity Calculation (veh/ln/hr)

2-15

Chapter 2

2014 Network Development

to identify the links with auxiliary lanes. So, if a freeway link has an auxiliary lane, the value under “AuxLane” is set to 1; otherwise, it’s zero. Only 8 links in the Genesee county network were identified with auxiliary lanes. Figure 2-9 shows a map of the locations were auxiliary lanes were coded in the network.

Capacities for each link are later adjusted in the GISDK code based on lane width, on-street parking availability, and existence of auxiliary lanes in freeways. Auxiliary lanes are used to connect freeway on-ramps to the downstream off-ramp in order to reduce the impact of merging and diverging traffic on through traffic. Through a mutual collaboration with GCMPC, a new attribute was added to the network Figure 2-9: Auxiliary Lanes in the Genesee County Network

Based on HCM 2010 methodologies, the following adjustments were made to the capacity values from the lookup table:

shows the percentage increase in capacity based on number of through lanes. Table 2-5: Average Percent Increase in Capacity by Adding an Auxiliary Lane

• Lane width: If lane width is less than 12 feet, the link capacity is reduced by 4%. • On-street parking: The capacity is reduced by 5% for on-street parking on one side of the street and 10% for parking on both sides of the street. • Auxiliary lanes: The capacity is increased if auxiliary lanes exist in freeways. Table 2-5

2-16

NUMBER OF THROUGH LANES

PERCENTAGE INCREASE IN CAPACITY

48.87

32.07

23.81

18.71

Chapter 2

2014 Network Development

ones used in Tri-County TDM were also used to calculate the time period capacities as shown in Table 2-6. It is also notable that these values will later be validated and modified, if necessary.

In this model, the user is given the option of overriding any computed speed and capacity value from HCM 2010 for any link. To do so, the modeler must manually enter the desired speed and/or capacity value under “Speed_Adj” and/or “Cap_Adj” in the network for that specific link, respectively.

Table 2-6: Capacity Calculation by Time Period

The hourly directional capacity for each link is then calculated by multiplying the number of lanes by the value from the lookup table and adjusting it based on the above-mentioned conditions. For time-of-day analysis, the capacity values must be calculated for each time period. The estimation of link capacity by time period is not as straightforward as multiplying the hourly capacity by the duration (in hours) of the time period. The process is complicated by the fact that the travel demand varies throughout the different time periods. The simplification of multiplying the capacities by the duration of the time period would, therefore, over-estimates the capacity for the time period and consequently under-estimates the congestion.

TIME PERIOD

CAPACITY

AM Peak (6am-9am)

Hourly adjusted capacity * 2.29

Midday (9am-3pm)

Hourly adjusted capacity * 4.82

PM Peak (3pm-6pm)

Hourly adjusted capacity * 2.69

Nighttime (6pm-6am)

Hourly adjusted capacity * 3.66

Based on the new methodology, the GISDK code was updated and tested to make sure that the model runs smoothly without any errors and the correct speed and capacity values are calculated. Appendix A describes the methodologies for capacity calculation in more details.

The methodology used in the model assumes that the peak period congestion is equivalent to the peak hour congestion, and the congestion for the off-peak period is equivalent to the congestion in the median off-peak hour. This process assumes that the trips during each time period are concentrated into a briefer time period, so that all these trips are subject to the peak level of congestion during the time period. For example, a time of day distribution may have a PM peak period of three hours during which 25 percent of the daily trips occur, and a peak hour during which 10 percent of the daily trips occurs.

Delays on Interrupted Facilities Free-flow speeds and roadway capacities estimated in the previous steps are adjusted to account for delays associated with traffic signals. The adjustment is made directionally according to the methodology described below. Traffic signals are entered in the network as link attributes with designations of approach prioritization and multiple signals. If the approach to the signalized intersection is a higher functional class than crossroad, it is coded as “high” priority. If it is on par with the crossroad, it is assumed to have “equal” priority. If it is a lower functional class than the crossroad, it is given “low” priority. The number of multiple upstream signals are coded to account for progression effect as a result of signal coordination.

The methodology calculates the peak period capacity as 2.5 times the peak-hour capacity, where 2.5 equals the peak period trip rate divided by the peak hour trip rate. The factor used to estimate the offpeak capacity (Night) is calculated by dividing the share of all off-peak trips by the share of trips during the median off-peak hour. Similar coefficients to the

2-17

Chapter 2

2014 Network Development

These speed and capacity adjustments due to traffic signals are made directionally. Thus, signal approach lane(s) and lane(s) in the other direction are estimated with different speed and capacity values. In the Genesee County node layer signal information is stored in the field: “TRAFFIC SIGNAL”. The presence of a signal is indicated by filling this filed with “Y”. A new convention is being developed for filling this field to consider the presence of left turn lanes by approach in the capacity calculations.

The speed and capacity adjustment for traffic signal delay followed a HCM methodology that uses the following equation: 2

 g d = 0.5C1 -  ⋅ PF  C where: d = delay per vehicle; g = effective green time; C = cycle length; and, PF = progression adjustment factor.

Observed Traffic Data Genesee County Road Commission (GCRC) and Michigan DOT staff provided traffic data to be added to the network which will ultimately be used to validate both the truck and total traffic on all roadways in Genesee County to the year 2014 for four time periods. The overall goal of the count integration process was:

Delay estimated from the above equation is added to the free-flow speed-based link travel-time to come up with an “adjusted” free-flow travel time. Based on the fact that the mainline road is given a higher priority than the lower-class crossroad, varying green time ratios (g/C) are assumed by the priority code of the signal approach. HCM provides the progression adjustment factor as a function of the green time ratio and the arrival type. The arrival type for the signal approach is assumed based on multiple signals coded in the network. With the assumed green time ratio and the arrival type, an appropriate progression factor in HCM is sought and used to estimate signal delay of the approach.

• To obtain counts on the major interstate and Michigan roadways to satisfy overall model accuracy in the highest traffic locations; • To include as many secondary roads in the county, particularly those state and county roads that are heavily used; • To incorporate some observed data from each city in the county, all external stations, and rural areas, particularly those where population is growing;

The capacity reduction methodology is based on travel-speed reductions resulting from delays on the flow-interrupted facilities. The service flow rate is a function of the travel time along a road segment. Increasing signal densities effectively reduces travel speeds, and, in turn, reduces the amount of traffic flow that is possible. The reduction in service flow is calculated by dividing the maximum service flow approximate based on free-flow speed by the maximum service flow approximates based on speeds with traffic signal delays.

• To integrate truck traffic data wherever available; • To integrate time of day data wherever possible; and, • To physically enter this count data by direction on the final Genesee County travel model network.

2-18

Chapter 2

One of the biggest challenges in this project was the lack of sufficient traffic count data by time of day and also by vehicle classifications. GCRC is the main source of traffic count data for Genesee county. They upload their traffic count database on â&#x20AC;&#x153;ms2softâ&#x20AC;? website 4. The main efforts on collecting data for this project took place during the course of the model update process. Therefore, due to time constraints, only data on a few count stations were updated with classification and Time-of-day counts. Also, Michigan DOT provided historic count data on some major roads from 2010 to 2014. Traffic count data from both sources were added to the Genesee county network in TransCAD.

2014 Network Development

Figure 2-10: FHWA Vehicle Classification Scheme

MDOT and GCRC provided vehicle classification counts on a subset of the highway segments that were counted. The vehicle classification referenced the standard 13-class FHWA scheme shown in Figure 2-10. Commercial vehicles (trucks) were defined as FHWA classification 5 and greater. Only 52 Genesee County network links have counts by Time-of-day and Classification for 2014.

http://gcmpc.ms2soft.com/tcds/tsearch.asp?loc=Gcmpc &mod= 4

2-19

Chapter 3

Traffic Analysis Zone Development

Chapter 3: Traffic Analysis Zone Development TAZ Development The study area of the Genesee county model was disaggregated into 639 traffic analysis zones (Figure 3-2). There are 37 external zones, and the TAZ layer consists of a total of 676 zones. The internalzone attributes include land area, county name/number, TAZ number and detailed categorization of population, households, vehicle ownership, mean household income, school enrollment, university enrollment, and employment by economic sector. These demographic and employment features are the inputs for trip generation. The TAZ layer contains the multi-year attribute data, including the data not only for the 2014 base year, but also for the future years. For details about TAZ attributes, refer to the Model Users Guide.

The purpose of this task was to update the TAZ structure. This included updating the geometry of the TAZ polygons, updating the data allocation model, and updating the household model. To do so, as the first step, GCMPC provided Corradino with all the required data needed for model runs for years 2014, 2020, 2025, 2035, and 2045 on population, employment, and school enrollment. The TAZ file was updated using this new dataset. In general, TAZs should be bounded by major roadways and should not be bisected by roadways that carry through traffic. Also, TAZs should be limited in size so that centroid connectors do not deliver unreasonably large point loadings where they join the roadway network. Finally, TAZs should be small where there is intense transit service, so that walk access trips can be reasonably represented. Therefore, Corradino reviewed the TAZ system with respect to the issues noted above, paying special attention to Census boundaries, any new major roadways built or planned since 2005, and anticipated transit system expansion. Obviously, some compromises between conflicting requirements – the ideal situation and data sources – are required. At this point, Corradino believes that the existing TAZ structure meets these requirements.

One of the attributes of each TAZ is the area type. There are five Area Types used in the Genesee TDM TAZ structure: 1) 2) 3) 4) 5)

CBD; Urban; Suburban; Fringe; and, Rural.

External stations are assigned area type of zero. Local GCMPC staff added the values for area type based on their understanding and knowledge of the area. Figure 3-3 shows the area type for each TAZ.

The model study area fully covers Genesee County (Figure 3-1). All roadway classes—which include interstates, major and minor arterials, major and minor collectors, and some local roads—are represented in the model’s coverage area. The zone structure of the county is detailed to address diverse and intense socioeconomic activities in the county.

3-1

Chapter 3

Traffic Analysis Zone Development

Figure 3-1: Genesee County Travel Model Study Area BURT $

]^\75 MONTROSE $

OTTER LAKE $

;!57

CLIO

;!57

OTISVILLE $

COLUMBIA $

MOUNT MORRIS

WLOTHROP

$ BEECHER

;!54

;!15

FLUSHING $

;!13

]^\75

DAVISON $

$ FLINT

]^\69

;!21

]^\475

LENNON $

\]^69

$ BURTON

SWARTZ CREEK $

GRAND BLANC $

GOODRICH

AND

GAINES $

Genesee County Water Area ORTONVILLE County (High) Flint Burton Beecher ;!15 Other Cities 2 4 6 $

LAKE FENTON $

\]^

`_23

BYRON $

LINDEN $

$ FENTON

ARGENTINE $

3-2

HOLLY $

Miles

Chapter 3

Traffic Analysis Zone Development

Figure 3-2: Genesee County TAZs 640

647

646

644 645

642

641

649

648

286 252

245 250 243 674

244 557 558

246 247 361

253 257 258

251 255 272 254 256 559 260 261 262

248

263 264265 266

267

249

268 269 270

271

275 276 280281

273

274

284

285 287 564

277

278

288 289

282

283

291

292

290

531 668

362

335 336 337

338

549 548 546 613 547 551 614 534 550 628 630 535 627 615 553 617 626 629 618 634 662 554 639 537 552 622 635 637 555 536 663 666 667 665 664

652

653 654 655

656

657

Traffic Analysis Zone TAZ Genesee County 0 2 4 Miles

3-3

651

293

296 312 294 295 309 310 311 565 673 317 299 343 567 313 314 316 298 297 340 341 345 346 364 302 301 319 321 322 300 339 315 365 363 342 347 348 326 325 349 350 351 352 353 320324 367 154 155 159 570 305 356 303 304 355 569 366 577 354 151 156 162 187 581 186 329 328 327 672 572573 579 150 147 164 308 330 334 306 307 357 358 359 144 146 368 332 333 369 185 182 165 136 419 418 582 195 196142 135 179 177 382383 384 417 372 371 192193 198 140 197 122 128 169 172 174 385 386 387 584 199 201 203 123 585960 388 389 390 7 423 422 421 420 4 26 56 636261 117 394 393 392 206 207 208 209 103 113 3953 428 374 5266 6770 395 396 373 210 211 212 213 98 92 91 4580 87 671 94 219 430 214 216 220222 90 88 8177 74 401 400398 397 427 377 848582 7675 402 411 413 415 432 426 431 223 226 229 236 240 405 403 378 608611232 375 89 189 434 376 604 609 233 239 242 407409 410 412 414 416 433 670 234 190188 380 602 605 379 606 448 450 451 452 453 437 495496 601 490 491 494 497445 447 436 435 524 528 465 466 669 501502 454457 461 438 440 462 592593 467 518 519 455 505 460 588 504 468 599 587 596 476 529 525 526 498 499 500 439 475 675 471 469 589590 508 506 477 472 474 444 509 483 441 479 484 478 516 517 507 510 482 527 442 530 520 485 443 512 513 514 515 489 488 487 511 523 521 543 659 539 541 658 676 661 660 538 545 542 544 540 533 532 360

650

Chapter 3

Traffic Analysis Zone Development

Figure 3-3: TAZ Area Types in Genesee County

3-4

Chapter 3

A numbering scheme was established that would be helpful throughout the project. The numbering scheme is sequential, running from 1 to 676. The numbers begin in downtown Flint, continue to the City of Flint, and then cover Flint Township. After Flint is handled, all the townships are numbered from northwest to southeast, skipping the cities. Then all the cities within Genesee County are numbered. Finally, the 37 externals received a number. Table 3-1 summarizes the numbering scheme and range of numbers applied to each type.

Traffic Analysis Zone Development

Table 3-1: Genesee County TAZ Numbering Scheme City or Township

Begin TAZ #

End TAZ #

# of Zones

% of Zones

Flint CBD

Flint Non-CBD

191

151

22%

Flint Township

192

242

Montrose Township

243

249

Vienna Township

250

271

Thetford Township

272

283

Forest Township

284

293

Richfield Township

294

308

Genesee Township

309

334

Mt. Morris Township

335

359

Flushing Township

360

370

Clayton Township

371

380

Household Model

Burton City

381

416

The previous Genesee County Travel Demand Model (GCTDM) used simple linear regression based on Census for Transportation Planning Package (CTPP) data to estimate the percentage of households in each TAZ in each market segment as listed below. A true joint distribution was not developed.

Davison Township

417

434

Atlas Township

435

444

Grand Blanc Township

445

489

Mundy Township

490

517

Gaines Township

518

530

Argentine Township

531

537

Fenton Township

538

555

Cities

556

639

12%

Externals

640

676

100%

As part of the model update effort, Corradino developed a new household joint distribution model using a more advanced logit model (equation on next page) based on the 2014 American Community Survey (ACS) and Public Use Microdata Sample (PUMS) data for year 2014. The new household model uses the same market segmentation as the old model. This minimized changes to the Trip Generation procedures. Therefore, four different sub-models were developed:

Total

• Vehicles per Household – 0, 1, 2, and 3+ Vehicles per Household; and, • Household Income – Low and High. For the new model, households with annual income less than $48,000 are categorized as the lower income group, and the rest are in the high income category. This new cut-off value is obtained from median value of 2014 PUMS data sample for Genesee county area and is reasonably higher than the value in the previous model, which was $42,500.

• Household Size – 1, 2, 3, and 4+ Persons; • Household Workers – 0, 1, 2, and 3+ Workers per Household;

3-5

Chapter 3

Traffic Analysis Zone Development

The independent variables, similar to the previous model, include:

• Vehicles Model – Average Number of Vehicles per Household; and, • Income Model – Zonal Average Income/Regional Average Income.

• Household Size Model – Average Household Size; • Workers Model – Average Number of Workers per Household;

The model coefficients and constants are reported in Tables 3-2, 3-3, 3-4 and 3-5.

Table 3-5: Household Size Model Coefficients and Constants Independent

Dependent HH1 HH2 HH3 HH4

Average Household Size -1.287 -0.021 0.5818 1.176

Average Vehicles per Households 0.2593 0.2946 0.2077 0.2796

Constant 0 -2.740367 -4.882311 -6.332183

Table 3-5: Workers Model Coefficients and Constants Independent

Dependent Wrk0 Wrk1 Wrk2 Wrk3

Average Workers per Household -3.0367 -1.5632 -0.39513 -0.195227

Average Vehicles per Households -0.711648 -0.664849 -0.775332 -0.581597

Constant 3.041797 1.537275 0 -2.217346

Table 3-4: Vehicles Model Coefficients and Constants Independent

Dependent Veh0 Veh1 Veh2 Veh3

Average Household Size 0.575214 0.575214 0.575214 0.575214

Average Vehicles per Households 0.002095 1.528581 2.528215 4.01488

Constant 5.889929 5.113672 3.434444 0

Table 3-5: Income Model Coefficients and Constants Independent Dependent LowInc HighInc

Ratio of Zonal Average Income to Regional Average Income -5.027904 -3.57515

3-6

Average Vehicles per Households

Constant

-94.237593 -93.704798

0 -2.447697

Chapter 3

Traffic Analysis Zone Development

maximize the log likelihood. This probabilistic method allows the stratification model to be sensitive to more than just one zonal variable, thereby better predicting the household classification.

As noted earlier, the stratification curves were developed using the disaggregate-level ACS PUMS data, supplemented by TAZ level Census data. Each household record from the survey was geocoded to its TAZ, and the TAZ attributes for average household size, workers per household, autos per household, and average income were added to the household record. The proportions of households in a given category–such as household size, auto ownership, number of workers, or income range–were estimated using the following logit equation:

Pcat =

Figure 3-4, on the next page, shows examples of fitted curves for each of the sub-models. Based on the new model specifications, the GISDK code for Genesee County TDM was updated and the household model parameters were saved in a .bin file. The new model was also tested to make sure that the expected results are generated. To check for any possible errors, different market shares were compared between the old and the new models, as well as the 2014 ACS data. The result of this comparison is reported in Table 3-6. This table shows that the new model has better estimates for all the household sub-models. Note that the market share comparison is based on the household sample data in the region and, also, the comparison is for the entire Genesee county, not at zone-level.

eucat1

∑e

x (u ) catx

U (cat1) = C1 + C2 × (AVGhhv1) + C3 × (AVGhhv2) Pcat = Proportion of households in category x (1 through 4) U(cat1) = Utility score for category 1 C1 = Constant

Table 3-6: Comparison of Market Shares between the Old and New Model

C2 = Coefficient for the household variable 1 C3 = Coefficient for the household variable 2

Sub-model

AVGhhv1 = TAZ average household variable 1 AVGhhv2 = TAZ average household variable 2

Household Size

Coefficients were estimated using the Maximum Likelihood Method, which maximizes the sum of the log likelihood computations for each observation. The log likelihood is computed by applying the log of the modeled proportion for a given category to a binary variable that indicates the membership status of an individual household for that variable (e.g., if the household is a member of income quartile 2, then INC2 =1). The procedure simultaneously minimized the region-wide aggregate error for each category (e.g., correct number of households with zero autos, etc.) taken directly from Census data. The calibration process finds the set of logit coefficients that

Workers

Vehicles

Income

3-7

Category

Market Share Percentage – Old Model

Market Share Percentage – New Model

2014 ACS

1 person

26.0

30.5

29.8

2 persons

33.2

33.8

34.0

3 persons

17.1

15.7

15.9

4+ persons

23.7

20.0

20.3

0 worker

29.3

39.2

1 worker

37.9

36.5

2 workers

27.0

20.5

3+ workers

5.8

3.8

0 vehicle

7.9

9.2

9.1

1 vehicle

33.5

38.2

2 vehicles

40.2

36.0

36.1

3+ vehicles

18.3

16.6

Low Income

49.7

52.9

53.2

High Income

50.3

47.1

46.8

Chapter 3

Figure 3-4: Household Stratification Models

Traffic Analysis Zone Development

3-8

Chapter 3

Traffic Analysis Zone Development

External Stations

Table 3-7: Example of Growth Factors

As mentioned earlier in this chapter, there are 37 external stations in the Genesee County Travel Demand Model network. The auto and truck trip growth factors for these stations as well as other discussions about external trips will be addressed in detail in Chapter 4: External Trips.

Land Use Allocation Model Corradino has developed a land use growth allocation model to allow the staff and consultant to quickly develop alternative future year scenarios. The procedure assumes that a TAZ file exists which contains zonal data for a base year and for at least one future year. The procedure can be used to create another future year scenario, where the growth is proportional to the specified base and future year scenario growth, but sums to a user-specified control total. The model is structured to allow the modeler to allocate growth separately by Minor Civil Division (MCD), as defined in the TAZ attribute LUGCODE.

Percent Change

POP

MAN_EMP

OTH_EMP

TRAN_EMP

FIN_EMP

RET_EMP

WHOL_EMP

SERV_EMP

GOV_EMP

specified that the new data fields should be 12% larger than the original data, except for population and manufacturing employment, which will be 13% larger. The modeler must also specify the base and future year alternative designation (like 14_, and 45_), and the designation of the new fields (like 45a_). Note that any combination of growth factors can be used. Also, decreases could be specified (e.g., “-5” means 5% decrease). Also, please note that this shows the control file for only one LUGCODE. There will be a separate column for each MCD.

The procedure operates on the following TAZ data: • • • • • • • • • •

Attribute

Households (HH); Population (POP); Manufacturing employment (MAN_EMP); Other employment (OTH_EMP); Transportation employment (TRAN_EMP); Financial employment (FIN_EMP); Retail employment (RET_EMP); Wholesale employment (WHOL_EMP); Service employment (SERV_EMP); and, Government employment (GOV_EMP).

5D Smart-Growth Planning Tool A customized 5-Dimension (5D) post-processing tool has been developed for the Genesee County Metropolitan Planning Committee (GCMPC). This tool can be used to facilitate the smart-growth oriented planning activities, based on available data in the Genesee County travel demand model. The following section describes the methodology of developing 5D variables and urban design score.

To use the model, the modeler must supply an input model .bin file with the desired change from the original base year file for the attributes listed above. The model will then reallocate the growth to the TAZs in proportion to the original growth, while ensuring that the regional total meets the growth specification, and will add the required fields to the TAZ data table. In the example Table 3-7, the modeler has

5D Variables It has been a popular research topic to identify urban design elements that contribute to smart growth in metropolitan areas that advocate compact, transit-

3-9

Chapter 3

Traffic Analysis Zone Development

d3 = land use diversity.

oriented, walkable, bicycle-friendly land use. Previous research efforts suggested a â&#x20AC;&#x153;5Dâ&#x20AC;? concept that is relevant to reduced reliance on auto travel. These five key factors are often referred to:

Design Variables Design variables describe aspects of the urban network. These measures describe the degree to which the urban network is interconnected, grid-like, and more conducive or inviting to walking/bicycling. The 5D tool incorporates three design variables described as below.

Density â&#x20AC;&#x201C; dwellings or jobs per square mile; Diversity â&#x20AC;&#x201C; mix of land uses in an area; Design of the urban environment; Destinations â&#x20AC;&#x201C; proximity to regional activity centers; and, â&#x20AC;˘ Distance to Transit stations and services.

â&#x20AC;˘ â&#x20AC;˘ â&#x20AC;˘ â&#x20AC;˘

Walkability The walkability variable (d4) is defined as the percentage of streets within a TAZ that are walkable. â&#x20AC;&#x153;Walkableâ&#x20AC;? links are typically a selection set of low functional class, low speed, low volume roads. For this specific project, walkable links include the following facilities:

Density Variables Density variables are used to measure the intensity of activity within a certain geographic space. Areas with higher levels of density and intensity are likely to make vehicular travel more expensive (time and parking cost) and more conducive to transit or nonmotorized travel. Typical variables used to measure this quality of an area are household density (HH/mile2) and employment density (Emp/mile2) which are both readily computed for a given TAZ.

â&#x20AC;˘ Minor arterials, collectors, and local roads with posted speed <= 25 mph; â&#x20AC;˘ Bike facilities (i.e., network â&#x20AC;&#x153;BikeFacil_13 >= 1); and,

Diversity Variables

â&#x20AC;˘ Pedestrian facilities (i.e., network attribute â&#x20AC;&#x153;PedFacil_13 >= 1).

Diversity variables measure the degree to which land uses are segregated. Urban design elements which promote the mixing of residential and employment are known to contribute to shorter and potentially fewer vehicular trips. The level of diversity is often measured using a jobs/housing ratio. In places where there is a large degree of land use segregation, the ratio is either very low or very high. For the Genesee County region, the diversity variable (d3) is expressed by the equation below, which is simple to compute for a given TAZ using available model data.

where:

đ?&#x2018;&#x2018;đ?&#x2018;&#x2018;3 = 1 â&#x2C6;&#x2019; ďż˝

attribute

Then a ratio is computed using walkable link distances versus the sum for all links for a given TAZ. Blockface Blockface is a geometric measure of the average blockface size within a TAZ. Average blockface is a good measure of how grid-like the street network is. A tight urban street grid pattern typically yields low blockface values, while a more open and less connected street pattern has much higher blockface values. The more connected the network, the more efficient walk or bike trips could be. This same arrangement has the opposite effect on vehicular travel by adding intersection delays, so it serves as a deterrent to auto travel. For a given TAZ, the blockface variable (d5) is expressed by equation below:

đ?&#x2018;&#x2018;đ?&#x2018;&#x2018;1 â&#x2C6;&#x2019; đ?&#x2018;&#x2018;đ?&#x2018;&#x2018;2 ďż˝ đ?&#x2018;&#x2018;đ?&#x2018;&#x2018;1 + đ?&#x2018;&#x2018;đ?&#x2018;&#x2018;2

d1 = population per square mile; d2 = employment per square mile; and,

3-10

Chapter 3

đ?&#x2018;&#x2018;đ?&#x2018;&#x2018;5 = 1 â&#x2C6;&#x2019;

Street Density

Traffic Analysis Zone Development

đ?&#x2018;&#x2026;đ?&#x2018;&#x2026;đ?&#x2018;&#x2026;đ?&#x2018;&#x2026;đ?&#x2018;&#x2026;đ?&#x2018;&#x2026;đ?&#x2018;&#x2026;đ?&#x2018;&#x2026;đ?&#x2018;&#x2026;đ?&#x2018;&#x2026;đ?&#x2018;&#x2026;đ?&#x2018;&#x2026;đ?&#x2018;&#x2026;đ?&#x2018;&#x2026; đ??śđ??śđ??śđ??śđ??śđ??śđ??śđ??śđ??śđ??śđ??śđ??śđ??śđ??śđ??śđ??śđ??śđ??śđ??śđ??ś đ?&#x2018;&#x20AC;đ?&#x2018;&#x20AC;đ?&#x2018;&#x20AC;đ?&#x2018;&#x20AC;đ?&#x2018;&#x20AC;đ?&#x2018;&#x20AC;đ?&#x2018;&#x20AC;đ?&#x2018;&#x20AC;đ?&#x2018;&#x20AC;đ?&#x2018;&#x20AC;đ?&#x2018;&#x20AC;đ?&#x2018;&#x20AC;đ?&#x2018;&#x20AC;đ?&#x2018;&#x20AC; đ?&#x2018; đ?&#x2018; đ?&#x2018; đ?&#x2018; đ?&#x2018; đ?&#x2018; đ?&#x2018; đ?&#x2018; đ?&#x2018; đ?&#x2018; đ?&#x2018; đ?&#x2018; đ?&#x2018;&#x153;đ?&#x2018;&#x153;đ?&#x2018;&#x153;đ?&#x2018;&#x153; đ??żđ??żđ??żđ??żđ??żđ??żđ??żđ??żđ??żđ??ż

good the transit service is around a TAZ. This is computed by calculating the transit stop density within a 5-mile radius of the TAZ. It is intended to be used as a simple indicator of what other locations can be accessed via transit. The underlying assumption is that transit can be a competitive substitute for auto travel with increasing levels of accessibility and coverage.

Street density (d6) is another geometric measure that is simply the centerline miles of streets within a given TAZ divided by the land area of the TAZ in square miles. The street density variable complements the other two design variables.

Urban Design Score A comprehensive urban design score can be derived for a given TAZ based on the 5D variables described above. In this study, the urban design score is calculated by the equation below.

Destination Variables Destination variables describe the level of regional vibrancy. Mixed land used patterns are frequently observed in prosperous areas where many trip purposes (e.g. work, shopping, or entertainment) can usually be accomplished without auto trips. The variable must be sensitive to the types of land use that are close enough for a non-motorized trip to be more likely chosen over an auto trip. In this project, destinations are measured using two variables:

5đ??ˇđ??ˇ đ?&#x2018;&#x2020;đ?&#x2018;&#x2020;đ?&#x2018;&#x2020;đ?&#x2018;&#x2020;đ?&#x2018;&#x2020;đ?&#x2018;&#x2020;đ?&#x2018;&#x2020;đ?&#x2018;&#x2020;đ?&#x2018;&#x2020;đ?&#x2018;&#x2020; =

where,

đ?&#x2018;&#x2018;đ?&#x2018;&#x2018;1 + đ?&#x2018;&#x2018;đ?&#x2018;&#x2018;2 + đ?&#x2018;&#x2018;đ?&#x2018;&#x2018;3 + đ?&#x2018;&#x2018;đ?&#x2018;&#x2018;4 + đ?&#x2018;&#x2018;đ?&#x2018;&#x2018;5 + đ?&#x2018;&#x2018;đ?&#x2018;&#x2018;6 + đ?&#x2018;&#x2018;đ?&#x2018;&#x2018;7 + đ?&#x2018;&#x2018;đ?&#x2018;&#x2018;8 + đ?&#x2018;&#x2018;đ?&#x2018;&#x2018;9 + đ?&#x2018;&#x2018;đ?&#x2018;&#x2018;10 8

d1 = population density; d2 = employment density; d3 = diversity variable;

â&#x20AC;˘ d7 â&#x20AC;&#x201C; number of service jobs within a 10-minute walk (about 1/6 mile); and,

d4 = walkability;

â&#x20AC;˘ d8 â&#x20AC;&#x201C; number of retail jobs within a 10-minute walk (about 1/6 mile).

d5 = average blockface; d6 = average street density;

Both variables are ways of describing the vibrancy of an area. Previous studies indicate the threshold of 10-minute walking distance (about 1/6 mile) allows for a more realistic differentiation among the TAZs.

d7 = number of service jobs within 10-minute walking distance; d8 = number of retail jobs within 10-minute walking distance;

Distance to Transit Variables

d9 = walk access to transit; and,

Distance to transit variables describe the degree to which the area is served by transit. The 5D tool incorporates two variables relevant to transit service. The first variable (d9) evaluates how easy to access transit service by walking from a TAZ. This can be measured by a ratio of number of transit stops within a halfmile radius to the total centerline mileage of the TAZ. The easier it is to walk to transit service, the more likely it is that a trip will be made by transit instead of by auto. The second variable (d10) evaluates how

d10 = accessibility via transit. The more likely a TAZ follows principles of smart growth, the higher its urban design score is. The urban design score can facilitate the scenario-based planning process to evaluate various strategies of smart growth. For example, area size, population, and employment covered by smart growth zones (urban design score > 0.8) in a region can be compared

3-11

Chapter 3

Traffic Analysis Zone Development

between various scenarios in terms of land use development, roadway network design, availability of bike/pedestrian facilities, transit service, etc.

Upon the completion of the 5D processor, the TAZ file is updated by automatically adding and calculating the following fields for each TAZ (Figure 3-8).

Table 3-8: New Urban Design Score Fields in the TAZ File Field Name

Description

Dens1

Population density per square mile

Dens2

Employment density per square mile

Diver1

Diversity variable

Design1

Walkability variable

Design2

Average blockface

Design3

Street density

Dest1

Service jobs within 10-min walking distance

Dest2

Retail jobs within 10-min walking distance

Dist1

Walk access to transit

Dist2

Accessibility via transit

Link mileage

LNK

Number of links

WLK

Walkable link mileage

NRSTP Score5D

Number of transit stop within a half-mile radius of a TAZ 5D urban design score

3-12

Chapter 4

External Trips

Chapter 4: External Trips Introduction and Overview methodologies explained later in this chapter. Then these external trip tables were adjusted to match the base year traffic counts at all external stations. The Michigan statewide model covers only major roads in Genesee county. The general assumption of this method is the external-external trips from one external station to another are significant only if there are trips between these two locations in the Michigan statewide model. Thus, all external trips on roadways that are not in the Michigan statewide model network are assumed to be EI or IE trips.

This memo deals with trips that have one or both ends outside Genesee County. Each trip has two ends, one is the origin and the other is the destination. Trips with one end in the study area are called External-Internal (EI) or Internal-External (IE) trips while trips with no ends in the study area, but passing through the area, are called through or ExternalExternal (EE) trips. The external trips include both EI & IE (EI-IE) trips and EE trips. The end point on the roadway outside the study area or on the roadway where the study area bound line is crossed is referred to as an external station. As shown in Figure 4-1, the Genesee travel demand model has 37 external stations.

MDOT staff also created a preliminary 2045 external passenger vehicle trip table using a similar process.

Base Year External Station Summary

Truck and auto vehicle classification is taken into account in the external trip estimation. A commercial vehicle with six tires or above belongs to the truck class (FHWA Classifications 5 to 13) while a motorcycle, a passenger car or a commercial vehicle with four tires (FHWA Classifications 1 to 4) belongs to the auto vehicle class.

Detailed information on the 37 external stations shown in Figure 4-1 is given in Table 4-1. It includes the name, location, functional class, daily traffic count, daily truck count and truck percent at each external station. Among those 37 external stations, thirty are in rural areas and seven are in urban areas. Six major external stations are on interstate, expressway and principal arterial, and ID numbers are 643, 655, 658, 660, 665 and 670. The Average Daily Traffic (ADT) counts vary from 284 to 60,702 and the ADT truck counts vary from 10 to 4,698. The highest truck percentage is 15.5%, the lowest truck percentage is 3.4% and the average truck percentage is 6.1% for all external stations.

The base year of the Genesee County Travel demand model is 2014. Michigan DOT staff used the TransCAD subarea extraction method in the Michigan Statewide Travel Demand Model to generate a preliminary external trip table for Genesee County. Only one external trip table including passenger vehicle trips was generated. This external trip table was later divided into auto trips and truck trips based on the

4-1

Chapter 4

External Trips

Figure 4-1: External Station Locations and IDs

4-2

Chapter 4

External Trips

Table 4-1: External Station Summary External Station TAZ ID

Name

Location

Functional Class

ADT Traffic Count

ADT Truck Count

Truck Percent

640

Sheridan Ave

North of Study Area

Rural Minor Arterial

4,590

240

5.23%

641

Nichols Rd

North of Study Area

Rural Major Collector

642

7.79%

642

Elms Rd

North of Study Area

Rural Major Collector

1,382

3.47%

643

I 75 North

North of Study Area

Rural Interstate

60,702

4,698

7.74%

644

Saginaw Rd

North of Study Area

Rural Minor Arterial

5,182

266

5.13%

645

Clio Rd

North of Study Area

Urban Minor Arterial

2,516

150

5.96%

646

Bray Rd

North of Study Area

Rural Major Collector

2,500

112

4.48%

647

Irish Rd

North of Study Area

Rural Major Collector

910

7.03%

648

State Rd

North of Study Area

Rural Minor Arterial

7,070

362

5.12%

649

Henderson Rd

North of Study Area

Rural Major Collector

708

5.65%

650

Lake Rd

East of Study Area

Rural Minor Arterial

3,500

193

5.51%

651

Columbiaville Rd

East of Study Area

Rural Major Collector

3,098

130

4.20%

652

E Mount Morris R

East of Study Area

Rural Major Collector

284

3.52%

653

Davison Rd

East of Study Area

Rural Minor Arterial

6,816

312

4.58%

654

Lapeer Rd

East of Study Area

Rural Major Collector

2,116

4.16%

655

I 69 East

East of Study Area

Rural Interstate

33,100

2,143

6.47%

656

Hill Rd

East of Study Area

Rural Major Collector

2,334

103

4.41%

657

Hegel Rd

East of Study Area

Rural Major Collector

1,493

3.68%

658

Ortonville RD

South of Study Area

Rural Principal Arterial

13,400

565

4.22%

659

Dixie Hwy

South of Study Area

Urban Minor Arterial

12,637

579

4.58%

660

I 75 South

South of Study Area

Rural Interstate

43,448

2,676

6.16%

661

N Holly Rd

South of Study Area

Rural Minor Arterial

10,078

405

4.02%

662

Main St

South of Study Area

Rural Minor Arterial

13,219

706

5.34%

663

S Holly Rd

South of Study Area

Urban Collector

5,828

300

5.15%

664

Adelaide St

SW of Study Area

Urban Collector

665

S US 23

SW of Study Area

Urban Expressway

666

Linden Rd

South of Study Area

667

Seymour Rd

SW of Study Area

668

Silver Lake Rd

West of Study Area

Rural Minor Arterial

3,628

156

4.30%

669

Lansing Rd

West of Study Area

Rural Minor Arterial

4,815

246

5.11%

670

I 69 West

West of Study Area

Rural Interstate

24,496

1,848

7.54%

671

M 21

West of Study Area

Rural Minor Arterial

7,484

490

6.55%

672

Pierson Rd

West of Study Area

Rural Major Collector

1,386

4.62%

673

W Mount Morris R

West of Study Area

Rural Major Collector

800

4.63%

674

Vienna Rd

West of Study Area

Rural Minor Arterial

4,900

266

5.43%

675

Grand Blanc

West of Study Area

Rural Major Collector

2,506

130

5.19%

676

Thompson Rd

East of Study Area

Urban Minor Arterial

554

15.52%

350,425

21,330

6.09%

3,300

112

3.39%

47,676

3,014

6.32%

Urban Collector

7,438

426

5.73%

Rural Minor Arterial

3,889

160

4.11%

Total

4-3

Chapter 4

External Trips

Base Year External Trip Estimation

The TransCAD MMA subarea function was used to create a subarea truck trip table for each trip purpose mentioned above. A preliminary truck trip table was created from auto trips: 20% NHBO and 10% NHBWB. The preliminary auto trip table is equal to the difference between all-vehicle trip table and the truck trip table. The inbound and outbound traffic volumes were unbalanced and therefore, matrix balancing was used to produce symmetrical inbound and outbound volumes.

The base year external trip estimation has three steps: subarea analysis, trip table adjustment, and EIIE trip calculation by trip purpose. Genesee County was selected as the subarea in the Michigan statewide travel demand model, and the subarea analysis generated preliminary auto external trip table. Finally, the EI-IE trips were calculated by three trip purposes: non-work, EI, and IE work purposes. This calculation used the split ratios obtained from the “MI Travel Counts” household travel survey and the Census Transportation Planning Package (CTPP). The following section introduces the Michigan statewide travel demand model.

Trip Table Adjustment The preliminary external trip tables from the subarea analysis must be adjusted to match the Average Daily Traffic (ADT) counts. The following steps for adjusting the external trip table were followed,

Michigan Statewide Travel Demand Model The statewide model does not separately identify autos, trucks, and transit trips. The trip tables include the following trip purposes: Home Based Work Business (HBWB), Home Based Social Recreation (HBSR), Home Based Other (HBO), Non-Home Based Work Business (NHBWB), and Non-Home Based Other (NHBO). In the statewide model, a trip table (matrix) of all trips, all purposes combined, is created for the trip assignment. Trips are assigned using a user equilibrium traffic assignment method.

1) Estimate daily Origin (O) and Destination (D) trips by autos and trucks at each external station from ADT counts; 2) Generate symmetric trip matrices (tables) if these matrices are asymmetric; 3) Split EI-IE and EE trips and compute the final EE O-D matrices for autos and trucks; and, 4) Calculate EI-IE O and D trips by autos and trucks.

The subarea extraction method uses the statewide network and trip tables to create files just for Genesee County.

The final EE O-D matrix computation in Step 3, a key part in the adjustment, was done using the Fratar model.

Subarea Analysis Genesee County is defined as a subarea in the Michigan statewide model. The trips in the statewide model enter and leave Genesee County through 29 gates. Two external stations on Grand Blanc Road (west of the study area) and Flint Street (east of study area) are not present in the Genesee county model. Eight external stations in the Genesee model are on lower-functional class roadways and do not exist in the statewide model. It is assumed that there are no external-external trips for these eight external stations.

Fratar Model The doubly-constrained growth factor method, also known as the Fratar model, creates a trip matrix for which row and column totals match a specified set of values—the expected number of EE trip ends at each external station. The goal is to solve the following equation:

4-4

Chapter 4

External Trips

Tables 4-2 and 4-3 show 2014 ADT counts, 2014 statewide model volumes, 2014 final EE O & D, and 2014 final IE/EI (O+D) for autos and trucks trips. Note that the total auto ADT counts are split among “EE O”, “EE D” and IE/EI counts. For instance, for external station 640, total auto ADT counts (4,350) is divided equally by 2 (O and D columns) and then split between EE and IE/EI counts. Tables 4-4 and 4-5 display the EE O-D trip tables for autos and truck.

(1) where: Tij = Output trips from zone i to zone j tij = Original trips from zone i to zone j

External–Internal and Internal-External (EI-IE) Trips by Trip Purpose

oi = Balancing factor for row

Genesee County not only serves as a bedroom community to several neighboring counties, but also attracts trips into the region for other purposes, including working. For this reason, three separate EI-IE auto purposes were defined:

dj = Balancing factor for column Oi = Origin trips of zone i Dj = Destination trips of zone j

• EI_Work (EI_W). The EI_W trips represent the inbound commute to work and return from work made by residents outside of Genesee County. Trip Productions (P) are assigned at external stations as a percent of total volumes based on MI Travel Counts and CTPP JTW and trip attractions are estimated at internal zones as function of HBW attractions

The following steps will be used when applying the Fratar Model to adjust the external trips by vehicle class: i.

Split the EE and EI-IE trips in the subarea trip table gotten from the statewide model

ii. Factor the preliminary EE and EI-IE O & D trips to match the base year traffic counts at each external station

• IE_Work (IE_W). The IE_W trips represent the outbound work commute and return from work made by residents inside Genesee County. Trip productions estimated to internal TAZs as a function of HBW and trip attractions using MI Travel Count data and attractions are assigned to external stations as a percentage of total outbound traffic

iii. Balance the factored EE O & D trips by the Weighted Sum (50% O to 50% D) method in TransCAD. The balancing process makes total EE O trips equal to total EE D trips at all external stations

• External NonWork (E_NW). The E_NW trips represent other external trips that are not related to work. Trip productions are assigned at external stations as a percent of total volume, and trip attractions estimated at internal zones as function of HBO and HBSH attractions.

iv. Obtain the final EE O-D table by applying the balanced EE O & D to the preliminary EE O-D table using the Fratar model. This is the process to adjust the preliminary EE O-D matrix got from the statewide model to replicate the current local traffic conditions v. Obtain the EI-IE (O +D) as (O+D) minus EE (O+D).

4-5

Chapter 4

External Trips

Table 4-2: 2014 External Auto Trip Estimation 2014 Auto External Trip Results

Name

2014 Auto ADT Count

2014 SW Model Auto Volume

640

Sheridan Ave

4,350

641

Nichols Rd

592

642

Elms Rd

1,334

667

1,334

643

I 75 North

56,004

55,690

28,002

20,608

14,788

644

Saginaw Rd

4,916

8,870

2,458

198

4,520

645

Clio Rd

2,366

2,384

1,183

2,190

646

Bray Rd

2,388

5,406

1,194

227

1,934

647

Irish Rd

846

423

846

648

State Rd

6,708

6,738

3,354

1,542

3,624

649

Henderson Rd

668

334

668

650

Lake Rd

3,307

3,882

1,654

783

1,742

651

Columbiaville Rd

2,968

2,962

1,484

446

2,076

652

E Mount Morris R

274

276

137

264

653

Davison Rd

6,504

5,556

3,252

6,444

654

Lapeer Rd

655

I 69 East

656

TAZ ID

4,306

2,175

844

IE/EI P+A 2,662

296

592

EE O

EE D

2,028

2,666

1,014

1,970

30,957

36,318

15,479

4,843

21,272

Hill Rd

2,231

5,678

1,116

2,056

657

Hegel Rd

1,438

3,228

719

1,284

658

Ortonville RD

12,835

15,004

6,418

192

12,452

659

Dixie Hwy

12,058

8,420

6,029

687

10,684

660

I 75 South

40,772

40,586

20,386

6,662

27,448

661

N Holly Rd

9,673

24,536

4,837

372

8,930

662

Main St

12,513

10,926

6,257

1,324

9,866

663

S Holly Rd

5,528

2,764

5,528

664

Adelaide St

3,188

12,386

1,594

3,070

665

S US 23

44,662

44,776

22,331

9,709

25,244

666

Linden Rd

7,012

3,506

7,012

667

Seymour Rd

3,729

8,088

1,865

442

2,846

668

Silver Lake Rd

3,472

5,430

1,736

710

2,052

669

Lansing Rd

4,569

536

2,285

4,570

670

I 69 West

22,648

22,542

11,324

5,833

10,982

671

M 21

6,994

7,082

3,497

888

5,218

672

Pierson Rd

1,322

661

1,322

673

W Mount Morris R

763

4,984

382

606

674

Vienna Rd

4,634

7,764

2,317

599

3,436

675

Grand Blanc

2,376

4,236

1,188

147

2,082

676

Thompson Rd

468

234

468

329,095

361,256

164,552

57,511

214,082

Total

4-6

Chapter 4

External Trips

Table 4-3: 2014 External Truck Trip Estimation 2014 Truck External Trip Results

Name

2014 Truck ADT Count

2014 SW Model Truck Volume

640

Sheridan Ave

240

120

IE/EI P+A 122

641

Nichols Rd

642

Elms Rd

643

I 75 North

4,698

4,670

2,349

1,654

1,390

644

Saginaw Rd

266

482

133

242

645

Clio Rd

150

154

136

646

Bray Rd

112

268

647

Irish Rd

648

State Rd

362

364

181

164

649

Henderson Rd

650

Lake Rd

193

228

651

Columbiaville Rd

130

652

E Mount Morris R

653

Davison Rd

312

256

156

310

654

Lapeer Rd

126

655

I 69 East

2,143

2,512

1,072

510

1,124

656

Hill Rd

103

264

657

Hegel Rd

122

658

Ortonville RD

565

662

283

542

659

Dixie Hwy

579

406

290

466

660

I 75 South

2,676

2,662

1,338

639

1,398

661

N Holly Rd

405

1,024

203

352

662

Main St

706

618

353

548

663

S Holly Rd

300

150

300

664

Adelaide St

665

S US 23

666

TAZ ID

EE O

EE D

112

436

108

3,014

3,026

1,507

936

1,142

Linden Rd

426

213

426

667

Seymour Rd

160

344

120

668

Silver Lake Rd

156

246

669

Lansing Rd

246

123

246

670

I 69 West

1,848

1,846

924

600

648

671

M 21

490

498

245

332

672

Pierson Rd

673

W Mount Morris R

246

674

Vienna Rd

266

448

133

174

675

Grand Blanc

130

234

110

676

Thompson Rd Total

21,330

22,552

10,669

4,987

11,364

4-7

Table 4-4: 2014 External-to-External Auto Trips

Chapter 4 External Trips

4-8

Table 4-5: 2014 External-to-External Truck Trips

Chapter 4 External Trips

4-9

Chapter 4

External Trips

Table 4-6 reports the external station locations, 2014 auto EI-IE trip productions and attractions, distribution percent by purpose, EI_Work trip productions, IE_Work trip attractions and External NonWork trip productions. The 2014 auto EI-IE trip productions and attractions are equal to the 2014 auto EI-IE trip origins and destinations in Table 4-2.

statewide model trip purpose. The preliminary external truck trip table was estimated by adding 20% NBHO and 10% NHBWB matrices together, and the preliminary auto external trip matrix is equal to the difference between all-vehicle trip matrix and the truck external trip matrix. There were adjustments to the statewide model results in the 2014 external trip estimation, and those adjustments were applied to the 2045 statewide model results as well.

Future Year External Trip Estimation The Michigan statewide travel demand model has two target years: 2014 and 2045. The 2045 preliminary external trip tables were generated by the subarea analysis in the 2045 Michigan statewide model. The total productions and attractions for 2045 preliminary external trip tables were balanced. Then the auto and truck annual growth rates for each external station were calculated from the total number of productions and attractions in the 2014 and 2045 external trip tables. The growth rate calculation method is explained later in this chapter. The annual growth rates were used to calculate the total external productions and attractions for future years, and the Fratar model was used to compute the future year external-external trip matrices for autos and trucks. The Michigan statewide model covers only 29 external stations of the Genesee county model. For the other eight external stations not in the statewide model, annual growth rates were assumed as explained in the next section of this memo.

The final 2045 external traffic volumes of autos and trucks are listed in Table 4-7. The trip growth calculation equation is given below: Vol2045=Vol2014*(1+r)(2045-2014)

(2)

where: Vol2045 is the 2045 external traffic volumes (total trip origins and destinations), Vol2014 is the 2014 external traffic volumes, and r is the annual growth rate. Table 4-7 shows the calculated growth rates of external stations for autos and trucks. These rates were calculated based on the equation below: r=exp{[Ln(Vol2045)-Ln(Vol2014)]/31}â&#x20AC;&#x201C;1

(3)

where: exp() is the function returning the value of the constant e raised to a power, Ln() is the function returning the natural logarithm of a number.

The Fratar approach for factoring EE trips was coded in the Genesee model and can be used to calculate the external trip table of any year between 2014 and 2045. Finally, the EI-IE trips were calculated by three trip purposes: external non-work, EI work and IE work.

If the absolute value of a growth rate was less than 0.1%, then a rate of 0.1% or -0.1% was used. Eight external stations that are not in the Michigan statewide model are on low functional classification roadways, such as collector and minor arterials. For these roads an annual growth rate of 0.5% was assumed for auto trips and 0.3% for truck trips.

Growth Rate Calculation An MMA subarea extraction was performed in TransCAD to get the external trip tables (matrix) for each

4-10

Chapter 4

External Trips

Table 4-6: 2014 EI-IE Trips by Purpose TAZ ID

2,662

EI_W P 21%

Percent IE_W A 21%

E_NW P 58%

592

20%

71%

2014 Auto IE/EI P+A

Name

EI_W Prod

IE_W Attr

E_NW Prod

552

558

1,552

118

420

640

Sheridan Ave

641

Nichols Rd

642

Elms Rd

1,334

20%

71%

267

120

947

643

I 75 North

14,788

21%

58%

3,136

8,516

644

Saginaw Rd

4,520

19%

62%

856

2,808

645

Clio Rd

2,190

23%

55%

497

495

1,199

646

Bray Rd

1,934

20%

61%

379

377

1,178

647

Irish Rd

846

20%

71%

169

601

648

State Rd

3,624

21%

58%

765

764

2,094

649

Henderson Rd

668

20%

71%

134

474

650

Lake Rd

1,742

22%

55%

391

390

961

651

Columbiaville Rd

2,076

16%

68%

336

338

1,401

652

E Mount Morris R

264

19%

62%

164

653

Davison Rd

6,444

18%

20%

63%

1,138

1,262

4,043

654

Lapeer Rd

1,970

19%

17%

65%

369

330

1,271

655

I 69 East

21,272

22%

21%

57%

4,608

4,550

12,114

656

Hill Rd

2,056

18%

64%

367

365

1,324

657

Hegel Rd

1,284

16%

67%

211

862

658

Ortonville RD

12,452

22%

56%

2,761

2,777

6,914

659

Dixie Hwy

10,684

13%

74%

1,377

1,362

7,944

660

I 75 South

27,448

31%

38%

8,467

8,468

10,513

661

N Holly Rd

8,930

15%

70%

1,343

1,344

6,244

662

Main St

9,866

21%

58%

2,085

2,086

5,695

663

S Holly Rd

5,528

10%

54%

36%

553

2,985

1,990

664

Adelaide St

3,070

12%

76%

372

2,326

665

S US 23

25,244

24%

53%

5,984

5,967

13,293

666

Linden Rd

7,012

10%

54%

36%

701

3,786

2,524

667

Seymour Rd

2,846

21%

59%

588

593

1,665

668

Silver Lake Rd

2,052

18%

64%

369

367

1,315

669

Lansing Rd

4,570

19%

16%

64%

890

751

2,929

670

I 69 West

10,982

25%

51%

2,719

2,714

5,549

671

M 21

5,218

27%

26%

47%

1,386

1,382

2,450

672

Pierson Rd

1,322

27%

11%

62%

357

145

820

673

W Mount Morris R

606

21%

58%

126

354

674

Vienna Rd

3,436

18%

64%

621

2,194

675

Grand Blanc

2,082

21%

58%

434

1,213

676

Thompson Rd

468

20%

71%

Total

214,082

4-11

332

45,570

50,316

118,195

Chapter 4

External Trips

Table 4-7: Annual Growth Rate of Auto and Truck Trips TAZ ID

Name

Rate Type

640

Sheridan Ave

Calculated

641

Nichols Rd

Assumed

642

Elms Rd

Assumed

643

I 75 North

Assumed

644

Saginaw Rd

Assumed

645

Clio Rd

Calculated

646

Bray Rd

Calculated

647

Irish Rd

Assumed

648

State Rd

Calculated

649

Henderson Rd

Assumed

650

Lake Rd

Calculated

651

Columbiaville Rd

Calculated

652

E Mount Morris R

Assumed

653

Davison Rd

Calculated

654

Lapeer Rd

Calculated

655

I 69 East

Calculated

656

Hill Rd

Calculated

657

Hegel Rd

Assumed

658

Ortonville RD

Assumed

659

Dixie Hwy

Calculated

660

I 75 South

Assumed

661

N Holly Rd

Assumed

662

Main St

Calculated

663

S Holly Rd

Assumed

664

Adelaide St

Calculated

665

S US 23

Calculated

666

Linden Rd

Assumed

667

Seymour Rd

Assumed

668

Silver Lake Rd

Assumed

669

Lansing Rd

Calculated

670

I 69 West

Calculated

671

M 21

Assumed

672

Pierson Rd

Assumed

673

W Mount Morris R

Assumed

674

Vienna Rd

Calculated

675

Grand Blanc

Calculated

676

Thompson Rd

Assumed

Annual Auto Growth Rate

-0.3% 0.5% 0.5% 0.1% 0.1% 0.2% 0.2% 0.5% 0.2% 0.5% 0.2% 0.2% -0.1% 0.2% 0.1% 0.2% 0.2% 0.1% 0.1% 0.1% -0.1% 0.1% 0.3% 0.5% 0.6% 0.1% 0.5% 0.1% 0.1% 0.5% 0.1% 0.1% 0.5% -0.1% -0.2% -0.2% 0.5%

4-12

Rate Type Calculated Assumed Assumed Calculated Assumed Assumed Assumed Assumed Assumed Assumed Assumed Assumed Assumed Assumed Assumed Calculated Assumed Assumed Calculated Calculated Assumed Calculated Calculated Assumed Assumed Calculated Assumed Assumed Assumed Assumed Calculated Assumed Assumed Assumed Assumed Calculated Assumed

Annual Truck Growth Rate

-0.2% 0.5% 0.5% 0.1% -0.1% 0.1% 0.1% 0.5% 0.1% 0.5% 0.1% -0.1% 0.5% -0.1% 0.1% 0.3% 0.1% -0.1% -0.3% 0.1% 0.1% 0.1% 0.2% 0.5% -0.1% 0.1% 0.5% 0.1% 0.1% 0.5% 0.2% 0.1% 0.5% 0.1% 0.1% -0.3% 0.5%

Chapter 4

External Trips

2045 External Trip Estimation of the 2045 external trip estimation. The 2014 external trip numbers were added to the table to provide a comparison. Overall, the number of truck trips increased 4.3% from 2014 to 2045 while the number of auto trips increased 4.2%.

Based on the annual growth rates in Table 4-7 and the 2014 external trips in Table 4-2, the number of 2045 External-Internal and Internal-external (EI-IE) trips can be calculated using Equation (2). Then 2045 EI-IE trips can be further disaggregated into the 2045 External-Internal Work (EI_W) trip productions, Internal-External Work (IE_W) trip attractions and External Non-Work (E_NW) trip productions using the distribution percentages in Table 4-6.

The Fratar model was used to generate the 2045 external-external trip tables. The inputs to the Fratar model are the 2014 external-external trip tables and the 2045 external-external trip origins and destinations by external station. Table 4-9 displays the 2045 external-external auto trip matrix and Table 4-10 displays the 2045 external-external truck trip matrix.

The number of 2045 external-external (EE) trips can be calculated using Equation (2) in a similar way. For each vehicle class, the 2045 external trips are the EE trips plus the EI-IE trips. Table 4-8 reports the results

4-13

Chapter 4

External Trips

Table 4-8: 2045 External Trips TAZ ID 640

Name Sheridan Ave

2014 Truck ADT Count 240

2045 Truck O+D 228

Increase (%) -5.1%

2014 Auto ADT Count 4,350

2045 Auto O+D 3,979

Increase (%)

2045 Auto IE/EI P+A

-8.5%

2,435

2045 EI_W Prod 505

2045 IE_W Attr

2045 E_NW Prod

511

1,420

641

Nichols Rd

16.7%

592

691

16.7%

691

138

491

642

Elms Rd

16.7%

1,334

1,557

16.7%

1,557

311

140

1,106

643

I 75 North

4,698

4,846

3.2%

56,004

57,766

3.1%

15,291

3,243

3,242

8,806

644

Saginaw Rd

266

258

-3.1%

4,916

5,071

3.1%

4,664

884

883

2,897

645

Clio Rd

150

155

3.1%

2,366

2,500

5.7%

2,315

525

523

1,267

646

Bray Rd

112

116

3.1%

2,388

2,514

5.3%

2,037

399

397

1,240

647

Irish Rd

16.7%

846

987

16.7%

987

197

701

648

State Rd

362

373

3.1%

6,708

7,151

6.6%

3,871

818

816

2,237

649

Henderson Rd

16.7%

668

780

16.7%

780

156

554

650

193

199

3.1%

3,307

3,536

6.9%

1,858

417

416

1,025

130

126

-3.1%

2,968

3,148

6.1%

2,200

357

358

1,485

16.7%

274

266

-3.1%

256

159

653

Lake Rd Columbiaville Rd E Mount Morris R Davison Rd

312

302

-3.1%

6,504

6,939

6.7%

6,874

1,215

1,346

4,314

654

Lapeer Rd

3.1%

2,028

2,098

3.4%

2,038

382

341

1,315

655

I 69 East

2,143

2,332

8.8%

30,957

33,164

7.1%

22,782

4,936

4,873

12,974

656

Hill Rd

103

106

3.1%

2,231

2,383

6.8%

2,195

391

390

1,413

657

Hegel Rd

-3.1%

1,438

1,483

3.1%

1,325

218

889

658

Ortonville RD

565

518

-8.3%

12,835

13,239

3.1%

12,842

2,848

2,864

7,130

659

Dixie Hwy

579

599

3.5%

12,058

12,568

4.2%

11,135

1,436

1,420

8,279

660

I 75 South

2,676

2,760

3.1%

40,772

39,526

-3.1%

26,598

8,205

8,206

10,188

661

N Holly Rd

405

420

3.7%

9,673

9,977

3.1%

9,211

1,385

1,386

6,440

662

Main St

706

751

6.4%

12,513

13,561

8.4%

10,695

2,260

2,262

6,174

663

S Holly Rd

300

350

16.7%

5,528

6,452

16.7%

6,452

645

3,484

2,323

664

Adelaide St

665

S US 23

666

651 652

112

109

-3.1%

3,188

3,891

22.0%

3,747

454

2,838

3,014

3,133

3.9%

44,662

46,538

4.2%

26,284

6,230

6,213

13,841

Linden Rd

426

497

16.7%

7,012

8,184

16.7%

8,184

818

4,420

2,946

667

Seymour Rd

160

165

3.1%

3,729

3,846

3.1%

2,939

607

612

1,720

668

Silver Lake Rd

156

161

3.1%

3,472

3,581

3.1%

2,113

380

378

1,355

669

Lansing Rd

246

287

16.7%

4569

5,333

16.7%

5,333

1,038

876

3,419

670

I 69 West

1,848

1,959

6.0%

22,648

23,488

3.7%

11,384

2,819

2,813

5,752

671

M 21

490

505

3.1%

6,994

7,214

3.1%

5,385

1,430

1,426

2,528

672

16.7%

1,322

1,543

16.7%

1,543

417

170

957

3.1%

763

734

-3.8%

582

121

340

674

Pierson Rd W Mount Morris R Vienna Rd

266

274

3.1%

4,634

4,363

-5.8%

3,235

585

2,066

675

Grand Blanc

130

117

-10.0%

2,376

2,214

-6.8%

1,940

405

1,131

676

Thompson Rd

100

16.7%

468

546

16.7%

546

109

388

21,330

22,252

4.3%

329,095

342,811

4.2%

224,307

47,331

52,869

124,107

673

Total

4-14

640 643 644 645 646 648 650 651 652 653 654 655 656 657 658 659 660 661 662 664 665 667 668 670 671 673 674 675 Sum

640 19.8 3.6 58.8 2.0 54.8 2.0 304.8 28.2 297.7 771.7

643 644 645 646 648 773.7 137.6 55.9 274.0 2.5 29.5 2,978.7 75.2 40.2 119.2 65.0 11.8 5.9 3.4 188.8 484.2 7.5 6.1 5,347.1 119.4 257.1 9.5 8.0 8,207.7 106.7 3,661.5 130.2 239.5 43.2 21,219.0 202.6 92.8 238.3 1,636.1 650 774.4 1.0 18.7 26.9 3.0 1.0 15.1 840.2

651 136.8 27.1 305.6 5.6 475.1

652 653 654 655 656 657 658 659 660 18.1 3.2 51.4 - 3,062.1 501.8 5,457.4 67.2 76.5 10.5 6.0 7.6 10.7 1.4 2.9 47.5 24.5 114.6 189.0 5.7 18.1 3.2 15.7 6.6 7.7 20.8 16.0 1.3 326.9 70.1 4.7 - 1,284.7 135.9 468.0 196.8 35.5 289.9 5.0 5.2 0.2 1.1 13.1 76.8 1.5 8.8 100.2 1.6 2.4 3.9 38.9 5.0 32.1 30.0 5,193.3 93.8 79.4 198.8 717.0 6,469.6

Table 4-9: 2045 External-to-External Auto Trips 661 662 664 665 667 668 670 671 673 674 675 Sum 1.7 47.3 1.5 343.9 25.5 279.2 772.0 260.9 8,274.6 - 3,419.1 239.0 41.0 - 21,256.0 204.0 93.0 3.8 137.9 239.0 7.4 89.3 133.4 1,644.0 0.9 16.8 26.6 2.6 1.1 15.3 837.0 5.5 473.0 5.0 5.0 32.0 8.8 30.0 20.3 304.0 - 1,356.2 189.1 5.4 80.0 5,188.0 16.0 1.3 70.4 4.7 1.4 94.0 0.3 1.9 79.0 198.0 144.3 36.1 1.1 9.4 8.0 716.0 460.1 295.6 14.2 108.7 47.5 6,458.0 32.9 18.4 0.6 4.4 4.7 383.0 70.9 1,085.2 278.9 1,435.0 70.7 72.0 - 1,082.1 108.0 6.8 59.1 75.1 10,116.0 452.4 0.6 455.0 277.5 449.8 732.0 36.9 6,049.0 17.7 104.2 915.0 0.5 6.2 0.5 76.0 4.0 54.2 564.0 4.6 85.5 137.0 383.6 1,430.3 72.3 10,137.9 451.9 736.0 6,054.5 914.4 76.0 563.5 136.7 59,252.0

Chapter 4 External Trips

4-15

640 643 644 640 643 644 645 646 648 650 651 8.4 653 654 655 2.0 193.7 656 657 3.6 658 659 0.4 26.6 660 6.2 475.8 661 0.2 14.0 662 664 665 7.5 771.2 667 0.1 668 670 27.8 217.8 671 0.7 6.7 673 1.2 674 11.3 675 Sum 56.0 1,706.9 12.0

645 646 1.8 0.1 6.4 2.6 0.3 0.3 0.1 0.6 10.9 0.9 7.0 17.0

648 650 651 8.3 0.7 43.0 11.6 42.9 10.4 0.9 10.6 10.6 0.4 0.7 0.2 14.3 3.8 11.1 3.3 0.2 0.1 1.5 0.4 102.0 52.0 21.0

653 0.2 0.8 1.0

Table 4-10: 2045 External-to-External Truck Trips 654 0.5 0.3 0.3 1.0

655 1.9 198.4 6.5 0.7 10.6 3.3 73.1 232.3 17.5 0.7 10.0 555.0

656 0.7 0.0 5.0 0.3 0.0 6.0

657 658 659 660 0.3 5.6 27.8 481.7 3.7 0.2 0.3 0.0 0.2 3.7 0.4 10.6 0.6 0.1 25.3 112.2 3.0 37.4 0.0 0.1 1.8 0.1 1.1 13.7 0.1 0.2 2.5 4.0 11.0 59.0 658.7

661 662 0.2 14.1 0.3 0.7 0.2 0.3 3.2 0.7 2.0 68.5 13.7 6.2 1.4 0.1 0.5 0.2 28.0 84.1

664 0.0 1.9 2.0

665 6.7 777.3 12.7 3.7 68.9 5.0 68.5 14.3 1.0 8.5 6.0 972.4

667 668 670 0.0 30.3 199.0 11.1 11.5 3.5 240.2 0.3 26.0 108.8 5.5 13.6 19.9 20.0 0.0 20.1 33.8 636.0

671 0.6 6.6 0.1 16.7 3.0 37.9 1.5 14.6 81.0

673 1.1 0.1 0.7 0.0 0.1 1.9 0.1 1.0 0.0 5.0

674 10.4 1.5 0.4 10.0 0.1 1.1 14.3 0.5 8.7 47.0

675 Sum 56.0 1,706.0 12.0 7.0 17.0 102.0 52.0 21.0 1.0 1.0 555.0 0.0 6.0 4.0 11.0 0.4 59.0 2.9 659.0 0.2 28.0 84.0 2.0 5.4 973.0 20.0 34.0 636.0 81.0 5.0 47.0 9.0 9.0 5,188.0

Chapter 4 External Trips

4-16

Chapter 5

Trip Generation

Chapter 5: Trip Generation Introduction based on household size, workers per household, vehicles per household and school age children per household.

The purpose of this memorandum is to document the structure of the trip generation model as part of the Genesee County Travel Demand Model. This memorandum discusses the major components of the trip generation model including design and use of “MI Travel Counts” for calibration. The memorandum will document the following areas: • • • • • • •

• Trip Productions: Productions will be calculated for each zone using a two dimensional cross classification of rates. • Trip Attractions: Attractions will be calculated for each TAZ by trip purpose as a function of employment and households.

Overview of Model Structure; Use of “MI Travel Counts”; Trip Purpose Taxonomy; Household Disaggregation; Production Rate Calibration; Attraction Rate Calibration; and, Treatment of External – Internal Trips.

• Trip Balancing: The final step in the generation model is the balancing of attractions to productions by purpose. The exception to this rule is for non-home based trip purposes. For these purposes, the attractions are scaled to the production total and then productions are set equal to attractions.

A detailed description of the calibration methodology used, including statistical analysis and presentation of the model parameters will be discussed later in the chapter on calibration/validation.

The treatment of external – internal trips to Genesee County will be done within the existing steps of the generation model.

Overview of Trip Generation Model The trip generation model will have the following components:

MI Travel Counts As defined by the scope of services, the “MI Travel Counts” household travel survey is the primary source of data for the calibration of the trip generation model. The dataset includes a significant amount of data about all aspects of travel. Some examples include the characteristics of each traveler in the household surveyed, the type of trip they are making, where the trip is going, and how many similar trips the household makes in a day. The dataset

• Household Disaggregation: This step will disaggregate the zonal households into a two dimensional distribution consistent with the trip production rates. The model will create single dimension curves and then combine dimensions to create two dimensional distributions. The disaggregation of the households will be

5-1

Chapter 5

Trip Generation

is organized into several tables describing household characteristics, person characteristics and finally the trips made. Each data file can be linked to the others using unique household, person, and trip identifiers.

• Home Based Work (Low Income and High Income); • Home Based Shopping; • Home Based Other; • Home Based School – K12 (including the drop off trip made by non-students); • Home Based School – Univ/College; • Non Home Based Other; and, • Non Home Based Work.

The survey includes households for the entire state of Michigan. In calibrating trip generation models with statewide data, it may be possible to use larger datasets for some calibration steps when records from different areas share similar characteristics. As an example, it may be possible to include all households in Transportation Management Areas (TMA) for calibration of production rates, rather than just those in Genesee County. The advantage to a larger dataset is increased confidence of the rates being estimated when the sample size is larger. A second advantage is greater confidence when dividing the trips into subcategories, including home based work by income groups, as an example. The results of the analysis will be reported in a separate chapter on Calibration/Validation.

The above list is consistent with state of the practice models being applied across the country.

Household Stratifications State of the practice in trip generation models is to apply trip production rates that are cross classified into a two dimensional distribution. This allows for the rate of trips produced to increase as the size of the household and wealth of the household (typical cross classification variables) also increase. Typically, household data is reported at the zonal level by the total number of households, population, mean income, total workers, etc. Research has shown that households are distributed around the average. Thus, not all households are of the same size in a zone. Some are smaller, and others are larger. Rather than applying the same rate to all households in a zone based on the average size and wealth, the households in a zone are disaggregated into size categories. For example, households are stratified into typically four bins: 1 person, 2 person, 3 person and 4 or more persons per household.

Trip Purpose Taxonomy Trip purposes define the types of trip-making activities that occur within a region on a daily basis. The typical trip purposes used in travel demand models include home based work (HBW), home based other (HBO) and non-home based trips (NHB). The disadvantage to this simplified set of purposes is that it is difficult to separate unique trip behaviors. An example is home based shopping trips when combined with HBO. When combined, HBO must include not only retail employment but other employment types to capture all HBO activities. By separating shopping as a unique purpose, there is confidence in the location of attractions in the region. Over simplification of trip purposes often requires the use of many special generators to adequately replicate travel patterns within a model.

With respect to the Genesee County model, cross class variables have been defined for each trip purpose (Table 5-1). Thus, the following one dimensional distribution of households will be applied in the model: • • • •

For the Genesee County model, the following trip purposes are used:

5-2

Household Size (1, 2, 3 and 4+); Workers per Household (0, 1, 2, 3+); K12 Students per Household (0, 1, 2, 3+); and, Vehicles per Household (0, 1, 2, 3+).

Chapter 5

Trip Generation

Table 5-1: Cross Classification Variables by Trip Purpose Trip Purpose Home Based Work Home Based Shopping Home Based Other Home Based School K12 Home Based School U/C Non Home Based Other Non Home Based Work

Size Variable Workers / Household Pop / Household Pop / Household Pop 18 years & younger/Household Pop / Household Pop / Household Workers / Household

Wealth Variable

Vehicles / Household

Production Rates

Attraction Rates

Unique production rates will be calibrated using the MI Travel Counts for each of the trip purposes defined above for all cells in the 4*4 cross classification matrix. The calibration will be done using two separate datasets. The first will include only those households that reside in Genesee County, and trips that are made internal to the county. The second dataset will include all households in TMA areas (except for SEMCOG) and include trips that are made internal to the TMA of the residence.

The attraction rates for each trip purpose will be calibrated using the expanded MI Travel Count data specific to Genesee County. The attraction end of each trip will be assigned to a TAZ in the model area. Because of the limited number of households in the dataset, the TAZs will be aggregated to a district system for this analysis. Once the trips are aggregated, the total number of attractions by trip purpose will be compared to the employment and households in each zone. Using stepwise regression techniques, an attraction model will be calibrated using the following employment categories:

The intent of the analysis is to show that the trip rates generated using a larger dataset are statistically the same as those from Genesee County alone. If this is proven to be true, there will be increased confidence in the trip rates coming from a larger number of observations in each cell.

• • • • • • • • •

A further improvement being proposed to the trip generation model is to stratify home based work trips by income. The advantage to this approach is that it better links the income of the trip makers to the types of jobs they are typically going to. In models without this approach, high income homes often generate work trips that are attracted to neighboring low income employment rather than the more distant high income jobs. This will eliminate that potential error in the model. Work trips will be divided into low and high income groups.

Total Employment; Manufacturing Employment; Other Employment; Transportation Related Employment; Financial Related Employment; Retail Employment; Wholesale Employment; Service Employment; and, Government Employment.

The final model for each purpose will be evaluated based on several criteria including the goodness of the statistical fit, logic of the size of the coefficients and mix of employment categories, and finally the scaling factor between the productions and attractions. Ideally the scaling factor between the unbalanced productions and attractions should be as close to one as possible.

5-3

Chapter 5

Trip Generation

External-Internal Trips o Productions assigned at external station as a percent of total volumes based on MI Travel Counts and CTPP Journey to Work (JTW) data; and,

The external-internal trips represent the interaction Genesee County has with the surrounding region. Genesee County is in a unique situation that it serves as a bedroom community to several neighboring counties, and also attracts trips into the region for other purposes. For this reason, three separate external– internal trip purposes were defined:

o Attractions estimated at internal zones as function of HBW attractions. • IE_Work

• EI_NonWork

o Represents the outbound work commute made by residents inside Genesee County;

o Production assigned at external station as a percent of total volume; and,

o Productions estimated to internal TAZs as a function of HBW Productions using MI Travel Count data; and,

o Attractions estimated at internal zones as function of HBO and HBSH attractions. • EI_Work

o Attractions assigned to external cordon as a percentage of total outbound traffic.

o Represents the inbound work commute made by residents outside of Genesee County;

5-4

Chapter 6

Trip Distribution

Chapter 6: Trip Distribution Introduction and Overview o Home Based College/University (HBU); and, o Home Based Other (HBO).

The Genesee travel demand model uses a four-step modeling process with a travel time feedback loop. These four steps are trip generation, trip distribution, mode choice, and traffic assignment. Trip distribution links the trip productions and attractions for each pair of Traffic Analysis Zones (TAZs) in Genesee County. The gravity model, which is a type of destination-choice model, is the most widely used model for trip distribution. This model estimates the relative number of trips of each trip purpose, proportional to the number of productions and attractions, made between two geographical areas (TAZs), and inversely proportional to a function of the travel impedance separating the TAZs. The Genesee gravity model now uses composite impedance instead of highway travel time. This major enhancement makes the distribution model sensitive to transit and nonmotorized travel times in addition to highway travel time.

• Non-Home Based (NHB): o Non-Home Based Work (NHBW); and, o Non-Home Based Other (NHBO). For external trips with only one end in Genesee County, trip purposes were categorized as: • External-Internal Work (EIW); • Internal-External Work (IEW); and, • External Non-Work (ENW). The following sections describe the input data, distribution model, and model calibration.

Trip Distribution Model The gravity model is the most widely used model for trip distribution. Based on Newton’s law of gravitation, it assumes that the trips (i.e., trip productions) from a Traffic Analysis Zone (TAZ) are distributed to any TAZ (i.e., trip attractions) in direct proportion to the number of trip attraction and in inverse proportion to the spatial separation between TAZs. In general, the number of trips attracted to a TAZ reflects the size of the attraction TAZ and the interzonal travel time of the spatial separation between the TAZs. A gravity model with friction factors was used for trip distribution.

For internal trips with both trip ends in Genesee County, trip purposes are categorized as: • Home Based Work (HBW): o Home Based Work – Low Income (HBWLo); and, o Home Based Work – Medium and High Income (HBW-Hi). • Home Based Non-Work (HBNW): o Home Based Shopping (HBS); o Home Based School (HBSCH);

6-1

Chapter 6

Trip Distribution

Shortest path travel time is used as the highway travel impedance between Traffic Analysis Zones (TAZs). The time impedance between TAZs includes the travel time on roadway and terminal time. The terminal time is the time using to walk to/from vehicle and start or park the vehicle. It is defined as below,

The gravity model is sensitive to changes in transportation network such as travel speed of roadway, and incorporation of a new facility, etc. In accordance with these changes, the gravity model re-estimates the trip interchange of person trips based on changes in the network link impedance. The gravity model is expressed as:

  A j Fij K ij Tij = Pi  zones   ∑ Ak Fik K ik  k =1

• Three minutes for the CBD and urban area (Area Type =1,2);

     

• Two minutes for the suburban and rural area (Area Type=3,4,5); and, • Five minutes for the external station. The external trips do not exactly start at the external stations. Therefore, a five-minute terminal time is assigned to the external stations to account for that. This will help adjust the IE/EI trip distribution as well.

where: Tij = O-D trips between TAZ i and TAZ j; Pi = total trip productions of TAZ I; Dj = total trips attractions of TAZ j; Fij = friction factor between TAZ i and TAZ j; and, Kij = socioeconomic factor between TAZ i and TAZ j.

The composite impedance incorporates highway travel distance and the logsum of the highway, transit, and non-motorized impedances, as output from TransCAD’s nested logit model macro. Thus, the trip distribution model is now sensitive to the transportation supply provided by all modes, not just highways. Also, note that to use the composite impedance, mode shares are calculated before trip distribution. This is usual practice in most advanced practice (activity-based) models. Composite impedance is used for all internal-internal trip purposes. Highway travel time is used for EIW, IEW, and ENW (external) trip purposes.

Input Data The trip distribution modeling process incorporated the following data inputs and modeling elements: • Production (P) and Attraction (A) trip ends by trip purpose from the trip generation model, and for each trip purpose. The TransCAD gravity model requires that for each trip purpose, the sum of production to equal the sum of attractions,

Before being used in the gravity model, the GISDK code scales the logsums so that their values are always positive so they can be used to select friction factors in the gravity model.

• Interzonal and intrazonal composite impedances computed from the Genesee County roadway network, transit network, and nonmotorized network.

The impedance used to distribute trips is based on the multimodal logsums as calculated in the nestedlogit mode choice model and the zone-to-zone distance. Travel time components of the logsums are:

• Friction factors calibrated for each trip purpose using gravity model procedures. • Gravity model applications by trip purpose using TransCAD procedures.

6-2

Chapter 6

Trip Distribution

trip distribution by trip length for each trip purpose, based on the distribution of observed trips from the household travel survey.

• The highway travel time, based on distance and congested speeds; • Transit travel times, based on the underlying highway speed, dwell times at the stops, wait times and transfer times;

The gravity calibration function in TransCAD is used to generate initial friction factor look-up tables using a base P-A matrix by trip purpose (travel survey), impedance matrix (composite impedance), and the model zone structure. The initial friction factors were further calibrated using the following equation until the average trip length and trip length frequency distribution for each trip purpose were within a desirable range.

• Walking times at three miles per hour, with the speed modified as a function of the suitability of the walkway; and, • Bicycle times at ten miles per hour, with the speed modified as a function of the suitability of the bikeway.

where:

The bike and walk mode travel speeds are functions of suitability scores along travel links (roadways, paths, bikeways, sidewalks, etc.). Walk and bike suitability score are calculated separately. Network attributes that affect the suitability scores are:

i +1 t

 Tt obs = F  i  Tt i t

  

Fti +1 = the friction factors for time interval t for

• Daily auto and truck volumes (from the highway assignment); • Functional classification (no walking or biking allowed on freeways) (NFC); • Vehicular speed along the link (AB/BA_FFSPD); • Roadway shoulder width (SHOULDER_WIDTH); • Type of bicycle facility (Bk_lns); and, • Type of pedestrian facility (sidewalk).

iteration i+1;

Fti = the friction factor for time interval t for iteration i;

Ttobs = the observed number of trips in time interval t; and,

Thus, the impedance for bikes is based on a speed less than or equal to ten miles per hour. Similarly, the impedance for walking is based on a speed less than or equal to three miles per hour.

Tti

= the estimated number of trips in time in-

terval t for iteration i. The calibrated friction factors used in the TDM (for first 45 minutes) are provided in Table 6-1. Figure 6-1 also provides a visual graphic of the calibrated friction factors for some trip purposes (note that logarithmic graph scale is used to account for higher values of friction factors at some impedances)

Friction Factors Friction factors were used to control the gravity model by regulating the trip lengths and trip length frequency distribution for each trip purpose. Friction factor values are associated with specific range of composite impedance (scaled to time), usually in one minute intervals and adjustments to friction factors reflect the change in the travel patterns across the region. The friction factors in the Genesee County TDM were calibrated to produce a general curve of

6-3

Chapter 6

Trip Distribution

Table 6-1: Friction Factors by Trip Purpose

6-4

Chapter 6

Trip Distribution

Figure 6-1: Logarithmic Graph of Friction Factors

1.00E+11 1.00E+10 1.00E+09 1.00E+08

Friction Factor

1.00E+07 1.00E+06 1.00E+05 1.00E+04 1.00E+03 1.00E+02 1.00E+01 1.00E+00

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Composite Impedance (Minutes) HBW

HBO

Using the Genesee County specific trip records, the average trip length and trip length frequency distribution were calculated for each trip purpose. For each record in the MI Travel Counts database, a TAZ was assigned for the origin and destination based on the geocoded coordinates provided by MORPACE. The origin and destination TAZs were then used to assign a skimmed travel time from the model network with free-flow travel time and terminal time. Then the skimmed travel time was aggregated and averaged to represent the actual (survey) trip length frequency distribution and average travel time. Table 6-2 compares the average travel time by different trip purposes between the model run and MI Travel Survey, and Tables 6-3 through 6-6 list the trip length frequency distribution data from the Model

HBSCH

HBS

HBU

for each period by trip purpose. Also, Figures 6-2 through 6-5 display the trip length frequency distribution curves by trip purpose for each period. Note that while the distribution model now uses composite impedance, the model was calibrated to match the observed MI Travel Counts trip length frequency distribution in minutes.

Table 6-2: Average Daily Travel Time by Purpose (Minutes) HBW

HBNW

NHB

EIW

IEW

ENW

Actual

17.73

14.17

14.46

16.74

18.35

16.65

Model

17.17

13.72

14.05

17.23

19.40

16.15

% Difference

3.2%

2.9%

5.6%

3.0%

6-5

Chapter 6

Trip Distribution

Table 6-3: Trip Length Frequency (Average Travel Time in minutes) for AM Peak TRAVEL TIME

HBW%

HBNW%

NHB%

EIW%

IEW%

ENW%

0.00

0.01

0.00

0.01

0.00

0.03

0.13

0.32

0.27

0.76

0.55

0.25

0.79

1.37

1.11

2.18

1.80

9.61

2.74

4.46

3.36

2.07

2.13

3.00

5.50

8.45

6.15

6.57

6.42

4.94

4.97

8.30

6.79

6.01

4.49

2.43

4.81

8.58

7.76

3.87

5.25

8.91

4.60

7.88

8.12

5.21

2.65

4.31

4.56

8.22

8.40

3.60

3.58

5.03

4.82

7.84

8.33

4.79

4.51

6.32

5.39

7.55

8.35

7.34

5.74

6.71

5.83

7.08

8.13

4.47

4.05

4.04

5.93

6.13

7.30

4.69

4.54

2.40

6.02

4.92

5.98

6.11

4.79

13.76

6.02

4.01

5.01

7.04

5.20

1.93

5.23

2.89

3.41

6.95

5.31

4.29

4.71

2.15

2.44

4.76

4.53

3.50

4.19

1.64

1.82

3.83

3.18

0.70

3.64

1.38

1.50

2.86

2.45

3.09

3.10

1.14

1.22

1.92

1.76

0.60

2.77

0.98

1.05

2.47

1.58

2.34

2.48

0.89

0.84

1.32

1.11

0.45

2.03

0.74

0.67

0.95

1.14

1.55

1.71

0.57

0.51

1.21

1.44

0.95

1.57

0.54

0.41

1.23

2.09

0.69

1.24

0.41

0.30

1.47

1.89

2.22

1.08

0.33

0.22

1.85

3.04

1.40

0.89

0.25

0.15

0.84

1.85

1.14

0.69

0.20

0.09

1.07

2.14

0.51

0.59

0.16

0.07

0.76

2.15

0.82

0.42

0.11

0.04

0.70

2.25

0.17

0.33

0.08

0.03

0.41

1.34

0.33

0.26

0.07

0.02

0.23

1.37

0.09

0.20

0.06

0.02

0.17

1.04

0.12

0.16

0.05

0.02

0.12

1.18

0.10

0.14

0.04

0.02

0.09

0.38

0.07

0.11

0.03

0.01

0.03

0.52

0.11

6-6

Chapter 6

Trip Distribution

Table 6-3: Trip Length Frequency (Average Travel Time in minutes) for AM Peak (continued) TRAVEL TIME

HBW%

HBNW%

NHB%

EIW%

IEW%

ENW%

0.09

0.02

0.01

0.02

0.13

0.01

0.07

0.02

0.01

0.16

0.02

0.05

0.02

0.01

0.00

0.06

0.27

0.03

0.01

0.00

0.06

0.42

0.03

0.02

0.01

0.00

0.06

0.15

0.02

0.01

0.00

0.03

0.13

0.02

0.01

0.00

0.03

0.04

0.01

0.00

0.01

0.00

0.01

0.03

0.01

0.00

0.01

0.00

0.01

0.00

0.01

0.00

0.01

0.00

Total

100.00

6-7

Chapter 6

Trip Distribution

Table 6-4: Trip Length Frequency (Average Travel Time in minutes) for Midday TRAVEL TIME 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40

HBW% 0.00 0.00 0.00 0.00 0.00 0.13 0.78 2.60 5.40 4.83 4.75 4.46 4.47 4.72 5.18 5.67 5.87 6.01 5.99 5.30 4.73 4.22 3.79 3.22 2.83 2.54 2.14 1.82 1.58 1.39 1.09 0.95 0.77 0.62 0.47 0.37 0.28 0.22 0.18 0.15 0.12

HBNW% 0.00 0.00 0.00 0.00 0.00 0.31 1.34 4.33 8.10 8.01 8.52 7.59 8.02 7.72 7.43 7.12 6.23 5.25 4.20 3.10 2.32 1.74 1.43 1.21 1.05 0.90 0.79 0.62 0.55 0.45 0.34 0.28 0.21 0.17 0.13 0.10 0.07 0.06 0.05 0.04 0.03

NHB% 0.00 0.00 0.00 0.00 0.01 0.27 1.09 3.25 5.84 6.55 7.51 7.80 8.18 8.09 8.18 8.10 7.41 6.32 5.27 3.71 2.66 1.95 1.57 1.30 1.10 0.90 0.72 0.56 0.44 0.34 0.24 0.17 0.11 0.08 0.05 0.03 0.02 0.02 0.02 0.02 0.02

6-8

EIW% 0.00 0.00 0.00 0.01 0.01 0.95 2.28 2.02 6.13 5.33 3.95 4.20 3.68 5.72 6.77 4.69 4.32 6.67 6.77 6.73 5.34 3.89 3.25 2.12 2.01 1.62 0.89 0.94 1.34 1.47 1.51 0.97 1.10 0.94 0.67 0.60 0.34 0.33 0.20 0.09 0.08

IEW% 0.00 0.00 0.00 0.00 0.00 0.56 2.02 2.21 5.73 5.01 4.56 2.86 3.44 4.78 5.43 4.02 4.60 4.95 4.93 5.14 4.44 3.82 2.38 1.93 1.71 1.21 1.04 1.49 1.83 2.61 2.91 1.96 2.19 1.98 1.89 1.52 1.32 1.18 0.88 0.55 0.35

ENW% 0.00 0.00 0.00 0.00 0.02 0.58 9.76 2.35 5.11 2.41 7.60 3.92 5.53 6.53 7.23 4.08 2.49 12.66 2.33 3.68 4.36 0.89 3.30 0.81 2.28 0.52 1.09 0.85 0.47 2.41 1.60 1.33 0.87 0.38 0.26 0.30 0.12 0.22 0.10 0.08 0.25

Chapter 6

Trip Distribution

Table 6-4: Trip Length Frequency (Average Travel Time in minutes) for Midday (continued) TRAVEL TIME 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 Total

HBW% 0.10 0.08 0.05 0.04 0.03 0.02 0.02 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 100.00

HBNW% 0.03 0.02 0.02 0.02 0.01 0.01 0.01 0.02 0.01 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 100.00

NHB% 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 100.00

6-9

EIW% 0.02 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 100.00

IEW% 0.17 0.12 0.06 0.08 0.05 0.04 0.02 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 100.00

ENW% 0.05 0.09 0.42 0.33 0.15 0.08 0.06 0.02 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 100.00

Chapter 6

Trip Distribution

Table 6-5: Trip Length Frequency (Average Travel Time in minutes) for PM Peak TRAVEL TIME 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40

HBW% 0.00 0.00 0.00 0.00 0.00 0.13 0.76 2.52 5.23 4.70 4.63 4.18 4.42 4.40 4.91 5.36 5.75 5.70 5.84 5.54 4.78 4.31 3.70 3.37 3.00 2.60 2.29 2.00 1.65 1.44 1.25 1.07 0.88 0.72 0.61 0.48 0.38 0.28 0.22 0.19 0.16

HBNW% 0.00 0.00 0.00 0.00 0.00 0.30 1.31 4.26 7.72 7.87 8.28 7.24 7.68 7.60 7.30 6.95 6.40 5.38 4.40 3.49 2.50 1.92 1.47 1.27 1.10 0.93 0.83 0.71 0.54 0.51 0.39 0.33 0.26 0.19 0.17 0.13 0.10 0.08 0.06 0.05 0.04

NHB% 0.00 0.00 0.00 0.00 0.01 0.27 1.06 3.18 5.55 6.26 7.28 7.34 7.84 7.94 7.92 7.86 7.52 6.42 5.53 4.32 2.90 2.15 1.64 1.39 1.14 0.99 0.78 0.65 0.48 0.40 0.30 0.22 0.16 0.11 0.08 0.05 0.04 0.03 0.02 0.02 0.02

6-10

EIW% 0.00 0.00 0.00 0.01 0.09 0.90 2.23 2.21 5.49 5.32 3.85 4.03 4.13 3.90 8.19 4.08 3.79 6.87 6.39 7.16 5.21 4.26 3.65 2.11 1.97 1.55 1.13 0.81 1.14 1.64 1.05 1.04 0.83 0.73 1.35 0.79 0.51 0.45 0.50 0.26 0.19

IEW% 0.00 0.00 0.00 0.00 0.00 0.58 1.95 1.97 5.37 5.00 5.19 2.26 3.47 4.51 4.12 5.33 3.96 5.34 5.15 4.94 3.85 3.85 3.05 2.13 1.34 1.20 1.20 0.87 1.22 1.39 2.61 1.81 1.87 1.68 2.02 2.58 1.67 1.43 1.35 0.82 1.28

ENW% 0.00 0.00 0.00 0.00 0.15 0.72 9.67 2.46 4.81 2.87 7.34 3.45 5.37 5.10 8.47 3.08 2.41 13.46 1.89 4.10 3.90 0.87 3.11 0.69 2.86 0.64 1.69 0.79 0.54 0.93 0.23 2.20 0.34 0.39 0.18 0.39 0.31 0.45 0.32 0.27 0.50

Chapter 6

Trip Distribution

Table 6-5: Trip Length Frequency (Average Travel Time in minutes) for PM Peak (continued) TRAVEL TIME 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 Total

HBW% 0.12 0.11 0.08 0.07 0.04 0.03 0.03 0.02 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 100.00

HBNW% 0.03 0.03 0.02 0.02 0.01 0.02 0.02 0.01 0.01 0.01 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 100.00

NHB% 0.02 0.02 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 100.00

6-11

EIW% 0.12 0.04 0.02 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 100.00

IEW% 0.36 0.52 0.27 0.13 0.09 0.04 0.08 0.06 0.03 0.02 0.01 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 100.00

ENW% 0.28 0.47 1.34 0.26 0.35 0.20 0.09 0.01 0.01 0.03 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 100.00

Chapter 6

Trip Distribution

Table 6-6: Trip Length Frequency (Average Travel Time in Minutes) for Off-Peak Travel Time 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40

HBW% 0.00 0.00 0.00 0.00 0.00 0.13 0.80 2.76 5.66 5.15 4.96 4.73 4.70 5.05 5.57 6.15 6.07 6.14 6.03 5.19 4.66 4.07 3.49 3.06 2.76 2.29 1.95 1.69 1.43 1.16 0.99 0.75 0.60 0.46 0.34 0.28 0.21 0.18 0.14 0.11 0.09

HBNW% 0.00 0.00 0.00 0.00 0.00 0.30 1.41 4.55 8.57 8.66 8.78 8.16 8.52 7.90 7.63 7.14 5.97 4.73 3.75 2.68 2.06 1.58 1.30 1.11 0.98 0.86 0.69 0.56 0.49 0.37 0.27 0.21 0.15 0.13 0.09 0.07 0.05 0.05 0.04 0.03 0.02

NHB% 0.00 0.00 0.00 0.00 0.01 0.28 1.14 3.43 6.25 7.13 8.01 8.29 8.81 8.39 8.51 8.28 7.15 5.86 4.67 3.11 2.28 1.73 1.41 1.20 1.01 0.77 0.61 0.47 0.35 0.25 0.18 0.11 0.07 0.05 0.03 0.02 0.02 0.02 0.02 0.01 0.01

6-12

EIW% 0.00 0.00 0.00 0.01 0.00 0.81 2.15 2.91 6.13 5.04 4.78 5.77 3.84 4.97 7.21 3.91 4.56 6.98 7.00 6.48 4.81 3.58 2.73 2.16 2.32 1.42 1.09 1.38 1.44 1.44 1.30 0.91 0.80 0.81 0.51 0.31 0.18 0.10 0.06 0.07 0.02

IEW% 0.00 0.00 0.00 0.00 0.00 0.61 1.92 2.66 5.87 4.98 4.48 3.19 3.52 4.34 6.49 4.23 4.44 5.14 4.98 5.44 3.89 3.36 2.45 2.06 1.97 1.37 1.60 1.78 2.72 3.18 3.29 1.94 1.73 1.53 1.40 1.45 0.83 0.40 0.29 0.16 0.11

ENW% 0.00 0.00 0.00 0.00 0.02 0.32 9.75 3.90 4.76 2.11 8.55 5.29 7.07 4.89 6.60 3.61 3.27 12.50 1.74 3.93 3.52 0.57 3.04 0.59 2.65 0.57 1.32 1.35 0.79 2.53 1.06 1.22 0.57 0.41 0.16 0.12 0.03 0.10 0.02 0.03 0.04

Chapter 6

Trip Distribution

Table 6-6: Trip Length Frequency (Average Travel Time in Minutes) for Off-Peak (continued) Travel Time 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Total

HBW% 0.07 0.05 0.03 0.02 0.02 0.02 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 100.00

HBNW% 0.02 0.02 0.01 0.01 0.01 0.01 0.02 0.01 0.01 0.01 0.01 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 100.00

NHB% 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 100.00

6-13

EIW% 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 100.00

IEW% 0.08 0.05 0.03 0.01 0.02 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 100.00

ENW% 0.01 0.00 0.39 0.30 0.14 0.10 0.00 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 100.00

Chapter 6

Trip Distribution

Figure 6-2: Trip Length Frequency Distribution (Travel Time) for AM Peak

AM Trip Length Frequency Distribution (4-minute groups) 35

HBW HBNW NHB EIW IEW ENR

Percentage of Trips

30 25 20 15 10 5 0

Length (Minutes)

Figure 6-3: Trip Length Frequency Distribution (Travel Time) for Midday

Midday Trip Length Frequency Distribution (4-minute groups) 35 HBW HBNW NHB EIW IEW

Percentage of Trips

30 25 20 15 10 5 0

Length (Minutes)

6-14

Chapter 6

Trip Distribution

Figure 6-4: Trip Length Frequency Distribution (Travel Time) for PM Peak

PM Trip Length Frequency Distribution (4-minute groups) 35 HBW HBNW NHB EIW IEW ENR

Percentage of Trips

30 25 20 15 10 5 0

Length (Minutes)

Figure 6-5: Trip Length Frequency Distribution (Travel Time) for Off-Peak

Off-Peak Trip Length Frequency Distribution (4-minute groups) 40

HBW HBNW NHB EIW IEW ENR

Percentage of Trips

30 25 20 15 10 5 0

Length (Minutes)

6-15

Chapter 6

Trip Distribution

Feedback Loop

External-External Trip Distribution

Traditionally, in the trip distribution step the impedance matrices are created using the uncongested free flow travel times. The travel time output from the trip assignment step on the other hand is congested travel times. This inconsistent use of travel time in the travel distribution and trip assignment steps is inaccurate and not desirable. A feedback loop was introduced in the model to solve this issue by feeding the congested travel times from the trip assignment back into trip distribution so that the next set of travel time skims reflects the updated congested travel times and speed. At each iteration, the composite impedances discussed earlier were recalculated to incorporate transit and non-motorized travel times.

The E-E auto and truck trips between the external zone pair were balanced using an iterative proportional factoring (Fratar) process. E-E auto and truck seed matrix obtained from MDOT were used to balance the E-E trips. The model processes the E-E trips for each external station with the E-E trip seed matrix using a doubly-constrained growth factor model to generate a new E-E trip matrix reflective of the current number of E-E trips at the external stations while maintaining the same basic distribution of E-E trips as the seed matrix. Refer to Chapter 4: External Trips for more details.

Time of Day Model A time of day model enables the analysis of both daily and peak period conditions. The estimation of travel over specific periods of the day is necessary for certain transportation planning studies, such as peak hour congestion, emission analyses, and transit services. A set of time-of-day parameters are then used to develop period-specific production-attraction trip tables from the daily trip tables. The daily trip tables in the Genesee County TDM are separated into the following time periods:

TransCAD’s Method of Successive Averages (MSAs) feedback method was used in the Genesee County TDM. In the MSA method, output volumes from trip assignment from previous iterations are weighted together to produce the current iteration’s link volumes. Adjusted congested times are then calculated based on the normal volume-delay relationship. This adjusted congested time is then fed back to the skimming procedures. This process is continued until a convergence criterion is met. In the Genesee County model RSME is calculated to measure the difference between congested travel times in successive iterations, and when the RMSE is less than 0.01, the feedback loop is terminated. In the Genesee County model the feedback convergence was checked for each skim by time period (AM, Midday, PM, and OffPeak).

• • • •

AM Period (6 a.m. and 9 a.m.); Midday Period (9 a.m. and 3 p.m.); PM Peak Period (3 p.m. and 6 p.m.); and, Night Period (6 p.m. and 6 a.m.).

The sum of these periods will be the 24-hour daily trips and flows. The time of day factors will be discussed in more details in Chapter 8: Time of Day Model. The truck time of day percentages will also be discussed in Chapter 9: Truck Model. The trip tables separated by time period are carried through the mode choice step.

6-16

Chapter 7

Mode Choice Model

Chapter 7: Mode Choice Model Introduction and Overview

The mathematical formulation of the nested multinomial logit model is as follows:

The trip generation models generate numbers of person trips and the trip distribution models allocate these trips for trip production zones to attraction zones for each trip purpose. These trips must be further divided into trips by various transportation modes and then converted to vehicle trips and passenger trips for the purpose of predicting vehicle flows on the roadway network and passenger flows on the transit routes. The Genesee County model divides the person trips into trips of five modes: car driver alone, car share ride, transit (bus), and nonmotorized (walk/bike). The nested logit model is decided to be used for the Genesee County model, and its structure is shown in Figure 7-1.

PDA = P( DA | Auto, M ) * P ( Auto | M ) * P ( M ) PSR = P ( SR | Auto, M ) * P ( Auto | M ) * P ( M )

PTr = P(Tr | M ) * P( M ) PWlak = P(Walk | NM ) * P( NM ) PBike = P( Bike | NM ) * P( NM ) where:

is the probability of choosing mode alterna-

tive i ;

Figure 7-1: Structure of Mode Choice Model

â&#x20AC;&#x153;iâ&#x20AC;? is Drive Alone(DA), Share Ride(SR), Transit(Tr), Walk or Bike;

Choice

P (i | Auto, M ) is the conditional probability Motorized

of choosing i from among DA and SR;

Non-Motorized

P ( j | M ) is the conditional probability of Auto

Transit

Bike

Pedestrian

choosing j from among Auto and Transit;

P( s | NM ) is the conditional probability of Drive Alone

choosing s from among Walk and Bike;

Share Ride

P (M ) is the probability of choosing Motorized

Only for HBW Trips

mode; and,

7-1

Chapter 7

Mode Choice Model

The utility expression for each available choice mode (i) is specified as a linear function:

P(NM ) is the probability of choosing a NonMotorized mode.

U i = b1 * IVTTi + b2 * OVTTi

eU M P (M ) = U e M + eU NM

+ b3 * Costi + b4 * SE1i + b5 * SE 2i + b6 * SE 3i + b0

and

P ( NM ) =

where:

eU NM eU M + eU NM

IVVTi is the In-Vehicle Travel time of mode alternative I;

U M and U NM are the Utilities of the motorized

OVTTi is the Out-Vehicle Travel Time of the

and non-motorized modes, and its expressions are,

description of alternative I;

U M = a1 + Logsum( M ) * ln(eU Auto +U Tr )

Cost is the fare related cost when choice is bus otherwise it is the distance related cost; and,

U NM = a 2 + Logsum( NM ) * ln(eU Walk +U Bike )

SE1i , SE2i and SE3i are the socio-eco-

P (j|M) =

nomic indicators of alternative i.

Calibration, Coefficients and Constants

e U Auto

The mode choice model calibration is based on the Travel Counts Household Travel Survey data and the 2014 bus on-board survey. The 2000 CTPP data is used as the reference for HBW trip as well. The indicators and coefficients mentioned above can be found in Table 7-1.

U Tr

and

eU s P ( s | NM ) = U e Walk + eU Bike

The mode choice constants were calibrated so that the model estimated mode shares matched the mode share targets developed using the MI travel survey and the 2014 MTA on-board survey. Table 7-2 shows the modeled mode shares. A more detailed mode choice summary by time period and trip purpose, produced by the mode choice step will be presented in the chapter on validation and calibration.

Logsum(M), Logsum(NM), a1 and a2 are constants.

U Auto is the Utility of the auto modes and its expression is,

U Auto = a3 + Logsum( A) * ln(eU DA +U SR ) eU i P (i | Auto, M ) = U e DA + eU SR

7-2

Chapter 7

Mode Choice Model

Table 7-1: Nested Logit and Travel Utility Parameters VARIABLE

AUTO/DA

TRANSIT

PEDESTRIAN

BIKE

mode constant in-vehicle time auto distance transit out-of-vehicle time transit fare ($) ped. walk impedance bike impedance home taz workers/hh home taz autos/hh home taz hh income motorized nest logsum non-motorized nest logsum

mKhbw civtt xdist covtt XCIVTT cwtt cbtt whhK vhhK mediK nc ncn

0 -0.023 -0.0383 X X X X -0.1 0.2 0.000001 0.8 0.7

-1.0346 -0.023 -0.0383 X X X X -0.2 0.1 0.000001 X X

-1.1465 -0.023 X -0.0575 -0.1917 X X X X X X X

-1.6498 X X X X -0.023 X X X X X X

-1.5197 X X X X X -0.115 X X X X X

mKhbo civtt xdist covtt XCIVTT cwtt cbtt whhK vhhK mediK nc ncn

0 -0.016 -0.0267 X X X X -0.1 0.1 0.000001 0.8 0.7

X X X X X X X X X X X X

-2.657 -0.016 X -0.04 -0.1333 X x X X X X X

-1.1495 X X X X -0.016 x X X X X X

-1.6023 X X X X X -0.08 X X X X X

mode constant in-vehicle time auto distance transit out-of-vehicle time transit fare ($) ped. walk impedance bike impedance motorized nest logsum non-motorized nest logsum

mKnhb civtt xdist covtt XCIVTT cwtt cbtt nc ncn

0 -0.018 -0.03 X X X X 0.8 0.7

X X X X X X X X X

-2.6714 -0.018 X -0.045 -0.1500 X X X X

-0.9771 X X X X -0.018 X X X

-1.3319 X X X X X -0.09 X X

PURPOSE

HBW

HBO

NHB

aoc auto operating ($/mile) 0.2 vot value of time ($/minute) 0.12 see Chapters 2 and 6 for the definition of the pedestrian impedance, which includes walk time and suitability see Chapters 2 and 6 for the definition of bike impedance, which includes bike time and suitability

7-3

Chapter 7

Mode Choice Model

Table 7-2: Modeled Mode Shares MODE

TOTAL VEHICLES

PERCENT VEHICLES

TOTAL PERSONS

PERCENT PERSONS

Auto

915,873

100

1,243,310

83.59

Transit

14,529

0.98

Pedestrian

165,661

11.14

Bike

63,784

4.29

TOTAL

915,873

100

1,487,284

100.00

Input and Output Files • Transit time skims by time-of-day; and, • Non-motorized impedance skims.

The mode choice model coefficients and constants are embedded in the GISDK code.

Outputs of the mode choice model are:

The input files required for the mode choice model include:

• Mode shares by time-of-day; and, • Logsums by time-of-day, used in the trip distribution composite impedance calculations.

• Zonal data (TAZ file); • Highway time skims by time-of-day;

7-4

Chapter 8

Time of Day Model

Chapter 8: Time of Day Model Introduction and Overview Table 8-2 shows the control file (PAOD_ToD.bin) for the time-of-day model. It is based on the percentage of trips by purpose and direction in each time period, reported in MI Travel Counts. Directional factors for transit trips also are shown in Table 8-2, but are not used, because in the updated model, mode choice and distribution are applied by time of day and trip purpose. Thus, transit trips inherit their temporal distribution from the underlying person trip tables.

Using the MI Travel Counts dataset for all TMA trips, a frequency distribution was calculated by departure hour for each trip purpose. The percent distribution is shown in Figure 8-1 and Table 8-1. The four Genesee County model time periods were identified based on observations of the hourly traffic counts available in the region. Those periods are: • • • •

AM Peak: 6:00am – 9:00am Midday: 9:00am – 3:00pm PM Peak: 3:00pm – 6:00pm Night: 6:00pm – 6:00am

Period factors were estimated from the Genesee County records in the MI Travel Count database. The factors represent the number of trips that depart in each period as defined above. Trips by period are summarized by P-A and A-P direction. The directional factors were developed from the MI Travel Counts using the travel direction reported in the survey. The home end of the trip is the production zone. Conversely, trips from work to home are in the A-P direction.

Figure 8-1: Time of Day Distribution of Trips

Since non-home-based trips do not have a home end, they are equally divided between P-A and A-P.

8-1

Chapter 8

Time of Day Model

Table 8-1: Hourly Distribution of Trips DEPARTURE HOUR

HBW

HBO

HBS

NHBO

NHBW

HBSCH

HBU

TOTAL

0.7%

0.2%

0.0%

0.1%

0.4%

0.0%

0.2%

0.4%

0.2%

0.5%

0.0%

0.4%

0.0%

0.2%

0.7%

0.1%

0.0%

0.2%

0.0%

0.1%

1.0%

0.1%

0.0%

0.6%

0.0%

0.2%

1.2%

0.3%

0.1%

0.0%

0.3%

6.0%

0.4%

0.0%

0.8%

0.0%

1.0%

8.3%

2.1%

0.8%

0.6%

1.6%

4.8%

0.0%

2.8%

13.1%

3.6%

2.5%

2.6%

7.0%

21.1%

11.1%

7.5%

9.5%

6.5%

2.1%

5.6%

7.0%

18.9%

7.4%

8.2%

3.0%

5.5%

6.0%

5.1%

4.7%

1.1%

3.7%

4.4%

1.9%

7.0%

8.9%

7.7%

2.5%

1.2%

9.3%

5.5%

1.6%

6.1%

6.0%

8.9%

12.1%

2.9%

3.7%

6.1%

2.0%

5.9%

7.1%

9.8%

12.3%

1.8%

9.3%

6.3%

3.5%

6.3%

6.9%

7.6%

5.3%

1.5%

7.4%

5.5%

5.8%

6.2%

8.3%

9.4%

11.3%

1.9%

7.9%

9.5%

7.6%

8.9%

10.0%

10.4%

19.6%

3.7%

10.7%

7.0%

9.2%

11.2%

9.5%

10.2%

4.7%

9.3%

8.6%

10.2%

8.8%

10.2%

8.0%

7.4%

3.6%

9.3%

8.0%

4.6%

8.3%

7.5%

6.8%

2.9%

2.5%

7.4%

6.0%

1.9%

5.5%

6.3%

4.5%

1.8%

2.5%

3.7%

4.1%

1.8%

5.1%

4.4%

3.0%

0.6%

1.6%

7.4%

3.3%

2.2%

3.3%

1.6%

1.3%

0.8%

0.7%

5.6%

1.9%

2.0%

0.9%

0.5%

1.0%

0.0%

0.8%

Table 8-2: TOD Directional Factors by Trip Purpose (PAOD_ToD.bin)

8-2

Chapter 9

Truck Model

Chapter 9: Truck Model Introduction and Overview trips. The discussion of the attraction of external-internal and inter-external trips can be found in that chapter.

The Genesee travel demand model update considers three vehicle classes: bus, truck and auto vehicle. Trucks are defined as commercial vehicles with six tires or above, and combination commercial vehicles consisting of a power unit (truck or tractor) and one or more trailing units. Auto vehicles include four-tire commercial vehicles, motorcycles and passenger cars. In other words, the four-tire commercial vehicle is a subclass of the auto vehicle.

Trip Generation The inputs to trip generation are the number of employees and the number of households by Traffic Analysis Zone (TAZ). The daily trip generation rates shown in Table 9-1 are for Origin (O) and Destination (D) trips. These rates were originally obtained from the Quick Response Freight Manual II. From the preliminary results of the model runs, these rates were further adjusted by a global factor 0.70 to replicate observed truck VMT in Genesee County. For example, the final combination truck rate per retail employee is 0.0455 that is equal to original rate 0.065 multiplied by 0.70.

Since there was no freight and truck survey available for model calibration, the commercial vehicle and truck model was developed based on the method recommended in Quick Response Freight Manual II (2007) to estimate trips for four-tire commercial vehicles and trucks. It is a four-step process and includes trip generation, trip distribution, choice of time-of-day, and trip assignment. Before trip assignment, the four-tire vehicle trip table is merged into the auto vehicle trip table. This chapter focuses on the first three steps and trip assignment will be discussed in Chapter 10: Traffic Assignment. In addition, a preliminary assignment of 2014 daily truck trips was conducted to check link truck volumes against daily truck counts for adjusting the model parameters and coefficients.

The number of trips generated for 4-tire commercial vehicles and trucks is summarized in Table 9-2. After the Origin (O) and Destination (D) trips were generated for each vehicle class, they were factored to match the weighted sum of total O and D trips in Genesee County. The weighted factors are 0.5 for both trip origins and destinations. Weighted sum balancing is a standard function in TransCAD. The trip balancing process makes the total trip origins for each vehicle class equal to the total trip destinations in Genesee County.

The base year of the Genesee County Travel demand model is 2014. The inputs to the commercial vehicle model are the Traffic Analysis Zone (TAZ) file with the socioeconomic data, the roadway network, and external trip tables. The external-external trip generation was discussed in Chapter 4 together with the production of external-internal and internal-external

The productions of External-Internal and Internal-External (EI-IE) truck trips are obtained from the external trip model. Since there is no freight and truck survey available for Genesee County, it is assumed that

9-1

Chapter 9

Truck Model

Table 9-2: Daily Trip Generation Rates (QRFM II)

Generator (Employment and Household) Agriculture, Mining and Construction Manufacturing, Transportation, Communications, Utilities & Wholesale Trade Retail Office and Services Households

Commercial Vehicle Trip Destinations (or Origins) per Unit per Day Four -Tire Vehicles Trucks (Single Unit 6+ Tires) Trucks (Combination) 0.174 1.11 0.289 0.938

0.242

0.104

0.888 0.437 0.251

0.253 0.068 0.099

0.065 0.009 0.038

Table 9-1: Summary of 2014 Trip Generation Trip Type 4-Tire Commercial Vehicle Truck EI-IE Truck

Origin (O) Destination (D) Origin (O) Destination (D) Production (P) Attraction (A)

Number of Trips Original Balanced 108,087 108,087 108,087 108,087 39,041 39,041 39,041 39,041 5,682 5,682 5,682 5,682

In this project, counts were available for only a small subset of links, too few to reliably use ODME. An effective sample must include enough measurements from widely dispersed parts of the network, preferably such that every O-D pair is connected by route(s) that pass through at least one counted link.

the EI-IE truck trip attractions are proportional to the truck destination trips. Initially, truck trip destinations are used as EI_IE truck trip attractions. The balancing process then scales total truck trip attractions to match the total truck productions, which are the total truck counts of all external stations. The EI-IE truck trips also are summarized in Table 9-2.

Genesee county TDM is a time-of-day model and therefore, to perform ODME process, truck counts by time-of-day (i.e. AM, Midday, PM and off-peak periods) were required. Genesee County Metropolitan Planning Commission, through Genesee County Road Commission, provided the consulting team with time-of-day traffic counts by different vehicle classes for only 54 locations, and these locations did not cover all the facility types. Thus, it was not possible to incorporate the ODME process.

A truck trip special generator was set up for the airport freight. Based on truck traffic counts, the airport tuck trip special generator rate was assumed to be six times the number of transportation employees at the airport.

Origin-Destination Matrix Estimation (ODME) The original intent for this model was to base the model on QRFM II, and then adjust the QRFM ii results using ODME matrix estimation procedures. ODME procedures require a rich truck traffic count database, by time period.

Trip Distribution Based on the method recommended in Quick Response Freight Manual II (2007), the gravity model

9-2

Chapter 9

Truck Model

For internal trips, friction factors recommended in the Quick Response Freight Manual II were used. The recommended friction factors have the following form.

was used to distribute truck trips. The gravity model is most widely used model for trip distribution. Based on Newton’s law of gravitation, it assumes that the trips from a TAZ (i.e., trip productions) are distributed to any TAZ (i.e., trip attractions) in direct proportion to the number of trip attraction and in inverse proportion to the spatial separation between adjacent TAZs. In general, the number of trips attracted to a TAZ reflects the size of the attraction TAZ and the interzonal travel time of the spatial separation between the TAZs.

Four-tire commercial vehicles:

Fij = e

and K ij = 1

−0.08tij

and K ij = 1

Trucks:

Fij = e

The EI-IE truck trips were classified as an individual type of trips so that their behavior could considered separately from internal-internal truck trips, and controlled by truck traffic counts at the external stations.

The form of the gravity model expression is:

  D j Fij K ij Tij = Oi  zones   ∑ Dk Fik K ik  k =1

−0.13tij

The average travel time of all trip types are given in Table 9-3. The four-tire commercial vehicle has the shortest average travel time of almost 17 minutes while the EI-IE truck has the longest travel time of 22.75 minutes. The trip length frequency distribution curves by time period are shown in Figure 9-1.

     

where: Tij = O-D trips between TAZ i and TAZ j; Oi = total trips originating from TAZ I; Dj = total trips destined to TAZ j; Fij = friction factor between TAZ i and TAZ j; and. Kij = socioeconomic factor between TAZ i and TAZ j. Table 9-3: Average Travel Time by Trip Type Trip Type

Average Travel Time (minutes) AM

MID

Daily

4-Tire Commercial Vehicle

16.81

17.03

17.38

16.55

16.99

Internal Truck

19.16

19.40

19.83

18.85

19.41

EI-IE Truck

22.48

22.71

23.30

22.03

22.75

9-3

Figure 9-1: Truck Trip Length Frequency Distribution

Chapter 9 Truck Model

9-4

Chapter 9

Truck Model

Time of Day Choice

odologies. As a model validation step for the commercial and truck trips, a daily multi-class assignment was employed to obtain the truck volumes and check it against base year truck traffic counts. Model adjustments were made based on the assignment results and truck traffic counts.

Time-of-day assignments were implemented to produce better model results. To facilitate it, the trip tables from trip distribution must be factored to reflect morning peak, evening peak and off-peak periods prior to trip assignment. The hourly time-of-day factors recommended in Quick Response Freight Manual II were aggregated into the periods defined in the following table and applied for the Genesee County Travel Demand Model.

Table 9-5 shows the results from the assignment, including number of truck count locations, RMSE, total truck counts, total assigned volumes with percent difference, total count VMT and model VMT with percent difference. At the beginning of test assignment, RMSE was over 100% and percent difference between assigned volumes and counts was over 200%. During truck model calibration, the trip rate adjustment factor was changed from 1 to 0.70 step by step. Finally, the results in Table 9-5 were obtained. These results were refined at the final model calibration step. The final assignment results are displayed in Figure 9-2. Note the close agreement in truck VMT estimated from counts and the model. Also, the percent RMSE is considered to be quite good, RMSE is judged on a sliding scale, depending on volumes, and the volume of trucks is relatively low when compared to total vehicles.

The internal trip tables are in O-D format while the EI-IE trip table is in P-A format. Since the O-D tables are required for the trip assignment, the P-A format trip table were converted to O-D format. The departure and return percentages were used for converting the P-A matrix into the O-D matrix. These percentages are listed in Table 9-4.

Model Validation Check and Adjustment The truck trips developed from this process are assigned to the roadway network together with auto vehicle trips by using the multi-class assignment for each time period. The discussion of assignments can be found in Chapter 10 describing assignment meth-

Table 9-4: Time of Day Factors PERIOD

4-TIRE COMMERCIAL VEHICLE

TRUCK

AM Peak (6am – 9am)

20%

PM Peak (3pm – 6pm)

EI-IE TRUCK TOTAL

DEPARTURE

RETURN

17%

10%

24%

17%

10%

Mid-day (9am – 3pm)

33%

42%

21%

Off-Peak (6pm – 6am)

23%

24%

12%

Table 9-5: Model Volumes vs. Truck Traffic Counts

Observations Total Count Total Model Flows Pct Diff Pct RMSE Count VMT Model VMT VMT Pct Diff 241,410 241,968 0.23% 219 205,191 223,037 8.70% 49.31%

9-5

Chapter 9

Truck Model

Figure 9-2: Results of Daily Truck Trip Assignment

9-6

Chapter 10

Traffic Assignment

Chapter 10: Traffic Assignment Introduction and Overview convergence is reached (doesn’t change between iterations), the travel model highway and traffic assignments will be reported by time periods and for the entire day. The GCMPC model will run quickly on a modern microcomputer – running time will not be an issue.

The assignment of trips to the network is the last step of the traditional sequential modeling processes. It provides the foundation for validating the model’s performance in replicating base-year travel patterns. Once the base-year assignment is validated, it is further used to forecast future traffic conditions on the network and to evaluate any transportation improvements in the future.

This section describes the assignment procedures and feedback loop of the model. Individual model components were put together to run trip assignment. Corradino wrote a batch program to automate the entire assignment and feedback process. The traffic assignment component of the Genesee County model has a number of changes.

The Genesee County model uses a time-of-day modeling procedure. In this procedure, auto and transit assignments are made for each of the four periods (AM-Peak, PM-Peak, Mid-Day and Off-Peak periods), and the periods are summed to produce daily assignments. For each time period, a two-step assignment procedure is implemented. The first step, which is referred to as “priority pre-loading”, is to assign the external-to-external auto trips and the truck trips onto the roadway network separately. Then the internal auto trips are assigned onto the network with considerations of these preloading volumes. The assignment method is an advanced variation of the user equilibrium assignment, called n-Conjugate User Equilibrium or “CUE”, which can reach a high level of convergence in few iterations. Finally, resulting travel times are then fed back to trip distribution, assuring consistent travel times in all model steps. The specific method used is the method of successive averages, or MSA, which modifies link loads after each feedback iteration, and then calculates the resulting travel time using the volume/delay function. These times are used the trip distribution model in the next feedback iteration. After the travel time feedback

• Improved Estimation of Free-Flow Speeds. Instead of using posted speed limits as a surrogate for free-flow speeds, free-flow speeds were estimated based on procedures in the Highway Capacity Manual 2010 (HCM 2010). A lookup table is used to calculate the free flow speeds based on link type and area type. This was explained in Chapter 2: 2014 Network Development in more detail. Small changes to the speed table were made to improve model validation metrics. • Improved Estimation of Link Capacities. Like free-flow speed calculations, a lookup table is used to estimate directional hourly capacities based on the methodologies in Highway Capacity Manual 2010 (HCM 2010). The estimated peak-hour capacities were then converted to am and pm peak, mid-day, off-peak period and daily capacities. This was explained in Chapter 2

10-1

Chapter 10

Traffic Assignment

and Appendix A (Capacity Calculation Methodology) in greater detail.

mated and used to redistribute trips in subsequent model runs, or feedback assignments. Recent research has shown that historically most highway assignment models were not run to a sufficient level of convergence. Thus, the Caliper’s latest highway assignment with warmstart algorithms are used in this model which will iterate the capacity restrained assignment to a relative gap of less than 0.00001. The final assignment results are obtained from the feedback assignment.

• Intersection Delays. Delays associated with controlled intersections were estimated to adjust directional link free-flow speeds and capacities. The HCM 2010 method of calculating vehicle delay that takes into consideration green time and progression effect was adopted. • Time-of-Day Models. The Genesee County model consists of four time-of-day (TOD) models: morning peak, evening peak, mid-day period, and off-peak periods. Modeling factors that are unique to each time period were derived from the MI Travel Counts household survey. Compared to a single daily model, the TOD modeling generates a more accurate travel model by treating each period uniquely.

This section describes the assignment procedures in Genesee County model. Individual model components of the Genesee County model were put together to run trip assignment. A batch program to implement the individual models was written to automate the whole assignment process.

Trip Assignment Procedures

• Truck Models. Travel patterns of trucks are different from those of passenger cars, thus it is desirable to have a separate truck mode in the model. In each of the four step processes, the Genesee County model maintains a separate truck model to address the unique travel characteristics of trucks. Truck trips are separately generated and distributed. Then, they are assigned to the network for each TOD simultaneously with the corresponding passenger car assignments.

Given a network and a demand matrix, traffic assignment estimates the traffic flow patterns and congestion on all links in the network. Traffic assignment is a key element in the urban travel demand forecasting process. The traffic assignment model predicts the network flows that are associated with future planning scenarios, and generates estimates of the link travel times and related attributes that are the basis for benefits estimation and air quality impacts. The traffic assignment model is also used to generate the estimates of network performance that are used in the mode choice and trip distribution stages of many models.

• Transit Assignment. The updated Genesee County model now includes a mode choice model, and carries transit trips, by time of day, all the way through the traffic assignment step. Transit assignments are done using a TransCAD route system generated transit network that is unique for each time of day. Transit assignments will be validated too.

Historically, a wide variety of traffic assignment models have been developed and applied. Equilibrium methods take account of the volume dependence of travel times, and result in the calculation of link flows and travel times that are mutually consistent. Equilibrium flow algorithms require iteration between assigning flows and calculating loaded travel times. Despite the additional computational burden, equilibrium methods will almost always be preferable to other assignment models.

• Feedback Loop. Link free-flow speeds derive the first phase of the model run, or initial assignment. It is used for network skimming, trip distribution and route choice. Following the first phase, link congested-speeds are esti-

10-2

Chapter 10

Traffic Assignment

The MMA model is a generalized cost assignment that assigns trips by individual modes or user classes to the network simultaneously. This method models the influence of toll facilities of all types as well as HOV facilities. Each mode or class can have different network exclusions, congestion impacts (passenger car equivalent values), values of time, and toll costs.

In many urban areas, there are many alternate routes that could be and are used to travel from a single origin zone to a single destination zone. Often trips from various points within an origin zone to various points in a destination zone will use entirely different major roads to make the trip. In some instances, reasonable alternate routes may be so numerous that they cannot be easily counted. For the traffic assignment model to be valid, it must correctly assign car volumes to these alternative paths.

Traffic Assignment Procedure As explained in the previous section, trip assignment for Genesee County model follows time-of-day procedures instead of running a single 24-hour assignment. For each of four time periods, a truck trip table developed for the respective time period was pre-assigned before an auto trip table was assigned. Then, an origin and destination auto trip table for the time period was assigned with truck trips preloaded. This process was repeated for all time periods. Total 24-hour link volumes were then obtained by aggregating the truck and auto loadings by time period. Each of these assignments utilized a user equilibrium method.

From a behavioral perspective, traffic assignment is the result of aggregating the individual route choices of travelers. Assignment models, not surprisingly, also differ in the assumptions made about how and which routes are chosen for travel. The key behavioral assumptions underlying the User Equilibrium assignment model are that every traveler has perfect information concerning the attributes of network alternatives, all travelers choose routes that minimizes their travel time or travel costs, and all travelers have the same valuations of network attributes. First proposed by Wardrop, at user equilibrium (UE), no individual travelers can unilaterally reduce their travel time by changing paths (Sheffi, 1985). A consequence of the UE principle is that all used paths for an O-D pair have the same minimum cost. This has become the most accepted and widely used highway assignment method. The CUE assignment used in the Genesee County Model is an advanced implementation of the UE method.

The congested travel time for each link is calculated by using the Bureau of Public Roads (BPR) form of the volume delay function with link specific parameters. The volume delay function is used to adjust the linkâ&#x20AC;&#x2122;s free-flow speed on the basis of its volume to capacity ratio to account for congestion related delay. The alpha and beta parameters for the BPR equation which are used in both the travel modelâ&#x20AC;&#x2122;s assignment procedure as well as the post-processing are coded on the network links. The default sets of volume-delay parameters are presented in Table 10-1. Note that these values may be adjusted during the calibration process.

MMA Assignments TransCADâ&#x20AC;&#x2122;s Multi-Modal Multi-Class Assignment (MMA) is a flexible master assignment routine designed for use in major metropolitan areas, and is directly applicable in statewide or interregional models. Note that, while most MMA models are just multi-modal, the model in TransCAD is multi-modal and multi-class.

Feedback Loop Steps in the travel demand model process require feedback iterations to reach systemic equilibrium. Feedback from trip assignment to trip distribution will provide more accurate travel times reflecting congestion. Considering that the inter-zonal travel

10-3

Chapter 10

Traffic Assignment

Table 10-1: Default Volume Delay Function Parameters by Roadway Class RURAL

FUNCTIONAL CLASS DESCRIPTION

URBAN

Principal Arterial - Interstate

α 0.95

β 5

α 0.95

β 5

Principal Arterial - Other Freeway/Expressway

0.99

0.72

2.7

Principal Arterial - Other

0.5

2.5

0.5

2.5

Minor Arterial

0.44

2.3

0.5

2.3

Major Collector

0.4

2.1

0.45

2.1

Minor Collector

0.4

2.1

0.45

2.1

Local

0.4

2.1

0.45

2.1

Ramp

0.68

time is input to the distribution stage, the feedback will improve the trip distribution results for providing more reasonable trip tables to trip assignment.

as link free-flow travel times, link peak and offpeak capacities, and link-specific BPR parameters.

In this model update, trip distribution, time-of-day choice and trip assignment were re-computed for each of the four modeled time periods. As noted earlier, the feedback process employed the Method of Successive Average (MSA). In the MSA method, assigned link volumes from previous iteration are weighted together to produce the current iteration’s link volumes; Adjusted congested time is then calculated based on the normal volume-delay function. This adjusted congested time is then fed back to calculate the travel time between each OD pair. This feedback process is kept until the maximum iteration equals 10 or the stop criterion is reached.

• Turn Restrictions. Turn prohibitions at intersections and interchanges where a certain movement(s) is prohibited.

Transit Assignment The method used in the Genesee Model for transit path calculations and assignment is the shortest path (SP) method. It finds the single best path from an origin to a destination that minimizes the total generalized travel cost. On any path segment only one transit line will be chosen, even if the segment is served by several transit lines with identical travel times. Fares can be used in finding the best generalized path.

Trip Assignment Data Inputs

The generalized costs are the combination of in vehicle travel time, access/egress time, waiting time, transfer time, dwelling time, transfer penalty and fare together with its weights. The network settings for finding Shortest Generalized travel cost Path include the following configurable settings:

The data inputs used in trip assignment and validation process included: • Origin-Destination Vehicle Trip Tables. Outputs from the trip distribution and subsequent matrix manipulation procedures. These tables are vehicle trip matrices by time-of-day.

• The travel time field to use to determine best paths, skim variables or perform assignments;

• Highway Network. The Genesee County Model highway network with key link attributes such

10-4

Chapter 10

Traffic Assignment

from a field in the mode table, and global values are entered in directly in the transit network settings dialog box. The route-level values have the highest priority. However, the route attribute may be missing, because "None" was chosen from parameter dropdown list or the value stored in the table is missing. In this case TransCAD will try to find the value in the mode table, if modes are defined in the Mode tab. If the value is also missing from the mode table, the global value will be used. The transit system configuration of the Genesee model is listed as follows:

• The network attributes containing route headways, transfer penalties, dwell times and layover times; • Limits on the number of transfers, maximum and minimum wait time, total trip cost, maximum transfer times, maximum access and egress times, and maximum modal travel times; • Weights to assign to waiting times, travel times, dwell times, non-transit times, and transfer times;

• Time Value ($/min.)................................... 0.2

• Fare structure information;

• Max Access Time (min.).............................. 30

• Mode-specific information; and,

• Max Initial Waiting Time (min.) .................. 30 • Max Egress Time (min.) .............................. 30

• Route-stop-specific information.

• Max Transfer Waiting Time (min.) ............. 30

There are five types of settings in the Shortest Path Transit Network Settings dialog box:

• Max Transfer Time (min.) ........................... 30 • Max Transfer Number .................................. 2

• General – Sets the travel time field, path method and maximum trip cost, transfer time, maximum number of transfers, and centroids.

• Transfer Penalty Time .............................. 1.5

• Mode – Sets the mode table and mode transfer table, and some mode specific restrictions and defaults. In the Genesee Model, there is no mode setting.

• Max Trip Time (min.) ................................ 120

• Fare – Sets the fare to be flat, zonal-based, or mixed.

• Wait Weight Factor ................................. 1.00

• Min Init Wait Time (min.) ........................ 2.00 • Min Transfer Wait Time (min.) ................ 2.00 • Max Trip Cost ........................................... 120 • Dwelling Time (min.) ................................. 0.2 • Walk Weight Factor ................................. 3.00 • Fare Weight Factor .................................. 1.00 • Link Time Weight Factor.......................... 1.00

• Weights – Sets the weighing factors to be used for all components of the transit network when determining the best path.

• Transfer Penalty Time Weight Factor...... 1.00 • Dwell Time Weight Factor ....................... 1.00 • Interarrival Parameter ............................. 0.15

• Other – Sets the headway, transfer, dwelling and layover time parameter, and sets minimum and maximum times for waiting, access, egress, and travel times.

• Use Park and Ride...................................... NO

Time of Day An important design feature of the Genesee County model is a time of day disaggregation. The daily trips generated and distributed in the model will be disaggregated into periods for traffic assignment. The

Many transit network settings can be specified at the route level, at the mode level, or globally for all the routes in the network. Route-level values come from a field in the route layer, mode level values come

10-5

Chapter 10

Traffic Assignment

Definition of TOD Periods

trips are disaggregated by direction meaning that factor is applied to the daily trip table to get trips in the P to A and A to P directions. The result is a better representation of peak flows, and flows by direction in congested areas. The MI Travel Counts database was utilized in the following ways:

Using the MI Travel Counts dataset for all TMA trips, a frequency distribution was calculated by departure hour for each trip purpose. For more details on TOD model refer to Chapter 8: Time of Day Model.

• Defined TOD Periods; • Defined TOD Factors (percent of trip in period by purpose); and, • Directional factors for each trip purpose by period.

10-6

Chapter 11

Model Validation/Calibration

Chapter 11: Model Validation/Calibration Introduction

The Genesee County travel demand model was vali‐ dated to replicate the observed traffic count data. Model and network parameters were adjusted so that final RMSE and volume/count summaries met MDOT guidelines. Trip lengths and trip distribution parameters were adjusted during the trip distribu‐ tion model calibration to ensure that the modeled trip lengths for each trip purpose were comparable to the observed trip lengths from household surveys. Mode choice constants also were adjusted to match observed highway, transit and non‐motorized shares.

Model folder should be saved directly under the “C:\” drive, and the root folder (“GeneseeTDM”) should contain subfolders and files as explained in the model’s users guide. The Reports subfolder in‐ cludes a few .xml reports (GeneseeLog.xml and Gen‐ eseeModel.xml) and .xsl files (log.xsl and report.xsl).

This model estimates travel by time‐of‐day (TOD). Model volumes are compared to observed counts by TOD and daily totals, but the validation is based on daily comparisons. Also, this comparison was per‐ formed at the network, area type, functional class, cutline, screenline, volume group and network link levels at a minimum. The daily model RMSE is at 31.57%. When evaluating % RMSE for groups of links disaggregated by volume ranges, relatively large er‐ rors are acceptable for low volume groups. But, the errors should become smaller as volume increases. The idea behind the RMSE guidelines is to ensure that model‐estimated volumes are within one lane of actual roadway needs. For the daily model, all valida‐ tion results fall within the MDOT target ranges. More details on validation and calibration are explained in the following sections of this chapter. The new Genesee County model incorporates Cal‐ iper’s standard User Interface and standardized structure to organize model files. The Genesee

GeneseeLog.xml and GeneseeModel.xml can be re‐ named to save the old outputs and allow the current reports to open faster. For example, the files could be renamed to GeneseeLog2.xml and Gene‐ seeModel2.xml after each model run. The model then will create new report files to contain the re‐ ports from subsequent model runs. All the validation and calibration results, which are generated after every model run, are stored in the Reports subfolder and mainly GeneseeModel.xml. This file can be opened using Internet Explorer. Once opened, sev‐ eral hyperlinks are displayed as shown in Fig‐ ure 11‐1. These links will direct the user to different model run results. The user can click on each link to view the details. The following sections of this chap‐ ter will discuss the main calibration and validation re‐ sults that can be found in the report file.

Trip Distribution Summary The Genesee County Trip distribution model is ap‐ plied for each TOD period and trip purpose. Table 11‐ 1 summarizes Trip Distribution model results. The distribution model matches the target observed trip length frequency distribution well (see Figures 6‐1 through 6‐4 of Chapter 6: Trip Distribution). Aver‐ age trip lengths are summarized again in Table 11‐2.

11‐1

Chapter 11

Model Validation/Calibration

Figure 11‐1: Model Report File

Table 11‐1: Average Trips Length Survey Versus Model (using updated times) Survey Model

HBWLO HBWHI 17.79 23.742 17.31

Average Travel Time by Purpose (Updated Skim Minutes) HBO HBSCH HBS HBU NHBW NHBO EIW 15.919 16.129 15.597 20.131 12.911 17.044 20.36 13.60 14.42 18.95

11‐2

IEW 20.84 19.45

ENW 24.62 22.34

Chapter 11

Model Validation/Calibration

Table 11‐2: Trip Distribution Summary Report (Time and Distance) Trip Distribution Report (TIME & DISTANCE) AM Purpose Trips

ATL(min) ATL(mi) %Intra Trips 2.07

ATL(min) ATL(mi) %Intra Trips

Daily

ATL(min) ATL(mi) %Intra Trips

ATL(min) ATL(mi)

HBW

32,502

17.182

7.575

20,874

17.6

7.628

2.074

27,567

18.019

7.673

2.074

32,502

16.646

7.517

2.077

113,444

17.309

HBNW

58,267

13.443

4.806

4.725 270,086

13.765

4.823

4.722 161,777

14.064

4.837

4.707 195,367

13.098

4.79

4.749

685,497

13.618

7.592 4.816

NHB

32,982

14.154

5.312

3.179 293,172

14.513

5.336

3.175 152,694

14.857

5.356

3.165 131,928

13.772

5.298

3.19

610,776

14.42

5.331

EIW

11,324

18.817

12.156

8,070

19.151

12.206

11,007

19.565

12.258

11,852

18.352

12.08

42,253

18.945

12.171

IEW

11,341

19.395

12.447

8,094

19.701

12.541

10,607

20.09

12.615

14,432

18.874

12.386

44,474

19.447

12.484

ENW

4,371

22.095

14.628

50,261

22.577

14.679

24,038

23.053

14.7

30,594

21.429

14.456

109,263

22.341

14.619

Total

150,786

15.506

6.925

2.967 650,557

15.047

6.093

3.458 387,690

15.536

6.278

3.358 416,673

14.549

6.344

3.399 1,605,707

15.079

6.281

Mode Choice Outputs

The mode choice model was calibrated for each time period. Table 11‐3 summarizes the mode choice tar‐ gets versus model results, showing that the model does a good job of replicating mode shares estimated

from the survey and MTA. Tables 11‐4 through 11‐8 show the validation reports for the mode choice model for AM, Midday, PM, Off‐peak periods as well as the daily results by trip purpose.

Table 11‐3: Target and Modeled Mode Shares

HBW

TARGET 72.48% 19.85% 1.88% 3.52% 2.24% 99.98%

MODEL 72.13% 19.68% 1.99% 3.8% 2.4% 100%

HBO

TARGET 82.33% 0.86% 12.13% 4.67% 99.99%

MODEL 81.63% 0.80% 12.66% 4.91% 100%

NHB

TARGET 82.33% 0.86% 12.13% 4.67% 99.99%

MODEL 82.6% 0.8% 12.12% 4.49% 100%

DA SR Transit Walk Bike TOTAL

Auto Transit Walk Bike TOTAL

11‐3

Chapter 11

Model Validation/Calibration

Table 11‐4: Mode Choice Summary – AM Peak

Mode Drive Alone Shared Ride Transit Pedestrian Bike TOTAL Mode Auto Transit Pedestrian Bike TOTAL Mode Auto Transit Pedestrian Bike TOTAL Mode Auto Transit Pedestrian Bike TOTAL

Total Vehicles 23,429 2,526 ‐‐ ‐‐ ‐‐ 25,955 Total Vehicles 33,734 ‐‐ ‐‐ ‐‐ 33,734

Total Vehicles 20,097 ‐‐ ‐‐ ‐‐ 20,097 Total Vehicles 79,786 ‐‐ ‐‐ ‐‐ 79,786

MODE CHOICE SUMMARY AM HBW Percent Total Percent Vehicles Persons Persons 90.27 23,429 72.09 9.73 6,391 19.66 ‐‐ 670 2.06 ‐‐ 1,232 3.79 ‐‐ 779 2.40 100 32,502 100 MODE CHOICE SUMMARY AM HBO Percent Total Percent Vehicles Persons Persons 100 47,565 81.63 ‐‐ 484 0.83 ‐‐ 7,361 12.63 ‐‐ 2,857 4.90 100 58,267 100 MODE CHOICE SUMMARY AM NHB Percent Total Percent Vehicles Persons Persons 100 27,131 82.57 ‐‐ 277 0.84 ‐‐ 3,975 12.10 ‐‐ 1,474 4.49 100 32,858 100 MODE CHOICE SUMMARY AM TOTAL Percent Total Percent Vehicles Persons Persons 100 104,516 84.54 ‐‐ 1,432 1.16 ‐‐ 12,569 10.17 ‐‐ 5,110 4.13 100 123,627 100

11‐4

Auto Occupancy AM HBW 1.15

Auto Occupancy AM HBO 1.41

Auto Occupancy AM NHB 1.35

Auto Occupancy AM Total 1.31

Chapter 11

Model Validation/Calibration

Table 11‐5: Mode Choice Summary – Midday Peak

MODE CHOICE SUMMARY MD HBW Total Vehicles

Mode Drive Alone Shared Ride Transit Pedestrian Bike TOTAL

Percent Vehicles

Total Persons

Percent Persons

15,052

90.27

15,052

72.11

1,623

9.73

4,106

19.67

‐‐ ‐‐ ‐‐ 16,675

‐‐ ‐‐ ‐‐ 100

421 794 501 20,874

2.02 3.80 2.40 100

Auto Occupancy MD HBW 1.15

MODE CHOICE SUMMARY MD HBO

Total Vehicles

Mode Auto Transit Pedestrian Bike TOTAL

Percent Vehicles

156,308 ‐‐ ‐‐ ‐‐ 156,308

100 ‐‐ ‐‐ ‐‐ 100

Total Persons

Percent Persons

220,394 2,198 34,220 13,273 270,086

Auto Occupancy MD HBO

81.60 0.81 12.67 4.91 100

1.41

MODE CHOICE SUMMARY MD NHB

Total Vehicles

Mode Auto Transit Pedestrian Bike TOTAL

Percent Vehicles

178,559 ‐‐ ‐‐ ‐‐ 178,559

100 ‐‐ ‐‐ ‐‐ 100

Total Persons

Percent Persons

241,055 2,423 35,458 13,136 292,072

Auto Occupancy MD NHB

82.53 0.83 12.14 4.50 100

1.35

MODE CHOICE SUMMARY MD TOTAL

Mode Auto Transit Pedestrian Bike TOTAL

Total Vehi‐ cles

Percent Vehicles

351,542 ‐‐ ‐‐ ‐‐ 351,542

100 ‐‐ ‐‐ ‐‐ 100

Total Persons 480,607 5,043 70,472 26,910 583,031

11‐5

Percent Persons 82.43 0.86 12.09 4.62 100

Auto Occupancy MD Total 1.37

Chapter 11

Model Validation/Calibration

Table 11‐6: Mode Choice Summary – PM Peak

MODE CHOICE SUMMARY PM HBW Mode Drive Alone Shared Ride Transit Pedestrian Bike TOTAL

Total Vehicles

Percent Vehicles

19,880 2,143 ‐‐ ‐‐ ‐‐ 22,023

Total Persons

90.27 9.73 ‐‐ ‐‐ ‐‐ 100

Percent Persons

19,880 5,423 550 1,051 663 27,567

72.12 19.67 2.00 3.81 2.40 100

Auto Occupancy PM HBW 1.15

MODE CHOICE SUMMARY PM HBO Mode Auto Transit Pedestrian Bike TOTAL

Total Vehicles

Percent Vehicles

93,606 ‐‐ ‐‐ ‐‐ 93,606

100 ‐‐ ‐‐ ‐‐ 100

Total Persons

Percent Persons

131,984 1,308 20,530 7,955 161,777

81.58 0.81 12.69 4.92 100

Auto Occupancy PM HBO 1.41

MODE CHOICE SUMMARY PM NHB Mode Auto Transit Pedestrian Bike TOTAL

Total Vehicles

Percent Vehicles

92,976 ‐‐ ‐‐ ‐‐ 92,976

100 ‐‐ ‐‐ ‐‐ 100

Total Persons

Percent Persons

125,517 1,252 18,505 6,847 152,121

82.51 0.82 12.16 4.50 100

Auto Occupancy PM NHB 1.35

MODE CHOICE SUMMARY PM TOTAL Mode Auto Transit Pedestrian Bike TOTAL

Total Vehicles 208,605 ‐‐ ‐‐ ‐‐ 208,605

Percent Vehicles 100 ‐‐ ‐‐ ‐‐ 100

Total Persons 282,804 3,111 40,086 15,464 341,465

11‐6

Percent Persons 82.82 0.91 11.74 4.53 100

Auto Occupancy PM Total 1.36

Chapter 11

Model Validation/Calibration

Table 11‐7: Mode Choice Summary – Night

Total Vehicles

Mode Drive Alone Shared Ride Transit Pedestrian Bike TOTAL

23,463 2,530 ‐‐ ‐‐ ‐‐ 25,993 Total Vehicles

Mode Auto Transit Pedestrian Bike TOTAL

113,228 ‐‐ ‐‐ ‐‐ 113,228 Total Vehicles

Mode Auto Transit Pedestrian Bike TOTAL

Mode

Auto Transit Pedestrian Bike TOTAL

80,503 ‐‐ ‐‐ ‐‐ 80,503

Total Vehicles 219,724 ‐‐ ‐‐ ‐‐ 219,724

MODE CHOICE SUMMARY OP HBW Percent Total Percent Vehicles Persons Persons 90.27 9.73 ‐‐ ‐‐ ‐‐ 100

23,463 6,401 627 1,231 779 32,502

72.19 19.69 1.93 3.79 2.40 100

MODE CHOICE SUMMARY OP HBO Percent Total Percent Vehicles Persons Persons 100 ‐‐ ‐‐ ‐‐ 100

159,651 1,494 24,647 9,575 195,367

108,679 995 15,865 5,893 131,432

298,194 3,117 41,743 16,246 359,301

11‐7

Auto Occupancy OP HBO 1.41

Auto Occupancy OP NHB

82.69 0.76 12.07 4.48 100

MODE CHOICE SUMMARY OP TOTAL Percent Total Percent Vehicles Persons Persons 100 ‐‐ ‐‐ ‐‐ 100

1.15

81.72 0.76 12.62 4.90 100

MODE CHOICE SUMMARY OP NHB Percent Total Percent Vehicles Persons Persons 100 ‐‐ ‐‐ ‐‐ 100

Auto Occupancy OP HBW

82.99 0.87 11.62 4.52 100

1.35

Auto Occupancy OP Total 1.36

Chapter 11

Model Validation/Calibration

Table 11‐8: Mode Choice Summary – Daily

MODE CHOICE SUMMARY ‐ DAILY HBW Mode Drive Alone Shared Ride Transit Pedestrian Bike TOTAL

Total Vehicles

Percent Vehicles

81,824 8,823 0 0 0 90,646

Total Persons

90.27 9.73 0 0 0 100

81,824 22,321 2,268 4,309 2,722 113,444

Percent Persons 72.13 19.68 2.00 3.80 2.40 100

Auto Occupancy 1.15

MODE CHOICE SUMMARY ‐ DAILY HBO Mode Auto Transit Pedestrian Bike TOTAL

Total Vehicles

Percent Vehicles

396,875 0 0 0 396,875

Total Persons

100 0 0 0 100

559,594 5,485 86,758 33,659 685,497

Percent Persons 81.63 0.80 12.66 4.91 100

Auto Occupancy 1.41

MODE CHOICE SUMMARY ‐ DAILY NHB Mode Auto Transit Pedestrian Bike TOTAL

Total Vehicles 372,135 0 0 0 372,135

Percent Vehicles

Total Persons

100 0 0 0 100

502,383 4,948 73,803 27,349 608,483

Percent Persons 82.56 0.81 12.13 4.49 100

Auto Occupancy 1.35

MODE CHOICE SUMMARY ‐ DAILY Mode Auto Transit Pedestrian Bike TOTAL

Total Vehicles 859,657 ‐‐ ‐‐ ‐‐ 859,657

Percent Vehicles

Total Persons

100 ‐‐ ‐‐ ‐‐ 100

1,166,121 12,702 164,870 63,731 1,407,424

11‐8

Percent Persons 82.86 0.90 11.71 4.53 100

Auto Occupancy 1.36

Chapter 11

Model Validation/Calibration

Highway Assignment

Calibration ensures accuracy of the travel forecast model. The basic goal in calibrating the model is to verify assigned traffic volumes replicate counted vol‐ umes on the actual roadway network for the same year. This is an iterative process in which assigned volumes were compared to actual ground counts. If results do not correspond to counted volumes, ad‐ justments were made to the model in several ways.

MDOT developed statistical guidelines to measure the level of calibration of TDM’s in Michigan. The Genesee County model was adjusted to better meet the MDOT guidelines. Tables 11‐9 and 11‐10 present the model estimated daily VMT versus observed daily VMT by area type and facility type and compare it against MDOT validation guidelines. The area type and NFC guidelines are met in every case.

Region‐wide issues might indicate need for adjusting trip rates in the trip generation model. Over or un‐ der‐assignments in specific locations might require reviewing and adjusting socio‐economic data at the TAZ level or including a special generator in the trip generation process. Problems at the link level might indicate need to adjust capacity, speed or other at‐ tributes of links with problems. This iterative process was repeated until assigned volumes replicate ground counts for the region.

When calibrating a model, it is important to strike a balance between matching traffic counts while main‐ taining the model’s forecasting power. Most model‐ ers acknowledge that it is more important for the model to provide logical relationships between travel supply and demand than it is to very closely match traffic counts. In fact, excessive model adjust‐ ments made just to match traffic counts may de‐ grade the model’s forecasting power. It is also im‐ portant to note that traffic count data are subject to

Table 11‐9: Count and VMT Comparison by Area Type AREA TYPE

COUNT

CBD (1)

FLOW

COUNT COUNT VMT FLOW VMT VMT RATIO VMT % DIFF. TARGET(%) #LINKS RATIO

158,603

130,444

0.822

26,437

27,767

1.05

5.00%

+/‐10%

Urban (2)

3,167,768

2,971,525

0.938

1,337,767

1,266,178

0.946

‐5.40%

+/‐10%

231

Suburban (3)

2,534,011

2,504,345

0.988

2,440,731

2,504,156

1.026

2.60%

+/‐10%

242

Fringe (4)

400,541

367,852

0.918

339,076

310,508

0.916

‐8.40%

+/‐10%

Rural (5)

720,893

757,072

1.05

787,690

832,676

1.057

5.70%

+/‐10%

119

6,981,816

6,731,239

0.964

4,931,701

4,941,285

1.002

‐0.20%

+/‐5%

671

VMT TARGET VMT % DIFF. RATIO (%)

# LINKS

TOTAL

Table 11‐10: Count and VMT Comparison by Functional Class (NFC) FUNCTIONAL CLASS (NFC) Interstate (1)

COUNT

FLOW

1,844,640 1,929,073

COUNT RATIO

COUNT VMT FLOW VMT

1.046

2,194,208

2,290,718

1.044

4.40%

+/‐6%

257,692

1.126

385,919

438,875

1.137

13.70%

+/‐6%

Oth Princ Arterial (3)

2,012,736 1,902,274

0.945

698,079

689,750

0.988

‐1.20%

+/‐7%

130

Minor arterial (4)

2,277,548 2,089,707

0.918

1,214,527

1,108,654

0.913

‐8.70%

+/‐10%

278

Oth Fwy Exp (2)

228,876

Major Collector (5)

268,255

221,461

0.826

240,229

207,619

0.864

‐13.60%

+/‐20%

Minor Collector (6)

334,094

315,031

0.943

178,702

184,449

1.032

3.20%

+/‐20%

TOTAL 6,966,149 6,715,237

0.964

4,911,664

4,920,065

1.002

‐0.20%

+/‐5%

669

11‐9

Chapter 11

Model Validation/Calibration

error, just as a model is subject to error. Variations in counts on the order of plus or minus 15% due only to ordinary daily variations are not uncommon. Traffic incidents, construction, special events, weather con‐ ditions, and other conditions may create even greater variations.

Screen line/Cutline Analysis An additional measure of model accuracy is a review of screen lines and cutlines in the study area model. A screen line is an imaginary line drawn through the region which divides the study area into two parts. The volume of traffic crossing that line in the model is then compared to actual ground counts. This gives a broad indication of whether overall volumes in the model are approaching correct values for major trip movements in the region. Eight screen lines were de‐ veloped for the Genesee County region. Additionally, cutlines were developed for major corridors in the region. Cutlines are short screen lines which give a more accurate indication of whether volumes in a specific corridor are within acceptable parameters. Four corridor cutlines were developed. Locations for screen lines and cutlines are shown in Figure 11‐2 and 11‐3. Table 11‐12 presents the count vs. volume comparison for all the screen lines. Table 11‐13 shows this comparison between the cutline corri‐ dors. More detailed information on each screenline and cutline can be found in the model report .xml file.

The Genesee County model achieves the required balance, providing logical relationships between traffic supply and demand, while matching traffic counts in in nearly every situation. This is shown in the comparison of link‐based %RMSE, tabulations by area type and facility type, and measures of model volumes versus counts at screenlines, cutlines, and in corridors. Based on a combination of these measures, it is clear that the Genesee County model is calibrated. Table 11‐11 presents the percent RSME comparison of daily VMT by volume group for the Genesee County travel demand model. Targets by volume group are met for all volume groups.

Table 11‐11: Percent RMSE by Volume Group COUNT RANGE

FLOW

COUNT RATIO

COUNT % DIFF

CALCULATED % RMSE

MDOT TARGET % RMSE

#LINKS

<1,000

12,661

13,150

1.04

3.863

72.38%

< 100%

1,000 ‐ 2,500

125,297

151,074

1.21

20.57

67.08%

< 100%

463,531

2,500 ‐ 5,000

480,146

1.04

3.585

47.12%

< 100%

124

5,000 ‐ 10,000

1,361,541 1,368,181

1.01

0.488

35.87%

< 45%

190

10,000 ‐ 25,000

3,478,245 3,198,186

0.92

‐8.052

27.34%

< 30%

224

25,000 ‐ 50,000

1,439,041 1,421,972

0.99

‐1.186

17.91%

< 25%

98,529

0.97

‐2.927

4.23%

< 20%

6,981,816 6,731,238

0.96

3.59

31.57%

<40%

671

> 50,000 all

COUNT

101,500

COUNT VMT =

4,931,701

FLOW VMT = FLOW VMT/COUNT VMT = Percent Difference

4,941,285

1.002 0.20% MDOT Standard (<5%)

11‐10

Chapter 11

Model Validation/Calibration

Figure 11‐2: Screen line Locations in Genesee County Travel Demand Model

11‐11

Chapter 11

Model Validation/Calibration

Figure 11‐3: Cutline Locations in Genesee County Travel Demand Model

11‐12

Chapter 11

Model Validation/Calibration

Table 11‐12: Summary of the Screenline Counts vs. Volumes NUMBER

NAME

# LINKS

VOLUME

COUNT

RATIO

Irish Road

107,969

90,734

1.19

Elms Hogan

122,668

122,398

1.00

Pierson Road

236,102

214,685

1.10

Hill Road

172,720

179,009

0.96

Ray Road

90,114

75,813

1.19

CBD Cordon

239,697

235,579

1.02

Flint River

170,873

173,072

0.99

External Cordon

302,971

302,980

1.00

Table 11‐13: Volume and Count Comparison for the Cutline Corridors NUMBER

NAME

# LINKS

VOLUME

COUNT

RATIO

I‐475

395,512

365,900

1.08

I‐69

685,327

636,890

1.08

I‐75

862,651

854,150

1.01

US‐23

257,692

228,876

1.13

Other model validation results can be found in the model report file (GeneseeModel.XML). This file in‐ cludes all the results, including the time taken by each model run step as well as other details on each step of the four‐step Travel Demand Model. For ex‐ ample, Table 11‐14 shows the model feedback re‐

port by time of day. As expected, the RMSE values decrease after each feedback iteration, and after the third iteration, link volumes change only slightly from the preceding iteration, indicating that travel times and volumes are consistent throughout the model.

Table 11‐14: Model Feedback Report by Time‐of‐Day ITERATION

AM RMSE%

MD RMSE%

PM RMSE%

179.72

170.57

173.35

188.43

1.01

3.16

2.89

0.45

3.16

0.36

1.04

0.97

0.16

1.04

0.21

0.53

0.45

0.08

0.53

11‐13

OP RMSE%

MAX RMSE%

Chapter 11

Model Validation/Calibration

Level of Service Analysis

One of the post processing steps in the model run is to generate Level of Service values for each link in addition to the V/C ratios and some other variables. This step calculates three major LOS values.

areas and 0.88 is assumed in rural areas. The peak hour volume is assumed to be 60.9% of the AM period loading or 34.53% of the PM period loading. The directional split from the model for the peak period is used.

HCM_LOS – LOS Based on Highway Capacity Manual 2010. LOS for freeways, express‐ ways and rural multilane highways is deter‐ mined by peak period flow density in terms of passenger cars per lane per mile. For, rural two‐lane roads and highways, level of ser‐ vice is determined by estimated vehicle headway percent of free flow speed. For ur‐ ban streets, level of service is determined by average speed alone. For all facility types, a peak hour factor of 0.92 is assumed in urban

2. LOSBYVC – LOS based on peak hour V/C ratio as shown in Table 11‐15.

3. LOS_CMS – Genesee County Congestion Management System Level of Service method, based on 24‐hr V/C ratio as shown in Table 11‐15. These values are added to the scenario network after the post processor runs. Table 11‐15 includes a de‐ scription of these fields.

Table 11‐15: Level of Service Related Fields Calculated Using the Model Post Processors FIELD MAXVC1 DAYVC AVGVC1 PKVC AVGSP AVGTT1 HCM_LOS

LOSBYVC

LOS_CMS

HRS_DELAY VHT VMT PKFLOW PKSPD PKFD LPTC LDELAY

DESCRIPTION Maximum of VC ratio over time periods Daily VC ratio (Total daily volume over total daily capacity) Average VC ratio PM Peak Hour VC ratio Average Daily Speed Average Travel Time Level of Service‐using HCM 2010 flow density method Level of Service‐using volume to capacity ratio thresholds A, if Peak VC ratio (PKVC) <= 0.26 B, if 0.26 < PKVC < 0.43 C, if 0.43 < PKVC < 0.62 D, if 0.62 < PKVC < 0.82 E, if 0.82 < PKVC < 1.00 F, if 1.00 < PKVC Congestion Management System Level of Service A, if Daily VC ratio (DAYVC) <= 0.44 B, if 0.44 < DAYVC < 0.67 C, if 0.67 < DAYVC < 0.85 D, if 0.85 < DAYVC < 1.00 E, if 1.00 < DAYVC < 1.25 F, if 1.25 < DAYVC Total Vehicle Hours of Delay Vehicle Hours Traveled Vehicle Miles Traveled Peak Hour Flow Peak Hour Speed Flow Density Percent Vehicle Hours of Delay = 100 * linkdelay (LDELAY) / VHT Link Delay

11‐14

Chapter 11

Model Validation/Calibration

Transit Ridership

Another output of this model is the Transit Ridership values by different TOD periods as well as total daily ridership for each route. The ridership values can be compared to the on‐board transit surveys for the base year. The transit boarding and alighting values

are reported by TOD in the scenario transit layer. However, a final daily ridership summary is also added to the report file (GeneseeModel.xml). Ta‐ ble 11‐16 shows the observed vs. modeled TOD rid‐ ership by route for the base year.

Table 11‐16: Transit Ridership Comparisons (Model vs. on Board Surveys) for the Base Year AM ROUTE_NAME

DAILY

SUR‐ VEY

MODE L

SUR‐ VEY

MODE L

SUR‐ VEY

MODE L

SUR‐ VEY

MODE L

SUR‐ VEY

MODEL

Lapeer Road_Outbound

191

531

354

230

219

250

188

1,202

855

Lapeer Road_Inbound

128

354

228

306

137

249

146

1,037

572

Lewis‐Selby_Outbound

129

111

348

194

Lapeer Rd_Inbound

128

354

173

306

113

249

1,037

405

ML King Avenue_Outbound

134

372

230

298

147

140

843

576

ML King Avenue_Inbound

134

372

254

298

152

138

843

620

Dupont_Inbound

119

318

308

244

188

198

169

879

752

Dupont_Outbound

119

318

319

244

195

198

184

879

783

Crosstown North_Outbound

155

170

104

332

414

Crosstown North_Inbound

155

216

134

139

332

534

Miller‐Linden_Outbound

191

627

696

396

409

276

379

1,394

1,675

Miller‐Linden_Inbound

159

627

651

396

426

276

363

1,394

1,598

Beecher‐Corunna_Out‐ bound

126

100

387

398

236

245

246

225

995

967

Fenton Road_Inbound

110

275

362

184

210

195

203

729

885

Beecher‐Corunna_Inbound

126

106

387

368

236

221

246

995

913

South Saginaw_Outbound

142

270

166

217

578

South Saginaw_Inbound

107

187

Richfield Road_Inbound

112

260

383

176

232

168

210

670

937

Richfield Road_Outbound

125

260

455

176

278

168

237

670

1,094

Franklin_Inbound

164

387

274

304

155

180

146

1,035

663

Franklin_Outbound

164

387

230

304

146

180

558

162

180

135

1,035

Downtown‐Campus_Loop

420

178

Civic Park_Outbound

143

353

335

276

223

222

191

993

822

North Saginaw_Outbound

149

434

268

282

175

276

163

1,140

668

North Saginaw_Inbound

149

434

274

282

160

276

158

1,140

679

Civic Park_Inbound

143

101

353

323

276

184

222

173

993

781

Lewis‐Selby_Inbound

129

111

348

215

South Saginaw_Inbound

246

410

166

226

102

375

594

1,046

South Saginaw_Outbound

246

191

166

119

102

344

594

690

Fenton Road_Outbound

107

275

445

184

281

195

256

729

1,089

3,195

2,463

9,203

8,848

6,570

5,443

4,828

5,174

23,790

21,928

TOTAL

11‐15

Chapter 11

Model Validation/Calibration

Trip Generation

The next sections describe the trip generation model. Please note that the same survey was used for this model as the preceding model. A review of the derivation of the trip generation model showed that it was valid, and thus it was not necessary to re‐ peat the statistical analysis. Thus, much of the expla‐ nation below was taken from earlier reports, and re‐ peated here for the reader’s convenience. The Gen‐ esee County Travel Demand Model was last updated in 2008. The following sections on trip production and attraction rate calibration are taken from the re‐ port prepared for the GCMPC by Wilbur Smith Asso‐ ciates in July 2008.

The survey stratification of households by trip pur‐ pose is as follows:  Household Vehicles (0 to 3+) x Household Size (1 to 4+): o HBSH: Home Based Shopping; o HBO: Home Based Other; o NHBO: Non‐Home Based Other; o HBSC1: Home Based School – K‐12; and, o HBSCU: Home Based School – College.  Household Vehicles (0 to 3+) x Household Workers (0 to 3+): o HBW; and, o NHBW.

Trip Production Rate Calibration As explained in Chapter 5: Trip Generation, all households in the entire state of Michigan were in‐ cluded in calibrating the trip generation model. Using the combined Transportation Management Area (TMA) dataset, trip production rates were calibrated for all trip purposes. Trip productions are calculated using a cross classification methodology based on the two‐dimensional stratification of households.

The trip records in the MI Travel Counts were joined to household records in the survey database. For each household, the household size, workers and ve‐ hicles owned were allocated to one of the four bins. Trip production rates were calculated by aggregating the total households in each cell and total trips by trip purpose made in the same cell. The trip rate is the average rate of all households, also estimated by dividing the total trips by total households in the cell.

Table 11‐17: Work Cross Classification Scheme

0 WORKERS

0 Vehicles

1 Vehicle

2 Vehicles 2

3+ Vehicles

1 WORKER

2 WORKERS

3+ WORKERS

Table 11‐18: Non‐Work Cross Classification Scheme

1 PERSON HH

0 Vehicles

1 Vehicle

2 PERSON HH

4+ PERSON HH

2 Vehicles 2 3+ Vehicles

3 PERSON HH

11‐16

Chapter 11

Model Validation/Calibration

Cell Compression Based on an analysis of the data, some cells were combined and averaged, as shown below. The rationale for combining cells is due to the num‐ ber of observations and households in each cell of the matrix. Typically, a minimum of 30 households is required for the resulting trip rate to be considered statistically significant. Additionally, based on the data, the variation in the variation between the cells is larger than the variation between the cells, indica‐ tion that differences in the trip rates are not signifi‐ cant. Home Based Work Income Stratification A design feature of the trip generation model is the stratification of home based work trips by income. The advantage to this approach is that it better links the income of the trip makers to the types of jobs they are typically going to. In models without this approach, high income homes often generate work trips that are attracted to neighboring low income employment rather than the more distant high in‐ come jobs. This will eliminate that potential error in the model. A quartile approach was first investigated where households were placed in one of four income cate‐ gories based on their reported household income from the MI Travel Counts. The quartiles were es‐ tablished using an analysis of the CTPP Household In‐ come Distribution. The median household income for Genesee County was found to be between 40,000 and 42,500 dollars (1999 dollars). The quartiles were then defined based on finding the median income in the lower and upper halves as defined by the county median. Following are the CTPP Income Categories that correspond to the limits of each quartile.    

Quartile 1: 0 to 22,500; Quartile 2: 22,500 to 42,500; Quartile 3: 42,500 to 100,000; and, Quartile 4: > 100,000.

Based on the CTPP quartile definitions, each house‐ hold record in the MI Travel Counts database was as‐ signed an income quartile based on the reported household income in the survey. The stratification of Home Based Work by income quartile were then evaluated based on the following criteria:  Statistically significant difference in trip rates between income quartiles; and,  Sufficient number of households in each cell. Number of Observations Table 11‐19 provides the distribution of the number of households included in the TMA dataset as part of the MI Travel Counts when the samples are stratified by number of vehicles per household, workers per household and income quartile. It becomes appar‐ ent that there are insufficient numbers of samples to accurately estimate production rates when using four income categories. The last two sections of the table show the number of observations when quar‐ tiles one and two are combined and three and four are combined. The results show that some cells still have insufficient numbers, but overall the distribu‐ tion is reasonable. Using two income categories will allow for the model to still recognize the advantages discussed above. Statistical Difference in Trip Rates When considering the four income quartiles, there is significant difference between the average HBW trip rates made between households in each quartile. The following is the average trip rate by quartile when weighting each household by the survey ex‐ pansion factor.    

11‐17

Quartile 1: 0.50 Trips / HH; Quartile 2: 1.07 Trips / HH; Quartile 3: 1.75 Trips / HH; and, Quartile 4: 1.95 Trips / HH.

Chapter 11

Model Validation/Calibration

Table 11‐19: Number of Sampled Households (Work Related)

VEH / HH VEH / HH VEH / HH VEH / HH VEH / HH VEH / HH

INC Q1 0 1 2 3 INC Q2 0 1 2 3 INC Q3 0 1 2 3 INC Q4 0 1 2 3 INC Q1/2 0 1 2 3 INC Q3/4 0 1 2 3

WORKERS PER HOUSEHOLD 0 1 2 71 23 74 66 8 12 16 2 2 2 0 0 1 21 6 1 113 190 36 57 96 68 23 41 29 0 0 1 4 1 1 17 65 4 22 84 132 5 47 65 0 0 1 1 11 20 3 13 69 178 2 33 104 0 1 2 92 29 1 187 256 44 69 112 70 25 43 29 0 0 1 4 2 1 28 85 7 35 153 310 7 80 169

11‐18

3 2 1 2 2 10 14 2 7 43 2 9 51 3 0 2 12 15 2 0 0 16 94

Chapter 11

Model Validation/Calibration

When an Analysis of Variance (ANOVA) is performed using the quartiles, the averages are statistically dif‐ ferent. Because of the limitations in sample size in each cell when using the quartiles, the difference be‐ tween the upper and lower quartiles was compared using ANOVA. The results showed that there was a statistical difference between average trip rates made by the upper and lower income households.

low income and high income groupings based on above and below the regional average income.

Recommendation

Tables 11‐20 through 11‐27 present the calibrated trip production rates using the cell compression and income stratifications described above. The trip pro‐ duction rates were calibrated using the combined TMA samples from the MI Travel Counts and were weighted by the expansion factors calculated by MORPACE. Because the MI Travel Counts survey covered a two‐day period, trip rates were calculated based on the two‐day period and then multiplied by 0.50 to create an average trip rate for each day.

Based on the review of the sample size and statistical analysis of trip rates, HBW trips were estimated for a

 Quartile 1 and 2: 0.92; and,  Quartile 3 and 4: 1.85.

Table 11‐20: HBW Low Income Trip Production Rates (Work Related)

0 1 2 3

VEHICLES PER HOUSEHOLD

WORKERS PER HOUSEHOLD 1 2 1.188 0.852 1.791 0.726 1.971 0.792 2.387

3 2.250 1.000 TOTAL

0.350

Table 11‐21: HBW High Income Trip Production Rates (Work Related)

0 1 2 3

VEHICLES PER HOUSEHOLD

WORKERS PER HOUSEHOLD 1 2 0.000 0.554 0.000 0.644 1.846 0.737 1.622

3 3.000 2.524 TOTAL

0.594

Table 11‐22: NHBW Trip Production Rates (Work Related)

0 1 2 3

VEHICLES PER HOUSEHOLD

WORKERS PER HOUSEHOLD 1 2 0.067 0.507 0.409 0.304 1.271 0.330 1.022

11‐19

3 0.400 1.195 TOTAL

0.483

Chapter 11

Model Validation/Calibration

Table 11‐23: HBO Trip Production Rates (Non‐work Related)

VEHICLES PER HOUSEHOLD

1 0.569 1.046

0 1 2 3

HOUSEHOLD SIZE 2 3 1.907 1.727 1.752 1.904 2.694 2.465

4 3.724 3.096 TOTAL

1.947

Table 11‐24: HBSH Trip Production Rates (Non‐work Related)

VEHICLES PER HOUSEHOLD

1 0.351 0.431

0 1 2 3

HOUSEHOLD SIZE 2 3 0.771 0.740 0.570 0.788 0.872 0.261

4 1.127 0.311 TOTAL

0.708

Table 11‐25: NHBO Trip Production Rates (Non‐work Related)

VEHICLES PER HOUSEHOLD

1 0.130 0.578

0 1 2 3

HOUSEHOLD SIZE 2 3 1.612 1.564 1.928 1.838 2.452 1.732

4 3.945 2.734 TOTAL

1.744

Table 11‐26: HBSC (K‐12) Trip Production Rates (School Trips)

0 1 2 3

VEHICLES PER HOUSEHOLD

1 0.000 0.028

HOUSEHOLD SIZE 2 3 1.715 0.397 3.180 0.131 1.373 0.838

4 3.534 2.951 TOTAL

1.169

Table 11‐27: HBSC (College ‐ Univ.) Trip Production Rates (School Trips)

0 1 2 3

VEHICLES PER HOUSEHOLD

1 0.000 0.000

HOUSEHOLD SIZE 2 3 0.111 0.000 0.042 0.020 0.011 0.218

11‐20

4 0.097 0.244 TOTAL 0.050

Chapter 11

Model Validation/Calibration

Combined Trip Rates Table 11‐28 represents a combined average trip rate for all purposes.

Table 11‐28: Aggregated Average Daily Production Rate HHVH

HBW

HBO

HBSH

NHBO

NHBW

HBSC(K12)

HBSC(U)

4,901

18,511

8,635

12,299

333

11,771

762

16,387

31,045

76,454

29,835

65,073

15,724

53,982

632

53,548

54,426

123,807

42,966

124,593

33,124

68,109

1,808

48,115

42,463

67,228

22,548

54,246

22,062

37,871

4,179

28,854

TOTAL

132,835

285,999

103,984

256,211

71,243

171,732

7,382

146,904

HHVH

HBW

HBO

HBSH

NHBO

NHBW

HBSC(K12)

HBSC(U)

TOTAL

0.03

0.13

0.06

0.08

0.00

0.08

0.01

0.39

0.21

0.52

0.20

0.44

0.11

0.37

0.00

1.86

0.37

0.84

0.29

0.85

0.23

0.46

0.01

3.06

0.29

0.46

0.15

0.37

0.15

0.26

0.03

1.71

TOTAL

0.90

1.95

0.71

1.74

0.48

1.17

0.05

7.01

Trip Attraction Rate Calibration

The attractions for each trip purpose are calculated using a linear regression model that was calibrated using the MI Travel Counts database with records specific to internal trips made within Genesee County.

The regions were defined to be consistent with polit‐ ical boundaries of the smaller communities and to combine areas of homogenous land uses. Table 11‐ 29 provides a brief description of each district.

District Aggregation Initial assessment of the survey data showed that zonal level data did not provide sufficient sample size to construct statistically significant regression mod‐ els. Thus, the data was combined into district level so that the models can be built on more samples in each level. Figure 11‐4 displays the regions used in the analysis.

Using the assigned origin and destination TAZ for each trip record, the attraction end of the trip was assigned to one of the 83 districts. For home‐based trips, the attraction end is the non‐home end of the trip including the work, shopping or school end. For non‐home based work trips, the work and non‐work ends were both assigned to a district. Trip ends for both ends of the non‐home based other trips were combined.

11‐21

Chapter 11

Model Validation/Calibration

Figure 11‐4: District Map

11‐22

Chapter 11

Model Validation/Calibration

Table 11‐29: District Description DISTRICT Argentine Twp‐RURAL‐RESIDENTIAL Argentine Twp‐SUBURBAN‐RESIDENTIAL Atlas Twp‐RURAL‐RESIDENTIAL Burton City‐SHOP Burton City‐SUBURBAN‐RESIDENTIAL Burton City‐URBAN‐RESIDENTIAL Central Flint City‐URBAN Clayton Twp‐RURAL‐RESIDENTIAL Clayton Twp‐SUBURBAN‐RESIDENTIAL Clio City‐URBAN‐RESIDENTIAL Davison City‐COMMERCIAL Davison City‐SHOP Davison City‐URBAN‐RESIDENTIAL Davison Twp‐RURAL‐RESIDENTIAL Davison Twp‐SUBURBAN‐RESIDENTIAL E Flint City‐URBAN Fenton City‐COMMERCIAL Fenton City‐INDUST Fenton City‐SHOP Fenton City‐SUBURBAN‐RESIDENTIAL Fenton City‐URBAN‐RESIDENTIAL Fenton Twp‐RURAL‐RESIDENTIAL Fenton Twp‐SUBURBAN‐RESIDENTIAL Flint City‐CBD‐COMMERCIAL Flint City‐COMMERCIAL Flint City‐INDUST Flint City‐SHOP Flint Twp‐COMMERCIAL Flint Twp‐INDUST Flint Twp‐SHOP Flint Twp‐SUBURBAN‐RESIDENTIAL Flint Twp‐URBAN‐RESIDENTIAL Flushing City‐COMMERCIAL Flushing City‐SHOP Flushing City‐SUBURBAN‐RESIDENTIAL Flushing City‐URBAN‐RESIDENTIAL Flushing Twp‐RURAL‐RESIDENTIAL Flushing Twp‐SUBURBAN‐RESIDENTIAL Forest Twp‐RURAL‐RESIDENTIAL Gaines Twp‐RURAL‐RESIDENTIAL Gaines Twp‐SUBURBAN‐RESIDENTIAL Genesee Twp‐RURAL‐RESIDENTIAL

DISTRICT # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42

DISTRICT Genesee Twp‐SUBURBAN‐RESIDENTIAL Genesee Twp‐URBAN‐RESIDENTIAL Goodrich Village‐SUBURBAN‐RESIDENTIAL Grand Blanc City‐COMMERCIAL Grand Blanc City‐INDUST Grand Blanc City‐SHOP Grand Blanc City‐SUBURBAN‐RESIDENTIAL Grand Blanc City‐URBAN‐RESIDENTIAL Grand Blanc Twp‐COMMERCIAL Grand Blanc Twp‐INDUST Grand Blanc Twp‐RURAL‐RESIDENTIAL Grand Blanc Twp‐SHOP Grand Blanc Twp‐SUBURBAN‐RESIDENTIAL Grand Blanc Twp‐URBAN‐RESIDENTIAL Linden City‐SUBURBAN‐RESIDENTIAL Montrose City‐SUBURBAN‐RESIDENTIAL Montrose Twp‐RURAL‐RESIDENTIAL Mt Morris City‐URBAN‐RESIDENTIAL Mt Morris Twp‐RURAL‐RESIDENTIAL Mt Morris Twp‐SUBURBAN‐RESIDENTIAL Mt Morris Twp‐URBAN‐RESIDENTIAL Mundy Twp‐COMMERCIAL Mundy Twp‐RURAL‐RESIDENTIAL Mundy Twp‐SHOP Mundy Twp‐SUBURBAN‐RESIDENTIAL Mundy Twp‐URBAN‐RESIDENTIAL NE Flint City‐URBAN NW Flint City‐URBAN Otisville Village‐SUBURBAN‐RESIDENTIAL Richfield Twp‐RURAL‐RESIDENTIAL Richfield Twp‐SUBURBAN‐RESIDENTIAL S Flint City‐URBAN Swartz Creek City‐INDUST Swartz Creek City‐SUBURBAN‐RESIDENTIAL Swartz Creek City‐URBAN‐RESIDENTIAL Thetford Twp‐RURAL‐RESIDENTIAL Thetford Twp‐SUBURBAN‐RESIDENTIAL Thetford Twp‐URBAN‐RESIDENTIAL Vienna Twp‐RURAL‐RESIDENTIAL Vienna Twp‐SUBURBAN‐RESIDENTIAL W Flint City‐URBAN

11‐23

DISTRICT # 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83

Chapter 11

Model Validation/Calibration

Model Estimation

4. The model selection process was not solely dependent on one statistics such as Adjusted R Square. Rather the process was based on combinational effects of the above statistics. For example, a model’s R Squared would in‐ crease as more independent variables are added, but it does not necessarily imply that the model is getting better. The perfor‐ mance of each of the entered variables need to be checked.

This section was also borrowed from the previous model update report by Wilbur Smith Associates in July 2008.The following logical steps were followed to develop the attraction equations: 1. The correlation between surveyed attrac‐ tions and available socioeconomic variables was investigated. The Pearson Correlation and the 2‐Tailed Level of Significance were examined. Supplemental to these statistics, nonparametric correlations such as Kendall’s tau_b and Spearman’s rho were also com‐ pared. From this analysis, significantly cor‐ related variables with attractions were se‐ lected as a pool of candidates for independ‐ ent variables. 2. Since the analysis involved numerous combi‐ nations of many socioeconomic variables, to be efficient, a stepwise regression technique was used. The stepwise technique is appro‐ priate to deal with multiple explanatory var‐ iables. In implementing the stepwise tech‐ nique, no constants were forced during the analysis since the model without a constant produced better result in most cases. 3. Regression results were analyzed for the fol‐ lowing main statistics: a. Adjusted R Square b. Overall model F‐statistics and its sig‐ nificance level c. Model coefficients (magnitude and signs) d. t‐statistics for each of entered varia‐ bles and its significance level e. Multicollinearity among entered variables

5. Besides the above statistics, logical judg‐ ments were made for appropriateness of each variable. For example, one shows sta‐ tistically significant, thus it is natural to in‐ clude the variable in the model since it im‐ proves the model. However, the variable may not make a logical connection to trip at‐ tractions for specific trip purpose. In this case, it was decided that the variable does not have reasonable explanatory power and the variable was subsequently removed from the model even though it sacrificed the model performance. Correlation Analysis As described above, SPSS was used to calculate the correlation between the attractions for each trip pur‐ pose to the socioeconomic variables in each district. The detailed employment variables, as well as the to‐ tal employment were used. In addition, total house‐ hold was tested as a variable. For the home based school purpose, k‐12 enrollment was not tested but was used as the independent variable in the regres‐ sion analysis. The results of the correlation analysis are shown in Table 11‐30. The pool of potential vari‐ ables used in the Step‐Wise Regression Analysis was based on these results.

11‐24

Chapter 11

Model Validation/Calibration

Table 11‐30: Correlation Analysis of Observed Trip Ends HBW_Z VARIABLE TOTAL MANUF OTHER TRANSP FINC RETAIL WHOLES SERV GOV HH VARIABLE TOTAL MANUF OTHER TRANSP FINC RETAIL WHOLES SERV GOV HH

PEARSON SIG. CORRELATION (2‐TAILED) 0.825 9.83063E‐22 0.349 0.001206609 0.534 1.97513E‐07 0.390 0.00027187 0.492 2.25871E‐06 0.707 8.28991E‐14 0.443 2.6906E‐05 0.819 2.83839E‐21 0.291 0.007639842 0.642 6.3666E‐11 NHBO PEARSON SIG. CORRELATION (2‐TAILED) 0.836 8.09512E‐23 0.400 0.000178715 0.680 1.57424E‐12 0.270 0.013603047 0.469 7.84888E‐06 0.822 1.70307E‐21 0.491 2.4492E‐06 0.829 3.91063E‐22 0.153 0.168352035 0.794 3.71824E‐19

HBO_Z PEARSON CORRELATION 0.830 0.392 0.677 0.223 0.495 0.779 0.469 0.821 0.188 0.801

SIG. (2‐TAILED) 3.32349E‐22 0.000248532 2.17671E‐12 0.04285578 1.99994E‐06 4.48302E‐18 7.80357E‐06 2.16369E‐21 0.087967695 1.0335E‐19

NHBW_W PEARSON SIG. CORRELATION (2‐TAILED) 0.779 4.22651E‐18 0.323 0.002906852 0.621 3.74632E‐10 0.442 2.91837E‐05 0.447 2.2494E‐05 0.717 2.56601E‐14 0.472 6.53312E‐06 0.776 6.90774E‐18 0.195 0.077496333 0.651 2.78481E‐11

HBSH_Z PEARSON SIG. CORRELATION (2‐TAILED) 0.681 1.34745E‐12 0.304 0.005162387 0.528 2.96917E‐07 0.235 0.032758661 0.378 0.000428053 0.835 1.06834E‐22 0.406 0.000142811 0.652 2.42493E‐11 0.091 0.412728639 0.651 2.7962E‐11 NHBW_O PEARSON SIG. CORRELATION (2‐TAILED) 0.804 6.12084E‐20 0.381 0.000386895 0.672 3.37598E‐12 0.332 0.002170822 0.421 7.31097E‐05 0.699 2.00946E‐13 0.484 3.46734E‐06 0.799 1.29755E‐19 0.201 0.068466087 0.779 4.40691E‐18

Regression Analysis Once the variables for use in Regression Analysis were selected based on the correlation analysis, Step‐Wise Regression was used to determine the best model for each trip purpose. As discussed above, R squared was used as one selection variable. The final model was selected based on a combination of R squared, logical variables and reasonableness of the coefficients. In Table 11‐31, the NHBO Coefficients are scaled to estimate the total number of trips ends which in

cludes both the production and attraction ends of the trip. In application, the NHBO attraction model is used to estimate only the attraction end of the trip. Thus, in application, the coefficients for the Genesee County model should be reduced by 0.50. The model was then applied to the district aggre‐ gated SE data. The comparison between the survey expanded trip ends by purpose are shown in Table 11‐32 and Figure 11‐5.

11‐25

Chapter 11

Model Validation/Calibration

Table 11‐31: Trip Attraction Step‐Wise Regression Results TOTAL MANUF OTHER TRANSP FINC RETAIL WHOLES SERV GOV HH K12 U R‐squared

HBW 0.590 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.898

HBO 0.000 0.000 0.000 0.000 0.000 3.069 0.000 0.961 0.000 0.624 0.000 0.000 0.931

HBSH 0.000 0.000 0.000 0.000 0.000 3.403 0.000 0.000 0.000 0.000 0.000 0.000 0.887

NHBO 0.000 0.000 0.000 0.000 0.000 7.567 0.000 1.499 0.000 0.797 0.505 0.000 0.938

NHBW_W 0.309 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.869

NHBW_O 0.000 0.000 0.000 0.000 0.000 0.504 0.000 0.336 0.000 0.171 0.000 0.000 0.893

HBSC1 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 1.838 0.000 0.887

HBSCU 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.209 0.900

Table 11‐32: Observed vs. Modeled Attractions by Purpose PURPOSE HBW HBO HBSH NHBO NHBW_W NHBW_O HBSC K12 HBSC U

OBSERVED 132,692 285,570 103,984 511,766 70,706 70,706 171,732 7,382

MODELED 127,183 390,102 137,800 514,775 68,840 68,840 240,390 8,193

Figure 11‐5: Total Observed vs. Modeled Attractions by Purpose

11‐26

Chapter 11

Model Validation/Calibration

Income Distribution of HBW Attractions HBW trips are treated differently in the Genesee County model. For all other purposes, the produc‐ tions and attractions are based on all households. HBW separates the low and high income households into separate sub purposes. Each sub purpose has a unique set of production rates. To estimate attrac‐ tions for the two sub‐purposes, different approaches were considered: 1. The first method is to calibrate unique re‐ gression models treating HBW low and HBW high as unique purposes. 2. The second method is based on the income distribution of industry type and apply the distribution to the zonal employment. Result

is that in each zone there is a distribution of low and high income HBW attractions based on the mix of zonal employment. The first approach was considered, but because of the limited number of HBW records in the Genesee County MI Travel Counts, it was felt that this method would be unreliable. The second approach was ap‐ plied to the Genesee County model. The source of the income distribution by employment type was borrowed from Toledo, Ohio. It was felt that Toledo and Flint share a similar economic base that applying the distribution was reasonable. Figure 11‐6 shows the distribution of employees in each industry type based on the household income broken down into the low and high categories.

Figure 11‐6: Distribution of Employees by Household Income Distribution of Employees by Income 0.9

0.8

0.7

Percent

0.6

0.5 Low High 0.4

0.3

0.2

0.1

0 MANUF

OTHER

TRANSP

FINC

RETAIL

WHOLES

SERV

GOV

Industry

11‐27

Chapter 11

Model Validation/Calibration

The calculation of each HBW sub purpose attractions is preformed using the following model:

HBW Coefficient = Regression equation rate of HBW trips per total employment;

HBW(L) = HBW Coefficient * EMP (1) * Percent Low (1) + … + HBW Coefficient * EMP (n) * Percent Low (n)

EMP(1) = Zonal Employment Category 1; and, Percent Low (1) = Percentage of employment in Category 1 that is low income.

where: The coefficients to be applied by income group for the zonal employment categories are presented in Table 11‐33.

HBW(L) = HBW low income;

Table 11‐33: Percent Distribution of Employment Type INCOME

EMPLOYMENT MANUF

OTHER

TRANSP

FINC

RETAIL

WHOLES

SERV

GOV

Low

0.247

0.238

0.23

0.29

0.41

0.23

0.355

0.19

High

0.753

0.762

0.77

0.71

0.59

0.77

0.645

0.81

External Models

o Productions assigned at external station as a percent of total volumes based on MI Travel Counts and CTPP JTW data; and,

External–internal trips represent the interaction Genesee County has with the surrounding region. As shown in Figure 11‐7, Genesee County is a bedroom community to several neighboring counties and at‐ tracts trips into the region for other purposes. Thus, three separate external – internal trip purposes were defined:

o Attractions estimated at internal zones as function of HBW attractions.  IE_Work:

 NonWork: o Production assigned at external station as a percent of total volume; and,

o Represents the outbound work com‐ mute made by residents inside Genesee County;

o Attractions estimated at internal zones as function of HBO and HBSH attrac‐ tions.

o Productions estimated to internal TAZs as a function of HBW Productions using MI Travel Count data; and,

 EI_Work: o Represents the inbound work commute made by residents outside of Genesee County;

o Attractions assigned to external cordon as a percentage of total outbound traf‐ fic.

11‐28

Chapter 11

Model Validation/Calibration

Figure 11‐7: External Regions to Genesee County

Distribution of Trip Purposes

Based on the above stratification of external to inter‐ nal movements, the MI Travel Counts database was analyzed to find records for all trips that had one trip end within Genesee County. In total, approximately nineteen hundred records were found which in‐ cluded travel on both days of the survey period. The first step in defining the distribution of trips entering

Genesee County was to define travel sheds for the major entry points into the county. The process was simplified by splitting the state into five regions: north, east, west, south, and southwest. The remain‐ ing Michigan counties were put into one of the five categories depending on its proximity to Genesee County and the major route between the counties. The regions are shown in Figure 11‐7.

11‐29

Chapter 11

Model Validation/Calibration

The northern region accesses Genesee County via In‐ terstate 75. Eastern and western counties assume In‐ terstate 69 as the predominate corridor. The south‐ ern region is the SEMCOG region which uses Inter‐ state 75. Southwestern counties are along the US 23 corridor.

would be 35% of the inbound traffic (100 / (100 + 362/2)). The outbound work flows (work outside of Genesee County is 20% (44 / (44 + 362/2)). This was done based on the MI travel counts in 2005. We as‐ sumed that this is still valid for the new model devel‐ opment effort.

Once each county was assigned a region, the trips in the travel survey were defined as work and non‐work based on the activity codes for each record. This ap‐ proach is similar to that used for the trip purpose definitions described above.

Average Trip Length and Distribution of Trip Ends Based on the internal end of the trip in Genesee County, and the external station, the average trip length was calculated for the inbound, outbound work trips and non‐work trips by region. An external zone was selected for each region.

Table 11‐32 reports the number of records in the da‐ tabase disaggregated based on inbound, and out‐ bound work and total non‐work related trips by re‐ gion. The volumes shown represent the total of the two days in the survey. The inbound trip refers to a trip made into the county for work and outbound re‐ fers to an outbound commute (works outside Gene‐ see County).

    

The model travel time was then used from the entry point to the TAZ. The average values are reported in Table 11‐35.

In application, the volumes in Table 11‐34 were con‐ verted to percentages to apply to existing travel counts to develop the control totals for each flow type entering the region. To do this it was necessary to assume that the inbound and outbound distribu‐ tion of non‐work trips is equal. So, for northern entry points, the inbound (work in Genesee County) flow

Table 11‐35 indicates small variations between the average trip length inside the model area by entry region. This is supported by the even distribution of locations of trip ends across the model area as seen in Figures 11‐8 and 11‐9.

Table 11‐34: MI Travel Counts Two Day External Counts

Table 11‐35: Average Trip Length in Minutes

WORK INBOUND OUTBOUND ENTRY REGION (TO WORK) (TO WORK) NON‐WORK North 100 44 362 East 59 26 212 West 94 37 217 South 43 230 154 Southwest 64 114 170 TOTAL 360 451 1,115

North: I‐75; East: I‐69; West: I‐69; South: I‐75; and, Southwest: US 23.

WORK ENTRY REGION INBOUND OUTBOUND (TO WORK) (TO WORK) North 19.10 20.28 East 15.26 14.67 West 15.34 15.70 South 13.83 17.54 Southwest 18.45 20.96 TOTAL 16.74 18.35

11‐30

NON‐WORK 18.59 15.09 15.20 15.57 17.23 16.65

Chapter 11

Model Validation/Calibration

Figure 11‐8: Location of Work Trip Ends

11‐31

Chapter 11

Model Validation/Calibration

Figure 11‐9: Location of Non‐Work Trip Ends

11‐32

Chapter 11

Model Validation/Calibration

Model Sensitivity The Genesee County travel demand model shows ap‐ propriate sensitivity to changes in transportation supply and travel demand as shown in Table 11‐36. The 2014 base year validation scenario, and the al‐ ternative setups for 2020, 2025, 2035 and 2045 were used as the test cases. In all cases, the model responded to changes in supply and demand at expected levels. The only unusual ob‐ servation is the low growth in zonal data (population, households, and employment), and the expansion of transportation facilities (road miles and lane miles) when compared to other urban areas in the US. Changes in demand are governed by the growth in population, households, and employment. In Gene‐ see County, population decreases between 2014 and 2025, and then begins a slow increase between 2025 and 2045. Note that between 2014 and 2045, popu‐ lation tends to decrease in the central area, and in‐ crease in the rural area (Figure 11‐10). Consistent with trends throughout the US, household sizes (per‐ son per households) decrease slightly in the years

between 2014 and 2045, so the number of house‐ holds remains almost constant between 2014 and 2025, and begins a slow increase between 2025 and 2045. The number of trips forecasted by the model roughly follows the changes in population, but starts a more substantial increase after 2025 because of forecasted increases in employment. Vehicle miles of travel (VMT) forecasted by the model closely follows the forecast of the number of trips, and shows a small decrease of between 2014 and 2045. Estimated highway delay increases be‐ tween 2014 and 2045 as forecasted VMT increases and as the central area become less dense. Note that this is not a linear relationship, because delays occur only after existing roadway capacities are exceeded. As would be expected in an area without major in‐ vestments in expanded transit service, transit rid‐ ership is forecasted to be nearly constant between 2014 and 2045. These results show that the model is appropriately sen‐ sitive to changes in transportation supply and demand.

Table 11‐36: Model Sensitivity to Changes in Changes in Transportation Supply and Demand MEASURE INPUTS

Population Households Employment Road Miles (non‐centroid) Lane Miles (Thru) MODEL RESULTS Internal +IE/EI person trips VMT Transit trips (linked) Estimated delay (daily hrs) CALCULATED VALUES Persons/Household Jobs/Household Trips/person Trips/Household Trips/Employee Average Link Load VMT/Person

2014

2020

YEAR 2025

412,899 165,199 191,484 1,062 2,643

405,550 166,254 209,887 1,064 2,645

402,263 166,929 214,685 1,064 2,645

402,688 169,340 217,610 1,064 2,645

407,869 172,223 221,643 1,064 2,645

1,605,707 10,753,449 12,637 16,174

1,600,905 9,863,788 12,406 16,225

1,599,018 9,933,049 12,193 16,736

1,607,773 10,022,876 11,928 17,311

1,633,779 10,174,328 11,918 18,236

2.50 1.16 3.89 9.72 8.39 10,126 26.04

2.44 1.26 3.95 9.63 7.63 9,270 24.32

2.41 1.29 3.98 9.58 7.45 9,336 24.69

11‐33

2035

2.38 1.29 3.99 9.49 7.39 9,420 24.89

2045

2.37 1.29 4.01 9.49 7.37 9,562 24.95

Chapter 11

Model Validation/Calibration

Figure 11‐10: Population Changes 2014‐2045

11‐34

Genesee Travel Demand Model

Final Report

APPENDIX A

Appendix A

Network Development

Capacity Calculation Methodology

Introduction1

Michigan Department of Transportation (MDOT) adopted a capacity calculator method as part of the Urban Model Improvement Program (UMIP). MDOT methodology considers a simplified capacity calcula‐ tion process by limiting the number of adjustments for roadway characteristics on a link by link basis and instead uses a lookup table to estimate highway ca‐ pacity based on link facility type and area type. The existing lookup table adopted by MDOT is based on Highway Capacity Manual (HCM) 2000. The Trans‐ portation Research Board (TRB) published the fifth edition of the Highway Capacity Manual (HCM 2010), an update to HCM 2000, which proposes updated ca‐ pacity calculation guidelines based on the latest re‐ search on highway capacity and quality of service.

Table A‐ 1: Recommended Facility Types Facility Type

The purpose of this document is to revise MDOT’s look‐up table capacities based on the methodology proposed in the updated HCM 2010. This document summarizes MDOT’s capacity calculator methodol‐ ogy and the process used to develop the new capac‐ ities using HCM 2010.

MDOT’s Capacity Calculator

Freeway

High Speed Ramps

On‐Ramps

Off‐Ramps

Principal Arterial w/TWLTL

Principal Arterial

One‐Way Minor Arterial

Minor Arterial w/TWLTL

Minor Arterial

One‐Way Collector

Collector w/ TWLTL

Collector

Local Roadway w/ TWLTL

Local Road

Centroid Connector

Trunk Principal Arterial w/TWLTL

Trunk Principal Arterial

Table A‐ 2: Recommended Area Types Area Type

The MDOT capacity calculator utilizes a look‐up table to estimate highway link capacity based on facility type and area type. Tables A‐1 and A‐2 show the fa‐ cility types (Link types) and area types recommended as part of this methodology. 1

This appendix is part of MDOT’s paper on their capacity calcu‐ lator methodology which was also adopted for Tri‐county Re‐ gional Planning Commission (TCRPC) travel demand model.

Code

A‐1

Code

High Density Urban Commercial

High Speed Ramps

Suburban

Fringe

Rural

Appendix A

Network Development

The HCM analysis details three area types (urban, suburban, and rural) compared to the five area types used in the Genesee County Metropolitan Planning Committee (GCMPC) model (HDUC, urban, subur‐ ban, fringe, and rural). HCM’s urban was assumed to be MDOTs urban, HCM’s suburban as MDOT’s subur‐ ban, and HCM’s rural as MDOTs rural. A set of default values were developed to account for parameters re‐ quired in the HCM 2000 capacity calculations. Capac‐ ities were calculated for freeway, principal arterial, minor arterial, collectors, and local roads. Capacities for the other two area types (HDUC and Fringe) and additional facility types were calculated using the fol‐ lowing assumptions: 

The capacities calculated using HCM 2000 were rounded to the nearest 50.



Capacities for the Fringe area type were calcu‐ lated as ((Rural –Suburban)/2) + Rural, rounded to the nearest 50.



Capacities for the HDUC area type were calcu‐ lated as Urban Capacity – 50.



One‐way and TWLTL link capacities were in‐ creased 20% and 5%, respectively.

 Off ramp capacities were reduced by 400. Table A‐3 shows the MDOT adopted Level of Service E capacities. The highlighted cells were calculated us‐ ing HCM/HCS 2000 and the other values were esti‐ mated using the above mentioned assumptions.

Table A‐ 3: Recommended Model Network Base Capacities FACILITY TYPE

AREA TYPE HDUC =1

URBAN=2

FRINGE=4

RURAL=5

Freeway

1,950

2,000

2,100

2100

High Speed Ramps

1,700

On‐Ramps

1,200

1200

1,200

Off‐Ramps

800

Principal Arterial w/TWLTL

950

1,000

1100

1,450

1,900

Principal Arterial

900

950

1050

1,400

1,850

One‐Way Minor Arterial

850

900

950

1,350

1,800

Minor Arterial w/TWLTL

750

800

850

1,200

1,550

Minor Arterial

700

750

800

1,150

1,500

One‐Way Collector

650

700

750

Collector w/ TWLTL

600

650

700

Collector

550

600

650

Local Roadway w/ TWLTL

500

Local Road

450

Centroid Connector

10,000

SUBURBAN=3

A‐2

Appendix A

Network Development

Further capacity adjustments were recommended to account for on‐street parking and lane width less than 12 feet. It was assumed that on‐street parking will reduce capacity by 5% for parking on one side and 10% for parking on both sides. Lane width less the 12 feet resulted in a capacity reduction of 4%.

BFFS = Base Free‐Flow Speed (mph; the default is 75.4)

Recommended/ Updated Capacity Calculator

Table A‐4 provides adjustments to free‐flow speed as a function of lane width.

The purpose of this effort is to revise MDOT’s capac‐ ity look up table based on the updated HCM 2010 methodology. The capacities are calculated for the highlighted cells in Table A‐3 and the remaining ca‐ pacities are estimated using the assumptions from MDOT’s capacity calculator. Once the base capacities are calculated, further adjustment to capacities is recommended on a link by link basis for on‐street parking and subpar lane width.

fLW

= Lane Width Adjustment (mph)

fLC

= Lateral Clearance Adjustment (mph)

TRD

= Total Ramp Density (ramps per mile)

Table A‐ 4: Adjustment to Free S for Average Lane Width AVERAGE LANE WIDTH (FT.)

REDUCTION IN FFS, FLW (MPH)

≥ 12

0.0

≥ 11 ‐ 12

1.9

≥ 10 ‐ 11

6.6

Source: HCM 2010, Exhibit 11‐8

The following sections from HCM 2010 are used to calculate the updated capacities.     

Basic Freeway Segments Freeway Merge and Diverge Segments (Ramps) Freeway Weaving Segments Multilane Highways Urban Streets

Freeway Facilities This section of the document discusses capacity cal‐ culations for freeway facilities; basic freeway seg‐ ments, ramps, and auxiliary lanes.

A default lane width of 12 feet is assumed in the above formula. Adjustment for substandard lane width is made at a later stage on a link‐by‐link basis. Table A‐5 shows the adjustments to free‐flow speed based on the right‐side lateral clearance. The right‐ side lateral clearance is measured from the right edge of the travel lane to the nearest lateral obstruc‐ tion. A default lateral clearance of 6 feet is assumed for all area types in the GCMPC model. Table A‐ 5: Adjustments to Free Flow Speed for Right‐ Side Lateral Clearance, fLC (mph) LANES IN ONE DIRECTION 2 0.0 0.6 1.2 1.8 2.4 3.0 3.6 Source: HCM 2010 Exhibit 11‐9 RIGHT‐SIDE LATERAL CLEAR‐ ANCE (FT.) ≥6 5 4 3 2 1 0

Basic Freeway Segment The capacity for a basic freeway segment is based on the segment free‐flow speed. The free‐flow speed can be estimated using the following formula, rec‐ ommended in HCM 2010: BFFS 3.22 ∗ ………………………HCM 2010 Eq. 11‐1

where:

A‐3

3 0.0 0.4 0.8 1.2 1.6 2.0 2.4

4 0.0 0.2 0.4 0.6 0.8 1.0 1.2

≥5 0.0 0.1 0.2 0.3 0.4 0.5 0.6

Appendix A

Network Development

Table A‐6 shows the total ramp density which is de‐ fined as the number of ramps (on and off ramps) lo‐ cated 3 miles north and 3 miles south of the study location divided by 6. Ramp density is used in capac‐ ity calculation to measure the impact of merging and diverging vehicles on free‐flow speed. The GCMPC model network was analyzed to develop the follow‐ ing default values for ramp density based on area type.

Table A‐ 6: Total Ramp Density Initial Default Values AREA TYPE

TOTAL RAMP DENSITY, TRD (RAMPS/MI)

Rural

0.4

Suburban

Urban

the freeway mainline. The free‐flow speed for free‐ way to freeway ramps is assumed to be ¾ times the mainline free‐flow speed. Service interchange ramps connect freeways to lower function class roadways (primarily principal ar‐ terials). The capacity for on‐ramp is based on its junc‐ tion with the freeway mainline. Off‐ramp capacity for a service interchange, on the other hand, is based on the traffic control at the end of the ramp. The free flow speed for freeway‐to‐arterial ramps is assumed to be 35 mph. The capacities for system interchange ramps and service interchange on‐ramp can be esti‐ mated using Table A‐8 provided in HCM 2010. Table A‐ 8: Ramp Roadway Capacities

Freeway capacities for the Level of Service E are esti‐ mated using the Table A‐7 from HCM 2010. Table A‐ 7: Basic Freeway Segment Capacities FREE‐FLOW SPEED (MI/H)

TARGET LEVEL OF SERVICE A

820

1,310

1,750

2,110

2,400

770

1,250

1,690

2,080

2,400

710

1,170

1,630

2,030

2,350

660

1,080

1,560

2,010

2,300

600

990

1,430

1,900

2,250

RAMP FFS

SINGLE‐LANE RAMPS

TWO‐LANE RAMPS

> 50

2,200

4,400

> 40 ‐ 50

2,100

4,200

> 30 ‐ 40

2,000

4,000

> 20 ‐ 30

1,900

3,800

< 20

1,800

3,600

Source: HCM 2010 Ex. 13‐10

Capacity for an off‐ramp depends on the traffic con‐ trol at the end of the ramp. When calculating off‐ ramp capacities, the downstream traffic control is as‐ sumed to be signalized. The off‐street ramp capacity is estimated using the following simplified signalized intersection capacity equation1.

Source: HCM 2010 Chapter 11

Ramps Freeway ramps can be categorized as system inter‐ change (freeway to freeway) or service interchange (freeway to arterial). HCM 2010 recommends ramp roadway capacities based on the ramp free‐flow speed (Exhibit 13‐10). Freeway to freeway ramps are high speed ramps and the ramp capacities are based on its junction with

………………………HCM 2010 Eq. 18‐15

Where: N

= Number of lanes

= Adjusted saturation flow rate (veh/h/ln)

= Green‐to‐cycle length ratio

The number of lanes at the end of the ramp is as‐ sumed to be two; one for right turn and one for left turn. The saturation flow rate is assumed to be 1,900

A‐4

Appendix A

Network Development

Passenger Cars per Hour per Lane (PCPHPL). A g/c ra‐ tio of 0.38 is assumed. The capacity calculation for signalized intersections is discussed later in this ap‐ pendix. Auxiliary Lanes Auxiliary lanes are used to connect freeway on‐ ramps to the downstream off‐ramp in order to re‐ duce the impact of merging and diverging traffic on through traffic. Auxiliary or weaving segments allow the merging and diverging traffic to cross each other (weave) and are usually not long enough to be con‐ sidered an independent lane. HCM 2010 recom‐ mends using one of the following two methods to calculate the weaving segment capacity. Method 1: Weaving segment capacity determined by density 438.2 1 119.8

0.0765

………..HCM 2010 Eq. 12‐5

Where: CIWL = Capacity of the weaving segment under equivalent ideal conditions per lane (pc/hr/ln) CIFL = Capacity of basic freeway segment VR = Volume ratio; weaving volume/ total vol‐ ume LS = Length of the weaving segment NWL = Number of lanes from which a weaving maneuver may be made with one or more lane changed (usually 2)

∗

…………HCM 2010 Eq. 12‐6

Where: Cw = Capacity of the weaving segment in vehicles per hour

fHV = Adjustment factor for heavy‐vehicle pres‐ ence fp = Adjustment factor for driver population Method 2: Weaving segment capacity determined by weaving demand flows (NWL=2) ……………HCM 2010 Eq. 12‐7 ∗

∗

…………………HCM 2010 Eq. 12‐8

Where: Cw = Capacity of all lanes in the weaving segment Cw = Capacity of the weaving segment fHV = Adjustment factor for heavy‐vehicle pres‐ ence fp = Adjustment factor for driver population In the above equations, the weaving and through freeway traffic needed to calculate the volume ratio is unknown. The Transportation Research Center (TRC) at the University of Florida conducted a re‐ search for Florida Department of Transportation (FDOT)2, to analyze the effect of auxiliary lanes on freeway segment volume throughput (capacity), by identifying the traffic volume level at which each level of service threshold density was met for condi‐ tions with and without the auxiliary lanes. Results from HCM 2010 freeway weaving analysis and COR‐ SIM micro simulation model were analyzed for 40 lo‐ cations in Florida and two adjustment equations were developed to estimate the percentage increase in volume throughput due to an auxiliary lane. The research identified that additional capacity realized due to an auxiliary lane primarily depend on the number of freeway lanes, and other factors (volume ratio, length of auxiliary lane) and have no effect on the percentage increase in the volume throughput.

2Transportation Research Center, University of Florida, “Investi‐

gation of Freeway Capacity: a) Effect of Auxiliary Lanes on Free‐

way Segment Volume Throughput, and b) Freeway Segment Ca‐ pacity Estimation for Florida Freeways”, Florida Department of Transportation.

A‐5

Appendix A

Network Development

Table A‐9 shows the average percentage increase in capacity by adding an auxiliary lane(s).

Table A‐ 10: Base Free Flow Speed for Arterial Segments Area Type/Facility Type

Table A‐ 9: Average Percent Increase in Volume by Add‐ ing an Auxiliary Lane Number of Through Lanes

Percentage Increase in Volume

48.87

32.03

23.81

18.71

Multilane Highways Multilane highways are defined as four‐lane road‐ ways either divided by a median or undivided. The capacities for the following facility and area types in the GCMPC model are estimated using the multilane highway methodology from HCM 2010.  

Rural Principal Arterial Rural Minor Arterial

For multilane highway segments, the free‐flow speed is estimated using the following formula:

…………………….HCM 2010 Eq. 14‐1

= Base Free‐Flow Speed (mph)

fLW

= Lane with Adjustment (mph)

= Total Lateral Clearance Adjustment fLC (mph) fM

= Median type adjustment (mph)

= Access‐point density adjustment (mph)

The base free‐flow speed for the GCMPC model links are obtained from MDOT’s free‐flow speed table (UMIP Task 1.2). Table A‐10 shows the base free‐ flow speed used for rural arterials segments.

Rural Minor Arterial

A default lane width of 12 feet is assumed in the above formula. Adjustment for substandard lane width is made at a later stage on a link‐by‐link basis. Total Lateral Clearance (TLC) can be calculated as the lateral clearance from the right edge of the travel lanes to the roadside obstructions plus the lateral clearance from the left edge of the travel lanes to ob‐ structions in the roadway median. Lateral obstruc‐ tions farther than 6 feet from the edge of the travel lanes are not expected to have any effect on the free‐ flow speed. Table A‐11 shows the recommended ad‐ justments to the free‐flow speed for four‐lane high‐ ways based on the TLC. A default TLC value of 12 feet is assumed for all multilane highways in the GCMPC model. Table A‐ 11 Adjustment to Free‐Flow Speed for Lateral Clearance TLC (ft.) REDUCTION IN FFS, fLC (MPH)

= Base Free‐Flow Speed (mph)

BFFS

Rural Principal Arterial

Where: FFS

BFFS

12 10

0.0 0.4

0.9

1.3

1.8

2 0 Source: HCM 2010 Exhibit 14‐9

3.6 5.4

Table A‐12 shows the recommended free‐flow ad‐ justment based on the median type. Rural principal arterials are assumed to have a divided median. Ru‐ ral minor arterials are assumed to be undivided. This is consistent with the assumptions made in the exist‐ ing MDOT capacity calculator.

A‐6

Appendix A

Network Development

Table A‐ 12: Adjustment to Free‐Flow Speed for Median Type MEDIAN TYPE

REDUCTION IN FFS, FM (MPH)

Undivided

1.6

TWLTL

0.0

Divided

0.0

Table A‐ 15: Multilane Highway Segment Capacities FREE‐FLOW SPEED (MI/H) 60 55 50 45

Source: HCM 2010 Exhibit 14‐10

Access‐point density is defined as the number of in‐ tersections and driveways per mile on the right side of the highway influencing the traffic flow in the di‐ rection of travel. Table A‐13 presents the default ac‐ cess‐point density values recommended in HCM 2010. Table A‐14 shows the recommended adjust‐ ment to free‐flow speed based on access‐point den‐ sity. Table A‐ 13: Access‐Point Density Recommended De‐ fault Values AREA TYPE

ACCESS POINTS PER MILE

Rural

Low‐Density Suburban

High‐Density Suburban

Source: HCM 2010, Exhibit 14‐16

TARGET LEVEL OF SERVICE A

660 600 550 290

1,080 990 900 810

1,550 1,430 1,300 1,170

1,980 1,850 1,710 1,550

2,200 2,100 2,000 1,900

Source: HCM 2010 Exhibit 14‐17

Urban Streets Urban street segment methodology in HCM 2010 is used to calculate capacity for urban and suburban ar‐ terials, collectors, and local roads in the GCMPC model. Capacity for an urban street segment is di‐ rectly related to the capacity of the downstream in‐ tersection. The capacity calculation is simplified by assuming that the downstream intersection signal‐ ized for all links. Intersection spillback is ignored. The capacity of the network link can be calculated as a function of the number of through lanes and the capacity of the signalized intersection(s) along the link. The capacity at the signalized intersection can be calculated using the following formula. ………………………………… HCM 2010 Eq. 18‐15

Table A‐ 14: Adjustment to FFS for Access Point Density (fA)

Where: ACCESS‐POINT DENSITY (ACCESS POINTS/MILE) 0 8 10 16 20 25 30 ≥ 40

REDUCTION IN FFS, FA (MPH) 0.0 2.0 2.5 4.0 5.0 6.25 7.5 10.0

Source: HCM 2010, Exhibit 14‐11

The capacity for multilane highway links in the model can be estimated using the table A‐15 provided in HCM 2010.

= Number of lanes serving the movement

= Adjusted saturation flow rate (veh/h/ln)

= Green‐to‐cycle length ratio

The number of lanes (N) is an attribute available as part of the GCMPC network database. Adjustment to capacity due to turning movements at the intersec‐ tion is considered minimal and is ignored. Saturation flow for a lane group can be defined as the maximum number of vehicles that can pass

A‐7

Appendix A

Network Development

through the intersection during one hour of continu‐ ous green under prevailing traffic and roadway con‐ ditions. The base saturation flow rate is the ideal sat‐ uration flow rate under idyllic traffic and roadway conditions, which can then be adjusted for the pre‐ vailing conditions using adjustment factors.

fp = adjustment factor for existence of a parking lane and parking activity adjacent to the lane group fbb = adjustment factor for blocking ef‐ fect of local buses that stop within the intersection area

HCM 2010 (Exhibit 18‐28) suggests a base saturation flow rate of 1,900 passenger cars per hour per lane (pc/h/ln) for metropolitan areas with population greater than 250,000. For everywhere else, default saturation flow rate of 1,750 passenger cars per hour per lane is recommended.

The following equation shows the corrections made to the saturation flow rate to derive the adjusted sat‐ uration flow rate.

fRT = adjustment factor for right‐turn vehicle presence in a lane group

fRpb = pedestrian‐bicycle adjustment factor for right‐turn groups

………………………….HCM 2010 Eq. 18‐5

Where:

fLU zation

= adjustment factor for area type = adjustment factor for lane utili‐

fLT = adjustment factor for left‐turn vehicle presence in a lane group

= pedestrian adjustment factor for fLpb left‐turn groups

s (veh/h/ln)

= adjusted saturation flow rate

so (pc/h/ln)

= base saturation flow rate

The information required for most of the adjustment factors are not available in the GCMPC model net‐ work. Some link attributes such as lane width, park‐ ing, and % percent heavy vehicles are available.

= adjustment for lane width

= adjustment for heavy vehicles in fHV the traffic stream fg grade

= adjustment factor for approach

The existing MDOT capacity calculator assumes sat‐ uration flow rate based on facility type and area type to estimate link capacity (Table A‐16). These satura‐ tion flow rates are used to estimate the link capaci‐ ties.

A‐8

Appendix A

Network Development

Table A‐ 16: Saturation Flow Rate by Facility Type and Area Type AREA TYPE Urban Suburban Rural

PRINCIPAL ARTERIAL 1750 1900*

FACILITY TYPE MINOR ARTERIAL COLLECTOR 1700 1700 1800 1700 1750

LOCAL ROAD 1650 1650 1700

*Based on HCM 2010 recommendation.

The green to cycle length (g/c) ratio represents the effective portion of the cycle length that serves the given lane group. The g/c ratio is expected to reduce as the functional class decreases. Table A‐17 shows

a default set of g/c ratios are assumed based on the functional class. The default value for g/c ratios are adopted from the Kentucky Statewide Travel Model (KYSTM)3.

Table A‐ 17 Default g/c ratio FACILITY TYPE Principal Arterial Minor Arterial Collector Local Roads

g/c RATIO 0.45 0.42 0.38 0.31

The above procedures were used to calculate the highlighted cells in Table A‐18. Table A‐ 18: Updated Capacity Look‐Up Table (HCM 2010 Calculation) FAC TYPE 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Freeway High Speed Ramps On‐Ramps Off‐Ramps Principal Arterial w/TWLTL Principal Arterial One‐Way Minor Arterial Minor Arterial w/TWLTL Minor Arterial One‐Way Collector Collector w/ TWLTL Collector Local Roadway w/ TWLTL Local Road Centroid Connector

HDUC =1 2100 2000 1350

Urban=2 2350 2100 2000 1350

AREA TYPE Suburban=3 2400 2100 2000 1350

800

900

700

750

600

500

3 Lexington Area Metropolitan Planning Organization (LAMPO)

Mode Documentation – Appendix A: Speed and Capacity Calcu‐ lator.

A‐9

Fringe=4 2100 2000 1350

Rural=5 2400 2200 2000 1350 2200

1900

600 500

Appendix A

Network Development

 If a two‐way left turn lane is present, the link capacity was increased based on area type; 20% for HDUC, 15% for CBD, 10% for suburban, and 5 % for fringe and rural.

The remaining cells in the table are filled based on the following assumptions:  All capacities were rounded to the nearest 50.

 For Trunk Principal Arterials with or without TWLTL, capacities of Principal Arterials with or without TWLTL are increased by 10%, respec‐ tively.

 Capacities for the Fringe area type were calcu‐ lated as ((Rural –Suburban)/2) + Rural, rounded to the nearest 50.  Capacities for the HDUC area type were calcu‐ lated as urban capacity – 50.

The updated model network capacities based on HCM 2010 are presented in Table A‐19.

 One‐way capacities were increased by 20% Table A‐ 19: Recommended Model Network Base Capacities FAC TYPE 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

HDUC =1 2300 2100 2000 1350 900 750 800 800 650 700 700 550 550 450 10000 990 825

URBAN=2 2350 2100 2000 1350 950 800 850 850 700 750 700 600 600 500 10000 1045 880

FRINGE=4 2400 2100 2000 1350 1650 1550 1650 1450 1350 750 650 600 550 500 10000 1815 1705

RURAL=5 2400 2200 2000 1350 2350 2200 2300 2000 1900 750 650 600 550 500 10000 2585 2420

 On‐street parking – The capacity is recom‐ mended to be adjusted downwards 5% for on‐ street parking on one side of the street or 10% for on‐street on both sides of the street.

Adjustments to base capacities are recommended based on the following roadway characteristics on a link‐by‐link basis.  Auxiliary lanes ‐ Discussed earlier in this appen‐ dix.  Lane width – If the lane width is less than 12 feet, the link capacity is recommended to be re‐ duced by 4%.

AREA TYPE SUBURBAN=3 2400 2100 2000 1350 1000 900 900 850 750 750 700 600 550 500 10000 1100 990

A‐10

Genesee Travel Demand Model

Final Report

APPENDIX B

Genesee County Bridge Technical Report

2045 Population Projections Methodology Report MAP-21 Performance Measures…………………………………………………………..12

2005 Base Year Population Data

Prepared by the Genesee County Metropolitan Planning Commission

Genesee County Population Projections Methodology Report Base Year Data and Factors…………………………………….…………..………1 Methodology for 2045 Projections………………………………………………….1 Methodology for the Local Units of Government Outside the City of Flint...2 City of Flint Population Projection Methodology……………………………..…3 Other Factors…………………..……………………………………………………….4 Comparison to Other Data Sources…………………….…………………………5 Population Projection Assumptions………………………………………………..5 2045 Genesee County Population Projections…………………………………..7

List of Figures Graph of City of Flint Projected Percent Reduction in Households…………4

Appendix A

Methodology Examples

Appendix B

Recovery Factors and Supporting Census Data

Genesee County 2045 Population Projections Methodology Base Year Data and Factors TAZ Level Data

The population projections for Genesee County were produced on a traffic analysis zone (TAZ) level where growth/decline was calculated for each TAZ which can then be aggregated up to the municipality level for all cities, townships and some villages. Genesee County is divided into 639 TAZ. 2014 Census estimates were used to calibrate 2014 base year population and housing data. The distribution of population and housing from the 2010 Census redistricting data was used to populate the 2014 TAZ with 2014 Census estimate data.

Building Permits and Demolitions

In the development of the 2035 projections building permit data (new builds and demolitions) was used to identify the areas of growth/decline in Genesee County. Building permit data was collected from every municipality, geolocated and aggregated to the TAZ level. Building permits include single-family residential, multi-family residential, and mobile homes all weighted equally per housing unit. Data was used from the years 2000 through 2006. Comparing the 1990 and 2000 Census and Genesee County building permit data for the same time period it was decided that a reduction factor of .42 would be used to compensate for building permits issued but not completed and vacancy rates. The factored net change was then averaged out from the seven years of data into an average yearly growth/decline factor that will be identified from this point on as the 2035 Annual TAZ Household Growth Factors. Comparing 2014 Census estimates for Genesee County to 2040 LRTP 2014 projections staff noted growth was 37% of what was projected. For the 2045 projections building factors were decreased to 37% of 2035 LRTP values to compensate for the actual growth realized through 2014. The adjusted factor will be identified from this point on as the Adjusted 2035 Annual TAZ Household Growth Factors. This factor is one of several factors used to project the 2045 base year data from the 2014 Census estimates.

Methodology for 2045 Population Projections

For the 2035 LRTP Population Projections all local units of Government in Genesee County (including the City of Flint) were projected using the same methodology. The 2040 projections use different methodology for the City of Flint than what was used for all other local units of government in Genesee County. The 2045 projections continue with the methodology of the 2040 projections in using separate methodology for the City of Flint. The primary reason for this separation is that the City of Flint is a unique case as it has lost on average 19,000 people per decade since 1980. No other local unit of government in Genesee County has a fraction of the continued loss realized in the City of Flint. The following sections of this report describe the methodology used for areas outside the City of Flint and for the City of Flint itself.

Methodology for the Local Units of Government outside the City of Flint A. New Construction

In the late 2000’s much of the growth realized earlier in the decade was halted with the crash of the housing market and the beginning of the national recession. These conditions resulted in the following: -An uncharacteristic number of foreclosures -An uncharacteristic number of short sales -An uncharacteristic number of abandoned homes -An uncharacteristic drop in housing values These conditions made it a lot cheaper and attractive to buy an existing home rather than building a new one. Many older homes were abandoned as homeowners were able to buy newer and larger homes for relatively the same monthly payment of their existing home. Residential development basically halted in the late 2000’s. In 2012/2013 the housing market began to stabilize and new residential development was starting throughout Genesee County. While seeing positive growth, the amount and the short timeframe of the recovery leading up to the 2045 population projections did not give a firm foundation to build growth factors from. As a result the main assumption that staff made moving forward is that Genesee County communities will eventually get back to a certain percentage of growth realized in the first half of the 2000’s. Comparing 2014 Census estimates for Genesee County to 2040 LRTP 2014 projections staff noted growth was 37% of what was projected. For the 2045 projections building factors were decreased to 37% of 2035 LRTP values to compensate for the actual growth realized through 2014. A large amount of infrastructure was put in place in the early 2000’s as seen in partially finished subdivisions throughout the County. It is assumed that factors such as infrastructure that made areas in the County attractive for growth before the housing market crash and the national recession will continue to attract growth as the recovery continues. To determine how a community may recover staff used Census data, specifically 2010 vacancy rates, percent change in population from the 2000 to 2010 Census, and a general trend in Census population numbers from 1980 to 2010 to develop a recovery factor for each community. Charts and maps of Census data used to create the Recovery Factors and of the Recovery Factors themselves can be found in Appendix B. The recovery factors were applied to 2035 Adjusted Annual TAZ HH Growth Factors creating the 2045 Annual TAZ HH Growth Factors. This allows each community to recover at its own pace until it reaches its’ projected annual growth potential. This growth/decline is represented as an annual change in households each year at the TAZ level. An example of how this is calculated is provided in Example 1 of Appendix A.

Vacancy

Every community in Genesee County had a higher 2010 Census vacancy rate as compared to the 2000 Census vacancy rate. Another assumption made by staff is that the same factors that have affected new construction have also affected vacancy and that in most communities many of the houses that were vacant in

2010 will be occupied returning the community to 2000 vacancy levels. A vacancy rate is hard to project into the future as demolitions and new construction each affect the rate. At this time the Genesee County population projections do not project vacant housing units into the future and thus a vacancy rate is not projected. To compensate for this staff identified a target number of houses in each community that will move from vacant to occupied in the future. The target was calculated by first applying the 2000 vacancy rate to the 2010 housing units. The difference in comparing the 2010 vacant units to the factored 2010 vacant units using the 2000 vacancy rate is the target. As with new construction each community will recover vacancy at a different rate so the target number of housing units is divided by the Recovery Factor to get an annual number of housing units that will move from vacant to occupied each year until the target number of units is reached. This is represented as an annual change in households each year at the TAZ level. An example of how this is calculated is provided in Example 2 of Appendix A.

Total Households

The combination of new construction households and households recovered from vacancy represents the growth in households for a TAZ for a given year. The households in a TAZ for a given year are multiplied by the projected persons per household for the TAZ for the representative year to calculate population. An example of how this is calculated is provided in Example 3 of Appendix A.

City of Flint Population Projection Methodology

The City of Flint has continued to see a steady loss in population over the past several decades averaging a loss of 19,000 persons per decade since 1980. At some point in the future this rate of loss should level out, however, this is hard to estimate given the consistency of population loss in the City even with significant investments made in the community over the past decade. Genesee County population projections are driven by changes to households. Staff used historic percent changes to households in the City of Flint to project future percent changes to households. The percent change in households increased each decade since 1980 leading up to the 2010 Census and the future projection reverses the pattern decreasing the percent change in households for the decades out to 2045. This approach tappers back the percent household reduction in the future. Comparing 2014 Census estimates for the City of Flint with 2040 LRTP 2014 projections it was noted that the Census estimates show less population decline than originally projected. The trend for the City was shifted a decade for the 2045 projections to account for this difference. An illustration of the change and difference between the 2040 and 2045 factors is provided on the chart on the next page.

This graph illustrates the pattern for existing and projected Percent Reduction in households for the City of Flint. 18.0% 17.0%

16.0%

17.0%

14.0% 12.0% 10.0%

9.6%

8.0% 6.0%

6.5%

4.0%

3.4%

2.0% 0.0% 1990

1.0% 2000

2010

2020

2040Â LRTP

2030

2040

2050

2045Â LRTP

Staff used information in the City of Flint Master Plan to identify areas and levels of growth and decline. This information was coded into the TAZ representing the City of Flint and used to distribute annual HH reductions. The projections also recognized areas of growth in the City such as Smith Village, student housing, and the Durant that were not accounted for or at least not fully accounted for in the 2010 Census. The projected households for each City of Flint TAZ are multiplied by the persons per household projections for each TAZ for the representative year. An example of how the City of Flint Population Projections are calculated is provided in Example 5 (a) and 5 (b) of Appendix A.

Other Factors A. Availability for growth

In high growth TAZs, availability of land was looked at to determine the number of housing units a TAZ can actually hold. Aerial imagery was used to determine available land and zoning ordinances were used to determine the number of units available in that area. These were applied to the high growth TAZ in the same method that was used in the previous two projections.

Household Size

Up to this point we are working with households not persons in our population forecasting. For each TAZ a person per household factor is derived from 2014 Census data. We know that the average household size is decreasing and that it is projected to continue to decrease in the future. The University of Michigan Institute for Research on Labor, Employment, and the Economy used Regional Economic Models Inc (REMI) 2040 population projection data as their base to

develop household projections for Genesee County out to the year 2040. This data is provided in five year increments and was developed for the Michigan Department of Transportation (MDOT). The information derived from this dataset for the Genesee County population projections is an annual projected change in household size. Persons per household (PPHH) is easily calculated from the UM/REMI projections by dividing the population by the number of households for each five year increment. This represents the projected UM/REMI average PPHH for Genesee County for each five year increment. The annual change in household size for years between each five year increment is calculated by dividing the difference in PPHH for two sequential five year increments by five. From this calculation each five year period is represented by an annual PPHH reduction factor that will be applied to each TAZ to project TAZ level reductions in annual household size. Years 2041 and 2045 are outside the range of the UM/REMI dataset so staff used the rate of change in the previous period to continue the declining PPHH trend. An example of how PPHH Reduction Factors are used at the TAZ level to project PPHH is provided in Example 4 of Appendix A.

Comparison to other data sources

As stated earlier the population projections are calculated at the TAZ level and then aggregated by local unit of government. The local unit of government data is further aggregated to County level projections. The County level projections are compared to and validate against other population projections such as the 2040 Genesee County LRTP Population Projections, 2040 Regional Economic Models Inc. (REMI) projections, and 2040 and 2050 Woods and Poole projections. 2040 Genesee County LRTP Population Projections: 2040 Woods and Poole: 2040 REMI: 2050 Woods and Poole: 2045 Genesee County LRTP Population Projections:

2045 429,774 423,557 400,519 395,692 407,870

A 2045 year was estimated for the 2040 Genesee County LRTP Population Projections, 2040 Woods and Poole, and 2040 REMI for comparison purposes. Population Projection Assumptions    

Data from the 2010 Census is accurate for each traffic analysis zone. Locations of building permits from 2000-2006 will represent the areas of future growth out to 2040. Locations of demolitions from 2000-2006 will represent areas of future decline out to 2040. All new building permits do not equal new housing units. The number of new housing units is a factor based on the difference between the number of new building permits between 1990 and 2000 compared to the

  



  

number of new households reported by the Census during that same time period. Density patterns of single-family residential will continue at the current densities now present in the local unit of governments’ master plan and zoning ordinances. Household size will continue to decline at the rates suggested in the 2040 University of Michigan Institute for Research on Labor, Employment, and the Economy/Regional Economic Models Inc (REMI) data. Interpolation of the five-year increments of household size in the 2040 University of Michigan Institute for Research on Labor, Employment, and the Economy/Regional Economic Models Inc (REMI) data can be analyzed to show household size changes for any given year out to 2045. Local planning knowledge of future development in Genesee County is a factor that is considered when applying statewide and national data to the local area and adjustments are made where known development is occurring that is not represented in the statewide and national datasets. Genesee County Local Units of Government will eventually get back to 37% of the levels of growth realized in the first half of the 2000’s. Genesee County Local Units of Government will eventually get back to the levels of vacancy realized in the first half of the 2000’s. Recovery Factors can be assigned to a community based on current and historic Census vacancy and population data and used to factor future construction and vacancy recovery.

Approved 2045 Long Range Transportation Plan (LRTP) Population Projections Local Unit Argentine Twp Atlas Twp Burton City Clayton Twp Clio City Davison City Davison Twp Fenton City Fenton Twp Flint City Flint Twp Flushing City Flushing Twp Forest Twp Gaines Twp Gaines Village Genesee Twp Goodrich Village Grand Blanc City Grand Blanc Twp Linden City Montrose City Montrose Twp Mt Morris City Mt Morris Twp Mundy Twp Otisville Village Richfield Twp Swartz Creek City Thetford Twp Vienna Twp Genesee County

2014 6,687 5,998 28,974 7,339 2,554 5,000 19,071 11,453 15,261 99,002 30,892 8,135 10,337 3,734 6,252 371 20,732 1,831 8,033 36,733 3,860 1,599 6,030 2,985 20,797 14,722 832 8,433 5,589 6,797 12,862 412,895

2020 6,716 5,940 28,733 7,339 2,505 4,904 19,071 11,728 16,115 90,854 30,502 8,048 10,281 3,694 6,284 371 20,505 2,207 7,930 37,196 3,910 1,575 5,994 2,984 20,737 14,974 831 8,382 5,603 6,791 12,849 405,553

2025 6,737 5,947 28,820 7,352 2,493 4,869 19,164 11,688 16,259 86,607 30,441 8,055 10,251 3,697 6,298 368 20,481 2,230 7,891 37,524 3,920 1,569 6,015 3,013 20,833 15,195 825 8,416 5,623 6,840 12,831 402,253

2035 6,844 6,016 29,463 7,455 2,485 4,872 19,562 11,840 17,059 80,851 30,726 8,015 10,298 3,702 6,387 366 20,682 2,297 7,908 38,830 4,085 1,557 6,141 3,125 21,244 15,793 820 8,568 5,727 7,030 12,938 402,689

2040 6,940 6,087 29,995 7,555 2,502 4,916 19,899 11,928 17,324 79,365 31,121 8,055 10,390 3,729 6,471 367 20,938 2,344 7,977 39,497 4,122 1,561 6,241 3,213 21,591 16,196 824 8,699 5,820 7,181 13,082 405,931

2045 7,004 6,131 30,120 7,619 2,503 4,932 20,130 11,950 17,504 78,538 31,352 8,050 10,430 3,738 6,525 367 21,087 2,380 8,001 39,963 4,138 1,558 6,261 3,288 21,835 16,516 823 8,788 5,881 7,299 13,157 407,870

2014‐2040 % Δ 3.8% 1.5% 3.5% 2.9% ‐2.0% ‐1.7% 4.3% 4.2% 13.5% ‐19.8% 0.7% ‐1.0% 0.5% ‐0.1% 3.5% ‐0.9% 1.0% 28.0% ‐0.7% 7.5% 6.8% ‐2.3% 3.5% 7.6% 3.8% 10.0% ‐1.0% 3.2% 4.1% 5.6% 1.7%

2014‐2045 % Δ 4.7% 2.2% 4.0% 3.8% ‐2.0% ‐1.4% 5.6% 4.3% 14.7% ‐20.7% 1.5% ‐1.0% 0.9% 0.1% 4.4% ‐1.0% 1.7% 30.0% ‐0.4% 8.8% 7.2% ‐2.5% 3.8% 10.1% 5.0% 12.2% ‐1.1% 4.2% 5.2% 7.4% 2.3%

2045 Genesee County Population Projections Dataset Comparison Datasets

2014 GCMPC 2040 LRTP 417,581 Woods & Poole 2040 421,581 Regional Economic Model Incorporated (REMI) 2040 419,664 Woods & Poole 2050 414,927 Census 2014 Estimates 412,895 GCMPC 2045 LRTP 412,895 Blue italicized numbers indicate interpolated or projected data.

2020 409,210 421,711 411,712 414,212 ‐ 405,553

2025 410,384 422,231 407,617 413,298 ‐ 402,253

2035 416,286 422,895 403,049 408,001 ‐ 402,689

2040 423,030 423,226 401,784 402,579 ‐ 405,931

2045 429,774 423,557 400,519 395,692 ‐ 407,870

2045 Population Projections SAGINAW COUNTY

Montrose Twp

TUSCOLA COUNTY

Vienna Twp

Forest Twp

Thetford Twp

Otter Lake

SAGINAW COUNTY

§ ¦ ¨ 23

£ ¤

Montrose

Clio

57 O P

Flushing Twp

57 O P

Mt. Morris Twp

Otisville

Richfield Twp

Genesee Twp Mt Morris

13 O P 475

§ ¦ ¨ 75

§ ¦ ¨

15 O P

£ ¤

Flint Twp

Davison Twp

LAPEER COUNTY

Clayton Twp

Flint

Davison

§ ¦ ¨

P O

Lennon

Burton

54 O P Swartz Creek

121

P O

Gaines Twp

§ ¦ ¨

Flint Twp

15 O P Grand Blanc Twp

Mundy Twp

Atlas Twp

Goodrich

Grand Blanc

Gaines Argentine Twp

§ ¦ ¨

OAKLAND COUNTY

Fenton Twp

OAKLAND COUNTY

SHIAWASSEE COUNTY

Flushing

£ ¤ Linden

Population Change from 2014 Census Estimates to 2045 Projections 5% and over 0% to 4.9% -0.1% to -4.9% Greater than -5%

Fenton

LIVINGSTON COUNTY

2.5

1.25

0 Miles

2.5

Appendix A Methodology Examples

1. Example Calculation for Annual Household (HH) Growth Factor for Areas outside the City of Flint Annual Distribution of Target 2035 Annual HH Adjusted 2035 Annual Growth Factor HH Growth Factor for for TAZ TAZ (37%) Community 1 TAZ 1

Community 1 TAZ 2

Community 1 TAZ 3

Community 1 Total

2045 Annual HH Growth Factor Recovery Year 1 Community 1 TAZ 1 Community 1 TAZ 2 Community 1 TAZ 3 Community 1 Total

Year 2 2.00 1.00 2.80 5.80

Recovery Factor for Community 1

   

= = = =

5 5 5 5

Year 3 4.00 2.00 5.60 11.60

Year 4 6.00 3.00 8.40 17.40

2045 Annual Recovery Factor for TAZ 2.00 1.00 2.80 5.80

Year 5 8.00 4.00 11.20 23.20

Year 6 10.00 5.00 14.00 29.00

Year 7 10.00 5.00 14.00 29.00

10.00 5.00 14.00 29.00

Year 6 940.00 574.00 582.00 2,096 29.00

Year 7 950.00 579.00 596.00 2,125 29.00

2045 Annual Recovery Factors for each TAZ are compounded each year until the Adjusted 2035 Annual HH Growth Factor is reached. 2045 HH Projection For Community 1 Using Only 2045 Annual HH Growth Factor (no Recovered Vacancy included) Year 1 Year 2 Year 3 Year 4 Year 5 Community 1 TAZ 1 902.00 906.00 912.00 920.00 930.00 Community 1 TAZ 2 555.00 557.00 560.00 564.00 569.00 Community 1 TAZ 3 528.80 534.40 542.80 554.00 568.00 Community 1 Total 1,986 1,997 2,015 2,038 2,067 Community 1 HH Growth 5.80 11.60 17.40 23.20 29.00 New households are added to existing households for each TAZ.

2. Example Calculation for Recovered Vacancy for Areas Outside the City of Flint Community 1 Information 2010 Vacant Houses 2010 Households 2010 Housing Units 2000 Vacancy Rate

220 1,980 2,200 6%

2010 Factored Vacant Houses Using Census 200 Vacancy Rate Difference = Target

132

2,200 x 6%=132

‐

220

132

The Target represents the number of housing units that will be moved from vacant to occupied through the timeframe of the projections. The rate at which this happens depends on the Recover Factor for the community the TAZ represents. Annual Distribution of Target Vacant Houses

Community 1 TAZ 1 Community 1 TAZ 2 Community 1 TAZ 3 Community 1 Total

100 71 49 220

Percent of Vacant Houses this TAZ represents for the community 45.5% 32.3% 22.3% 100.0%

Distribution of Target

Recovery Factor

Annual Recovery Factor

40.00 28.40 19.60 88

5 5 5

8.00 5.68 3.92 17.60

The Target is distributed based on the Percentage of Vacant Houses the TAZ represents for the community and is then divided by the Recovery Factor to get an Annual Recovery Factor for each TAZ. 2045 Annual Vacancy Recovery Year 1 Community 1 TAZ 1 Community 1 TAZ 2 Community 1 TAZ 3 Community 1 Total

8.00 5.68 3.92 17.60

Year 2

Year 3 8.00 5.68 3.92 17.60

8.00 5.68 3.92 17.60

Year 4

Year 5 8.00 5.68 3.92 17.60

Year 6

Year 7

8.00 5.68 3.92 17.60 88

0 0 0 0

Year 5 940.00 582.40 545.60 2,068 88 The housing units that are newly occupied from vacant houses in a given year are added to the existing households in each TAZ

Year 6 940.00 582.40 545.60 2,068

Year 7 940.00 582.40 545.60 2,068

The Annual Recovery Factor is applied to each year until the Target of housing units is reached for the TAZ. 2045 HH Projection For Community 1 Using Only Recovered Vacancy (no New Build Housing included) Year 1 Year 2 Year 3 Year 4 916.00 924.00 932.00 Community 1 TAZ 1 908.00 Community 1 TAZ 2 559.68 565.36 571.04 576.72 Community 1 TAZ 3 529.92 533.84 537.76 541.68 Community 1 Total 1,998 2,015 2,033 2,050

3. Example Calculation Combining Recovered Vacancy and Household Growth Factor for Areas Outside the City of Flint 2045 Annual Vacancy Recovery Year 1 Community 1 TAZ 1 8.00 Community 1 TAZ 2 5.68 Community 1 TAZ 3 3.92 Community 1 Total 17.60

Year 2 8.00 5.68 3.92 17.60

Year 3 8.00 5.68 3.92 17.60

Year 4 8.00 5.68 3.92 17.60

Year 5 8.00 5.68 3.92 17.60

Year 6 0.00 0.00 0.00 0.00

Year 7 0.00 0.00 0.00 0.00

Year 3 6.00 3.00 8.40 17.40

Year 4 8.00 4.00 11.20 23.20

Year 5 10.00 5.00 14.00 29.00

Year 6 10.00 5.00 14.00 29.00

Year 7 10.00 5.00 14.00 29.00

2045 Combined Annual Vacancy Recovery and Annual HH Growth Factor Recovery Year 1 Year 2 Year 3 Year 4 Year 5 Community 1 TAZ 1 10.00 12.00 14.00 16.00 18.00 Community 1 TAZ 2 6.68 7.68 8.68 9.68 10.68 Community 1 TAZ 3 6.72 9.52 12.32 15.12 17.92 Community 1 Total 23.40 29.20 35.00 40.80 46.60

Year 6 10.00 5.00 14.00 29.00

Year 7 10.00 5.00 14.00 29.00

Projected Households for Community 1 Year 1 Year 2 Community 1 TAZ 1 910.00 922.00 Community 1 TAZ 2 560.68 568.36 Community 1 TAZ 3 532.72 542.24 Community 1 Total 2,003 2,033

Year 6 980.00 602.40 601.60 2,184

Year 7 990.00 607.40 615.60 2,213

2045 Annual HH Growth Factor Recovery Year 1 Year 2 Community 1 TAZ 1 2.00 4.00 Community 1 TAZ 2 1.00 2.00 Community 1 TAZ 3 2.80 5.60 Community 1 Total 5.80 11.60

Year 3 936.00 577.04 554.56 2,068

Year 4 952.00 586.72 569.68 2,108

Year 5 970.00 597.40 587.60 2,155

New households from recovered vacancy and new builds are added to existing households.

4. Example Population Projections Combining All Factors for Areas Outside the City of Flint Projected Households (HH) for Community 1 Year 1 Community 1 TAZ 1 HH 910.00 Community 1 TAZ 2 HH 560.68 Community 1 TAZ 3 HH 532.72 Community 1 Total HH 2,003.40

Year 2 922.00 568.36 542.24 2,032.60

Year 3 936.00 577.04 554.56 2,067.60

Year 4 952.00 586.72 569.68 2,108.40

Year 5 970.00 597.40 587.60 2,155.00

Year 6 980.00 602.40 601.60 2,184.00

Year 7 990.00 607.40 615.60 2,213.00

Projected Persons Per Household (PPHH) Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 Year 7 PPHH Reduction Factors ‐0.01056 ‐0.01056 ‐0.01056 ‐0.01056 ‐0.01056 ‐0.00992 ‐0.00992 Projected PPHH Comm 1 TAZ 1 2.489436 2.478873 2.468309 2.457745 2.447182 2.437263 2.427345 Projected PPHH Comm 1 TAZ 2 2.589436 2.578873 2.568309 2.557745 2.547182 2.537263 2.527345 Projected PPHH Comm 1 TAZ 3 2.289436 2.278873 2.268309 2.257745 2.247182 2.237263 2.227345 The PPHH Reduction Factor for the County for a given year is subtracted from the previous years PPHH calculation for the TAZ. This is repeated each year for each TAZ. Projected Population for Community 1 Year 1 Community 1 TAZ 1 Pop 2265.387 Community 1 TAZ 2 Pop 1451.845 Community 1 TAZ 3 Pop 1219.629 Community 1 Total Pop 4936.861

Year 2 2285.521 1465.728 1235.696 4986.945

Population = Persons Per Household x Households.

Year 3 Year 4 Year 5 Year 6 Year 7 2310.337 2339.774 2373.766 2388.518 2403.072 1482.017 1500.68 1521.686 1528.447 1535.109 1257.913 1286.192 1320.444 1345.938 1371.154 5050.268 5126.646 5215.897 5262.903 5309.334

5 (a). Factors for City of Flint Household (HH) Reduction Projected Precent Reduction in Households (HH) between the years: 2021 to 2030 2031 to 2040 6.50% 3.40%

2011 to 2020 9.6%

2041 to 2050 1.00%

2014 Flint HH 41,078

2020 Flint HH (Projected) 37,805

2030 Flint HH( Projected) 35,348

2040 Flint HH( Projected) 34,146

2014 to 2020 HH Reduction 2,303

2021 to 2030 HH Reduction 2,457

2031 to 2040 HH Reduction 1,202

2041 to 2050 HH Reduction 341

For each period the City of Flint combined households are multiplied by the Percent Reduction in Households to calculate the HH Reduction for the represented decade. 2014 to 2020 HH Reduction Per Year 383.83

2021 to 2030 HH Reduction Per Year 245.73

2031 to 2040 HH Reduction Per Year 120.18

2041 to 2045 HH Reduction Per Year 34.15

The Household Reduction for the represented decade is divided by 10 to get an Annual Reduction Per Year. 18.0%

17.0%

16.0%

17.0%

14.0% 12.0% 10.0%

9.6%

8.0% 6.0%

6.5%

4.0%

0.0% 1990

3.4%

2.0%

1.0% 2000

2010 2040 LRTP

2020

2030

2040

2050

2045 LRTP

This graph illustrates the pattern for existing and projected Percent Reduction in households for the City of Flint.

5 (b).Example of How Household (HH) Reduction Factors for the City of Flint Change HHs at the TAZ Level

Percent of HH Change This TAZ Represents Example Flint TAZ 1 Example Flint TAZ 2 Example Flint TAZ 3 Example Flint TAZ 4

20% 30% 40% 10% 100%

Year 1 76.77 115.15 153.53 38.38 383.83

Year 2 76.77 115.15 153.53 38.38 383.83

Year 3 76.77 115.15 153.53 38.38 383.83

Year 4 76.77 115.15 153.53 38.38 383.83

Year 5 76.77 115.15 153.53 38.38 383.83

Year 6 76.77 115.15 153.53 38.38 383.83

Year 7 76.77 115.15 153.53 38.38 383.83

Year 8 76.77 115.15 153.53 38.38 383.83

Year 9 76.77 115.15 153.53 38.38 383.83

Year 10 76.77 115.15 153.53 38.38 383.83

Year 11 49.15 73.72 98.29 24.57 245.73

Year 12 49.15 73.72 98.29 24.57 245.73

In the chart above the HH Reduction Per Year for the City of Flint from 5 (a) is multiplied by the Percent of HH Change the TAZ Represents to get HH reduction per year per TAZ Year 1 Example Flint TAZ 1 8138.83333 Example Flint TAZ 2 12208.25 Example Flint TAZ 3 16277.6667 Example Flint TAZ 4 4069.41667 40,694

Year 2 8062.07 12093.10 16124.13 4031.03 40,310

Year 3 7985.30 11977.95 15970.60 3992.65 39,927

Year 4 7908.53 11862.80 15817.07 3954.27 39,543

Year 5 7831.77 11747.65 15663.53 3915.88 39,159

Year 6 7755.00 11632.50 15510.00 3877.50 38,775

Year 7 7678.23 11517.35 15356.47 3839.12 38,391

Year 8 7601.47 11402.20 15202.93 3800.73 38,007

Year 9 7524.70 11287.05 15049.40 3762.35 37,624

Year 10 7447.93 11171.90 14895.87 3723.97 37,240

Year 11 7398.79 11098.18 14797.57 3699.39 36,994

In the actual City of Flint projections new construction projects were manually added to the representing TAZ but were not included as part of this example. The City of Flint is represented by 191 TAZ in the Genesee County Transportation Model

Year 12 7349.64 11024.46 14699.28 3674.82 36,748

Appendix B Recovery Factors and Supporting Census Data

Recovery Factors for the 2045 Population Projections Local Unit Argentine Twp Atlas Twp Clayton Twp Davison Twp Fenton City Fenton Twp Flushing Twp Gaines Twp Goodrich Village Grand Blanc City Grand Blanc Twp Linden City Mundy Twp Otisville Village Richfield Twp Swartz Creek City Vienna Twp Clio City Flushing City Forest Twp Gaines Village Montrose City Burton City Montrose Twp Davison City Flint Twp Genesee Twp Mt Morris City Mt Morris Twp Thetford Twp

Recovery Factor 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 15 15 15 15 15 20 20 25 25 25 25 25 25

Summary Med Vac and Med Growth:Pos Growth 1980 Low Vac and Med Growth: Pos Growth 1990 Med Vac and Med Growth:Pos Growth 1980 Med Vac and High Growth: Pos Growth 1980 Med Vac and High Growth: Pos Growth 1980 Med Vac and High Growth: Pos Growth 1980 Low Vac and Med Growth: Pos Growth 1990 Low Vac and Med Growth: Pos Growth 1990 Med Vac and High Growth:Pos Growth 1980 Low Vac and Med Growth:Pos Growth 1980 Med Vac and High Growth:Pos Growth 1980 Med Vac and High Growth:Pos Growth 1980 Low Vac and High Growth:Pos Growth 1980 Med Vac and Mild Loss:Pos Growth 1980 Low Vac and Med Growth:Pos Growth 1980 Med Vac and High Growth:Pos Growth 1980 Med Vac and Med Growth:Pos Growth 1980 High Vac and Med Growth:Flat/Neg Growth 1980 Med Vac and Med Growth: Flat/Neg Growth 1980 Low Vac and Mild loss: Flat/Pos Growth 1980 Med Vac and Med Growth:Neg Growth 1980 Med Vac and Med Growth:Flat/Neg Growth 1980 Med Vac and Mild Loss:Flat Growth 1980 Med Vac and Mild Loss:Pos Growth 1980 Med Vac and High Loss: Neg Growth 1980 High Vac and High Loss:Neg Growth 1980:Neg Growth 1980 High Vac and High Loss:Neg Growth 1980 High Vac and Mild Loss:Neg Growth 1980 High Vac and High Loss:Neg Growth 1980 High Vac and High Loss:Neg Growth 1980

% Vacant 9.9% 4.1% 6.5% 6.5% 9.1% 9.1% 5.7% 5.0% 6.4% 5.8% 8.1% 8.4% 5.9% 9.8% 5.5% 7.1% 7.3% 10.5% 6.3% 4.0% 9.9% 8.0% 8.5% 8.2% 8.6% 10.4% 12.2% 12.5% 13.8% 10.8%

Pop Change 392 229 28 1853 1164 2584 410 329 507 34 7681 1130 2891 ‐18 560 656 147 163 41 ‐18 14 38 ‐347 ‐112 ‐363 ‐1724 ‐2535 ‐117 ‐2224 ‐1228

% Pop Change 6.0% 3.9% 0.4% 10.5% 11.0% 19.9% 4.0% 5.1% 37.5% 0.4% 25.8% 39.5% 23.7% ‐2.0% 6.9% 12.9% 1.1% 6.6% 0.5% ‐0.5% 3.8% 2.3% ‐1.1% ‐1.8% ‐6.6% ‐5.1% ‐10.5% ‐3.7% ‐9.4% ‐14.8%

2045 Population Projections SAGINAW COUNTY

Montrose Twp

TUSCOLA COUNTY

Vienna Twp

Forest Twp

Thetford Twp

Otter Lake

SAGINAW COUNTY

§ ¦ ¨ 23

£ ¤

Montrose

Clio

57 O P

Flushing Twp

57 O P

Mt. Morris Twp

Otisville

Richfield Twp

Genesee Twp Mt Morris

13 P O 475

§ ¦ ¨ 75

§ ¦ ¨

15 P O

£ ¤

Flint Twp

LAPEER COUNTY

Clayton Twp

Flint

Davison

Davison Twp

§ ¦ ¨

P O

Lennon

Burton

54 O P Swartz Creek

121

P O

Gaines Twp

§ ¦ ¨

Flint Twp

15 O P Grand Blanc Twp

Mundy Twp

Atlas Twp

Goodrich

Grand Blanc

Gaines

Argentine Twp

§ ¦ ¨

OAKLAND COUNTY

Fenton Twp

Recovery Factor

OAKLAND COUNTY

SHIAWASSEE COUNTY

Flushing

£ ¤ Linden

Factor: 10 Factor: 15 Factor: 20 Factor: 25 No Factor

Fenton

LIVINGSTON COUNTY

2.5

1.25

0 Miles

2.5

Vacancy Data for the 2000 to 2010 Census Local Unit Forest Township Atlas Township Gaines Township Richfield Township Flushing Township Grand Blanc City Mundy Township Flushing City Goodrich Village Clayton Township Davison Township Swartz Creek City Vienna Township Montrose City Grand Blanc Township Montrose Township Linden City Burton City Davison City Fenton City Fenton Township Otisville Village Argentine Township Gaines Village Flint Township Clio City Thetford Township Genesee Township Mount Morris City Mount Morris Township 2010 Percent Vacant: 10% to 14% 2010 Percent Vacant: 6% to 10% 2010 Percent Vacant: Less than 6%

Summary of 2010 Percent Vacant

2010 Percent Vacant

2000 Percent Vacant

2010 Vacant

2000 Vacant

Low Vacancy Low Vacancy Low Vacancy Low Vacancy Low Vacancy Low Vacancy Low Vacancy Medium Vacancy Medium Vacancy Medium Vacancy Medium Vacancy Medium Vacancy Medium Vacancy Medium Vacancy Medium Vacancy Medium Vacancy Medium Vacancy Medium Vacancy Medium Vacancy Medium Vacancy Medium Vacancy Medium Vacancy Medium Vacancy Medium Vacancy High Vacancy High Vacancy High Vacancy High Vacancy High Vacancy High Vacancy

4.02% 4.05% 4.67% 5.45% 5.67% 5.76% 5.85% 6.34% 6.36% 6.46% 6.49% 7.09% 7.34% 7.99% 8.07% 8.22% 8.44% 8.50% 8.56% 9.07% 9.14% 9.76% 9.90% 9.94% 10.42% 10.48% 10.82% 12.25% 12.49% 13.77%

2.99% 3.95% 2.43% 5.06% 4.24% 4.91% 3.39% 3.46% 6.08% 4.93% 5.07% 5.18% 5.25% 6.44% 5.28% 5.00% 4.98% 5.26% 5.88% 5.12% 6.94% 7.05% 8.02% 7.74% 6.00% 9.29% 3.16% 7.38% 6.42% 7.42%

76 89 115 187 241 218 381 242 44 200 570 195 409 58 1,295 196 143 1,111 222 505 605 37 282 17 1,548 140 324 1,181 188 1,310

42 80 53 158 165 183 171 123 32 143 398 122 273 43 657 110 61 649 156 234 364 26 200 12 892 112 97 733 90 706

2045 Population Projections SAGINAW COUNTY

Montrose Twp

TUSCOLA COUNTY

Vienna Twp

Forest Twp

Thetford Twp

Otter Lake

SAGINAW COUNTY

§ ¦ ¨ 23

£ ¤

Montrose

Clio

57 O P

Flushing Twp

57 O P

Mt. Morris Twp

Otisville

Richfield Twp

Genesee Twp Mt Morris

13 O P 475

§ ¦ ¨ 75

§ ¦ ¨

15 O P

£ ¤

Flint Twp

Davison Twp

LAPEER COUNTY

Clayton Twp

Flint

Davison

§ ¦ ¨

21 P O

Lennon

Burton

54 O P Swartz Creek

121

P O

Gaines Twp

§ ¦ ¨

Flint Twp

15 O P Grand Blanc Twp

Mundy Twp

Atlas Twp

Goodrich

Grand Blanc

Gaines Argentine Twp

§ ¦ ¨

OAKLAND COUNTY

Fenton Twp

2010 Percent Vacant

OAKLAND COUNTY

SHIAWASSEE COUNTY

Flushing

£ ¤ Linden

Less then 6% 6% to 9.9% 10% to 14% Greater than 14%

Fenton

2.5 LIVINGSTON COUNTY

1.25

0 Miles

2.5

Population Change from 2000 to 2010 Census Area Name Linden City Goodrich Village Grand Blanc Township Mundy Township Fenton Township Swartz Creek City Fenton City Davison Township Richfield Township Clio City Argentine Township Gaines Township Flushing Township Atlas Township Gaines Village Montrose City Vienna Township Flushing City Grand Blanc City ClaytonTownship Forest Township Burton City Montrose Township Otisville Village Mt. Morris City Flint Township Davison City Mt. Morris Township Genesee Township Thetford Township Growth: 10% and over Growth: 0 to 9.9% Loss: -0.1% to -4.9% Loss: -5% and higher loss

Summary of Percent Population Change High Growth High Growth High Growth High Growth High Growth High Growth High Growth High Growth Medium Growth Medium Growth Medium Growth Medium Growth Medium Growth Medium Growth Medium Growth Medium Growth Medium Growth Medium Growth Medium Growth Medium Growth Mild Loss Mild Loss Mild Loss Mild Loss Mild Loss High Loss High Loss High Loss High Loss High Loss

Percent Change 39.5% 37.5% 25.8% 23.7% 19.9% 12.9% 11.0% 10.5% 6.9% 6.6% 6.0% 5.1% 4.0% 3.9% 3.8% 2.3% 1.1% 0.5% 0.4% 0.4% -0.5% -1.1% -1.8% -2.0% -3.7% -5.1% -6.6% -9.4% -10.5% -14.8%

Change 1,130 507 7,681 2,891 2,584 656 1,164 1,853 560 163 392 315 410 229 14 38 147 41 34 28 -18 -347 -112 -18 -117 -1,724 -363 -2,224 -2,535 -1,228

2045 Population Projections SAGINAW COUNTY

Montrose Twp

TUSCOLA COUNTY

Vienna Twp

Forest Twp

Thetford Twp

Otter Lake

SAGINAW COUNTY

§ ¦ ¨ 23

£ ¤

Montrose

Clio

57 O P

Flushing Twp

57 O P

Mt. Morris Twp

Otisville

Richfield Twp

Genesee Twp Mt Morris

13 O P 475

§ ¦ ¨ 75

§ ¦ ¨

15 O P

£ ¤

Flint Twp

Davison Twp

LAPEER COUNTY

Clayton Twp

Flint

Davison

§ ¦ ¨

P O

Lennon

Burton

54 O P Swartz Creek

121

P O

Gaines Twp

§ ¦ ¨

Flint Twp

15 O P Grand Blanc Twp

Mundy Twp

Atlas Twp

Goodrich

Grand Blanc

Gaines Argentine Twp

§ ¦ ¨

OAKLAND COUNTY

Fenton Twp

OAKLAND COUNTY

SHIAWASSEE COUNTY

Flushing

£ ¤ Linden

Population Change from 2000 to 2010 Census 10% and over 0% to 9.9% -0.1% to -4.9% -5% to -15% Greater than -15%

Fenton

LIVINGSTON COUNTY

2.5

1.25

0 Miles

2.5

Historic Genesee County Census Populations Local Unit Argentine Township Atlas Township Burton City Clayton Township Clio City Davison City Davison Township Fenton Township Fenton City Flint Township Flint City Flushing Township Flushing City Forest Township Gaines Township Genesee Township Grand Blanc Township Grand Blanc City Linden City Montrose Township Montrose City Mount Morris City Mount Morris Township Mundy Township Richfield Township Swartz Creek City Thetford Township Vienna Township Gaines Village Goodrich Village Otisville Village Genesee County

Summary of Population Trends Since 1980 Positive Growth Since 1980 Positive Growth Since 1980 Flat Growth Since 1980 Positive Growth Since 1980 Flat/Negative Growth Since 1980 Negative Growth Since 1980 Positive Growth Since 1980 Positive Growth Since 1980 Positive Growth Since 1980 Negative Growth Since 1980 Negative Growth Since 1980 Positive Growth Since 1980 Flat/Negative Growth Since 1980 Flat/Positive Growth Since 1980 Positive Growth Since 1980 Negative Growth Since 1980 Positive Growth Since 1980 Positive Growth Since 1980 Positive Growth Since 1980 Positive Growth Since 1980 Flat/Negative Growth Since 1980 Negative Growth Since 1980 Negative Growth Since 1980 Positive Growth Since 1980 Positive Growth Since 1980 Positive Growth Since 1980 Negative Growth Since 1980 Positive Growth Since 1980 Negative Growth Since 1980 Positive Growth Since 1980 Positive Growth Since 1980 Negative Growth Since 1980

Pop 1980

Pop 1990

4,180 4,096 29,976 7,269 2,669 6,087 13,708 9,570 8,098 35,405 159,611 9,246 8,624 3,573 4,769 25,065 24,413 6,848 2,174 6,164 1,706 3,246 27,928 10,786 6,895 5,013 8,499 12,914 440 795 682 450,449

4,651 4,635 27,437 7,368 2,629 5,693 14,671 10,073 8,434 34,072 140,925 9,223 8,542 3,685 4,964 24,093 25,392 7,760 2,407 6,236 1,811 3,292 25,198 11,536 7,271 4,851 8,333 13,210 427 916 724 430,459

Pop 2000 6,521 5,904 30,346 7,553 2,483 5,536 17,722 12,968 10,582 33,653 124,943 10,230 8,348 3,856 6,125 24,116 29,827 8,242 2,861 6,336 1,619 3,203 23,725 12,191 8,170 5,102 8,277 13,108 366 1,353 882 436,148

Pop 2010 6,913 6,133 29,999 7,581 2,646 5,173 19,575 15,552 11,746 31,929 102,434 10,640 8,389 3,838 6,440 21,581 37,508 8,276 3,991 6,224 1,657 3,086 21,501 15,082 8,730 5,758 7,049 13,255 380 1,860 864 425,790

City of Flint Draft Master Plan Population Change 2000 ‐2010 Map Page 13

Genesee Travel Demand Model

Final Report

APPENDIX C

2045 Employment Projections Methodology Report

2014 Base Year Employment Data 2045 Employment Projections Methodology Report

April 2016

Prepared by the Genesee County Metropolitan Planning Commission Staff 1

Executive Summary The overall employment in Genesee County is projected to improve steadily over the next 30 years. With significant improvements planned on the transportation system around the Flint-Genesee metropolitan region, communities can expect to see the creation of new employment and expansion of businesses. The purpose of this update is to establish 2014 as the base year for the 2045 Employment Projections, which is the first year of the new Genesee County Transportation Model. In calculating the projections for Genesee County, staff began with 2010 employment figures, validated through the 2040 Long Range Transportation Planâ&#x20AC;&#x2122;s (LRTP) Socioeconomic Projections. To project forward, staff used the Regional Economic Models, Inc. (REMI) growth rates for each 5-year period and interpolated the yearly growth rate, per employment sector & traffic analysis zone, for each year out to 2045. From 2041-2045, REMI growth rates were held constant. To increase the accuracy of the projections, the 2013 employment data was validated against the 2013 Bureau of Economic Analysis (BEA) data. Any locally significant economic impacts were applied directly to year, sector and traffic analysis zone in the final step. Table 1 reflects the final 2045 Employment Projections for Genesee County.

Genesee County 2045 Employment Projections by Sector

Employment Sector 2014

2020

2025

2035

2040

2045

Manufacturing Other Transportation and Public Utilities Finance, Insurance and Real Estate Retail Trade Wholesale Trade Services Government

13,090 10,487 5,362 19,981 24,602 6,277 89,533 22,210

14,309 12,023 5,822 21,523 23,701 6,014 103,929 22,836

13,431 12,068 5,823 21,114 23,200 5,970 109,972 23,107

12,504 11,670 6,134 20,184 22,373 5,750 115,393 23,602

12,013 11,416 6,387 19,787 22,072 5,556 118,523 23,789

11,547 11,168 6,645 19,411 21,780 5,372 121,743 23,977

Total

191,542

209,887

214,685

217,610

219,543

221,643

Table 1

The following document will take readers through a step-by-step approach, including methodology used by staff during projections. Graphs and maps are provided at the conclusion of this report depicting the individual and overall trends from 2010 to 2045.

Step One: Calculating Preliminary Employment Figures Preliminary Employment Data Methodology The Genesee County Metropolitan Planning Commission (GCMPC) utilized the projected 2013 employment figures, originating from the validated 2010 base year of the 2040 Long Range Transportation Planâ&#x20AC;&#x2122;s (LRTP) Socioeconomic Projections, as the preliminary starting year of 2045 projections. The first year of the new Genesee County Transportation Model will be 2014. GCMPC staff chose the year 2013 as the next significant year to validate the employment projection data to as it is the latest year among available, confirmed datasets. The calibrated Transportation Model that was utilized during the development of the 2040 Long Range Transportation Plan (LRTP), supplied staff with geographically located employers in Genesee County, their number of employees, and industry codes. GCMPC staff coded the employees, based on the North American Industrial Classification System (NAICS) codes, into eight sectors using the same categories and definitions as the previous employment estimates from the 2040 LRTP. Results of this report will assist with the development of the new Transportation Model, to be used during the development of the 2045 LRTP. Table 2 on the next page shows GCMPC sectors and their comparable SIC and NAICS codes to allow for the data to be easily comparable between plans.

GCMPC Model Employment Sectors & Corresponding SIC and NAICS Codes GCMPC Sectors

1. Manufacturing

2. Other

3. Transportation, Warehousing & Public Utilities 4. Finance, Insurance & Real Estate 5. Retail Trade 6. Wholesale Trade

SIC Categories

Durables Non-Durables Mining Construction Agriculture, Forestry &Fishing Farm Transportation &Public Utilities Finance, Insurance & Real Estate Retail Trade Wholesale Trade

NAICS Codes 33 31-32 21 23

8. Government

Manufacturing Manufacturing Mining Construction Agriculture, Forestry, Fishing and Hunting

11 11 48-49 22 52 53 44-45 42 51 54 55

7. Service

NAICS Titles

56 61 62 71 72

Service

81 92 92 92

State and Local Federal Civilian Federal Military

Table 2

Agriculture, Forestry, Fishing and Hunting Transportation and Warehousing Utilities Finance and Insurance Real Estate, Rental and Leasing Retail Trade Wholesale Trade Information Professional, Scientific and Technical Services Management of Companies and Enterprises Administrative, Support, Waste Management and Remediation Services Educational Services Health Care and Social Assistance Arts, Entertainment and Recreation Accommodation and Food Services Other Services (except Public Administration) Public Administration Public Administration Public Administration

Step Two: Comparing Preliminary 2013 Employment Figures To increase the accuracy of Genesee County’s preliminary 2013 employment data; staff took into account other available data sources as illustrated in Table 3. Comparison of Genesee County Employment Data to Other Data Sources by Industry Preliminary Census Woods Employment Five-Year BEA & Poole Employment Sector Data for Estimates 2013 2013 GCMPC 2013 2013 type of code

Manufacturing Other Transportation and Public Utilities Finance, Insurance, Real Estate Retail Trade Wholesale Trade Services Government Total

NAICS

10,434 10,423

NAICS

13,065 10,289

4,601 16,314 23,939 5,774 92,553 24,355 188,393

5,350 19,770 24,773 6,278 87,984 22,316 189,825

189,147

Table 3

Definitions of Data Sources: GCMPC Projections 2013 – The Genesee County Metropolitan Planning Commission employment data for 2013 was calculated using REMI growth factors to project the 2010 data forward to 2040. Census: American Community Survey 5-Year Estimates 2013 – This data is the 2009-2013 Selected Economic Characteristics from the U.S. Census website. Woods & Poole 2013 – Woods & Poole Economics, Incorporated is an independent firm that specializes in long-term economic and demographic projections at the county level. This dataset is based on the NAICS code system. The historical data year for this dataset is 2013. BEA 2013 – Bureau of Economic Analysis (BEA) is part of the U.S. Department of Commerce and provides regional economic information by industry. The dataset is based on the NAICS code system. The historical data year for this dataset is 2012

Step Three: Finalizing Preliminary Employment Figures Staff determined that the 2013 Genesee County employment data was slightly low overall when compared to other available datasets and would need to be factored to reduce the gap in total employment. The finance, insurance, real estate employment sector in particular held far less individuals employed than any other dataset. Bureau of Economic Analysis (BEA) data originates from the U.S. Department of Commerce and provides regional economic information by employment sector. Of the datasets available, the BEA dataset seemed most consistent with the preliminary GCMPC 2013 projections and is from a reliable source. Additionally, the use of BEA is consistent with the previous employment projections methodology. For these reasons, the BEA dataset was used to validate the 2013 GCMPC employment data. The percent change was calculated between the 2013 Genesee County employment data to the BEA 2013 data. The resulting factors were applied to each of the 639 TAZ in the model for each of the eight employment sectors. After factoring the adjusted employment estimates for each sector, employment figures are within a few employees of the BEA 2013 totals (see Table 4). BEA Adjusted 2013 Genesee County Employment Data

Employment Sector Manufacturing Other Transportation and Public Utilities Finance, Insurance, Real Estate Retail Trade Wholesale Trade Services Government Total

Preliminary Employment BEA Data for 2013 GCMPC 2013 10,434 13,065 10,423 10,289 4,601 16,314 23,939 5,774 92,553 24,355 188,393

5,350 19,770 24,773 6,278 87,984 22,316 189,825 Table 4

% Difference BEA to GCMPC 2013 2,631 0.2522 -134 -0.0129

Difference BEA to GCMPC 2013

749 3,456 834 504 -4,569 -2,039 1,432

0.1628 0.2118 0.0348 0.0873 -0.0494 -0.0837

Adjusted Employment Data for GCMPC 2013 13,082 10,272 5,324 19,758 24,722 6,276 88,029 22,323 189,786

Difference new data to BEA 2013 17 -17 -26 -12 -51 -2 45 7 -39

Employment adjustments following 2013 are further discussed in step four. Table 5 shows the 2013 employment figures in Genesee County by employment sector. This data will be used to calculate future employment estimates for the 2045 Genesee County Long Range Transportation Plan and the new Urban Travel Demand Model. Final 2013 Genesee County Employment Data 2013 Genesee County Employment 13,082 10,272 5,324 19,758 24,722 6,276 88,029 22,323 189,786

Sector Manufacturing Other Transportation and Public Utilities Finance, Insurance, Real Estate Retail Trade Wholesale Trade Services Government Total Table 5

Step Four:

Projecting Employment Figures out to 2045

Genesee County 2045 Employment Projections Methodology As previously stated, the Regional Economic Models, Inc. (REMI) data includes a countywide total for employment and by employment sector in 5-year increments out to the year 2040. For use in our employment projections we calculated growth rates for each 5-year period and interpolated the yearly growth rate, per employment sector, for each year from 2011-2045. Since the REMI data is only projected out to 2040, the growth rates were held constant from 2041-2045. The calculated growth rates are shown in Table 6 below. REMI 5-year Growth Rates 2011-2045

Employment Sector Manufacturing Other Transportation & Public Utilities Finance, Insurance & Real Estate Retail Trade Wholesale Trade Services Government Total

GCMPC Adjusted 2013 13,082 10,272

20112015 5-year change 0.0040 0.1058

20162020 5-year change* -0.0271 0.0476

20212025 5-year change -0.0447 0.0080

20262030 5-year change -0.0333 -0.0131

20312035 5-year change -0.0394 -0.0274

203620412040 2045 5-year 5-year change change^ -0.0408 -0.0408 -0.0253 -0.0253

5,324

0.0379

0.0160

0.0007

0.0197

0.0371

0.0426

19,758 24,722 6,276 88,029 22,323 189,786

0.0579 -0.0251 0.0033 0.0858 -0.0250

0.0373 -0.0088 -0.0049 0.0775 0.0606

-0.0201 -0.0227 -0.0113 0.0206 0.0125

-0.0264 -0.0271 -0.0195 0.0201 0.0099

-0.0205 -0.0116 -0.0248 0.0288 0.0125

-0.0211 -0.0153 -0.0377 0.0274 0.0083

*Adjusted REMI 2016-2020 Growth Rate applied due to local economic impact Table 6 ^Held REMI 2041-2045 Growth Rate constant

To determine the accuracy of the new dataset and as a validation measure, Genesee County Metropolitan Planning Commission (GCMPC) staff attempted to contact all employers with over 100 employees to determine if the number of employees represented in the dataset were accurate, and if the employees were located in the correct location. Not all employers could be reached or were willing to provide the information. Out of the 180 employers contacted, any reported differences to their employee numbers or to their locations was corrected in the employment database. GCMPC staff made a special attempt to contact the top 10 employers in Genesee County to get accurate estimates of their employment. These were also adjusted to the dataset. Some duplicate entries were removed from the dataset and some employers were no longer operating businesses in Genesee County. Prior to finalizing the employment projections, staff took into account any significant increase or decrease in jobs that were publicized in recent news articles or revealed through local development plans. Approximately 4,100 jobs would be added between 8

years six and eleven of the operations phase (Economic Impact of Genesys Health Park Campus Expansion Plans, prepared by the Anderson Economic Group, LLC, 2012). These jobs and others were located to the exact employment sector, TAZ, and applied to the nearest projected year ending in 5 or 0. Facilities built within the Health Park Campus will not be limited to hospital functions alone but is proposed to include an area for research & development, a learning institution, and senior living complexes. The Genesee County Freight and Connectivity Study is projecting for the Genesys expansion (in part with the proposed Dort Highway Extension) to bring 15,000 support jobs to the region (Genesee County Freight and Connectivity Study, prepared by the Corradino Group of Michigan, Inc., 2011). Based on the location of the health park campus, staff felt the number of support jobs created within Genesee County would be less than the projected 15,000. After recalculating to account for the location, approximately 7,300 support jobs is projected within Genesee County. Since the exact location and amount of jobs in each TAZ is unknown, the 7,300 jobs were proportionally applied based on the existing distribution of employment in each TAZ in REMI year 2020. Staff was able to calculate a new 5-year growth rate from 2016-2020 and apply the corresponding growth rates to each employment sector resulting in Genesee Countyâ&#x20AC;&#x2122;s final employment projections. After all adjustments and calculations were complete, the jobs from each traffic analysis zones, in each of the eight employment sectors were tallied to create the 2045 Genesee County Employment Projections. Genesee County 2045 Employment Projections by Sector

Employment Sector 2014

2020

2025

2035

2040

2045

Manufacturing Other Transportation and Public Utilities Finance, Insurance and Real Estate Retail Trade Wholesale Trade Services Government

13,090 10,487 5,362 19,981 24,602 6,277 89,533 22,210

14,309 12,023 5,822 21,523 23,701 6,014 103,929 22,836

13,431 12,068 5,823 21,114 23,200 5,970 109,972 23,107

12,504 11,670 6,134 20,184 22,373 5,750 115,393 23,602

12,013 11,416 6,387 19,787 22,072 5,556 118,523 23,789

11,547 11,168 6,645 19,411 21,780 5,372 121,743 23,977

Total

191,542

209,887

214,685

217,610

219,543

221,643

Table 7

Conclusion While the manufacturing, wholesale, finance, and retail trade sectors are projected to experience a gradual decline in employment, the services sector is projecting a substantial growth in the next 30 years. Following a recession and overall decrease in jobs prior to 2010, Genesee County has and is projected to continue to see modest signs of improvement in years to come. As stated in the Flint & Genesee County Comprehensive Economic Development Strategy, “to begin to replace the jobs lost, we must understand economic development can no longer happen by ‘chance’, but rather, through deliberate actions and strategies on the part of Genesee County and its component communities.” Looking forward, Genesee County’s total employment is projected to increase and we can conclude that job creation will vary between employment sectors. Percent Change of Genesee County Employment Sectors Employment Sector

2014

2045

Manufacturing

13,090

11,547

-11.8%

Other

10,487

11,168

6.5%

5,362

6,645

23.9%

Finance, Insurance, and Real Estate

19,981

19,411

-2.9%

Retail Trade

24,602

21,780

-11.5%

6,277

5,372

-14.4%

Services

89,533 121,743

36.0%

Government

22,210

Transportation and Public Utilities

Wholesale Trade

Table 8

23,977

% Change

8.0%

Trend

2014 vs. 2045 Total Employment Change by Sector 121,743

140,000

89,533

120,000

80,000

2014

60,000

2045

23,977

22,210 6,277

5,372

21,780

24,602

19,981

19,411

6,645

11,168

5,362

20,000

10,487

11,547

40,000

13,090

# of Employees

100,000

0 Manufacturing

Other

Transportation and Public Utilities

Finance, Insurance and Real Estate

Retail Trade

Employment Sector

Wholesale Trade

Services

Government

2014 vs. 2045 Percent Employment by Sector 60.0

50.0

40.0

30.0

20.0

10.0

0.0 Manufacturing

Other

Transportation and Finance, Insurance Public Utilities and Real Estate 2014

Retail Trade 2045

Wholesale Trade

Services

Government

Total Employment # of Employees

230,000 220,000

209,887

214,685

217,610

219,543

221,643

2035

2040

2045

12,504

12,013

11,547

2035

2040

2045

210,000 200,000

191,542

190,000 180,000 170,000 2014

2020

2025

Year

# of Employees

Manufacturing 16,000 14,000 12,000 10,000 8,000 6,000 4,000 2,000 -

13,090

2014

14,039

2020

13,431

2025

Year

Other # of Employees

12,500

12,023

12,068

12,000

11,670

11,500 11,000

11,416

11,168

10,487

10,500 10,000 9,500 2014

2020

2025

2035

2040

2045

Year

Transportation and Public Utilities # of Employees

7,000 6,000

5,362

5,823

6,134

6,387

6,645

5,822

2020

2025

2035

2040

2045

5,000 4,000 3,000 2,000 1,000 2014

Year

# of Employees

Finance, Insurance, and Real Estate 22,000 21,500 21,000 20,500 20,000 19,500 19,000 18,500 18,000

21,523 21,114 20,184

19,981

19,787 19,411

2014

2020

2025

2035

2040

2045

Year

Retail Trade # of Employees

25,000

24,602 23,701

24,000

23,200

23,000

22,373

22,072

22,000

21,780

21,000 20,000 2014

2020

2025

2035

Year

2040

2045

# of Employees

Wholesale Trade 6,277

6,400 6,200 6,000 5,800 5,600 5,400 5,200 5,000 4,800

6,014

5,970 5,750

5,556 5,372

2014

2020

2025

2035

2040

2045

Year

Services # of Employees

140,000 120,000 100,000

109,972

115,393

118,523

121,743

103,929

2020

2025

2035

2040

2045

89,533

80,000 60,000 40,000 20,000 2014

Year

Government # of Employees

24,500 24,000

23,602

23,500 22,836

23,000 22,500

23,789

23,977

23,107

22,210

22,000

21,500 21,000 2014

2020

2025

2035

Year

2040

2045

2045 Employment Projections SAGINAW COUNTY

TUSCOLA COUNTY

Vienna Twp

108

£ ¤

144

347

575

P O

276

178

283

206 224

714

Clio

140

263

Flushing Twp

654

115

P O 323

123

113

273

114

168

SHIAWASSEE COUNTY

384 35 16

120 258 67 313 19 353 377 162 23 455 161 29

Clayton Twp

867 1381

267

269

778

969

945

Flint Twp

384

396

21 P O

Lennon

165

133

348

319

§ ¦ ¨

Gaines Twp

120 63

504

342

431

189

664

165

435

379

357

148

245

315

255

171

121

P O

950

128

Fenton Twp

102

165

274 121

66 642

497

868

189

462

426

229

376

106

148

328

407

741

379

186

273

236

38 139

273

198

196 225 196 685 1110 898 193 68 267 23

289

448

734

731 1030

325

411

272

182

132

335

123 891

Court

423

164

169

141

269 1005

308

840 892

204

162

307

777

237

101

LIVINGSTON COUNTY

254

131

134

150

Atlas Twp

170

1695

Grand Blanc 195

382 545

48 120

1618

248

123 190

307

251

Goodrich 841

740

249 306

143

128

354

15 O P

613

360

697

207

§ ¦ ¨ 86

116

101

340

4303

751

216 878

180

1103

357

283

819

207

1137

966

324

385

448 135

215

403

156 142

§ ¦ ¨

174

154

OAKLAND COUNTY

2014 Total Employment 1 - 200 201 - 500 501 - 1000

1001 - 2000 2001 - 3000 3001 - 4303

123 = Number of employees

257

203

25 Burton

411

550 437 115 322

Davison Twp

115

Grand Blanc Twp

Davison

152

Fenton

337 550

408

135

148

210

Davison

178

0 39

232

918

129

152

592

981

332

15 O P

314

259

578

1087 932

743

592

507

147

141

328

P O

322 250 163 309 131

423

239

135

11 58

349

455

Flint

164

249

147

700

940

107

211

194 372 282

466 1184

£ ¤

183

108

Linden

210

587

1180

220

942

344

197

519

290

315

208

1617

Argentine Twp

167

Flint Twp

5 17

26 0

Gaines

124

258

1328

27 13

161

429

158

3497

166

283

85 78

100 112 110

72 5

267

Richfield Twp

100

161

563

672

Mundy Twp

143

124

578

605

542 1310 109 237

0 17

310

2150 115 956

915

370

323

668 443

221

658

Swartz Creek

680

1795 252 1370

490

498 3233 109 2 11 261 349 682 72 244 Robert T Longway 3 1030 41 829 335 49 9 1279 134 220 737 312 0 97 51 20 382 406 195 165 80 129 0 72 3 0 65 141 1338 0 27 496 133 10 193 1083 1065 249 2 14 34 082 2348 2329 178 397 10 0 8 259 27 95 181 62 97 72 499 405 65 387 225 316 90 40 265 78 277 659 847 86 237 79 166 260 130 697 297 115 862 543 94 54 464 50 121 1906

1174

391

439

123 385

152

406

220 134

246

704

218 146

11 245

488

353

428

213 209

125

£ ¤

1143

336 87

337

1112

§ ¦ ¨

Flushing

101

475

§ ¦ ¨

161

291

289

255

320

Mt Morris

13 O P

141

105

Genesee Twp

130

OAKLAND COUNTY

Otisville

153

Mt. Morris Twp

105

106

111 15

51 23

122

LAPEER COUNT Y

153

528

§ ¦ ¨

150 160

Montrose

Otter Lake

310

Averill

SAGINAW COUNTY

Forest Twp

Thetford Twp

Center

Montrose Twp

2.5

1.25

0 Miles

2.5

2045 Employment Projections SAGINAW COUNTY

TUSCOLA COUNTY

Vienna Twp

136

SAGINAW COUNTY

624

§ ¦ ¨

155

157

£ ¤

168

630

P O

336

215

329

Clio

178

Flushing Twp

703

132

122

P O 351

143

136

281

137

197

SHIAWASSEE COUNTY

440 43 17

148 332 88 353 21 395 409 206 24 494 176 26

Clayton Twp

1514

Flint Twp 23

337

1200

121

856

1154

727

267

431

225

169

156

499

21 P O

Lennon

199

159

356

270

§ ¦ ¨

Gaines Twp

131 77

355 572

105

373

488

240 196

755

196

459

470

296

163

277

369

323

207

121

P O

1100

152

123

187

310 145

67 807

590

907

224

512

495

246

162

262

429

388

705

393

404

490 211

124

178

376

262

485

814

151

1067 1208

378

42 175

293

297

267 184 741 1171 846 210 84 325 24

1087

301

190

263

513

890

384

336

786

840

265

183

379

974

138

207

279

166

209 316

984

262

27 Burton

244

152

1109

224

369 1267

1375

353

850

260

§ ¦ ¨ 476

323 172

1016

1068

249

103

751 135

Atlas Twp 161

127

198

393

393 2024

Grand Blanc 215

237

10497

894

155

1907

276 614

284

127 207

353

309

Goodrich 1078

908

271 364

407

142

714

468

15 O P

230

254

440

174 171

§ ¦ ¨

218

196

OAKLAND COUNTY

2045 Total Employment 1 - 200 201 - 500 501 - 1,000

1,001 - 3,000 3,001 - 6,000 6,001 - 10,497

123 = Number of employees

476

618 492 142 18 374 6

287

125

Grand Blanc Twp

Davison Twp

222

Davison

181

Fenton

LIVINGSTON COUNTY

476

208

155

165

228

167

658

421 618

331

15 O P

353

308

713

277

574

230

225 36

170

802

649

1255

P O

429

396 207 317 250 160

668

11 60

240

131

206

198

555

Flint

942

1030

123

190

256

231 425 313

117

501

1421

£ ¤

202

100

Linden

263

719

1365

222

Fenton Twp

773

402

227

546

370

362

113

235

1785

Argentine Twp

204

Flint Twp

5 19

33 0

Gaines

153

261

1639

36 14

194

2848

302

Richfield Twp

204

103 87

180

327

179 126 127

104

82 5

468

104

147

625

818

Mundy Twp

164

158 487

624

786

632 1542 109 333

0 19

262

2086 127 993 1020

429

354

799 440

742

734

Swartz Creek

11 225

1476

1807 253 1470

581

123 2 13 309 392 755 67 296 403 2318 3 501 1133 36 998 60 1476 171 240 816 362 0 11 108 61 22 433 198 525 244 89 151 0 95 0 3 80 174 1705 700 0 31 153 10 230 1359 306 1319 2 16 44 073 2162 2550 189 599 10 329 12 0 36 111 222 72 109 72 500 492 84 378 267 360 108 329 49 281 733 943 102 282 79 207 301 168 734 337 144 1005 669 115 54 549 51 132 2069 186

513

344

46 113

309

842

239

474

11 316

553

253

482 271

238

151

£ ¤

1287

412 96

389

1223

§ ¦ ¨

Flushing

122

475

§ ¦ ¨

178

339

332

258

377

Mt Morris

13 O P

167

108

Genesee Twp

145

OAKLAND COUNTY

Otisville

170

Mt. Morris Twp

113

138 30

104

228 257

784

61 300

106

145

438

Otter Lake

510

159

Montrose

Forest Twp

Thetford Twp

LAPEER COUNT Y

Montrose Twp

2.5

1.25

0 Miles

2.5

2045 Employment Projections SAGINAW COUNTY

Montrose Twp

TUSCOLA COUNTY

Vienna Twp

Forest Twp

Thetford Twp

Otter Lake

SAGINAW COUNTY

§ ¦ ¨ 23

£ ¤

Montrose

Clio

57 O P

Flushing Twp

57 O P

Mt. Morris Twp

Otisville

Richfield Twp

Genesee Twp

Mt Morris

13 P O 475

§ ¦ ¨ 75

§ ¦ ¨

15 P O

£ ¤

Flint Twp

Davison Twp

LAPEER COUNT Y

Clayton Twp

Flint

Davison

§ ¦ ¨

P O

Lennon

Burton

54 O P Swartz Creek

121

P O

Gaines Twp

§ ¦ ¨

Flint Twp

15 O P Grand Blanc Twp

Mundy Twp

Atlas Twp

Goodrich Grand Blanc

Gaines

Argentine Twp

§ ¦ ¨

OAKLAND COUNTY

Fenton Twp

Percent Change in Employment 2014 - 2045

OAKLAND COUNTY

SHIAWASSEE COUNTY

Flushing

£ ¤ Linden

-45% - -25% -24.99% - -10% -9.99% - 0%

Fenton

LIVINGSTON COUNTY

0.01% - 10% 10.01% - 25% 25.01% and above

+/- 123 = Change in Employment 2

0 Miles

2045 Employment Projections SAGINAW COUNTY

TUSCOLA COUNTY

Vienna Twp

Otter Lake

SAGINAW COUNTY

§ ¦ ¨ 23

Montrose

£ ¤

102

156

Clio

57 O P

Flushing Twp

167

631

Mt. Morris Twp

13 P O

201

1067

Otisville

Richfield Twp

164 883

475

§ ¦ ¨

403

Lennon

57 O P

§ ¦ ¨

Flushing

Clayton Twp

119

Genesee Twp

Mt Morris

15 P O

£ ¤

Flint

Flint Twp

208

Davison Twp

2998

P O

352 69

1329 54 O P

§ ¦ ¨ 710

Burton

Swartz Creek

Gaines Twp

§ ¦ ¨

434

Davison

7818

121

P O

Flint Twp

15 O P Grand Blanc Twp

Mundy Twp

Atlas Twp

100 54

407

1046

Goodrich

295

Grand Blanc

8583 75

Gaines

§ ¦ ¨

OAKLAND COUNTY

Fenton Twp

Argentine Twp

£ ¤ 218

Linden

176

Increase in Employment 2014 - 2045

OAKLAND COUNTY

SHIAWASSEE COUNTY

Forest Twp

Thetford Twp

LAPEER COUNT Y

Montrose Twp

321

0 - 500 501 - 1000 1001 and above

Fenton

123

951

LIVINGSTON COUNTY

0 Miles

= Increase in Employment 2

2045 Employment Projections SAGINAW COUNTY

Montrose Twp

TUSCOLA COUNTY

Vienna Twp

Forest Twp

Thetford Twp

Otter Lake

5 10

SAGINAW COUNTY

§ ¦ ¨

Montrose

£ ¤

Clio

P O

6 27

Otisville

3 12

Flushing Twp

Mt. Morris Twp 31

Mt Morris

13 O P

Richfield Twp

Genesee Twp 1

475

§ ¦ ¨ § ¦ ¨

182

Clayton Twp

1 58

12 11

Swartz Creek 223

Gaines Twp

§ ¦ ¨

121

P O

2305

12 7

15 P O Atlas Twp

82 34

204

Goodrich

Gaines

588

£ ¤

Linden 4

1 4

§ ¦ ¨

OAKLAND COUNTY

244

2045 Manufacturing Employment 1 - 25 26 - 100 101 - 200

52 1

LIVINGSTON COUNTY

201 - 500 501 - 1000 1001 - 2551

123 = Number of employees

Fenton 1

6 1

219

9 3

Fenton Twp

Grand Blanc

Argentine Twp

§ ¦ ¨

123

3 17

Grand Blanc Twp

725

Burton

54 P O

Davison

3 4

181

325

Flint Twp 4

4 11

2551

517 3 3 3 1 13

Mundy Twp

3 4

27 8

21 O P

Lennon

Davison Twp 1

711

140

Flint Twp

Flint

15 O P

£ ¤

OAKLAND COUNTY

SHIAWASSEE COUNTY

LAPEER COUNT Y

Flushing

2.5

1.25

0 Miles

2.5

2045 Employment Projections SAGINAW COUNTY

TUSCOLA COUNTY

Vienna Twp

SAGINAW COUNTY

§ ¦ ¨

£ ¤

P O

Flushing Twp

§ ¦ ¨

144

Clayton Twp 15

2 20

21 O P

Lennon

27 32

59 4

§ ¦ ¨

Gaines Twp

121

Flint Twp

1 9

16 89

313

P O

Gaines

Fenton Twp

Argentine Twp

Linden 21

7 42

£ ¤

2 27

19 19

13 82

§ ¦ ¨ 19

2 20

Atlas Twp

106

Goodrich 1

4 16

36 19

2 7

23 7

15 O P

32 35

§ ¦ ¨

OAKLAND COUNTY

2045 Other Employment 1 - 25 26 - 50 51 - 100

101 - 200 201 - 300 301 - 380

123 = Number of employees

LIVINGSTON COUNTY

27 54

Fenton 9

Davison 13

144

9 27

Davison Twp

Grand Blanc Twp

13 2

3 Burton

187

191

115

Grand Blanc

15 O P

104

112

1 1

213

8 61 910 12 27 20

P O

Mundy Twp

2 9

22 2

14 9 13

9 10 34 4 51 14 9

Swartz Creek 9

3 8

21 13

44 102 4

4 18

7 32

9 23

Flint

153

1 2

Flint Twp 9

£ ¤

380

Richfield Twp

OAKLAND COUNTY

SHIAWASSEE COUNTY

2 4

475

§ ¦ ¨

Flushing

13 19

Genesee Twp

Mt Morris

13 O P

Mt. Morris Twp

7 4

Otisville

9 16

7 Clio

P O

Otter Lake

Montrose

Forest Twp

Thetford Twp

LAPEER COUNT Y

Montrose Twp

2.5

1.25

0 Miles

2.5

2045 Employment Projections SAGINAW COUNTY

Montrose Twp

TUSCOLA COUNTY

Vienna Twp

Forest Twp

Thetford Twp

Otter Lake

2 2

£ ¤

Montrose 2 15

Clio

57 P O

48 2

Flushing Twp

13 O P 475

§ ¦ ¨ § ¦ ¨ 45

206

Clayton Twp

14 60

210

121

P O

10 7

Gaines Twp

Swartz Creek

§ ¦ ¨

3 73

24 1

Flint Twp

2 18

Grand Blanc Twp 175

Atlas Twp 35

18 35 5

15 O P

Burton 27

§ ¦ ¨

54 O P

657

31 40

291

Mundy Twp

23 27 15 2

Davison

3 124

611

Davison Twp

984

21 O P

Lennon

40 27

7 5

P O

Flint

1 1

Flint Twp

£ ¤

53 32

Flushing

Richfield Twp 2

Mt Morris

Goodrich

Grand Blanc

Gaines

§ ¦ ¨

Argentine Twp

Fenton Twp

£ ¤

Linden

1 - 25 26 - 50 51 - 75

LIVINGSTON COUNTY

76 - 100 101 - 250 251 - 984

123 = Number of employees

Fenton

OAKLAND COUNTY

2045 Transportation & Public Utilities Employment

OAKLAND COUNTY

SHIAWASSEE COUNTY

Genesee Twp

Mt. Morris Twp

Otisville

LAPEER COUNT Y

SAGINAW COUNTY

§ ¦ ¨

2.5

1.25

0 Miles

2.5

2045 Employment Projections SAGINAW COUNTY

Montrose Twp

TUSCOLA COUNTY

Vienna Twp

SAGINAW COUNTY

Otter Lake

11 13

§ ¦ ¨

38 6

Montrose

£ ¤

Forest Twp

Thetford Twp

P O

Clio

110

P O

8 69

Otisville

Mt Morris

15 5

§ ¦ ¨

282

Clayton Twp 8

21 P O

Lennon

46 88 17

23 77

Swartz Creek 21 48

§ ¦ ¨

18 11

8 17

219

71 83

120

Mundy Twp

253

121

Flint Twp 32

297

6 4

P O

33 192

17 5

122

157

11 88

23 11 20 21

74 109

Grand Blanc Twp 25

§ ¦ ¨ 23

123

Goodrich

321

92 160

Grand Blanc 2

Atlas Twp

97 88

§ ¦ ¨

OAKLAND COUNTY

1 - 25 26 - 50 51 - 75

76 - 100 101 - 250 251 - 3217

123 = Number of employees

LIVINGSTON COUNTY

203

20 206 23 6

132

15 O P

Fenton 206

317

Burton

2045 Finance, Insurance, & Real Estate Employment

OAKLAND COUNTY

£ ¤ 17

Davison

Linden

114

Fenton Twp

Argentine Twp

123

Davison Twp

Gaines

4 24

6 44

238

124

110 1

32 170

11 25

15 O P

65 18 5 4 6 28 81 21 8

P O

2 94

35 8

183 19 1 6

Flint

129

5 6

180

85 5

163

114

127 725

132

156

13 38

151

Gaines Twp

Flint Twp

£ ¤

135 67

11 91

Flushing

475

§ ¦ ¨

SHIAWASSEE COUNTY

15 2

Richfield Twp

Genesee Twp

13 O P

Mt. Morris Twp

LAPEER COUNT Y

Flushing Twp

2.5

1.25

0 Miles

2.5

2045 Employment Projections SAGINAW COUNTY

Montrose Twp

TUSCOLA COUNTY

Vienna Twp

Forest Twp

Thetford Twp

Otter Lake

31 25

§ ¦ ¨

9 89

Montrose

£ ¤

161

P O

Clio

166

P O 56

Flushing Twp

£ ¤

36 71

206

Clayton Twp

Flint Twp 11

225

§ ¦ ¨

Gaines Twp

47 1

58 112

Swartz Creek

11 400

1266

678

41 115

48 7

111

121

P O

Flint Twp 85

588

14 75

181

84 10

10 137

Fenton Twp

67 18

27 13

648

11 Burton

202

711 323

44 8

LIVINGSTON COUNTY

224 5

180 58

211

158

§ ¦ ¨ 9

19 9

15 O P

Atlas Twp 1

8 4

111

Goodrich 2

151

§ ¦ ¨

37 39

OAKLAND COUNTY

2045 Retail Employment 1 - 50 51 - 100 101 - 200

201 - 500 501 - 1000 1001 - 1266

123 = Number of employees

16 48 21 25

Fenton 48

Davison

Davison Twp

194

112

361

Grand Blanc

£ ¤ 2

Linden

1 3

24 400

Argentine Twp

94 357

667

Gaines

1 4

Grand Blanc Twp

3 18

8 1

123

173

100

2 11

123

P O

15 O P

19 9 21 8

Flint

3 66

239

13 13

Mundy Twp

210

476

117

203

21 P O

100

16 36 14 18 123 1 4 11 3 3 18 20 6 3 7 21 8 27 29 4 3 3 122 25 7 3 3 87 4 2 9 9 3 9 3 14 29 27 36 24 114 38 21 11 10 16 18 18 32 16 31 44 225 36 36 154 63 32 54 19 41 492

106

100

43 3

Lennon

7 7

14 106

163

143

Richfield Twp

OAKLAND COUNTY

SHIAWASSEE COUNTY

§ ¦ ¨

Flushing 67

475

§ ¦ ¨

48 21

Genesee Twp

Mt. Morris Twp

Mt Morris

13 O P

Otisville

23 11

14 11

135

3 1

LAPEER COUNT Y

SAGINAW COUNTY

2.5

1.25

0 Miles

2.5

2045 Employment Projections SAGINAW COUNTY

Montrose Twp

TUSCOLA COUNTY

Vienna Twp

Forest Twp

Thetford Twp

Otter Lake

£ ¤

Montrose 1

Clio

P O

57 P O

Otisville

Flushing Twp

Mt. Morris Twp 29

Mt Morris

13 P O

Clayton Twp

Flint Twp 9

P O

11 2

4 2

Gaines Twp

Linden 8

3 7

3 13

2 3

Burton

15 O P

Grand Blanc Twp

Atlas Twp

1 14

4 8

Goodrich

§ ¦ ¨

OAKLAND COUNTY

2045 Wholesale Trade Employment 1 - 25 26 - 50 51 - 75

76 - 100 101 - 200 201 - 282

123 = Number of employees

LIVINGSTON COUNTY

16 28

Fenton

1 3

§ ¦ ¨

282

Davison Twp

£ ¤

Fenton Twp

Argentine Twp

Davison

Gaines

Grand Blanc

54 O P

13 207

116

120

Flint Twp 1

121

P O

Mundy Twp

7 21 2 1 13 4

115 23

Swartz Creek

§ ¦ ¨

121

3 25

151

3 54

Flint

15 P O

Lennon

3 2 1

OAKLAND COUNTY

SHIAWASSEE COUNTY

£ ¤

§ ¦ ¨

475

§ ¦ ¨

Flushing

Richfield Twp

Genesee Twp

LAPEER COUNT Y

SAGINAW COUNTY

§ ¦ ¨

2.5

1.25

0 Miles

2.5

2045 Employment Projections SAGINAW COUNTY

TUSCOLA COUNTY

Vienna Twp

106

SAGINAW COUNTY

387

§ ¦ ¨

£ ¤

123

340

P O

243

167

236

201 215

330

Clio

236

P O 189

Flushing Twp 36

216

SHIAWASSEE COUNTY

287 26 2

§ ¦ ¨

95 45

127 88 72 204

315

£ ¤

614 218

180

776

221 781

111

Clayton Twp 294

200

714

481

Flint Twp 14

309

65 75

126

152

§ ¦ ¨

Gaines Twp

242 363

715

443 688

102

156

232

404

142

284

296

153

78 32

145

121

P O

445

265

210

453

703

116 1010

Fenton Twp

Linden

182 91

16 712

536

102

229

136

267

269

112

269

183

217

498

156

215

379 767

179 112

146

389

184

112

369 368 220 125 65 267 3

109

159

349

232

258

102

312

569

121

189 122

216

262

114

50 13 Burton

166

270

LIVINGSTON COUNTY

1122

100

530

221

§ ¦ ¨ 400

105

11 269

1272

Grand Blanc 89

118

146

10485

11 37

514

187

15 O P

Atlas Twp

259

140

1320

172 419

123

757

42 87

240

149

Goodrich 1051

63 130

§ ¦ ¨

205

175

OAKLAND COUNTY

2045 Service Employment 1 - 100 101 - 200 201 - 500

501 - 1000 1001 - 2000 2001 - 10,485

123 = Number of employees

325

354 167 106 172

79 387

776

157

120

189

173 255

127

Davison 159

Grand Blanc Twp 123

Davison Twp

141

Fenton

365 354 589

344

262

157

105

663

389

552

200 50

114

124

61 17

166

172

320

£ ¤ 60

189

162

508

285

324

327

15 O P

235

546

142

P O

3 8

123

327 171 47 91 119

217

118

168

438

Flint

229 399 112

540

230

1162

313

176

5 3

Richfield Twp

224

Argentine Twp

255

151 221

210

342

1399

Gaines

165

151

143

33 2

112

Flint Twp

162

268

203

79 4

37 3

1109 1

222

160

220

264

514

481

Mundy Twp

269

190

368

632

235

199

475

939

67 331

Swartz Creek

231 259 2 10 3 220 184 284 11 3 250 34 344 452 50 720 4057 11 10 1289 144 112 466 13 175 514 219 99 51 77 95 510 700 120 49 3 8 77 83 59 130 2 1080 1306 223 44 14 88 47 2 92 171 2 253 10 294 12 30 60 160 50 53 15 141 398 84 114 204 172 78 33 294 60 231 162 226 70 13 162 210 76 167 277 180 128 704 584 76 54 373 9 72 1276

389

65 608

379

100

215

21 O P

Lennon

142 450

1321

437

272

142

465

157

637

215

5 306

318

112

293 51

239

§ ¦ ¨

Flushing

475

199

121

200

171

102

Genesee Twp

121

Mt Morris

13 O P

Mt. Morris Twp

Otisville

OAKLAND COUNTY

114

159

33 111

361

Otter Lake

415

Montrose

Forest Twp

Thetford Twp

LAPEER COUNT Y

Montrose Twp

2.5

1.25

0 Miles

2.5

2045 Employment Projections SAGINAW COUNTY

Montrose Twp

TUSCOLA COUNTY

Vienna Twp

Forest Twp

Thetford Twp

SAGINAW COUNTY

§ ¦ ¨

Otter Lake

£ ¤

Montrose 57

198

57 O P

Clio

57 O P

126

Otisville

149

111 136

Mt. Morris Twp

Mt Morris 167

13 O P

281

196

503

Clayton Twp

347

111

258

165

Swartz Creek

Gaines Twp

146

557

500

60 32

121

P O

104

54 O P

622

§ ¦ ¨

171

419

Burton

15 O P

519

Mundy Twp

Flint Twp

189

Davison

Davison Twp

24 1089 466 177 192 144 36 6 163 4 29 75 13 97 2474 1802 24 28

P O

99 4

361

Flint

201

329

Flint Twp

15 O P

121

£ ¤

548

Grand Blanc Twp

Atlas Twp

106

Goodrich

454

Grand Blanc

106

74 54

Gaines 21

§ ¦ ¨

Fenton Twp

Argentine Twp

£ ¤ 32

Linden

121

67 32

1 - 25 26 - 50 51 - 100

217

143

LIVINGSTON COUNTY

101 - 500 501 - 1000 1001 - 2,474

123 = Number of employees

Fenton 13

OAKLAND COUNTY

2045 Government Employment

OAKLAND COUNTY

SHIAWASSEE COUNTY

108

1148

§ ¦ ¨

205

166

475

Flushing

§ ¦ ¨

104

Lennon

Richfield Twp

Genesee Twp

LAPEER COUNT Y

Flushing Twp

2.5

1.25

0 Miles

2.5

Genesee County Metropolitan Planning Commission 1101 Beach Street, Room 223 Flint, MI 48502-1470 (810) 257-3010 www.GCMPC.org