Spatial pattern of drought-induced tree mortality and canopy opening from mixed forest patches in Memorial Park, Houston Texas Xavier A. Jaime, Rosemary Dwight, and Dr. X. Ben Wu MPC- Houston Livestock Show and Rodeo
Urban Ecosystem Services • Green Spaces (27% of the Nation’s total tree canopy cover)
– Urban pine-hardwood forests, savanna and savanna prairies
• Blue Spaces – Streams (Buffalo Bayou), ponds, and storm water retention ponds
Study Site
574 ha Park Boundary 418 ha Forest Cover
Questions 1. What were the mortality rates and percentage of live tree cover before and after the drought of 2011? 2. How were they spatially arranged in the landscape, within each soil types before and after the drought? 3. How did distance from impervious surface and water source influenced the distribution of tree mortality?
Hypotheses 1.
2.
We posited that mortality rates after the drought were significantly higher, with larger patches covering larger (%) of the landscape, geometrically complex and spatially aggregated within neighboring mortality patches in the landscape. Different soil types will be an influencing factor on their spatial pattern, while distribution of dead trees will be higher in proximity to roads and tracks, yet lower in proximity to water.
Data and Methodology
Metrics Used • Spatial arrangement of tree mortality and live tree cover within 5 soil types: – Landscape composition: •
Contagion index (Spatial aggregation)
– Mortality vs Live Tree Cover: • • •
Percentage of total area cover (PLAND) Mean patch size (AREA_MN) Mean shape index (SHAPE_MN)
– Proximity Analysis: •
Frequency distribution to distance from: – Roads and Tracks – Water
B e f o r e D r o u g h t
A f t e r
Tree Mortality
3.1%
43.4%
Live Tree Cover
83.7%
65.0%
D r o u g h t
Overall estimate of tree mortality within 5 major soil types before and after the drought
CONTAGION INDEX
90 80
60 50
80 60 40 20 0 Fine Sandy Loam 2006-2010
2010-2013
100 80 60 40 20 0 Silty Clay &Urban 2006-2010
2010-2013
30 20 10 0 LANDSCAPE 2006-2010
2010-2013
100 80 60 40 20 0 Loam 2006-2010
2010-2013
CONTAGION INDEX
40
CONTAGION INDEX
CONTAGION INDEX
70
100
CONTAGION INDEX
Spatial Aggregation Mosaic Landscape
100 80 60 40 20 0 Very Fine Sandy Loam 2006-2010
2010-2013
1.3
100 90 80 70 60 50 40 30 20 10 0 2006-2010 2010-2013 Live Tree Cover
S P A T I A L
C O M P L E X I T Y
Mean Shape Index
Percentage of total area (%)
1.32 1.28 1.26 1.24 1.22 1.2 1.18 1.16 1.14 2006-2010 2010-2013 Mortality
Mean Shape Index
P A T T E R N
100 90 80 70 60 50 40 30 20 10 0 2006-2010 2010-2013 Mortality
Percentage of total area (%)
S P A T I A L
1.32 1.3 1.28 1.26 1.24 1.22 1.2 1.18 1.16 1.14 1.12 2006-2010 2010-2013 Live Tree Cover
SCU VFSL 2006-2010 Tree Mortality
SCU VFSL
150
L
100
SCU
50
VFSL
0
250 200
FSL
150
L
100
SCU
50
VFSL
0 2006-2010 Tree Cover
Mean Shape Index
FSL
Mean Patch Size (m2)
200
SCU VFSL
1.65 1.55
FSL
1.45
L SCU
1.35
VFSL 1.25
2010-2013 Tree Cover
2006-2010 Tree Cover
FSL L SCU VFSL 2010-2013 Tree Mortality
1.75
300
250
L
1.6 1.55 1.5 1.45 1.4 1.35 1.3 1.25 1.2
2006-2010 Tree Mortality
2010-2013 Tree Mortality
300 Mean Patch Size (m2)
L
FSL
Mean Shape Index
L
FSL
1.6 1.55 1.5 1.45 1.4 1.35 1.3 1.25 1.2
Mean Shape Index
FSL
40 35 30 25 20 15 10 5 0
Mean Shape Index Mean Shape Index
40 35 30 25 20 15 10 5 0
Mean Patch Size (m2)
Mean Patch Size (m2)
Mean Patch Size
1.75 1.7 1.65 1.6 1.55 1.5 1.45 1.4 1.35
FSL L SCU VFSL 2010-2013 Tree Cover
Spatial distribution of live and dead trees in proximity to water source 7000000 Freq. distribution
6000000 5000000 4000000 3000000
2006-2010 All 2006-2010 Mortality 2006-2010 Live Trees R² = 0.8858
2000000 1000000 0
Euclidean Distance from roads and tracks (m) 7000000 Freq. distribution
6000000 5000000 4000000 3000000
2010-2013 All 2010-2013 Mortality 2010-2013 Live Trees R² = 0.8615
2000000 1000000 0 Euclidean Distance from water (m)
Conclusion • The drought of 2011 undoubtedly lead to a noticeable loss of overall tree canopy cover and increased landscape fragmentation • Soils are shown to be significant influencing factors in how drought would impact tree cover, while impervios surface and manmade structures had no influence. However, hydrographic proximity will have an influence in the spatial pattern of tree mortality. • These findings may not only have implications to ecosystem functions and services provided by these urban ecosystems, but would at least give a better insight on where the most affected parts of the forest are in order to effectively restore and manage it efficiently. • Further plans should aim for an inclusive restoration and management strategy considering the historical conditions of this natural prairie system with alternated forest patches better suited to environmental stressors.
Acknowledgements • Jay Daniels and Carolyn White from the Memorial Park Conservancy Staff for their insights, collaboration in the logistics and data support • Dr. Ben Wu and Dr. Frederick Smeins for their overall support and expertise • Cassidy Ince for GIS-Remote Sensing data cataloging support • James M. Carder Assistantship fund for urban forest ecology and restoration