Climate Change in the tropical Andes - Part 1 by Proyecto PRAA

Climate Change in the tropical Andes - Impacts and consequences for glaciation and water resources Part I: The scientific basis A report prepared by

MATHIAS VUILLE with contributions from RAYMOND S. BRADLEY BERNARD FRANCOU GEORG KASER BRYAN G. MARK

Climate Change in the tropical Andes â&#x20AC;&#x201C; Impacts and consequences for glaciation and water resources Part I: The scientific basis A report for CONAM and the World Bank prepared by

MATHIAS VUILLE (University of Massachusetts)

with contributions from RAYMOND S. BRADLEY (University of Massachusetts) BERNARD FRANCOU (IRD) GEORG KASER (University of Innsbruck) BRYAN G. MARK (Ohio State University)

Amherst, Massachusetts, 24. January, 2007

Cover photo: Cordillera Vilcanota, June 2006 (photo credit: M. Vuille)

TABLE OF CONTENTS SUMMARY ......................................................................................................................2 1) INTRODUCTION ...................................................................................................5 2) OBSERVED GLACIER VARIATIONS ..........................................................7 2.1. Venezuela ..................................................................................................7 2.2. Colombia .....................................................................................................7 2.3. Ecuador .....................................................................................................10 2.4. Peru.............................................................................................................12 2.4.1. Quelccaya Ice Cap.......................................................................13 2.4.2. Cordillera Blanca ..........................................................................14 2.4.3. Coropuna.......................................................................................17 2.5. Bolivia .........................................................................................................18 2.5.1. Charquini, Chacaltaya and Zongo glaciers ..............................19 2.5.2. Cordillera Occidental (Sajama)..................................................22 2.6. Summary ..................................................................................................22

3) GLACIER MASS BALANCE ............................................................................24 3.1 General characteristics of tropical glacier m. balance ...........24 3.2. Mass balance - climate relationships ...........................................26 3.3. Mass balance and large-scale forcing .........................................28 4) GLACIER SURFACE ENERGY BALANCE ..............................................31 4.1. Sensitivity studies .................................................................................31 4.2. The Surface Energy Balance (SEB) .............................................33 5) OBSERVED AND PROJECTED CLIMATE CHANGE.........................37 5.1. Observed 20th century climate change ........................................37 5.1.1. Temperature .................................................................................37 5.1.2. Precipitation ..................................................................................41 5.1.3. Humidity.........................................................................................42 5.1.4. Cloud cover ...................................................................................43 5.1.5. Atmospheric circulation ...............................................................44 5.2. Projected future climate change .....................................................46

6) IMPLICATIONS FOR WATER RESOURCES .........................................49 6.1. Antizana ....................................................................................................51 6.2. Cordillera Blanca ...................................................................................52 6.3. Cordillera Real........................................................................................57 7) CONCLUSIONS ....................................................................................................59 REFERENCES............................................................................................................61 1

SUMMARY Observations on glacier extent from Venezuela, Colombia, Ecuador, Peru and Bolivia give a detailed and unequivocal account of rapid shrinkage of tropical Andean glaciers since the Little Ice Age (LIA). The retreat appears to have been non-uniform throughout the 20th century with periods of stronger recession interrupted by phases of more stable conditions or even minor readvances such as in the 1970s and at the end of the 20th century. In general however, there is clear evidence ongoing shrinkage since the mid-1970s. Many smaller, low-lying glaciers are completely out of equilibrium with current climate and may disappear within a few decades. Mass balance records from Bolivia and Ecuador similarly show a very coherent picture, with a generally negative mass balance, which appears to be driven by the same background forcing throughout the region. Superimposed on this negative trend are interannual variations with occasional periods of equilibrated or even positive mass balance, in particular during prolonged La Niña events, such as between 1999 and 2001. Glaciers grow or shrink as a reaction to changes in their mass balance, an obvious reaction being the advance or retreat of their tongues. The mass balance describes where and how a glacier is gaining or losing mass due the predominance of accumulation or ablation processes, which in turn are determined by climatic variables such as temperature, precipitation, solar radiation, humidity, etc. Because of the lack of a pronounced thermal seasonality but a clear differentiation between dry and wet seasons, tropical glacier mass balance and its sensitivity to climate change are fundamentally different from mid- and high latitude glaciers. Inner tropical glaciers have a very negative mass balance below their Equilibrium Line Altitude (ELA) due the exposure of the ablation zone to melt and sublimation 365 days per year. In the subtropics ablation is much reduced, mostly because sublimation dominates over melt due to generally low humidity. Because of these characteristics the ELA is generally close to the 0˚C-line in the inner tropics and hence the ELA will react sensitive to changes in temperature. In the outer tropics the ELA is usually located considerably above the 0˚C-line, and a temperature increase will not have such an immediate effect. The ELA instead is more sensitive to changes in precipitation and humidity, which determines the ratio of melting to sublimation. In general the glacier mass balance in the outer tropics reflects the variability in wet season accumulation and ablation, while mass turn over is minor during the dry season. In the inner tropics on the other hand, mass net ablation remains quite constant throughout the year. The most dominant forcing factor on interannual timescales is associated with ENSO. During El Niño years subtropical glaciers experience reduced accumulation, a lowered albedo, combined with an increase in incoming shortwave radiation due to reduced cloud cover. All these factors contribute to a very negative mass balance. In the inner tropics the glacier response to ENSO is very similar but for different reasons. Here the impact of El Niño is through increased air temperature, which favors rain over snowfall, and to a lesser degree due to sporadic snowfall, insufficient to maintain a high glacier albedo, low wind speeds, which limit the transfer of energy from melting to sublimation, and reduced cloud cover, which increases the incoming shortwave radiation.

Several terms of the energy balance, show pronounced seasonality with very different behavior during dry and wet seasons. In the outer tropics both mass loss and runoff show a marked seasonality, which can not be explained by the sensible heat transfer, which show little seasonality and is generally small. Instead net radiation and the latent heat flux dominate the energy balance. Incoming energy is quite constant throughout the year and instead it is the partitioning of this energy into melt and sublimation, controlled by humidity, which causes the much higher mass loss during the wet season. When humidity is high (wet season), the available energy is directly consumed by melting, while the enhanced vapor pressure gradient during the dry season favors sublimation, which is energetically inefficient and therefore leads to reduced mass loss. These peculiarities of the energy balance make subtropical glaciers highly sensitive to a) changes in atmospheric humidity which governs sublimation, b) precipitation, whose variability, particularly during the rainy season induces a positive feedback on albedo and c) cloudiness, which controls the incoming long-wave radiation. In the inner tropics the absence of thermal seasonality exposes the ablation zone to oscillations of the 0˚C isotherm throughout the year. These small fluctuations in temperature determine the rain-snow line on the ablation zone and hence have a major impact on the albedo. Consequently air temperature significantly influences the energy balance in the inner tropics, albeit not through the sensible heat flux, as commonly thought, but indirectly through changes in albedo and net radiation receipts. The high sensitivity of tropical glacier mass and energy balance to climate change, which is under way and well documented, leaves little room for doubt that the observed glacier retreat is occurring in response to a changing climate. Temperature in the Andes has increased by approximately 0.1°C/decade, since 1939, with the bulk of the warming occurring over the last 2 decades. Since the mid 1970s’ the rate of warming almost tripled to 0.3°C/decade. The eastern slopes show a much subdued warming, while the warming on the Pacific side is strongest. Higher elevations have experienced an intermediate, albeit still significant warming. On average about 50-70% of the observed temperature change in the Andes, can be attributed to a temperature increase in the tropical Pacific. Observations suggest that precipitation has slightly increased in the second half of the 20th century in the inner tropics and decreased in the outer tropics. However, trends at individual stations are weak and mostly insignificant. Nonetheless the general pattern of moistening in the inner tropics and drying in the subtropical Andes is dynamically consistent with observed changes in the large-scale circulation. Both satellite information and reanalysis data seem to suggest a strengthening of the tropical atmospheric circulation. Changes in the inner tropics are characterized by enhanced lowlevel convergence, upper-level divergence, and as a result enhanced upward motion, increased convective activity, and more humid conditions. In the subtropics the opposite trends prevail, with increased subsidence and reduced convective activity leading to potentially drier conditions. Model projections of future climate change in the tropical Andes indicate a continued warming of the tropical troposphere throughout the 21st century, with a temperature increase that is enhanced at higher elevations. By the end of the 21st century, following the SRES A2 emission scenario, the tropical Andes may experience a massive warming on the order of 4.5˚-5˚C. The Special Report on Emission Scenarios (SRES) A1B scenario reaches about 80-90% of the warming displayed in the SRES A2 scenario

at the end of the century, while the more moderate SRES B1 path displays only about half of the warming of SRES A2. All emission paths tend to show the same pattern of warming, but they differ in amplitude. Predicted changes in precipitation include an increase in precipitation during the wet season and a decrease during the dry season, which would effectively enhance the seasonal hydrological cycle in the tropical Andes. If tropical glaciers continue to retreat and eventually disappear from certain catchments, the change in streamflow, due to the lack of a glacial buffer during the dry season, will significantly affect the availability of drinking water, water for hydropower production, mining and irrigation. In the tropical Andes the problem is exacerbated when compared with mid latitude mountain ranges because ablation and accumulation seasons coincide, which precludes the development of a long-lasting seasonal snow cover outside the glaciated areas. Glaciers are therefore the only major seasonally changing water reservoir in the tropical Andes. Tropical Andean catchments show a high correlation between their capacity to store precipitation and their percentage of glaciated area. As glaciers retreat and lose mass, they add to a temporary increase in runoff. Downstream users will quickly adapt to this temporary increase in water supply, which raises serious sustainability concerns. In the Cordillera Blanca at least 10%, and potentially as much as 20%, of the annual discharge stems from volume loss of stored glacier ice. Simulations based on different IPCC scenarios for 2050 and 2080 indicate that glacier volume in the Cordillera Blanca will be significantly lowered, but glaciers in most catchments do not completely disappear. Simulations further suggest that the overall discharge may not change very much, but that the seasonality intensifies significantly. Dry season runoff is reduced, in particular in the A2 scenario, but during the wet season discharge is higher, since the larger glacier-free area leads to enhanced direct runoff. In general the results of the A2 scenario are much more dramatic in 2050 than they are 30 years later, in 2080 under the more moderate B1 scenario. These results illustrate how uncertain the future extent of glaciation and therefore the changes in runoff really are; they clearly depend on which emission path we will ultimately follow. In order to improve our knowledge and to enhance our understanding to a level where useful decisions regarding adaptation and mitigation can be made, a number of scientific and institutional improvements are necessary. These include a better equipped and denser on-site monitoring network, enhanced use of the available remote sensing and GIS technologies, more adequate modeling studies which take into account the topographic and climatic peculiarities of the tropical Andes, a better collaboration among the scientists and institutions involved, and finally, a better dissemination of results to local stakeholders and decision-makers.

1) INTRODUCTION In the arid and semiarid regions of the tropics and subtropics more than 80% of the freshwater supply originates in mountain regions, affecting more than half of the earth’s population (Messerli, 2001). Much of this water is initially stored as ice in mountain glaciers and then gradually released over time. Mountain glaciers, such as those found in the tropical Andes, therefore act as a critical buffer against highly seasonal precipitation and provide water at times when rainfall is low or even absent. At the same time these glaciers are particularly sensitive to climate change because they are constantly close to melting conditions. They are arguably the most visible indicator of climate change, due to their fast response time, their sensitivity to climate variations and the clear visibility of their reaction (glacier growth or shrinkage) to the public. Indeed the existence and the shrinkage of glaciers in the tropical Andes are deeply rooted in the perceptions of local indigenous people living in the Andes. For example a mountain in the Cordillera Blanca that was once called ‘sleeping lion’ due to the shape of its glacier is now called ‘lion has left’ (i.e. ‘leon dormido’ has become ‘leon se ha ido’; Young and Lipton, 2006). For indigenous people in the Cuzco region local myth holds that ‘when the snow disappears from the tops of the mountains, it will herald the end of the world’ (Regalado, 2005). While the ramifications of disappearing glaciers may in reality not be quite that dramatic, major repercussions will be felt throughout Andean countries, which rely on fresh water from glaciated basins during the dry season. Glaciers retain water that falls on the glacier as snow and release it later so that water is available for domestic, agricultural or industrial use even at times when rainfall is low or absent. This looming threat of changes in water supply associated with tropical glacier retreat has received little attention so far, mostly because the climate change community is very much focused on observed and projected large climatic changes at high northern latitudes. However, these global projections rely on models, which due to their coarse resolution, are inadequate to resolve the steep topography of long and narrow mountain chains, such as the Andes. As a consequence, climate change at high tropical locales is not well simulated in these models. Indeed, when considering the rate of warming in the free troposphere (e.g. Bradley et al., 2004, 2006) rather than at the surface, it becomes evident that warming in the tropical Andes is likely to be of similar magnitude as in the Arctic, and with consequences that may be felt much sooner and which will affect a much larger population. This report provides the scientific basis regarding climate change and glaciers in the tropical Andes. It is meant to facilitate a discussion amongst policy makers and local stakeholders, who are in charge of implementing adaptation and mitigation strategies. Such measures, designed to alleviate the magnitude and impacts of climate change, can only be put in place with an accurate and detailed knowledge on how future climate change will affect glaciological and hydrological systems in the Andes. Therefore it is important to first summarize the current state of knowledge regarding climate change, tropical glaciology and potential future impacts on Andean water supply. This report first documents observed historical changes in glaciation in all tropical Andean countries (section 2). A detailed description of how glaciers interact with and respond to changes in climate by adjusting their mass balance (section 3) and how the energy received at the glacier surface is consumed by melting and sublimation processes (section 4) is given

next. A review of observed changes in climate during the 20th century as well as projections of how climate might change during the 21st century (section 5) is needed to put the observed cryospheric changes into a climatic context. Only after a thorough review and discussion of all these factors is it possible to assess the potential ramifications of the observed and projected future glacier retreat for glacier discharge and downstream water supplies (section 6). Finally this reports ends with some concluding remarks and some recommendations as to how the scientific network in the Andes should be expanded (section 7).

2) OBSERVED GLACIER VARIATIONS In the following sections a review of glacier fluctuations in the tropical Andes is provided, for Venezuela, Colombia, Ecuador, Peru and Bolivia. Although Chile also has a few glaciers which can be considered tropical in the broadest sense in its northernmost corner, these glaciers are all located along the border with Bolivia in the Cordillera Occidental, and they are therefore covered in the Bolivian chapter. We limit our analysis to fluctuations since the Little Ice Age (LIA) when glaciers in most of the tropical Andes reached a maximum extent. Most of the discussion however, will be restricted to the 20th century, with a special emphasis on the last 3-4 decades. Since the 1970s much more detailed information has become available, thanks to the initiation of several glacier monitoring programs on selected glaciers and the availability of various new satellite products. This chapter is purely descriptive, showing where and when glaciers changed in size, length and volume (mass), but it does not address the climatic causes or the potential impacts of the observed change. These issues will be dealt with in later sections.

2.1. Venezuela Glaciers in Venezuela are currently restricted to only 3 peaks in the Sierra de Merida: Pico Bolivar (5002 m), Humboldt (4942 m) and Bonpland (4839 m). Combined the 5 remaining cirque glaciers cover less than 2 km2 and extend down to elevations of 4450 m. Glaciers at these locations have been continuously retreating during historical times and are currently not in equilibrium with the modern climate. A comparison of early sources (reports and paintings) with present-day conditions indicates a rapid glacier retreat during the last 100 years. Reports by Schubert (1992, 1998) show that these glaciers have lost more than 95% of their glacier-covered area since the mid-19th century. Their extent was estimated at approximately 10 km2 in 1910 and about 3 km2 in 1952. Of the 10 glaciers mapped in 1952, 4 have completely or almost completely disappeared, 1 has disintegrated into firn patches, and the remaining 5 are substantially smaller.

2.2. Colombia Glacier monitoring in Colombia grew mostly out of concern for natural hazards. The combination of glaciers, active volcanoes and earthquakes poses a significant threat. The 1985 lahar on Nevado del Ruiz was one of the deadliest ever recorded, with 23,000 deaths and most of Armero, a city of about 25,000, disappearing under a mudflow (Hoyos-Patino, 1998). In June 1994, an avalanche originating in the Nevado del Huila killed at least 1,500 people (Hoyos-Patino, 1998). Nonetheless, glaciology is still a very young science in Colombia and all information on glacier change is based on glacier length observations. Mass-balance studies are just starting and results from these programs are not yet available (Ceballos et al., 2006). During the LIA, glaciers in Colombia advanced to about 4200 m a.s.l. in the south and to about 4600 m a.s.l. in the north, occupying a total area of about 374 km2 (Florez,

1992). Dramatic glacier recession took place in the 20th century, most pronounced from the mid-1980s onward. However, various figures have been given for the total extent of the remaining glacierized area in Colombia. Jordan et al. (1989) estimated a total of 246 glaciers on 9 mountains, with a total glacierized area of 109 km2. This work was based on fieldwork and aerial photography dating from 1957 to 1978. Hoyos-Patino (1998) measured the extent of ice-and-snow areas on Landsat-MSS images from the early 1970's and determined a total area of 104 km2. Thouret et al., (1996) gave a range from 100 to 112 km2 for the presently glacierized areas of Colombia. In 2003, according to Ceballos et al. (2006), glacier termini had retreated to 4700 – 4900 m a.s.l. and the total glacierized area had shrunk to 55.4 km2. Today six different mountain ranges still have glacier coverage, while eight Nevados (formerly glacierized areas) have lost their glaciers entirely during the 20th century (Ceballos et al., 2006). The currently glaciarized mountains are: the Sierra Nevada de Santa Marta, Sierra Nevada del Cocuy and Nevados del Ruiz, de Santa Isabel, del Tolima and del Huila. The glacierized areas on these mountains vary from 1 to 20 km2. The regions most affected by glacier shrinkage are the Sierra Nevada de Santa Marta and the Nevado de Santa Isabel (Figure 1). In the former, a 50% area loss has been observed in the last 20 years, while the latter showed a 50% loss in just 15 years (Ceballos et al., 2006). Sierra Nevada del Cocuy and Nevados de Tolima and del Ruiz have also been greatly affected, losing 35–45% of their glacier area in the last 15–17 years of the 20th century. Glaciers on Nevado del Huila showed a comparatively moderate shrinkage, with about a 20% loss in the last 20 years (all estimates from Ceballos et al., 2006).

Figure 1: Glacier extent on Nevado de Santa Isabel as seen from aerial photography in 1959 and 1996 and a Landsat Thematic Mapper satellite image in 2002 (from Ceballos et al., 2006).

Glacier length measurements conducted since the 1980s reveal a similar picture of rapid retreat. On Nevado de Santa Isabel, glacier termini retreated 170–250 m in only 15 years (1988– 2003; Figure 2), which translates into an annual retreat rate of 11–17 m. In the Sierra Nevada del Cocuy, where the total glacier area is about six times larger than on Nevado de Santa Isabel, the retreat at some glacier tongues was nearly 500 m in the last

18 years (Figure 3). This is a very large value since the total lengths of individual glaciers in the Cocuy region are about 1â&#x20AC;&#x201C;1.5 km (all estimates from Ceballos et al., 2006).

Figure 2: Cumulative retreat of different glacier tongues of Nevado de Santa Isabel (from Ceballos et al., 2006).

Figure 3: Cumulative retreat of different glacier tongues in Sierra Nevada del Cocuy (from Ceballos et al., 2006).

The small elevation range of most Colombian glaciers makes them particularly vulnerable to climate change and therefore future disappearance of several glaciers is projected for the next few decades (Hoyos-Patino, 1998; Ceballos et al., 2006). In addition, edge effects may become increasingly more dominant, especially for the smaller ice caps, such as on Nevado de Santa Isabel (Figures 1-2). The extraordinary loss of almost 40% of its area in only 6 years from 1996 to 2002 supports this hypothesis and makes a further rapid shrinkage likely (Ceballos et al., 2006). The larger and compact ice caps on the active volcanoes Nevados del Ruiz and del Huila have experienced slightly

(Ruiz) and significantly (Huila) smaller glacier retreat in the last 20 years (Ceballos et al., 2006).

2.3. Ecuador Ecuador's glaciers are situated closer to the equator than any other Andean glaciers and are therefore considered typical for examples of continental tropical glaciation (Hastenrath, 1981; Jordan and Hastenrath, 1998). The glaciers are restricted to the highest peaks, mostly of volcanic origin, and do not contain large contiguous ice fields, such as those found in Peru, or Bolivia. Instead the glaciers occur as ice caps that feed numerous outlet glaciers and are confined to the limited summit areas (Jordan and Hastenrath, 1998). The glaciers in Ecuador are located on two mountain chains, the Cordillera Occidental and the Cordillera Oriental. According to Jordan and Hastenrath (1998) 4 mountains are glacierized in the Cordillera Occidental, and 13 in the Cordillera Oriental. Glaciers are more common in the Cordillera Oriental because it is exposed to the moisture supply from the Amazon basin. The total glacierized area in Ecuador in 1998 was 97.21 km2, of which 21.92 km2 was located in the Cordillera Occidental and 75.29 km2 in the Cordillera Oriental (Jordan and Hastenrath, 1998). Various historical sources indicate a rather extensive glaciation from the 1500's to the first part of the 1800's, followed by a drastic ice recession starting around the middle of the past century and continuing to the present time (Hastenrath, 1981). More information is available for the last few decades, thanks primarily to the tropical glacier monitoring program, spearheaded by the French IRD (Pouyaud, et al., 1995). Detailed measurements on Antizana 15 glacier, including monitoring of glacier length changes and glacier mass balance have been conducted since 1994 (Semiond et al., 1997; 1998). This ice cap, located only 40 km east of Quito is of special interest given the use of its glacial runoff for the capitalâ&#x20AC;&#x2122;s water supply (Francou et al., 2004). Aerial photogrammetry, starting in 1956 has been used to put the on-site monitoring, which started in 1995, in a longer-term perspective (Francou et al., 2000; 2004). Results show that the glacier retreated 7-8 times faster between 1995 and 2000 than during the previous period 1956-1993 (Figure 4; Francou et al., 2000). This period was influenced by strong El NiĂąo events and a later study (Francou et al., 2004) subsequently confirmed the strong sensitivity of the glacier mass balance on Antizana (Figure 5) to ENSO extremes. Aerial photography was also used on Cotopaxi volcano to reconstruct glacier extent and recession since the mid 1950â&#x20AC;&#x2122;s (Jordan et al., 2005). Results from this study show that glaciers on Cotopaxi were almost stagnant between 1956 and 1976 and then lost approximately 30% of their surface area between 1976 and 1997 (Figure 6). The calculated total mass (thickness) loss on selected snouts of Cotopaxi between 1976 and 1997 equals 78 m, or 3-4 m w.e. yr-1, consistent with similar values obtained on Antizana.

Figure 4: Antizana Glacier 15: terminus fluctuations over the last four decades (from Framcou et al., 2000).

Figure 5: Monthly and cumulative mass balance on Antizana 15a glacier (ablation zone) between 1995 and 2002 (from Francou et al., 2004).

Figure 6: Cotopaxi glaciers in orthophoto (left) and retreat between 1976 (red) and 1997 (blue). (from Jordan et al., 2005).

2.3. Peru The Peruvian Andes contain the largest fraction of all tropical glaciers, and glaciers in Peru are among the best studied in the tropical Andes. In 1988 the total ice covered area was estimated at 2,600 km2 (Morales-Arnao, 1998, see Figure 7).

Figure 7: Glacierized mountain ranges in Peru (from Morales-Arnao, 1998).

The largest single glacier is the Quelccaya ice cap, located in the Cordillera Vilcanota, with 8 of the next 11 largest glaciers located in the Cordillera Blanca (Morales-Arnao, 1998). We refrain from giving a detailed description of every single glacierized mountain region in Peru and refer the reader to the compilation in MoralesArnao (1998), which gives a nice regional overview and a detailed description of all glacier massifs. Here we focus primarily on the two glacierized areas, which have received the most attention and where by far the most scientific research has been done, the Quelccaya Ice cap and the Cordillera Blanca.

2.4.1. Quelccaya Ice Cap The Quelccaya ice cap is situated in the Cordillera Vilcanota in the eastern branch (Cordillera Oriental) of the Peruvian Andes (Figure 7). It a low lying (summit elevation of only around 5,700 m) glacierized plateau near the drop-off to the wet Amazon basin and is among the few large ice plateaus in the tropics. In 1998 its size was estimated at 54 km2 (Hastenrath, 1998). The Quelccaya ice cap has been the target of detailed research by Prof. Lonnie Thompson and his group from Ohio State University for over 20 years. The main aim of this research is of paleoclimatic nature (reconstructing climate of the past by analyzing ice cores drilled on the summit). In parallel, however, the extent of a large outlet glacier to the west, Qori Kalis, has been monitored annually. This outlet glacier has retreated throughout the entire monitoring period and a large lake formed in its former snout area (Fig. 8). Brecher and Thompson (1993) first noted the accelerated recession, with a retreat rate that was nearly 3 times as fast between 1983 and 1991 as between 1963 and 1978, and a rate of volume loss, which was nearly seven times as great. More recent analyses show that the retreat rate was 10 times faster (~60 m yr-1) between 1991 and 2005 than in the initial measuring period, 1963-1978 (Thompson et al., 2006). However, the recent retreat rate may not reflect climate forcing alone, but also depend on the glacier bed geometry.

Figure 8: Retreat of Qori Kalis glacier (from Thompson et al., 2006)

2.4.2. Cordillera Blanca The Cordillera Blanca is the worldâ&#x20AC;&#x2122;s most extensively glacier-covered tropical mountain range (Morales-Arnao, 1998), with several peaks above 6,000 m (Figure 9). A total of 722 individual glaciers were recognized in the Cordillera Blanca based on air photos from 1962 to 1970, covering an area of 723.4 km2 (Ames et al., 1989). 530 glaciers were identified as west-sloping (covering an area of 507.5 km2) and 192 glaciers (covering an area of 215.9 km2) were facing east in the assessment by Ames et al., (1989). The estimated ELAs were generally higher in the west than in the east, reflecting the east-west precipitation gradient. The majority of glacierized watersheds in the Cordillera Blanca discharge toward the Rio Santa, the second largest river in Peru discharging into the Pacific. Four hydroelectric power plants are situated along the river between the Cordillera Blanca and the Pacific coast (Mark, 2007).

Figure 9: Glaciers in the Cordillera Blanca ((from Morales-Arnao, 1998).

The initiation of glacier monitoring networks in the Cordillera Blanca was in part motivated by several catastrophic outbursts and floods from proglacial lakes and ice avalanches. In May 1970 the city of Yungay was destroyed by an ice avalanche from Huascaran Norte, with an estimated 18,000 people dead (Patzelt, 1983; Portocarrero, 1995; Ames, 1998; Morales-Arnao, 1998; Mark, 2007). The recent glacier retreat in the Cordillera Blanca has lead to the formation of a number of proglacial lakes dammed by moraines, many of which appear to be increasing in size (Ames, 1998). Glacier shrinkage in the Cordillera Blanca probably started at the end of the Little Ice Age (LIA), in the middle of the 19th century (Ames, 1998; Kaser, 1999 and references therein; Kaser and Osmaston, 2002; Solomina et al., 2007). There is evidence for an advance in the mid 1920s at least in parts of the Cordillera Blanca, followed by a rapid retreat in the 1930s (Ames, 1998; Kaser, 1999). More detailed information becomes available only later in the 20th century thanks to better maps, aerial photographs and eventually a regular on-site monitoring. Kaser et al. (1990) showed, based on aerial photographs and ground measurements that a general retreat occurred over the period 1940-1990. Much of the glacier shrinkage and tongue retreat may have occurred in the time between 1930 and 1950, when a significant rise of the ELA was observed (Kaser and Georges, 1997). The general retreat of the 20th century was again interrupted by minor advances in the late 1970s, which Kaser and Georges (1997) associated with generally cooler and wetter conditions. On glacier Artesonraju a retreat of 1140 m was observed between 1932 and 1987 and glacier Broggi retreated an average of 17.4 m yr-1 between 1932 and 1994, despite a slight advance around 1977 (Ames, 1998). Glacier Pucaranra similarly lost 690 m in length between 1936 and 1994 and glacier Urushraju retreated 675 m since 1932 (Ames, 1998). A similar analysis on glacier Yanamarey showed that its tongue retreated by 350 m between 1948 and 1988 (Hastenrath and Ames, 1995) and an average of 8.9 m yr-1 between 1932 and 1994 (Ames, 1998). Today glacier retreat on Yanamarey continues unabated with an estimated 20 m yr-1 (average 19772003), four times the speed of 5 m yr-1 observed between 1948 and 1977 (Mark et al., 2005). The loss of ice volume from 1948 to 1982 was estimated at 22x106 m3 and at 7x106 m3 between 1982 and 1988 (Hastenrath and Ames, 1995). On HuascaranChopicalqui the glacier extent decreased from 71 km2 in 1920 to 58 km2 in 1970 (Kaser et al., 1996a). This amounts to an increase in the ELA of ~95 m. Mark and Seltzer (2005a) estimated a glacier volume loss of 57x106 m3 between 1962 and 1999 in the Queshque massif of the southern Cordillera Blanca. This translates into a glacier thinning of 5 â&#x20AC;&#x201C; 22 m and an estimated ELA rise of 25-125 m, depending on the aspect of the glacier. A comprehensive overview over the entire Cordillera Blanca was provided by Georges (2004) based on new and revised data, including new estimates from SPOT satellite images. The revised numbers suggest an overall decline in glacierized area from ~ 850-900 km2 during the LIA to 800-850 km2 in 1930, 660-680 km2 in 1970 and 620 km2 in 1990. The ice coverage at the end of the 20th century was estimated to be slightly less than 600 km2. Georges (2004) also indicated that some glaciers had again stopped their retreat or even slightly advanced between 1999 and 2002, most likely as a consequence of persistent cold conditions in the tropical Pacific. Slightly different numbers were published by Silverio and Jaquet (2004). On the basis of Landsat TM data

they detected 643 km2 of glacierized area in 1987, and 600 km2 in 1991. As shown by Raup et al. (2006) these differences are mostly due to different methodologies and definitions as to what should be included as â&#x20AC;&#x2DC;glacierâ&#x20AC;&#x2122; (e.g peripheral snow fields, snow covered ground above the glacier accumulation zone, stagnant debris-covered ice, etc.). Raup et al. (2006), in a detailed study of the Huandoy-Artesonraju massif found a reduction in glacier size of 20% between 1962 and 2003, and a rise of the glacier termini of approximately 60 m (Figure 10).

Figure 10: Hypsometric distribution of glacierized area in the Huandoy-Artesonraju massif in 1962 and 2003. The total glacier area is reduced by 20% and the shift of the glacier to higher elevation is evident (from Raup et al., 2006).

While variations in glacier length and size are quite easy to observe, they are not that easy to interpret because the underlying climate signal may be delayed and filtered through glacier flow dynamics. Mass balance measurements, on the other side, are much harder to obtain, but provide a direct climate signal at the glacier surface. Mass-balance measurements were started in the Cordillera Blanca in 1966 on Pucahirca glacier and then extended between 1977 and 1983 by the glaciology department of ElectroperĂş to three other glaciers, Uruashraju, Yanamarey, and Santa Rosa (Kaser et al., 1990). These measurements were mainly restricted to the ablation areas. Kaser et al. (2003) were later able to reconstruct the annual mass balance history from the Cordillera Blanca back to 1953 based on glacial runoff data. Their data show a general negative trend, characterized by strong mass loss in the 1950s and 60s, interrupted by a short period of mass gain in the early 1970s. Afterwards the negative mass balance resumed again (Figure 11).

Figure 11: Cumulative glacier mass balance and individual glacier terminus variations in the Cordillera Blanca (from Kaser et al., 2003).

2.4.3. Coropuna Studies on glacier changes in Peru outside the Cordillera Blanca and the Quelccaya regions are virtually non-existent. One of the few exceptions is a recent analysis of glacier size and thickness change on Coropuna in the Cordillera Ampato in southwestern Peru, based on a comparison of older aerial photography with recent Shuttle Radar Topography Mission (SRTM) and Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) satellite data (Racoviteanu et al., 2007). In their study the authors found a significant decrease in glacier extent from 82.6 km2 in 1962 to 60.8 km2 in 2000. The glacier thickness appears to have changed as well, with a glacier thickening in the summit region and a thinning at lower elevations (Figure 12).

Figure 12: Glacier thickness change between 1955 and 2000 on Coropuna based on aerial photography derived maps (1955) and SRTM data (2000). The thinning at lower and thickening at higher elevations is apparent (from Racoviteanu et al., 2007).

2.5. Bolivia Glaciers in Bolivia are restricted to the highest peaks of the Andes. They can be found in two main mountain ranges, the Cordillera Occidental along the western border with Chile and the Cordilleras Apolobamba, Real, Tres Cruces and Nevado Santa Vera Cruz in the east (Jordan, 1998). Glaciers in the Cordillera Occidental are limited to Nevado Sajama (6,542 m) and its neighboring volcanoes, with a total surface area in 1998 of about 10 km2 (Jordan, 1998). They consist of small summit ice caps on extinct volcanoes, which due to the dry conditions have the highest minimum elevation on earth, with an ELA several hundred meters above the 0Ë&#x161;C isotherm (Messerli et al., 1993). Most glaciers, however, are located in the eastern Cordilleras and consist of ice caps, valley and mountain glaciers. Their extent was estimated at 550 km2 in 1998 (Jordan, 1998). Because of the limited precipitation, no glaciers exist today in southern Bolivia. The first compilation of a glacier inventory in Bolivia was started in 1980 (Jordan et al., 1980; Jordan, 1983; 1985). In the early 1990s the French IRD (formerly ORSTOM) started a detailed monitoring program on several glaciers in the Cordillera Real, namely glaciers Charquini, Chacaltaya and Zongo (Pouyaud et al., 1995). Results from these studies are discussed in more detail below.

2.5.1. Charquini, Chacaltaya and Zongo glaciers While early studies focused on a better understanding of glacier mass balance and the dominant influence of ENSO on interannual variability of mass balance and runoff (e.g. Francou et al., 1995a, 1995b, 1998; Ribstein et al., 1995b), subsequent studies started to focus on longer-term behavior and trends. Rabatel et al. (2005, 2006) mapped and dated LIA moraines on the Charquini massif and were able to show that the maximum glacier extent had been reached at the second half of the 17th century. Afterwards glaciers started to retreat almost unabated until today (Figure 13). This interpretation is consistent with the current decay of a rock glacier in southern Bolivia, which by the time of the LIA was still ice covered but now shows severe signs of degradation (Francou et al., 1999). Glaciers on Charquini have by now lost between 6578% (depending on aspect) of their LIA size and the ELA rose by approximately 160 m between the LIA maximum extent and 1997 (Rabatel et al., 2006). The authors further showed that recession rates have increased by a factor of 4 over the last decades and that Charquini glaciers experienced an average mass deficit of 1.36 m w.e. yr-1 between 1983 and 1997. This led the authors to speculate that all glaciers in the Cordillera Real located below 5300 m are in imbalance with current climate and may completely disappear in the near future (Rabatel, 2006). 100 80

Charquini Sur Charquini Sureste Charquini Noreste Charquini Norte

Charquini Oeste

ELA change (m)

40 20 0 -20 -40 -60 -80 -100 1600

1700

1800 Date AD

1900

2000

Figure 13: Change in ELA determined from dated moraines on Charquini using the AAR method. Current AAR was assumed constant over time. For the last 50 years, glaciers have been generally unbalanced and consequently situated at higher elevation than these estimates. The zero represents the average altitude of the ELA over the whole period (modified from Rabatel et al., 2006).

Ribstein et al. (1995b), based on a reconstruction of glacier discharge since the early 1970s, showed that Zongo glacier is losing more mass than is replenished by precipitation. 11 out of the 17 years analyzed showed an extremely negative hydrologic balance. In 2000 Francou et al. (2000) reported first results from the glacier mass balance studies on Chacaltaya. They showed that the glacier had lost 62% of its mass between 1940 and 1983 and that its remaining size in 1998 was only 7% of the extent in 1940. In the early 1990s the glacier still functioned as a small ski resort. The results by Francou et al. (2000) further suggested 3-5 times higher ablation rates in the 1990s than in previous decades, with an average loss of 1400 mm w.e. yr-1, and they made the prediction that the glacier might completely disappear within 10 years. Chacaltaya is a very small and lowlying glacier, and therefore particularly vulnerable to climate change. The retreat of such small glaciers is accelerated once they reach a critical size, below which â&#x20AC;&#x2DC;edge effectsâ&#x20AC;&#x2122; become important. At the edge of tropical glaciers air temperature above surrounding rocks can exceed 20Ë&#x161;C (Francou et al., 2003), and hence advection of warm air above the glacier can become very important. Nonetheless Chacaltaya must be considered representative of many glaciers in the region, since more than 80% of all glaciers in the Cordillera Real are less than 0.5 km2 in size (Francou et al., 2000). Ramirez et al. (2001), based on new data from ground penetrating radar (GPR) concluded that the glacier had lost 40% of its thickness in only 6 years from 1992 to 1998, as well as two thirds of its total volume. The low elevation of the glacier essentially meant that it had lost its accumulation zone, as the ELA had moved above the uppermost reaches of the catchments (Ramirez et al., 2001). According to the calculations by the authors, a 200 m lowering of the ELA would be required to stabilize the glacier. Francou et al. (2003) took a closer look at mass balance variability and trends on both Chacaltaya and Zongo glacier and showed that both glaciers featured very similar interannual variability and longer term trends, although data from Zongo is from the ablation zone only (Figure 14).

Figure 14: Cumulative mass balance from Zongo and Chacaltaya glaciers, Bolivia between 1991 and 2001 (from Francou et al., 2003).

The negative trend on Chacaltaya was interrupted briefly in 1993, 1996 and 2000. On the other hand the glacier lost a third of its entire volume (6 m w.e.) in an exceptional 18

month period between 1997 and 1999 (Figure 14). Today the glacier has essentially disappeared and disintegrated into a few small stagnant ice fields (Coudrain et al., 2005; Figure 15).

Figure 15: Evolution of the limits of Chacaltaya Glacier from 1963 to 2005. The outer border is probably dating from the second part of the 17th century, as on Charquini Sur glacier, where this moraine was dated by lichenometry ( from Berger et al., 2006).

2.5.2. Cordillera Occidental (Sajama)

Outside the Cordillera Real very little work on glacier variations has been done. Arnaud et al. (2001), based on aerial photographs and Landsat data, observed as steady rise in the snowline of a glacier on Sajama in the western, arid Cordillera Occidental, between 1963 and 1998. The number of images analyzed, however, was limited and the dates analyzed were strongly influenced by interannual variability due to ENSO, both of which inhibited the authors from drawing any general conclusions regarding the significance of the observed rise in snowline.

2.6. Summary The previous sections on glacier size and mass balance variations from Venezuela, Colombia, Ecuador, Peru and Bolivia give a detailed and unequivocal account of rapid and accelerated retreat of tropical Andean glaciers since the LIA. While the climatic forcing behind this retreat may have varied over time and may not necessarily be the same everywhere, evidence for a coherent regional pattern is beginning to emerge. Measurements of glacier area and length on 10 glaciers in Ecuador, Peru and Bolivia for example provide a clear picture of a region-wide glacier shrinkage, which appears to have accelerated over the last few decades (Francou and Vincent, 2007; Fig. 16). 0

-50000

-400

-100000 Antizana 15a

-600

Antizana 15b

-150000

Yanamarey -800

-200000

Broggi Pastoruri

-1000

-250000

Uruashraju Cajap

-1200

-300000

Cumulative area evolution in mÂ˛

Cumulative length evolution in m

-200

Zongo (area) -1400

Charquini-S (area)

-350000

Chacaltaya (area) -1600 1930

1940

1950

1960

1970

1980

1990

2000

-400000 2010

Figure 16: Change in length and surface area of 10 tropical Andean glaciers from Ecuador (Antizana 15a and 15b), Peru (Broggi, Pastoruri, Urushraju, Cajap) and Bolivia (Zongo, Charquini, Chacaltaya) between 1930 and 2005. (from Francou and Vincent, 2007).

Mass balance records from Bolivia and Ecuador similarly show a very coherent picture, with a generally negative mass balance, which appears to be driven by the same background forcing throughout the region (Figure 17). Superimposed on this general retreat are interannual variations with occasional periods of equilibrated or even positive

mass balance, in particular during prolonged La NiĂąa events, such as between 1999 and 2001. cumulative

per year 2000

Specific net mass balance [mm of w.e.]

Cumulative mass balance [mm of w.e.]

0 -2000 -4000 -6000 -8000 -10000 -12000 -14000

Chacaltaya Zongo

-16000 -18000

Antizana15a Charquini S.

-20000

-2000 Chacaltaya Zongo Antizana15a Charquini S. -4000

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

Hydrological years

Figure 17: Comparison of cumulative and annual mass balance on glaciers in Bolivia and Ecuador. Note that the hydrological year is September-August in Bolivia, and January-December in Ecuador (from Berger et al., 2006).

In order to understand what is causing this retreat and what factors may have been responsible in the past, we first have to review how tropical glaciers respond to climate change. This entails a review of mass and energy balance studies (sections 3 and 4) which will discuss the peculiarities of tropical glaciers and how their response and their sensitivities to climate differ from mid- or high latitude glaciers. Only then, in combination with a review of observed 20th century climate change in the Andes, can we obtain a complete picture of tropical glacier â&#x20AC;&#x201C; climate interactions and attempt to understand what has caused this dramatic retreat.

3) GLACIER MASS BALANCE 3.1 General characteristics of tropical glacier mass balance The glacier mass balance provides the most immediate link between a glacier and its surrounding climate. The mass balance describes where and how a glacier is gaining or losing mass due the predominance of accumulation or ablation processes, which in turn are determined by climatic variables such as temperature, precipitation, solar radiation, humidity, etc. (e.g. Kaser, 2002). Because of the lack of a pronounced thermal seasonality (temperatures stay more or less constant throughout the year) but a clear differentiation between dry and wet seasons (one each in the outer tropics, almost constant precipitation near the equator), tropical glacier mass balance and its sensitivity to climate change is fundamentally different from mid- and high latitude glaciers (see Kaser and Osmaston (2002) for a detailed review). Because of these peculiarities of tropical climate, accumulation is confined to one wet season (outer tropics) or occurs throughout the year (inner tropics), while ablation can take place all year long. So unlike mid-latitude glaciers where accumulation and ablation are separated into a winter accumulation and a summer ablation season, in the tropics ablation and accumulation can occur at the same time (Figure 18). Also, because temperature does not change much throughout the year, ablation occurs predominantly in the ablation zone, below the ELA and accumulation is restricted to regions above the snow-rain line, which remains more or less constant throughout the year (e.g. Kaser, 1995; Kaser et al., 1996; Kaser and Georges, 1999).

Figure 18: Schematic comparison of glacier mass balance in mid latitudes, inner and outer tropics (from Kaser and Georges, 1999).

Due to these differences in the mass balance, inner tropical glaciers have a much stronger vertical net mass balance gradient. This vertical mass balance profile describes the distribution of the specific net mass balance with altitude along a vertical profile from the glacier snout to its highest point (Kaser, 1995; Kaser et al., 1996). Figure 19 shows typical vertical mass balance profiles for the mid latitudes, the inner tropics and the subtropics. It is immediately apparent that inner tropical glaciers have a much higher negative mass balance (higher ablation), than mid-latitude glaciers below their ELA. This also means that the vertical distance between ELA and the glacier snout is generally less in the inner tropics. This characteristic of inner tropical glaciers is mostly due to the lack of thermal seasonality and the exposure of the ablation zone to melt and sublimation 365 days per year (Kaser, 1995). In the subtropics on the other hand ablation is much reduced, mostly because sublimation dominates over melt due to generally low humidity (Kaser, 2001). The profiles in Figure 19 are simulated using some idealized assumptions, such as 100 days of ablation in mid-latitudes and 365 days in the inner tropics, and having all available energy consumed by sublimation instead of melting in the subtropics (see section 4). Nonetheless they compare very favorably with actual measured mass balance profiles (Kaser, 2001).

Figure 19: Modeled vertical mass balance profile at a mid-latitude, subtropical and tropical glacier. Note that the 0Ë&#x161;C-line may differ from the ELA (above the ELA in the tropics, below the ELA in the outer tropics; from Kaser and Georges, (1999)).

Because of the differences in their vertical mass balance profiles, the sensitivity of the ELA to climate change is quite different in the humid inner tropics (e.g. Ecuador) as opposed to the drier subtropics (e.g. Bolivia). In the inner tropics the ELA is generally close or slightly below the 0Ë&#x161;C-line and hence the ELA will react to changes in temperature because a temperature increase would have immediate impacts on ablation, but also change the accumulation due to a shift in the rain-snow line. In the dry outer tropics, where the ELA is usually located considerably above the 0Ë&#x161;C-line, a temperature increase will not have such an immediate effect. Instead the ELA will be more sensitive to changes in the vertical gradient of accumulation and hence changes in precipitation,

but also to humidity changes, which determines the ratio of melting to sublimation (Kaser, 2001). Glacier terminus variations are often used to assess how glaciers respond to climate, as they are very easy to monitor (see previous section). The terminus of a glacier is determined by the balance between net ablation and downward mass flux. Where the two are equal, defines the end point of the glacier. Because of the strongly negative mass balance in the ablation zone and the close proximity to the ELA, glacier tongue variations are generally more rapid and more pronounced in the tropics (Kaser, 1995). The unique mass balance characteristics of tropical glaciers of course have significant ramifications for the catchmentsâ&#x20AC;&#x2122; downstream hydrology. In contrast to mid latitudes, the runoff regime is smoothed in glacierized catchments, especially in the outer tropics, where glacial melt water often provides the only runoff during the dry season (see section 6.).

3.2. Mass balance - climate relationships In the following we discuss mass balance characteristics from a few regions in the tropical Andes, where continuous measurement programs are under way. Before discussing these results in detail, it should be emphasized that most mass and energy balance studies, due to logistical reasons, stem from glaciers that are easily accessible and safe to work on. In effect this means that they are usually rather small and relatively lowlying. This in turn may affect the results to some degree, as there are indications that small and low-lying glaciers have more negative mass balances and have experienced greater loss than colder glaciers, located in higher, inaccessible terrain (Francou et al., 2005; Kaser et al., 2006). In addition mass balance measurements are often limited to the ablation zone, because of safety concerns, but also because the lack of a clear annual stratigraphy makes accumulation measurements very difficult (e.g. Kaser and Georges, 1999). Finally the limited sampling density (estimates for entire glaciers are often derived form a limited number of stakes) can lead to inaccurate or biased results, especially during extremely negative or positive years (Sicart et al., 2006). These limitations should be kept in mind during the following discussion. On Zongo and Chacaltaya glaciers in Bolivia mass balance measurements with stake networks were initiated in the early 1990s (Francou et al., 1995a, b). As shown by Francou et al. (2003) there is a strong seasonality associated with mass balance variability; that is, the largest differences from year to year occur during the summer months Octoberâ&#x20AC;&#x201C;April. This implies that the accumulation and ablation seasons coincide, as discussed in the previous section. During the dry and cold winter months Mayâ&#x20AC;&#x201C; September on the other hand, mass balance is always near the equilibrium and does not display any significant variations from year to year (Figure 20, left). Accordingly, the annual mass balance largely reflects the variability in summertime accumulation and ablation. The largest fraction of year-to-year mass balance variability on Chacaltaya can be attributed to the three summer months DJF, which alone account for 78% of the total variance of the annual mass balance (Francou et al., 2003). Continuous monthly mass balance measurements from the ablation zone of Antizana 15 glacier in the Andes of Ecuador between January 1995 and December 2002

200

100

net balance [mm w.e.]

(Francou et al., 2004) reveal quite a different picture. Here, on seasonal timescales, mean ablation rates remain at a quite constant level all year round (Figure 20, right), although the periods Februaryâ&#x20AC;&#x201C;May and September show much larger variations from year to year. The large variability during those months can be explained by the dominant influence of ENSO on the glacier mass balance and the large differences that occur in the seasonal cycle during the two opposite phases of ENSO (see next section).

-100 -200 -300

-400

-500

-600

-600 JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

Figure 20: Average monthly mass balance on left) Chacaltaya (1991-2001,modified from Francou et al., 2003) and right) Antizana (1995-2002, modified from Francou et al., 2004). Vertical bars indicate 1 standard deviation.

On interannual time scales Francou et al. (2003) found a clear inverse relationship between mass balance on Chacaltaya and temperature (Figure 21, left). Periods of negative (positive) mass balance values coincide with positive (negative) temperature anomalies. This is in seeming contradiction to earlier statements that glaciers in the outer tropics are more sensitive to changes in precipitation and humidity than temperature. However, as shown by Francou et al. (2003) temperature in this region is strongly correlated with humidity, cloudiness and precipitation, especially on interannual timescales. Since temperature integrates all the fluxes, it appears to be significantly correlated with mass balance on longer timescales, but the apparent correlation between temperature and mass balance does not reflect the real physical processes present at the glacier surface (Francou et al., 2003). On Antizana, on interannual time scales mass balance is equally closely related to temperature variations (Figure 21, right). Here however, the impact of temperature appears to be more direct. Higher temperatures in the inner tropics, where Antizana is located, directly influence the mass balance through changes in the snow-rain line, which reduces accumulation in the lower zone of the glacier, exposes the snout to rain as opposed to snow and thereby lowers the albedo (Francou et al., 2004). Besides increased temperature, weak and sporadic snowfall, insufficient to maintain a high glacier albedo, low wind speeds, which limit the transfer of energy from melting to sublimation, and reduced cloud cover, which increases the incoming short-wave radiation, were additional factors found by Francou et al. (2004), which negatively affect glacier mass balance in the inner tropics.

1.20

300 T500-12m BAL-12m

100

0.00

feb-01

feb-00

feb-99

feb-98

-300

feb-97

-1.20

feb-96

-200

feb-95

-0.80

feb-94

-100

feb-93

-0.40

Mass balance anomaly [mm w.e.]

200

0.40

feb-92

Temperature anomaly (°C)

0.80

Figure 21: Apparent anti-correlation between mass balance and temperature on interannual time scales. Left: Chacaltaya glacier (1992-2001, modified from Francou et al., 2003) and right Antizana 15 glacier (1995-2002, modified from Francou et al., 2004).

3.3. Mass balance and large-scale forcing Francou et al. (1995a, 1995b, 2000), Ribstein et al. (1995a), Wagnon et al. (2001) and Ramirez et al. (2001) first identified the significant role played by ENSO, with El Niño years featuring a strongly negative mass balance and La Niña events producing a nearly balanced or even slightly positive mass balance. These results do not come as a surprise since climate in the tropical and subtropical Andes is significantly influenced by ENSO on interannual timescales (Vuille, 1999; Vuille et al., 2000a, b; Garreaud et al., 2003; Vuille and Keimig, 2004), with La Niña years tending to be wet, while dry conditions usually prevail during El Niño years. In conjunction with dry conditions, the Andes also experiences above average temperatures during El Niño events (Vuille and Bradley, 2000; Vuille et al., 2003). On average, near-surface summer temperatures are 0.7˚–1.3˚C higher during El Niño as compared to La Niña (Vuille et al., 2000a). Tropical glaciers, such as Chacaltaya, thus do not only experience a deficit of summer precipitation and consequently reduced accumulation and a lowered albedo during El Niño events, but are also exposed to higher temperatures and an increase in incoming shortwave radiation due to reduced cloud cover (Wagnon et al., 2001; Francou et al., 2003). Indeed as shown in Figure 22, Chacaltaya mass balance is significantly correlated with tropical Pacific sea surface temperatures (SSTs), in a way which is reminiscent of the canonical ENSO mode. The causal mechanism linking tropical SSTA with glacier mass balance in the Andes is the same as for precipitation, described in previous publications (Vuille et al., 2000a, Garreaud et al., 2003). This result is of relevance because it shows that glacier mass balance is linked to SSTA in the tropical Pacific and that this linkage is being translated through changes in precipitation. It follows that mass balance anomalies on Chacaltaya are largely governed by climatic conditions in the tropical Pacific domain (Francou et al., 2003). This is consistent with observations indicating an accelerated negative mass balance and glacier retreat in many tropical Andean locations after the mid 1970s, concurrent with the 1976/77 Pacific climate shift.

Figure 22: Correlation of Chacaltaya mass balance with tropical SST. White contours indicate 95% significance levels (modified from Francou et al., 2003).

Results from Antizana similarly indicate a strong dependence on ENSO. Over the 8-year period investigated, mass balance was negative all year round during El Ni単o periods but remained close to equilibrium (positive anomalies) during La Ni単a events (Figures 23-24).

Figure 23: Correlation of Antizana mass balance with tropical SST. White contours indicate 95% significance levels (modified from Francou et al., 2004).

While the response to ENSO-related climate variability is very similar on Antizana to what is observed on glaciers in Bolivia, the seasonal dependence and the physical mechanisms linking ENSO with mass balance variations are very different

(Favier et al., 2004a). Unlike in Bolivia, the impact of El NiĂąo is primarily through increased air temperature, which favors rain over snowfall, but to a lesser degree also due to weak and sporadic snowfall, insufficient to maintain a high glacier albedo, low wind speeds, which limit the transfer of energy from melting to sublimation, and reduced cloud cover, which increases the incoming short-wave radiation. La NiĂąa events on the other hand are characterized by colder temperatures, higher snowfall amounts, and to a lesser degree, more constant winds, factors which increase albedo and sublimation and therefore preclude melting at the glacier surface (Francou et al., 2004).

Figure 24: Antizana mass balance anomalies stratified by ENSO events. Individual monthly measurements (small circles), the mean (large circles), and Âą1 standard deviation (vertical bars) are indicated (modified from Francou et al., 2004).

In the Cordillera Blanca, mass balance is also dependent on the phase of ENSO, but the influence is generally not as strong. Kaser et al. (2003) found a significant positive relationship between the Southern Oscillation Index (SOI) and reconstructed glacier mass balance using a 41 year long mass balance time series, but they also pointed out that this relationship does not hold in all years. Vuille et al. (2007) were able to show that this relationship is relatively weak because the Cordillera Blanca is located in an intermediate zone, where the ENSO influence on temperature and precipitation is not as strong as to the north on Antizana (temperature) or to the south on Chacaltaya (precipitation).

4) GLACIER SURFACE ENERGY BALANCE The previous chapter discussed how climate affects the mass balance of tropical glaciers through changes in accumulation and ablation (melting and sublimation). To understand how these changes are transmitted from the atmosphere to the glacier, however, requires detailed knowledge of the surface energy balance (SEB) at the glacieratmosphere intersection. The SEB describes the amount and direction of the various energy fluxes from and to the glacier surface. A detailed SEB can only be established through accurate measurements on the glacier itself, which requires the installation and maintenance of automated weather stations (AWS) for a period long enough to properly understand diurnal and seasonal cycles, as well as interannual variations. Such detailed monitoring programs are relatively new and the results from these programs will be discussed in section 4.2. First we review some previous studies which are based on simpler sensitivity calculations (section 4.1.).

4.1. Sensitivity studies The first energy balance studies focused on quantifying the amount of energy needed to produce an observed change in glacier mass balance, or the changes required to stabilize a retreating glacier. For example Hastenrath and Ames (1995b) estimated the average surface lowering on Yanamarey glacier in the Cordillera Blanca between 1977 and 1988 to be 1.5 m yr-1. Given the latent heat of melting, Lm = 33×104 J kg-1, the energy required to produce such a change corresponds to a positive energy balance (directed toward the glacier) of 16 W m-2. Sublimation, which is much less energy-efficient (latent heat of sublimation Ls = 284×104 J kg-1) was not considered a factor in this study. In theory, of course, the easiest way to stabilize the glacier would be to increase precipitation by the amount of imbalance, in this case, 1.5 m. This, however, is a completely unrealistic change. Alternatively Hastenrath and Ames (1995b) calculated that a cloudiness increase of 10%, a temperature reduction of 2˚C, a specific humidity decrease of less than 1 g kg-1 or any combination of the above, could lead to a stabilization of the glacier. On Urushraju glacier (also in the Cordillera Blanca) Ames and Hastenrath (1996) used the same sensitivity study approach to diagnose the energy imbalance of the glacier between 1977 and 1987. The results were very similar to what they previously found on Yanamarey, with an average surface lowering of 1.04 m yr-1, equivalent to an energy surplus of 12 W m-2. Their calculations suggested that the energy decrease required to stabilize Urushraju could be produced by a cloudiness increase of less than 10%, a temperature decrease of 1.5˚C, a decrease in specific humidity of less than 1 g kg-1 or any combination of these (Ames and Hastenrath, 1996). Kaser et al. (1996a) performed a similar sensitivity study in the Cordillera Blanca, calculating the energy surplus responsible for the observed ELA rise of 95 m between 1920 and 1970. In addition they compared their results with actual observed changes in climate to assess what combination of climatic changes was most likely responsible for the ELA shift. According to their results either a temperature increase of 0.51˚C, a precipitation decrease of 1177 mm, a global radiation increase of 1.076 MJ m-2 d-1 or an increase in evaporation of 157 mm yr-1 could have caused the change in ELA position.

Comparing their results with actual observed changes in climate, Kaser et al. (1996a) concluded that a temperature increase could only explain about half of the observed rise in ELA and that precipitation changes were too small to have had a significant impact. The remaining half of the observed ELA change was therefore attributed to a combined effect of the remaining factors. Similarly Kaser and Georges (1997) attributed the observed ELA increase in the Cordillera Blanca between 1930 and 1950 only in part to a temperature increase, but more so to decreased humidity and its indirect effects of reduced precipitation, reduced cloud cover and hence increased incoming radiation. Kaser and Georges (1997) also observed large differences in ELA rise across the Cordillera Blanca from east to west. They argued that this differential ELA rise was inconsistent with a temperature forcing, which should have had a similar impact throughout the region. Vuille and Bradley (2000) and Vuille et al. (2003), however, later showed that temperature trends in the 20 century have indeed been vastly different on the eastern and western Andean slopes (see section 5.1). On Chacaltaya glacier Ramirez et al. (2001) estimated the average melting (assuming no sublimation) since 1983 to be equal to an average increase in heat supply of 10 W m-2. According to their calculations it would take a ~200 m drop in the ELA to make up for this increase in heat supply and to again stabilize the glacier. Such a scenario would bring the ELA down to a level near 5220 m, close to the average ELA in the period 1940-1963, before accelerated glacier recession began (Francou et al., 2005). Mark and Seltzer (2005a) performed a similar sensitivity analysis on Nevado Queshque, also in the Cordillera Blanca, to assess the causes of the observed ice thinning (353 mm yr-1) between 1962 and 1999. Assuming that 20% of the mass loss was due to sublimation and the remaining 80% due to melt, they concluded that 9.3 W m-2 was required to produce the observed mass loss. According to their results this energy surplus was most likely caused by a temperature rise of 1Ë&#x161;C, combined with a specific humidity increase of 0.14 g kg-1, as this scenario was most in line with actual observed changes in climate over this period. They also pointed out that this estimate was inconsistent with the observed average ELA rise of only 72m, which they explained with the fact that the glaciers are not in equilibrium and lag behind the climate forcing. Another, rather simple but effective way to qualitatively attribute observed changes in glacier extent to various climate forcings, is to assess how the glacier geometry changes over time and where glaciers tend to lose the most mass. For example if changes in cloudiness and hence solar radiation were the main cause of changes in glacier extent, one would expect to see a differential thinning in accordance with the radiation geometry of the glacier (e.g. Mark and Seltzer, 2005a, b). While tropical Andean glaciers do indeed reside preferentially in locations that receive less shortwave radiation than non-glacierized areas at similar elevation (Klein et al., 1996), there are no studies indicating that a preferential thinning has taken place on tropical Andean glaciers, which could be attributed to changed radiation receipts. If glaciers react to increased temperatures on the other hand, one would expect to see larger mass loss at lower elevation glaciers, and a mass loss dependence on the glacier hypsometry, that is the vertical distribution of glacier mass and area. Mark and Seltzer (2005a, b) indeed detected such changes on Nevado Queshque in the Cordillera Blanca, which lead them to the conclusion that the observed mass loss was primarily caused by a temperature increase.

4.2. The Surface Energy Balance (SEB) While the sensitivity studies discussed in the previous section are very interesting and insightful in their own right, they fall short of clearly attributing observed changes in glacier mass and extent to one or several factors. There is always a variety of potential combinations which could explain the energy imbalance at the glacier surface. The way to solve this problem is to actually measure all the relevant fluxes at the surface over a long enough period of time. The energy balance of a melting ice surface can be written as: SW↓ (1-α) + LW↓ + LW↑ + H + C + Ls S + Lm M = 0 where SW↓ is the incoming shortwave radiation, α the surface albedo (reflectivity), LW↓ and LW↑ the long wave radiation (emission) toward and from the glacier surface, H the sensible heat flux, C the subsurface conductive heat flux and Ls S and Lm M the mass consuming terms, with Ls and Lm the latent heat of sublimation and melt respectively and S and M the rate of sublimation and melt. All energy fluxes are considered positive if directed toward the surface and negative if they are directed away from it. All these terms in turn depend on a number of climatic factors. SW↓ for example depends largely on cloud cover, α is controlled by the amount and timing of snowfall and LW↓ depends on atmospheric humidity and cloud cover. The mass consuming term Ls S is controlled by the vapor pressure gradient between the glacier surface and the air above, which in turn is related to air humidity and wind speed. The terms of the equation also show pronounced seasonality on tropical and especially subtropical glaciers, with very different behavior during dry and wet seasons. SW↓ for example is reduced during the wet season because of increased cloud cover and α is very high due to frequent snowfall. At the same time LW↓ and LW↑ almost cancel each other and sublimation is reduced because of the high humidity and low wind speeds. This effectively means that most of the surplus energy is consumed by the very energy-efficient melting. As a consequence high melt and high accumulation during the wet season lead to a large mass turnover (e.g. Sicart et al., 2003; Kaser et al., 2005). During the dry season on the other hand, the LW balance is negative (away from the surface) as the large vapor pressure gradient leads to enhanced sublimation, which is 8 times more energy-intensive than melting. As a result melting is reduced and mass turnover limited. These peculiarities of the tropical glacier energy balance make tropical glaciers highly sensitive to changes in atmospheric humidity which governs sublimation, precipitation, whose variability, particularly during the rainy season induces a positive feedback on albedo, and cloudiness, which controls the incoming longwave radiation (Francou et al., 2003). Some of the first energy balance studies were performed on Quelccaya Ice cap in the 1970s (Hastenrath, 1978, 1997). They were however very limited in terms of instrumentation and the length of operation. Hastenrath (1978), based on a few days worth of shortwave and longwave radiation measurements during austral winter, concluded that due to the high albedo and constant subfreezing temperatures virtually no energy was available for ablation. The limited sublimation that took place during the day was approximately compensated by nighttime deposition (Hastenrath, 1997). Hence it

was concluded that net ablation only took place at lower elevations of the ice cap (Hastenrath, 1978, 1997). More recent detailed energy balance measurements with an AWS installed permanently on the summit, however, paint a very different picture, with temperature above freezing and melting taking place throughout much of the year (D. Hardy, 2006, pers. comm.), However, it is difficult to assess exactly, how much of the difference between the early studies and todayâ&#x20AC;&#x2122;s measurements is due to a much more sophisticated, year-round monitoring at the summit, and how much is due to an actual change in climate between the mid 1970s and today. Wagnon et al. (1999a, b) provided the first in-depth SEB study from a tropical glacier based on measurements from an AWS installed on Zongo glacier. They were able to show that mass loss and runoff show a marked seasonality, which can not be explained by the sensible heat transfer, which shows little seasonality and is generally small. Instead net radiation and the turbulent fluxes dominate the SEB (Figure 25). They also demonstrated that the incoming energy is quite constant throughout the year and that instead it is the partitioning of this energy into melt and sublimation, controlled by humidity, which causes the much higher mass loss during the wet season. When humidity is high, the available radiative energy is directly consumed by melting (wet season), while the enhanced vapor pressure gradient during the dry season favors sublimation, which is energetically inefficient and therefore leads to reduced discharge. The calculations based on the measurements of the various terms were verified by comparison with mass balance measurements from a stake network, sublimation measurements using lysimeters and proglacial stream discharge measurements. As shown by Wagnon et al. (1999a), the SEB based calculations compared favorably with the ablation, sublimation and runoff measurements, increasing the confidence in the results. In summary, the initial studies on Zongo showed how important the humidity is on these subtropical glaciers, as it determines the partitioning of the available energy into melt and sublimation. As shown by Wagnon et al. (1999a), sublimation consumed 63% of the total available energy during the hydrologic year 1996-97 to produce only 17% of the mass loss. This shows how sensitive tropical glaciers are toward changes in specific humidity and that increasing humidity is the most effective climate change scenario to enhance glacier retreat, at least in the subtropics. In addition, as emphasized by Sicart et al. (2005), long-wave radiation, directly linked to humidity but also to cloud cover, is another key factor of the energy balance, as it equally drives the seasonal changes of energy available for melt. On interannual time scales the glacier mass balance is strongly influenced by ENSO (see section 3.3). Using the same AWS measurements from Zongo, Wagnon et al. (2001) demonstrated that the much larger mass loss during the El NiĂąo 1997/98 as compared to previous year was not due to the admittedly higher temperatures (and hence higher sensible heat flux), but due to a significant precipitation deficit. Below average precipitation, very common during El NiĂąo in the Bolivian Andes (e.g. Vuille, 1999; Vuille et al., 2000a; Garreaud et al., 2003) leads to low-albedo bare ice exposed for a much longer time of the year and over a much larger glacier surface (the ELA was 450 m higher in 1997/98 than during the previous year). Absorption of shortwave radiation was therefore greatly enhanced, translating into a dramatic increase in net all-wave radiation. As a consequence glacier melting was significantly enhanced and measured runoff at the glacier snout was two thirds higher than normal (Wagnon et al., 2001). These results

were later applied to the much smaller Chacaltaya glacier and the results there confirmed that cloudiness, which controls the incoming long-wave radiation, precipitation, which has a strong feedback on albedo, and humidity, which is responsible for sublimation, are the key variables explaining variations of the glacier mass balance (Francou et al., 2003).

Figure 25: Daily cycle of the various energy balance terms on Zongo glacier during the dry (above) and wet seasons (below). Net radiation (green is the main term of the energy balance in both seasons, but during the dry period it is almost entirely consumed by sublimation (red). During the wet season this term is much reduced, leaving a lot of energy available for melting (from Wagnon et al., 1999b).

In the inner tropics near the equator the various terms of the energy balance vary somewhat differently. Measurements of the SEB on Antizana glacier in Ecuador reveal that it is equally governed by net radiation and hence albedo (Favier et al., 2004b). However, because of the absence of thermal seasonality, the ablation zone is exposed to oscillations of the 0Ë&#x161;C isotherm throughout the year. These small fluctuations in temperature determine the rain-snow line on the ablation zone and hence have a major

impact on the albedo. Consequently air temperature significantly influences the energy balance in the inner tropics, albeit not through the sensible heat term of the SEB equation, as commonly thought, but indirectly through changes in albedo and net radiation receipts (Favier et al., 2004a, b; Francou et al., 2004). As already outlined in section 3.3., interannual variations of the SEB on Antizana are strongly affected by the ENSO cycle. During El NiĂąo, increased air temperature, which favors rain over snowfall, weak and sporadic snowfall, insufficient to maintain a high glacier albedo, low wind speeds, which limit the transfer of energy from melting to sublimation, and reduced cloud cover, which increases the incoming short-wave radiation, are the main factor affecting the SEB and causing high melt rates. La NiĂąa events on the other hand are characterized by colder temperatures, higher snowfall amounts, and to a lesser degree, more constant winds, factors which increase albedo and sublimation and therefore preclude melting at the glacier surface (Francou et al., 2004). For reasons discussed in section 3, most SEB studies stem from melting surfaces in the glacier ablation zone. To date there is only limited information available on the energy balance from the accumulation zone at high-altitude, where no melting takes place. Wagnon et al. (2003) performed some experiments and measured the SEB for a short time in austral winter near the summit of Illimani, at 6340 m in the Cordillera Real. Their results show that here the net radiation balance is mostly negative due to a constantly high albedo and the reduced LWâ&#x2020;&#x201C;, resulting from generally clear skies. The latent heat flux was always negative, which indicates significant sublimation. Wagnon et al. (2003) estimated the mass loss due to sublimation at ~ 1 mm w.e. d-1 at this high elevation site.

5) OBSERVED AND PROJECTED CLIMATE CHANGE The glacier sensitivity and SEB studies from the previous section have outlined how climate change can affect glaciers in the tropical Andes by altering the various terms of the energy budget. To accurately attribute glacier retreat to a particular climate forcing, however, requires detailed knowledge and understanding of the climatic changes that have actually taken place in the 20th century. In this section we will first review observed 20th century climate changes (section 5.1.) and then discuss what the most likely scenarios of future climate change in the 21st century look like (section 5.2.).

5.1. Observed 20th century climate change 5.1.1. Temperature Probably the most detailed analysis of near-surface temperature trends was presented by Vuille and Bradley (2000), based on a compilation of 268 station records between the 1°N and 23°S. Their results showed that near-surface air temperature has significantly increased over the last 60 years. Their analysis documented a temperature rise of 0.10 - 0.11°C/decade, since 1939, with the bulk of the warming occurring over the last 2 decades. Indeed the rate of warming almost tripled since the mid 1970s’ (0.32 – 0.34°C/decade (Figure 26). Vuille et al. (2003a), based on an updated data set, later confirmed these results and showed that other data sets (CRU05, and ECHAM4 model data) show similar warming trends in the Andes.

Figure 26: Annual temperature deviation from 1961-90 average in the tropical Andes (1°N-23°S) between 1939 and 1998. Black line indicates 11-yr running mean (modified from Vuille and Bradley, 2000).

One noteworthy result of the Vuille and Bradley (2000) and Vuille et al. (2003a) studies was the astonishing dependency of the temperature trend on elevation, and even more so on the location of the stations considered. Stations located on the eastern slope showed a much subdued warming, with trends that were close to zero and insignificant at the lowest elevations, while the warming on the Pacific side was strongest and most significant. Higher elevations showed an intermediate, albeit still significant warming (Figure 27).

Figure 27: Temperature trend as a function of elevation and slope. Vertical bars indicate elevation zone (1000 m) for which trend is valid. Horizontal bars indicate 95%-confidence limits, i.e. all trends whose horizontal bars do not intersect with the 0˚-line are significant at 95% (modified from Vuille and Bradley, 2000).

Vuille et al. (2003a) were able to reproduce the observed spatial differences, with large warming to the west and little to no warming to the east of the Andes with the ECHAM4 GCM (Figure 28, left). By regressing the temperature anomalies onto the Niño-4 index and then multiplying the regression coefficients by the trend in the Niño-4 index, they obtained the warming trend which was linearly congruent with a contemporaneous change in the tropical Pacific. This procedure yields an estimate of the fraction of the trend which can be attributed solely to the warming in the Niño-4 region. The signature of the Niño-4 index (Figure 28, right) is clearly reflected in both the spatial pattern and the sign of the simulated trend (Figure 28, left). Hence the east-west difference in the Andean temperature trends can in part be traced back to the warming in the central equatorial Pacific (Vuille et al., 2003a). On average about 50-70% of the observed temperature change in the Andes, can be attributed to a temperature increase in the tropical Pacific (Vuille et al., 2003a).

Figure 28: Annual mean temperature trend (˚C/decade, left) and the contribution of the Niño-4 index to the trend (˚C/decade, right) in the ECHAM-4 model between 1979 and 1998 (modified from Vuille et al., 2003a).

On a more regional scale a number of additional studies provide evidence for significant warming over the last decades of the 20th century. In Colombia, according to Ceballos et al. (2006), temperature at a high elevation station (4150 m) increased by roughly 1˚C during the 25 years prior to 2000. In central Peru (9° - 11°S), Mark (2002) and Mark and Seltzer (2005a), based on 29 low and high-elevation stations, found a temperature increase of 0.35°-0.39°C/decade between 1951 and 1999. Vuille et al. (2000) based on a principal component analysis (PCA) of station data found a significant warming trend since the mid 1970’s in southern Bolivia and northernmost Chile. Toumi et al. (1999) reported significant warming (0.20°C/decade between 1954 and 1987) at La Quiaca, a high elevation (3462 m) station at the border between Bolivia and Argentina by utilizing station pressure change as an indicator of warming. In addition to the analysis of mean temperature change, a number of studies focused on changes in temperature extremes and minimum and maximum temperatures. Quintana-Gomez (1997), based on daily temperature records from seven stations in the central Andes of Bolivia, showed that both minimum and maximum temperatures have increased between 1918 and 1990, but the trend for the minimum temperatures was much larger, thereby effectively reducing the daily temperature range (DTR). Similar results were found in Colombia and Venezuela, where the DTR was equally reduced over the last 25 years because of a sustained rise in minimum temperature (Quintana-Gomez, 1999). In general both the minimum temperature increase and the DTR decrease were more pronounced in Colombia than in Venezuela (Quintana-Gomez, 1999). However, only one of the 14 stations used in this analysis was located above 2500 m. These results were later confirmed by Aguilar et al. (2005), although they based their results for Venezuela and Colombia on only 6 and 2 stations respectively. Finally a repeat of this study in Ecuador, based on 15 stations between 1961 and 1990, revealed a similar pattern of increasing minimum and maximum temperatures, with a decrease of the DTR 39

(Quintana-Gomez, 2000). Vincent et al. (2005) later confirmed these results with a limited data set from South America, showing that the DTR decreased from Ecuador to Chile. They also reported that the coldest nights are getting warmer and that the percentage of extremely cold (warm) nights is decreasing (increasing), thereby confirming the strong nighttime warming. A seasonal analysis of their data indicated that these changes were more pronounced between December and May than during the rest of the year (Vincent et al., 2005). Further evidence for recent warming comes from the Quelccaya ice cap, where shallow ice cores reveal that the seasonal δ18O signal is no longer being preserved. This indicates that melting is taking place on the summit and that meltwater percolates through the snowpack, thereby destroying the seasonal δ18O signal (Thompson et al., 1993; Thompson, 2000). This hypothesis is consistent with recent temperature recordings by an automated weather station at the summit, indicating that temperatures reach above freezing for several months every year (Hardy, pers. comm., 2006). Thompson et al. (2003, 2006) have also reported a significant enrichment of the δ18O in their cores during the 20th century, which they have interpreted as a sign of rising temperatures. The correct interpretation of this signal, however, is controversial and most evidence suggests that it is in fact not related to a temperature increase (Bradley et al., 2003; Hardy et al., 2003; Vuille et al., 2003b,c; Hoffmann et al., 2003; Vuille and Werner, 2005; Vimeux et al., 2005). Diaz et al. (2003) analyzed changes in freezing level height (FLH) over the American Cordillera and the Andes, based on NCEP-NCAR reanalysis data. They found an increase in FLH of 73 m between 1948 and 2000 and of 53 m between 1958 and 2000, a period for which the data is considered more reliable. A simple regression analysis with the Niño-3 index shows that a 1°C warming of tropical Pacific SST equals about 76 m rise of the FLH. Diaz et al. (2003) concluded that the increase in tropical Pacific SST accounts for about half of the observed rise in FLH, consistent with the results discussed above by Vuille et al. (2003a). All these results provide compelling evidence that significant warming took place in the tropical Andes in the 20th century and that the warming has accelerated over the past few decades. These studies are therefore in stark contrast to reports from radiosonde measurements, indicating no warming but a slight cooling of the lower tropical troposphere since 1979 (Gaffen et al., 2000), after an initial warming occurred between 1970 and 1986 (Diaz and Graham, 1996). Vuille et al. (2000) suggested that free tropospheric and near-surface temperatures are decoupled from one another and that free tropospheric trends and data should not be used to assess climate change in tropical mountain regions. Seidel and Free (2003) later investigated this claim be analyzing pairs of nearby surface and tropospheric data. Indeed they found that temperature variations at mountain locations differ significantly from those at nearby lowland stations throughout the lower troposphere. In addition there are large structural uncertainties with radiosonde estimates and residual errors appear to affect many tropospheric data sets (radiosonde and satellite data, Santer et al., 2005). More recent evidence indeed suggests a warming of the free tropical troposphere, even relative to the surface (Fu and Johanson, 2005), consistent with an increase in tropopause height (Santer et al., 2003) and tropospheric water vapor content in the tropics (Wentz and Schabel, 2000).

5.1.2. Precipitation Changes in precipitation are much less notable than the changes in temperature described in the previous chapter. There is however, also a noticeable lack of long and high-quality precipitation records, which would allow for a detailed assessment of longterm trends. Vuille et al. (2003a) used 42 station records between 1950 and 1994 to analyze trends in precipitation in the Andes of Ecuador, Peru, Bolivia, and northernmost Chile and Argentina. Besides determining the statistical significance, Vuille et al. (2003a) also analyzed the trends for spatial coherence and elevation dependence. There is a tendency for increased precipitation north of ~11Ë&#x161;S, in Ecuador, northern and central Peru both on annual time scales (Figure 29 left) and at the height of the austral summer precipitation season (DJF, Figure 29 right). At most locations (except for the Cordillera Blanca region), however, the trends are not statistically significant. In southern Peru and along the Peru/Bolivia border on the other hand, most stations indicate a precipitation decrease. Even though this appears to be a quite coherent regional signal, individual station trends are mostly insignificant. Of the 42 stations analyzed only 5 (2) show a significant increase (decrease) in the annual precipitation amount. Furthermore one needs to take into account that several stations are located very close to one another and are thus not truly independent records, because they capture the same local climatic signal. The results are consistent with a more recent evaluation by Haylock et al. (2006), who also found a change toward wetter conditions in Ecuador and northern Peru, and a decrease in southern Peru, albeit based on much fewer stations. precipitation trend (mm/yr)

precipitation trend (mm/yr)

2.5 to 7.5

0.5 to 2.5

0 to 0.5

-0.5 to 0

-2.5 to -0.5

-7.5 to -2.5

Figure 29: Trends in station precipitation (mm yrâ&#x2C6;&#x2019;1) between 1950 and 1994 for (left) annual sum and (right) DJF (modified from Vuille et al., 2003a). Upward (downward) pointing triangles indicate an increase (decrease) in precipitation. Filled (open) triangles indicate that the trend is (not) significant at the 95%-confidence level.

Aguilar et al. (2005), looking at precipitation trends in Colombia and Venezuela between 1961 and 2003 found no coherent pattern of significant changes. Vuille et al. (2003a) also investigated whether the precipitation trends showed any signs of elevation dependence. They found that positive as well as negative trends occurred at all elevations between 0 and 5000 m, although unfortunately no station data was available in the 1000 – 2000 m elevation range. The JJA dry season was the only time of year where a slight bias toward wetter conditions was evident, with 27 (13) stations indicating a positive (negative trend), but even then no dependence on elevation was apparent (Vuille et al., 2003a). A few studies have looked at precipitation trends on a regional scale and found notable precipitation increases over time. It is interesting that all these reports come from the eastern slopes of the Andes, or even the lowlands to the east, such as the reported increase along the eastern slopes of the Andes in Ecuador during the MAM rainy season (Vuille et al., 2000b), along the eastern slopes in NW-Argentina (Villalba et al., 1998) or the lowlands of Bolivia (Ronchail, 1995).

5.1.3. Humidity Changes in humidity are very relevant in the context of glacier variations because of the significant impact humidity has on the partitioning of the available energy into melt and sublimation (see section 4.2.). A significant increase in near-surface and tropospheric humidity over the last decades has been reported from both the eastern and western tropical Pacific (Gutzler, 1992), tropical South America (Curtis and Hastenrath, 1999a, b) and the tropical troposphere in general (Wentz and Schabel, 2000). Nonetheless it is still difficult to accurately assess such long-term trends. In the Andes no long and continuous in-situ records exist, to document such changes. New data sets of tropospheric water vapor have become available thanks to improved remote sensing techniques, but they have their own calibration issues and the records in general are still too short to meaningfully distinguish between low-frequency variability and actual trends. Reanalysis data, such as NCEP-NCAR or ERA-40, equally have large biases, in particular in their moisture fields and over the tropics, which makes them ill suited for any kind of trend analysis (Trenberth and Guillemot, 1998; Trenberth et al., 2001). Vuille et al. (2003a) in their assessment of near-surface humidity changes in the Andes therefore relied on CRU05 data, which is based on station observations, spatially interpolated on a regular 0.5˚ × 0.5˚ grid (New et al., 2000). They found a significant increase in relative humidity between 1950 and 1995 of up to 2.5%/decade, with the most prominent positive trend in northern Ecuador and southern Colombia, while in southern Peru, western Bolivia and northernmost Chile the increase was more moderate (0.5–1.0%/decade). Overall Vuille et al. (2003a) found that the trends are very similar in all seasons with only slight modulations during winter and summer. Given the significant increase in temperature and the rising relative humidity levels, it follows that vapor pressure (or specific humidity) has increased significantly throughout the Andes as well. To the east of the Andes the trends were much lower or even negative. However, since the CRU05 data is partially interpolated from synthetic data (New et al., 2000) these trends should be interpreted with caution.

5.1.4. Cloud cover The most comprehensive data set available to assess changes in cloud cover over the past decades is the International Satellite Cloud Climatology Project (ISCCP), which has been used in the past to study cloud cover variations in the central Andes (Vuille and Keimig, 2004). Unfortunately it is not very well suited for linear trend analysis because of its short duration (start in July 1983) and the lack of an independent confirmation of the long-term calibration (e.g., Rossow and Schiffer, 1999). Another commonly used data set is based on measurements of the outgoing longwave radiation (OLR) emitted by the earth’s surface and the overlying atmosphere and constantly monitored by a number of polar orbiting satellites since 1974. OLR is sensitive to the amount and height of clouds over a given region and time and has been applied in a number of studies to investigate tropical convection and convective cloud cover over tropical South America (e.g., Chu et al., 1994; Kousky and Kayano, 1994; Aceituno and Montecinos, 1997; Liebmann et al., 1998; Chen et al., 2001; Vuille et al., 2003a). In the presence of deep convective clouds, the satellite sensor measures radiation emitted from the top of the clouds, which are high in the atmosphere and thus cold, leading to low OLR values. In the case of clear sky conditions on the other hand, high OLR values reflect radiation emitted from the earth’s surface and the lower atmosphere. In the absence of convective clouds, OLR is thus strongly influenced by other processes, such as changes in surface temperature, low-level cloud cover or water vapor content. According to Vuille et al. (2003a) OLR has decreased in the inner tropics (north of ~ 10˚S) since 1974, which indicates an increase in convective cloud cover over that region. The trends shown in Figure 30 are an update from the study by Vuille et al. (2003a) through 2005. The largest changes have taken place during austral summer, DJF, when OLR has significantly decreased over the tropical Andes and to the east over the Amazon basin (Figure 30, right side). This observed increase in convective activity and cloud cover in the inner tropics is consistent with earlier studies over the Amazon basin reaching similar conclusions (e.g., Chu et al., 1994; Chen et al., 2001). In the outer tropics (south of ~10˚S) the trend is reversed, featuring an increase in OLR. This pattern however is more difficult to interpret because OLR in the outer tropics is only a good proxy for convective activity and cloud cover during the rainy season (DJF). However, OLR has indeed significantly increased in the outer tropics during DJF as well, suggesting that cloud cover is indeed trending downward. The overall pattern is consistent with changes in precipitation reported in the previous section, with changes in vertical motion associated with the Hadley circulation (see next section) and with results by Wielicki et al. (2002) and Chen et al. (2002), who reported intensified upward motion and cloudiness in equatorial-convective regions and drier and less cloudy conditions in subtropical subsidence regions in the 1990s. However, questions have been raised about the reality of the variations in these latter two reports (Trenberth, 2002). The marked increase in tropical convective activity over and to the east of the Andes may be partially responsible for the observed differences in temperature trends to the east and the west of the Andes. It is reasonable to assume that increased convective cloud cover to the east of the Andes may have dampened the near-surface warming through a reduction of incoming shortwave radiation, while this effect is of minor

importance over the arid western slopes of the Andes. Indeed Vuille et al. (2003) were able to show that this mechanism was responsible for the simulated east-west difference in temperature in the ECHAM-4 model.

Figure 30: Trends in OLR (in W m-2 yr-1; 1974-2005) for annual mean (left) and DJF (right). Small insets in upper right show regions where increase (decrease) in OLR, shaded in light (dark) gray, is significant at the 95% level (modified and updated from Vuille et al., 2003).

5.1.5. Atmospheric circulation Although the significance of the observed trends varies widely, the previous sections on precipitation and cloud cover seem to suggest that the inner tropics are getting wetter and cloudier while the outer tropics are becoming drier and less cloudy. This behavior could be explained through an intensification of the meridionally overturning tropical circulation (the regional Hadley circulation), with more vigorous vertical ascent, favorable for convective activity, in the tropics, balance by enhanced subsidence and accordingly clear skies in the subtropics. A trend analysis of the vertical motion (Ω) and the meridional wind field (v) along a north-south transect at 65˚W in South America does indeed support this notion (Figure 31). The result shows that over the period analyzed (1950 – 1998), the regional Hadley circulation has indeed intensified, with more vigorous ascent in the tropics between ~10˚S and 10˚N and enhanced descending motion in the subtropics, in particular in upper- and mid-tropospheric levels between 10˚S and 30˚S. This pattern is robust both in the annual mean as well as in the seasonal analysis, although the trends are larger and more significant in the northern hemisphere.

Figure 31: Trend of vertical motion (Ω, in hPa s-1 yr-1) and the meridional wind field (v, in m s-1 yr-1) along a transect from 40˚N to 40˚S at 65˚W between 1950 and 1998. Trends of enhanced ascending (descending) motion, significant at the 95% significance level, are shaded in blue (red). Left Figure shows annual mean circulation trend, right is for DJF (M. Vuille, unpublished).

The upper and lower tropospheric divergent wind field has been subjected to similar analyses and the trends are dynamically consistent with the above results. As shown in Figure 31, the upper tropospheric circulation shows a clear trend towards enhanced divergence over tropical South America, while the lower level (850 hPa) features increased convergence. Both of these fields act to reinforce and sustain enhanced upward motion and convective activity in the tropics, consistent with the notion that there has indeed been a trend toward a strengthening of the tropical circulation.

Figure 31: Trend of velocity potential (χ, in 105 m2 s-1 yr-1) and divergent wind field (udiv, vdiv, in m s-1 yr-1) between 1950 and 1998. Divergent wind vectors are only shown where the trend is significant at the 95% significance level. Left Figure shows 250 hPa level, right Figure shows 850 hPa (M. Vuille, unpublished).

Chen et al. (2001) come to very similar conclusions, showing that the regional hydrological cycle over the Amazon basin has strengthened, and vapor flux convergence over the basin has increased. Chen et al. (2002) also report a strengthening of the Hadley circulation, based on satellite observations, which show that equatorial convective regions have intensified in upward motion and moistened, while both equatorial and subtropical subsidence regions have become drier and less cloudy. In summary, the analysis of changes in the large-scale circulation shows trends which are dynamically consistent with observational data and helps to explain why the Andes in the inner tropics appear to have become wetter and the subtropical Andes are becoming drier. In-situ observational data, satellite information and reanalysis data all seem to indicate a strengthening of the tropical atmospheric circulation. Changes in the inner tropics are characterized by enhanced low-level convergence and upper-level divergence, resulting in enhanced upward motion, increased convective activity and more humid conditions. In the subtropics the opposite trends prevail, with increased subsidence and reduced convective activity leading to potentially less humid conditions. So while individual observed trends at the surface (e.g. precipitation) are weak and often insignificant, the observed spatial pattern of change is dynamically consistent with changes in the large-scale atmospheric circulation. Finally it is noteworthy that there are also indications of significant changes in the zonally overturning (Walker) circulation over the tropical Pacific domain. Vecchi et al. (2006) have shown that the Walker Circulation in the tropical Pacific has significantly weakened throughout the 20th century, with sea level pressure (SLP) increasing in the western and decreasing in the eastern tropical Pacific. They attributed this change to increased anthropogenic greenhouse gas forcing. The observed SLP changes in the Pacific lead to a weakened zonal SLP gradient and hence to a shift of the mean conditions toward a more El Niño-like state. This is relevant in the context of our discussion of Andean glaciers because of the strong dependence of the glacier mass balance on tropical Pacific SSTs and ENSO in particular. Of course a change in the mean state is different from a change in El Niño intensity or frequency, and it is not clear what the exact ramifications of these studies for Andean glaciers are. Nonetheless future changes in tropical Pacific climate need to be carefully observed, given the tight coupling between tropical Pacific SST and tropical Andean glacier mass balance (Francou et al., 2003, 2004).

5.2. Projected future climate change Studies on future climate change in the tropical Andes and, more generally, South America, have focused on the two parameters temperature and precipitation. Bradley et al. (2004) analyzed free tropospheric temperature changes along the American Cordillera based on simulations from 7 different General Circulation Models (GCMs), forced with 2 × CO2 concentrations. Compared with control runs the temperature changes were large and increased with elevation. In the tropical Andes the projected temperature changes were on the order of 2.5-3˚C year-round. Bradley et al. (2006) in a similar analysis investigated how annual free-air temperatures changed in 8 different GCMs between 1990-1999 and 2090-2099 along that same transect from Alaska to Patagonia. All eight

GCMs included in their assessment were using CO2 levels from IPCC-scenario SRES A2 (Nakicenovic and Swart, 2000). The results indicate a continued warming of the tropical troposphere throughout the 21st century, with a temperature increase that is again enhanced at higher elevations (Figure 32). By the end of the 21st century, following this SRES A2 emission scenario, the tropical Andes experience a massive warming on the order of 4.5˚-5˚C. 9000

9000

8000

7000

6000

7 6.5 6 5.5

5000

4.5 4

5000

3.5 4000

4000

3000

2000

1000

°C

3 2.5 2 1.5 1 0.5

-50

-40

-30

-20

-10

-50

0 -40

-30

-20

-10

Latitude

Latitude 9000

9000

7 8000

8000

6.5 6

7000

5.5 5

6000

5000

4.5 4

5000

3.5

4000

3000

2000

1000

°C

3 2.5 2 1.5 1 0.5

-50

-40

-30

-20

-10

0 -40

-20

Latitude

Figure 32: Projected changes in mean annual free air temperatures for 2026-2035 (upper left), b) 20462055 (upper right), c) 2066-2075 (lower left) and d) 2090-2099 (lower right). All Panels show departure from 1990-1999 average along a transect from Alaska (68°N) to Patagonia (50°S), following the axis of the American Cordillera mountain chain. Results are the mean of eight different general circulation models used in the 4th assessment report (AR4) of the Intergovernmental Panel on Climate Change (IPCC) using CO2 levels from scenario SRES A2. Black line denotes mean elevation along transect; white areas have no data (surface or below in the models). From Vuille and Bradley (2007).

Boulanger et al. (2006) performed a similar analysis but focused on changes in surface rather than free tropospheric temperature. Their results suggest that tropical South America will warm more (by 3-4˚C in the SRES A2 scenario) than the southern part of the continent. By using different scenarios (SRES A1B, A2 and B1), they were further able to show that the choice of the emission path significantly affects the outcome of the results. While SRES A1B reaches about 80-90% of the warming displayed in the SRES A2 scenario at the end of the century, the more moderate (and optimistic) SRES B1 path displays only about half of the warming of SRES A2. Hence the scenario chosen by Bradley et al. (2006) is admittedly a worst-case scenario. Nonetheless, all emission paths

tend to show the same pattern of warming, but they differ in amplitude (Boulanger et al., 2006). The only study of precipitation changes in South America at the end of the 21st century relied on the same models from the IPCC-AR4 (Vera et al., 2006). In their study they analyzed the change in precipitation in the SRES A1B scenario during the period 2070-2099, compared with the control period 1970-1999. In the tropical Andes most models predict an increase in precipitation during the wet season and a decrease during the dry season, which would effectively enhance the seasonal hydrological cycle. However, intra-model differences, unlike the results for temperature, are large for precipitation (Vera et al., 2006). In summary, it is clear that the climatic changes under way and described in the previous section will have (and already have had) a profound impact on Andean glaciers. The temperature changes discussed in section 5.1.1., probably affected the inner tropical glaciers more than glaciers in the outer tropics. As described in the previous sections, glaciers in Ecuador for example are very sensitive to changes in temperature, as the rise of the 0˚C-isotherm has an immediate impact on the rain snow line and thus on albedo and the net radiation balance (Favier et al., 2004b). This impact is seen interannually through the immediate and severe impact of El Niño events on the net mass balance (Francou et al., 2004). In the outer tropics, changes in temperature are generally less relevant, given that glaciers are mostly located above the freezing line. In fact the difference between the ELA and the 0˚-isotherm, which is a good indicator of the sensitivity of tropical glaciers to global warming (Favier et al., 2004a), rises significantly from below 0 in the inner tropics to several hundred meters in the outer tropics of Bolivia or northernmost Chile. Glaciers in the outer tropics on the other hand are more sensitive to changes in precipitation, which governs the albedo and thus the radiation balance (Francou et al., 2003). The presented changes in precipitation and cloud cover (sections 5.1.2. and 5.1.4.), albeit still rather weak and mostly statistically insignificant, are consistent with glacier retreat in the outer tropics, while in the inner tropics they should have worked against the negative impact of higher temperature. Similarly the increase in humidity would reduce the vapor pressure gradient between snow and air and thereby enhance overall ablation rates in the subtropics through changes in the melt/sublimation ratio. Of course if the projected warming of several ˚C (see section 5.2) becomes reality, as projected in Bradley et al. (2006), the impact will certainly be felt throughout the tropical Andes. Nonetheless it is important to keep in mind that glaciers may respond differently (both in sign and magnitude) to changes in temperature, precipitation, humidity and cloud cover, depending on their location and their sensitivity to certain climate parameters.

6) IMPLICATIONS FOR WATER RESOURCES The previous chapters have shown that dramatic changes are taking place in the tropical Andean cryosphere (section 2). The high sensitivity of tropical glacier mass and energy balance (sections 3 and 4) to climate change, which is under way and well documented (section 5), leaves little room for doubt that the observed glacier retreat is occurring in response to a changing climate. A future increase in greenhouse gas concentrations is inevitable, no matter which emission path we decide to follow (Wigley, 2005). Hence future warming in the tropical Andes, whether of the magnitude shown in Figure 32 or less, will occur and will have a negative impact on glacier mass balance throughout the Cordillera. Glaciers will retreat and many may completely disappear, with significant consequences for local populations. People living in the region have long made use of and depended on glacier resources. Near Cusco in Peru indigenous groups have traditionally been cutting ice from glaciers and hauled them back to their villages, as part of a traditional religious ceremony. As reported by Regalado (2005) this tradition has stopped because of the groupsâ&#x20AC;&#x2122; worry about the rapid glacier retreat. Cutting ice is now taboo. In other parts, such as in and around La Paz, ice used to be cut and brought back to the city for cooling purposes (Jordan, 1983, and Figure 33)

Figure 33: Ice from Glacier Chacaltaya is transported to La Paz for cooling purposes in this picture from August 1979 (from Jordan, 1983).

More important than the lack of ice for religious ceremonies or cooling purposes, however, is the worry that the change in streamflow, due to the lack of a glacial buffer during the dry season, will significantly affect the availability of drinking water, water for hydropower production, mining and irrigation (e.g. Barry and Seimon, 2000; Barnett et al., 2005; Coudrain et al., 2005; Francou and Coudrain, 2005; Bradley et al., 2006; Diaz et al., 2006; Vuille, 2006; Vuille and Bradley, 2007). In the tropical Andes the problem is exacerbated when compared with mid-latitude mountain ranges, such as the Rockies or the Alps, because ablation and accumulation seasons coincide, which precludes the

development of a long-lasting seasonal snow cover outside the glaciated areas. Glaciers are therefore the only major seasonally changing water reservoir in the tropical Andes (e.g. Kaser et al., 2005). As shown by Kaser et al. (2003) tropical Andean catchments show a high correlation between their capacity to store precipitation and their percentage of glaciated area (Figure 34).

Figure 34: Seasonal cycle of precipitation (dashed line) and runoff (solid line) coefficients in catchments with different glacier coverage in the Cordillera Blanca, Peru. Coefficients indicate mean monthly departure from long-term mean. The shaded area represents the reservoir capacity of the respective catchments. The Figure illustrates that catchments with higher glaciation are able to retain precipitation and smooth the seasonal cycle of runoff, while runoff in catchments with low glaciation closely tracks the seasonal cycle of precipitation (from Kaser et al., 2005).

The potential impact of glacier retreat on water supply for human consumption, agriculture and ecosystem integrity is thus of grave concern (Buytaert et al., 2006; Young and Lipton, 2006). On the Pacific side of Peru 80% of the water resources originate from snow and ice in the Andes (Coudrain et al., 2005). Many large cities in the Andes are located above 2,500 m and thus depend almost entirely on high altitude water stocks to complement rainfall during the dry season. As these populations grow and water demands increase the problem is further aggravated (Vรถrรถsmarty et al. 2000). In addition, hydropower constitutes the major source of energy for electricity generation in most Andean countries. As these resources are impacted by reductions in seasonal runoff, these nations may have to shift to other (fossil fuel) energy sources, resulting in large capital outlays, higher operational and maintenance costs and carbonization of their power mix (Vergara, 2005). To further complicate the matter, as glaciers retreat and lose mass in many places in the tropical Andes, they add to a temporary increase in runoff (Coudrain et al., 2005). This increase however, will not last very long, as the frozen water stored in glaciers becomes less and less. Yet downstream user will quickly adapt to enhanced water availability, which raises sustainability concerns as runoff will become less and more seasonal as glacier melt progresses (e.g. Mark, 2007). There are signs emerging that these are not just future runoff scenarios but real changes which are already taking place and being observed by the local population (Young and Lipton, 2006).

In the following sections we discuss in more detail the current state of knowledge regarding the impact of glacier retreat on meltwater contribution and runoff on Antizana, (section 6.1) the Cordillera Blanca (section 6.2.) and the Cordillera Real (section 6.3.)

6.1. Antizana The glaciers on Antizana, a volcano, located only 40 km east of Quito have been retreating for decades, interrupted by some minor readvances (see section 2.3.) This is of concern, because glacial runoff from Antizana provides a significant part of the capitalâ&#x20AC;&#x2122;s water supply (Francou et al., 2004). Efforts are now under way to assess the feasibility of taping new water reservoirs and to provide for alternative water supplies. Figure 35 shows a map of the Quito-Antizana region outlining potential water supply alternatives and their feasibility.

Figure 35: Map of the Quito water supply system. Quito is shown in the upper right, the volcanoes Antizana and Cotopaxi are located to the southeast and east, respectively (from Vergara, 2006; presentation to the World Bank).

6.2. Cordillera Blanca The Cordillera Blanca and its drainage basin, the Rio Santa valley are probably the best-studied glacio-hydrologic complex in the tropical Andes. As it is the world’s most extensively glacier-covered tropical mountain range (Morales-Arnao, 1998), it comes as no surprise that is glacial runoff is used for multiple purposes. In the intensely cultivated Rio Santa valley itself, water is used for agriculture (Young and Lipton, 2006), but there is also large urban development and infrastructure, which depend on uninterrupted water supply (Figure 36). Four hydroelectric power plants are situated along the river between the Cordillera Blanca and the Pacific coast (Mark, 2007). The river has the second largest discharge volume of all Peruvian rivers draining toward the Pacific coast and maintains the most regular flow with the least month-to month variability (Mark et al., 2005). CUENCA RIO SANTA - PERU

Safuna Aporte de agua 1975

Broggi Pérdida de área desde 1972: 53.1 ha

253 lts/seg

Laguna

Cuenca Rio Quitaracsa

Entre 1972 y 1997 pierde 29x10^6 m3 de agua

Yanamarey Disminucion superficie a 1997 65.4 ha. Y perdida de volumen 63.5x10^6 m3 de agua

Uruasraju Retroseso de -16,5 mm dede 1980 Disminucion de área hasta 1997 de -22.9 m.

Retroceso de Laguna

-20.3 m/año

Cuenca del Llanganuco Area glaciar 1991 33,7 km2 % gla 0,39

Rio Yanayacu Cuenca Yanamarey Area glaciar (1991) 1,6 km2 % gla 0,73

Area glaciar 1991 30 km2 Rio Negro

% gla 0,08

Cuenca Olleros Area glaciar (1970) 28,5 km2

C H Cañon del pato 260 MW

Caraz 25.204 hab.

Yungay 20.149hab.

Carhuaz 12.905 hab.

Installed power: 300 MW

rio ervo ,65 Res go 0 Die San 3 m MM Huaraz

114.242 hab. Consumo agua 160 l/hab/dia

% gla. 0,16 RIO SANTA Recuay Orellos

5.501 hab

3.665 hab.

Water usage Rio Santa: 500,000 m3d Agro Industry Population Mine 91% 2%

Figure 36: Map of the Rio Santa Valley water supply and use (from Vergara, 2006; presentation at Quito conference).

An increase in stream discharge due to glacier thinning was first noted in 1985 on Yanamarey glacier (Ames, 1985). Later studies concluded that about half of the mean annual glacier water discharge of ~80 l s-1 is supplied by non-renewed glacier thinning (Hastenrath and Ames, 1995b). Similar studies on nearby Urushraju glacier confirmed these numbers (Ames and Hastenrath, 1996) and raised concern about the future water supply in the Rio Santa valley (Portocarrero, 1995). Mark and Seltzer (2003), based on water balance computations, estimated that in the two Rio Santa subcatchments Yanamarey and Urushraju 35% of the discharge is from non-renewed glacier melt. Using a different approach, based on oxygen isotope end-member mixing, they estimated a

contribution between 30-45%. When this method was applied to the entire Rio Santa watershed, Mark and Seltzer (2003) found that at least 10%, and potentially as much as 20%, of the annual discharge stems from volume loss of stored glacier ice. To make matters worse, Mark et al. (2005) later showed that during the dry season, when water is the most needed, this number is considerably higher and that an estimated 40% of the discharge of the Rio Santa, comes from melting ice that is not replenished by annual precipitation. In the Yanamarey subcatchment they calculated a 58% contribution from glacier volume loss over the time period 2001-2004, a number which is 23% higher than during 1998-99. The evidence provided By Mark and Seltzer (2003) and Mark et al. (2005) clearly shows that current runoff in the Rio Santa Valley is partially fed by non-renewed glacier storage, as the ice volume continues to shrink. This situation raises a number of serious sustainability concerns, especially in the light of the rapidly increasing population in the valley (Kaser et al., 2003). With continued retreat this glacial buffer will shrink significantly and in some watersheds disappear completely. As a consequence streamflow will become more variable and there will be less dry-season runoff. How and when these changes will take place, depends on which climate change scenario will become reality, and on the site-specific response of glacierized catchments to climate change. There is no uniform broad-brush answer to these questions, but instead they need to be investigated on a case by case basis in individual watersheds. Pouyaud et al. (2005) applied estimates of future temperature change to various glacial catchments in the Cordillera Blanca to assess how glaciation and runoff will change throughout the 21st century and beyond (Figure 37). Their results, however, should be interpreted cautiously as the glacier evolution in their model depends solely on temperature, assuming that sensible heat transfer alone can adequately represent the energy balance of a tropical glacier (see also section 3 and 4).

Figure 37: Simulated change in runoff in 4 glacierized catchments of the Cordillera Blanca (from Pouyaud et al., 2005).

By applying a uniform and admittedly very conservative linear warming rate of only ~1Ë&#x161;C per century, Pouyaud et al. (2005) simulated the runoff coefficient for several catchments in the Cordillera Blanca until the year 2300 (Figure 37). According to their results, runoff will continue to increase for another 25-50 years in these catchments before a decreased eventually sets in. Juen (2006) and Juen et al. (2007) used a more sophisticated approach to assess how the seasonality of runoff in the Cordillera Blanca changes, by using a tropical glacier-hydrology model (ITGG 2.0-R). The model was first calibrated over a ten year period (1965-75) and then validated with measured runoff time series from the Llanganuco catchment. Their results show that the model performs very well on both seasonal and interannual timescales and that it shows no long-term drift (Figure 38).

Figure 38: a) Measured (orange) and simulated (black) monthly runoff in the Llanganuco catchment between 1953 and 1997; b) difference and c) cumulative differencebetween measured and simulated runoff (from Juen et al., 2007).

In a next step Juen et al. (2007) simulated changes in future runoff characteristics by using estimates from four different IPCC scenarios (A1, A2, B1 and B2) taken from the third assessment report (TAR). These simulations are all based on the respective steady-state of the glacier, i.e. the glacier area is gradually reduced until, under the new climate background conditions, the specific mass balance is zero. Dynamic effects and response times, however, are not accounted for and therefore this approach excludes an exact prediction in time. As expected, glacier volume is significantly lowered in all scenarios, but glaciers in the Llanganuco catchment do not disappear completely. By 2050, a new steady-state glacier has an extent which is reduced between 38% (B1, â&#x20AC;&#x2DC;best

case’ scenario) and 60% (A2, ‘worst case’ scenario), when compared to 1990. In 2080 these values range from 49% (B1) to 75% (A2) reduction in glacier size (Juen et al., 2007). Consequently dry season runoff is significantly reduced, in particular in the A2 scenario (Figure 39). During the wet season, however, runoff is higher, which can be explained by the larger glacier-free areas and therefore enhanced direct runoff. In summary the results by Juen et al. (2007) suggest, that the overall discharge may not change very much, but that the seasonality intensifies significantly.

Fig. 39: Changes in monthly runoff in the Llanganuco catchment (in % departure from 1961-90 mean) in four IPCC climate change scenarios (A1, B1, A2 and B2) and for the years 2050 and 2080 respectively (from Juen et al., 2007).

Since the various catchments in the Cordillera Blanca have a different percentage of glaciated area, a different catchment hypsometry etc., climate change will affect their glacier extent and hence their runoff behavior in different ways. Juen (2006) simulated the change in glaciation and runoff response for five additional catchments, using the same approach as discussed above (Figure 40). In several of them the impacts under the A2 scenario are more severe than in the earlier example shown above. The Pachacoto catchment for example will be ice free by 2080 under the A2 assumption and Quilcay and Paron will have lost 92% and 95% or their glacier extent respectively (Juen, 2006). In this context it is noteworthy that the results of A2 are much more dramatic in 2050 than they are 30 years later, in 2080 under the more moderate B1 scenario. These results illustrate how uncertain the future extent of glaciation and therefore the changes in runoff really are; they clearly depend on which emission path we will ultimately follow. The IPCC never made a statement favoring one scenario over another nor did they discuss

that one path was more likely than the other. In this sense the results by Juen (2006) do not give more credence to one outcome or another, but the results can certainly put some upper and lower bounds on the changes that are to be expected 50-80 years from now.

Figure 40: Simulated change in monthly runoff in 2050 and 2080 (in %, compared with 1961-90 average) in 5 different catchments of the Cordillera Blanca based on the IPCC climate change scenarios B1 and A2 (from Juen, 2006).

In summary, the results presented above by Juen (2006) and Juen et al. (2007) show that glaciers in the tropical Andes will continue to retreat, no matter which scenario we will follow. As a consequence, glacier ice volume will continue to shrink, which is a looming threat for future water shortage during the dry season, with potentially severe consequences for the availability of drinking water, and water for irrigation, mining and hydropower production. The lack of a glacial buffer will likely lead to a more pronounced seasonal amplitude of runoff, characterized by an increase during the wet season and a decrease during the dry season. In addition the continued economic and population growth will probably also lower the overall water quality. The increased water demands will therefore likely lead to social and economic conflicts, in particular during the dry season, when water availability is low (Figure 41).

. Figure 41: Schematic view of how opposing trends in future water demand and water availability might lead to socio-economic problems (modified from Kaser et al., 2004).

6.3. Cordillera Real On Chacaltaya Ramirez et al. (2001) documented that there is a small albeit significant increase in the proglacial discharge, supplied by the progressive glacier thinning. They estimated that a complete disappearance of the glacier will reduce the overall runoff by about 30% and that its regime will become completely dependent on the precipitation variability. On Zongo glacier proglacial runoff has been monitored since the early 1990s, with earlier gauge readings by an electric power plant going back to 1973 (Ribstein et al., 1995a, b; Francou et al., 1995b; Caballero et al., 2004). Measurements show that runoff is generally higher than catchment precipitation, indicating that the glacier is losing more mass than is being replenished through snow accumulation. For example in 1991/92 916 mm of precipitation fell in the catchment, yet the runoff was 1793 mm. The following year the balance was more even with 1080 mm of precipitation and 1060 mm of runoff (Ribstein et al., 1995a). On average Ribstein et al. (1995a) estimate a 410 mm deficit (1062 mm precipitation versus 1472 mm runoff) between 1973 and 1993. They further note that sublimation was not accounted for in this study, which would have increased the deficit even further. Highest discharge occurs during warm El Ni単o years, despite dry conditions during those years. This is further confirmation for the role of glaciers as very efficient buffers, which tend to smooth runoff compared to precipitation, not just on seasonal time scales but also interannually (Ribstein et al., 1995a, Francou et al., 1995b). Nonetheless runoff from Zongo is quite seasonal, with 72% (76%) of the annual runoff occurring during the six months from October-March in 1991/92 (1992/93). Precipitation however is even more concentrated, with 82% (1991/92) and 84% (1992/93) of the annual total falling during these months (Ribstein et al., 1995a). The Zongo River supplies a whole cascade of hydroelectric power plants downstream, built and operated 57

by the Bolivian Power Company COBEE, to supply La Paz with electricity (Caballero et al., 2004). A total of 10 plants with a total capacity of 174.6 MW exist today (Figure 42). The system may not be too vulnerable to climate change related changes in seasonal runoff, because several dams already exist upstream, which retain water during the rainy season and release it during the dry season to allow for continued hydropower production.

Figure 42: Schematic map of the hydroelectric system in the Zongo valley (from Caballero et al., 2004)

7) CONCLUSIONS This report has outlined the current knowledge of climate change and its impacts on tropical Andean glaciers and hydrology. It provides a scientific basis that can be used in support of the decision-making process to find the best adaptation and mitigation strategies for the region. While some practical measures to adapt and prepare for future changes in runoff behavior should be implemented without delay (e.g. conservation, shift to less water-intensive agriculture, creation of water reservoirs, etc., see Vergara, 2005), it is also clear that much remains unknown and significant progress needs to be made on the scientific front. In order to monitor the ongoing changes in climate and cryosphere, a much better network is needed, which does not only focus on certain target areas, but provides for an actual connected network of sites along climatic gradients from north to south as well as across the Andes from east to west (Francou et al., 2005; Coudrain et al., 2005; Kaser et al., 2005; Casassa et al., 2007). A high elevation observational network of automated weather stations is needed to monitor climate change at the elevation of the glaciers and not simply at low elevations, where the changes are likely to be much less dramatic (Bradley et al., 2004). Many of the weather stations and stream gauges currently operating were installed in the middle of the 20th century (Mark and Seltzer, 2005). They are old and outdated and need to be replaced with more up-to date instrumentation. It is hard to believe but true, that we are currently not in a position to accurately monitor and document the rapid changes taking place at high elevation sites in the tropical Andes. Such a network of stations could be installed on glaciers along the tropical Andes, which would also allow to gain much needed additional information on tropical glacier energy balance along a north-south and east-west transect. The network installed and maintained by the IRD seems a logical starting point from where to expand. Networks of mass and energy balance measurement sites are important and needed but they are also costly, labor-intensive and by their very nature limited in space. They should therefore be complemented by increased use and application of available remote sensing techniques and data sets from space. New advances in combining digital elevation models, SRTM data, GPS and satellite data such as Landsat, ASTER and SPOT, offer the opportunity to give a more detailed large-scale picture of changes in both the atmosphere and the cryosphere. While they are no substitute for on-site measurements, they can provide a much needed complementary picture. The Peruvian Andes, for example have been selected as a priority site to monitor glaciers with ASTER data under the Global Land Ice Measurements from Space (GLIMS) umbrella (Mark and Seltzer, 2005). Apart from the needed improvements in on-site and remote monitoring, it has also become increasingly clear that we need better and more detailed scenarios of future climate change in this region of steep and complex topography. Output from GCMâ&#x20AC;&#x2122;s can at best provide us with a broad-brush perspective. High-resolution regional climate models, which allow for a better simulation of climate in mountain regions, coupled with tropical glacier-mass balance models, such as the one used by Juen et al. (2007) will help us to better understand and predict future climate changes and their impacts on tropical Andean glaciers and associated runoff. By running ensemble simulations under different SRES emissions paths (Nakicenovic and Swart, 2000) we can then define a probability

distribution of what encompasses the most likely climate change outcome. These simulations can further be used to feed tropical-glacier climate models, specifically designed to account for the peculiarities of energy and mass balance on tropical glaciers. Vuille (2006) recently proposed such a modeling strategy, which should allow us to establish robust projections of how glaciation and runoff will change in this region at the end of the 21st century. This has far-reaching implications for the anticipated future water shortage in the region and will provide much needed information for policy- and decision-makers. Finally it is equally vitally important that greater collaboration and integration, and more information and data-sharing between agencies and institutions takes place (Mark and Seltzer, 2005). After all the glacier-climate research in the tropical Andes is relevant not only from a purely scientific stand point but has very direct and immediate applications in the region. A number of recent workshops have tried to address this issue and to provide a starting point for such a collaborative network (Francou and Coudrain, 2005, Diaz et al., 2006, Casassa et al., 2007). Hopefully this report can contribute to this goal as well. This improved collaboration, in order to be truly relevant, should not be limited to scientific discussion and debate, but needs to include exchange with local stakeholders and decision makers. As outlined by Young and Lipton (2006) there is often a scale-related disjuncture between scientific and technical studies examining hydrologic resources, the national institutions involved in water management and the demands and needs of the local population. After all the problems surrounding a future water shortage in Andean countries are not only climatic in nature but to some extent a result of the economic and social developments in the region. One of the challenges to scientists is therefore to provide scientific information which is not only scientifically relevant but also socially applicable (Mark 2007). On the other hand, adequate mitigation and adaptation strategies can only be put in place with a better and much more detailed scientific knowledge on how future climate change will affect glaciological and hydrological systems in the Andes. Therefore it is of the highest priority to rapidly improve our understanding of both mechanisms and consequences related to the disappearance of tropical Andean glaciers.

REFERENCES Aceituno, P. Montecinos, A., 1997: Patterns of convective cloudiness in South America during austral summer from OLR pentads. Preprints, 5th Int. Conf. on Southern Hemisphere Meteorology and Oceanography, Pretoria, South Africa, Amer. Meteor. Soc., 328–329. Aguilar, E., Peterson, T. C., Obando, P. Ramírez, Frutos, R., Retana, J. A., Solera, M., Soley, J., García, I. González, Araujo, R. M., Santos, A. Rosa, Valle, V. E., Brunet, M., Aguilar, L., Álvarez, L., Bautista, M., Castañón, C., Herrera, L., Ruano, E., Sinay, J. J., Sánchez, E., Oviedo, G. I. Hernández, Obed, F., Salgado, J. E., Vázquez, J. L., Baca, M., Gutiérrez, M., Centella, C., Espinosa, J., Martínez, D., Olmedo, B., Espinoza, C. E. Ojeda, Núñez, R., Haylock, M., Benavides, H., Mayorga, R. 2005: Changes in precipitation and temperature extremes in Central America and northern South America, 1961-2003. J. Geophys. Res., 110, D23107, doi:10.1029/2005JD006119. Ames, A., Dolores, S., Valverde, A., Evangelista, C., Javier, D., Ganwini, W., Zuniga, J., 1989: Glacier inventory of Peru, Part I. Hidrandina S.A. Huaraz, Peru. Ames, A., Hastenrath, S., 1996: Mass Balance and Ice flow of the Urushraju glacier, Cordilera Blanca, Peru. Zeitschrift fuer Gletscherkunde und Glazialgeologie, 32, 83-89. Ames, A., 1998: A documentation of glacier tongue variations and lake development in the Cordillera Blanca, Peru. Zeitschrift fuer Gletscherkunde und Glazialgeologie, 34 (1), 1-36. Arnaud, Y., Muller, F., Vuille, M., Ribstein, P., 2001: El Niño - Southern Oscillation (ENSO) influence on a Sajama volcano glacier from 1963 to 1998 as seen from Landsat data and aerial photography. J. Geophys. Res., 106, 17773-17784. Barnett, T.P., Adam, J.C., Lettenmaier, D.P., 2005: Potential impacts of a warming climate on water availability in snow-dominated regions. Nature, 438, 303-309. Barry, R.G., Seimon, A., 2000: Research for mountain area development: Climatic fluctuations in the mountains of the Americas and their significance. Ambio, 29 (7), 364-370. Berger, T., Mendoza,J., Francou, B., Rojas, F., Fuertes, R., Flores, M., Noriega, L., Cinthya Ramallo, Ramirez, E. Baldivieso, H., 2006: Glaciares Zongo – Chacaltaya – Charquini Sur – Bolivia 16°S. Mediciones Glaciológicas, Hidrológicas & Meteorológicas, Año Hidrológico 2004-2005. Great Ice report, Bolivia, IRD-IHH-SENMAHI-COBEE, 171 p. Boulanger, J.-P., Martinez, F., Segura, E.C., 2006: Projection of future climate change conditions using IPCC simulations, neural networks and Bayesian statistics. Part 1: Temperature mean state and seasonal cycle in South America. Climate Dyn., 27, 233-259. Bradley, R. S., Vuille, M., Hardy, D.R., Thompson, L.G., 2003: Low latitude ice cores record Pacific sea surface temperatures. Geophys. Res. Lett., 30 (4), 1174, doi: 10.1029/2002GL016546. Bradley, R.S., Keimig, F.T., Diaz, H.F., 2004: Projected temperature changes along the American Cordillera and the planned GCOS network. Geophys. Res. Lett., 31, L16210, doi:10.1029/2004GL020229.

Bradley, R.S., Vuille, M., Diaz, H.F., Vergara, W., 2006: Threats to water supplies in the Tropical Andes. Science, 312, 5781, 1755-1756. Brecher, H.H., Thompson, L.G., 1993: Measurement of the retreat of Qori Kalis glacier in the tropical Andes of Peru by terrestrial photogrammetry. Photogrammetric Engineering Rem. Sensing, 59(6), 1017-1022. Buytaert, W., Celleri, R., De Bievre, B., Cisneros, F., Wyseure, G., Deckers, J., Hofstede, R., 2006: Human impact on the hydrology of the Andean paramos. Earth Sci. Rev., 79, 53-72. Caballero, Y., Chevalier, P., Gallaire, R., Pillco, R., 2004: Flow modeling in a high mountain valley equipped with hydropower plants: Rio Zongo Valley, Cordillera Real Bolivia. Hydrol. Proc., 18 (5), 939-957. Casassa, G., Haeberli, W., Jones, G., Kaser, G., Ribstein, P., Rivera, A., Schneider, C., 2007: Current status of Andean glaciers. Global Planet. Change, in press. Ceballos, J.L., Euscategui, C., Ramirez, J., Canon, M., Huggel, C., Haeberli, W., Machguth, H., 2006: Fast shrinkage of tropical glaciers in Colombia. Ann. Glaciol., 43, 194-201. Chen, J., Carlson, B. E., Del Genio, A. D., 2002: Evidence for strengthening of the tropical general circulation in the 1990s. Science, 295, 838–841. Chen, T.-C., Yoon, J. H., St. Croix, K. J., Takle, E. S., 2001: Suppressing impacts of the Amazonian deforestation by the global circulation change. Bull. Amer. Meteorol. Soc., 82, 2209–2216. Chu, P.-S., Yu, Z.-P., Hastenrath, S., 1994: Detecting climate change concurrent with deforestation in the Amazon basin: Which way has it gone? Bull. Amer. Meteorol. Soc. 75, 579–583. Coudrain, A., Francou, B., Kundzewicz, Z.W., 2005: Glacier shrinkage in the Andes and consequences for water resources. Hydrol. Sci. J., 50(6), 925-932. Curtis, S., Hastenrath, S., 1999a: Long-term trends and forcing mechanisms of circulation and climate in the equatorial Pacific. J. Climate, 12, 1134–1144. Curtis, S., Hastenrath, S., 1999b: Trends of upper-air circulation and water vapour over equatorial South America and adjacent oceans. Int. J. Clim., 19, 863–876. Diaz, H.F., Graham N.E. 1996: Recent changes in tropical freezing heights and the role of sea surface temperature. Nature, 383, 152-155. Diaz, H.F., Eischeid, J.K., Duncan, C., Bradley, R.S., 2003: Variability of freezing levels, melting season indicators and snow cover for selected high elevation and continental regions in the last 50 years. Climatic Change, 59, 33-52. Diaz, H.F., Villalba, R., Greenwood, G., Bradley, R.S., 2006: The impact of climate change in the American Cordillera. Eos Trans. Amer. Geophys. Union, 87, 32, 315-316. Favier, V., Wagnon, P., Ribstein, P., 2004a: Glaciers of the outer and inner tropics: A different behavior but a common response to climatic forcing. Geophys. Res. Lett., 31, L16403, doi:10.1029/2004GL020654. Favier, V., Wagnon, P., Chazarin, J.-P., Maisincho, L., Coudrain, A., 2004b: One-year measurements of surface heat budget on the ablation zone of Antizana glacier 15, Ecuadorian Andes. J. Geophys. Res., 109, D18105, doi:10.1029/2003JD004359. Florez, A. 1992. Los nevados de Colombia, glaciales y glaciaciones. Anal. Geogr., 22, 15–55.

Francou, B., Ribstein, P., Sémiond, H., Portocarrero, C., Rodriquez, A., 1995a: Balances de glaciares y clima en Bolivia y Perú: Impacto de los eventos ENSO. Bull. Inst. fr. études andines, 24,(3), 661-670. Francou, B., Ribstein, P., Saravia, R., Tiriau, E., 1995b: Monthly balance and water discharge of an inter-tropical glacier: Zongo glacier, Cordillera Real, Bolivia, 16°S. J. Glaciol., 41, (137), 61-67. Francou, B., Ramirez, E., Mendoza, J., Guereca, J., Miranda, G., Noriega, L., 1998: El glaciar de Chacaltaya (Cordillera Real, Bolivia), Investigaciones glaciológicas (1991-1997). Informe No. 56, ORSTOM - Bolivie, 98p. Francou, B., Fabre, D., Pouyaud, B., Jomelli, V., Arnaud, Y., 1999: Symptoms of degradation in a tropical rock glacier, Bolivian Andes. Permafrost Periglacial Proc.. 10, 91-100. Francou, B., Ramirez, E., Caceres, B., Mendoza, J., 2000: Glacier evolution in the tropical Andes during the last decades of the 20th century: Chacaltaya, Bolivia and Antizana, Ecuador. Ambio, 29 (7), 416-422. Francou, B., Vuille, M., Wagnon, P., Mendoza J., Sicart, J.E., 2003: Tropical climate change recorded by a glacier in the central Andes during the last decades of the 20th century: Chacaltaya, Bolivia, 16°S. J. Geophys. Res., 108, D5, 4154, doi: 10.1029/2002JD002959. Francou, B., Vuille, M., Favier, V., Cáceres, B., 2004: New evidence for an ENSO impact on low latitude glaciers: Antizana 15, Andes of Ecuador, 0°28’S. J. Geophys. Res., 109, D18106, doi:10.1029/2003JD004484. Francou, B., Coudrain, A., 2005: Glacier shrinkage and water resources in the Andes. EOS Trans. Amer. Geophys. Union, 43, 415. Francou, B., Ribstein, P., Wagnon, P., Ramirez, E., Pouyaud, B., 2005: Glaciers of the tropical Andes: Indicators of global climate variability. In: Huber, U., H. K. M. Bugmann, and M. A. Reasoner (eds.): Global Change and Mountain Regions: An overview of current knowledge, 23, 197-204, Springer, Dordrecht. Francou, B., Vincent, C., 2007: Les glaciers vont-ils disparaître? Glaciers face aux changement climatique. IRD and CNRS Editions, Paris, in press. Fu, Q., Johanson, C.M., 2005: Satellite-derived vertical dependence of tropical tropospheric temperature trends. Geophys. Res. Lett., 32, L10703, doi:10.1029/2004GL022266. Gaffen, D.J., Santer, B.D., Boyle, J.S., Christy, J.R., Graham, N.E., Ross, R.J., 2000: Multidecadal changes in the vertical temperature structure of the tropical troposphere. Science, 287, 1242-1245. Garreaud, R., Vuille, M., Clement, A.C., 2003: The climate of the Altiplano: Observed current conditions and mechanisms of past changes. Palaeogeogr. Palaeoclimatol. Palaeoecol., 194, 5-22. Georges, C., 2004: The 20th century glacier fluctuations in the tropical Cordillera Blanca, Peru. Arctic, Antarctic Alpine Res., 36 (1), 100-107. Gutzler, D. S., 1992: Climatic variability of temperature and humidity over the tropical western Pacific. Geophys. Res. Lett., 19, 1595–1598. Hardy, D.R., Vuille, M., Bradley, R.S., 2003: Variability of snow accumulation and isotopic composition on Nevado Sajama, Bolivia. J. Geophys. Res., 108, D22, 4693, doi: 10.1029/2003JD003623

Hastenrath, S., 1978: Heat-budget measurements on the Quelccaya ice cap, Peruvian Andes. J. Glaciol., 20 (82), 85-97. Hastenrath, S., 1981: The glaciation of the Ecuadorian Andes. A.A. Balkema, Rotterdam, 159p. Hastenrath, S., Ames, A., 1995a: Recession of Yamarey glacier in Cordillera Blanca, Peru during the 20th century. J. Glaciol., 41 (137), 191-196. Hastenrath, S., Ames, A., 1995b: Diagnosing the imbalance of Yanamarey glacier in the Cordillera Blanca of Peru. J. Geophys. Res., 100, 5105-5112. Hastenrath, S., 1997: Measurements of diurnal heat exchange on the Quelccaya Ice cap, Peruvian Andes. Meteorol. Atmos. Phys., 62, 71-78. Hastenrath, S., 1998: Cordillera Blanca on Landsat imagery and Quelccaya Ice cap. In: Williams, R.S., Ferrigno, J.G., (Eds.): Satellite Image atlas of the World- South America. USGS Professional paper 1386-I Haylock, M.R., Peterson, T.C., Alves, L.M., Ambrizzi, T., Anunciacao, M.T., Baez, J., Barros, V.R., Berlato, M.A., Bidegain, M., Coronel, G., Corradi, V., Garcia, V.J., Grimm, A.M., Karoly, D., Marengo, J. A., Marino, M.B., Moncunill, D.F., Nechet, D., Quintana, J., Rebello, E., Rusticucci, M., Santos, J.L., Trebejo, I., Vincent, L.A., 2006: Trends in total and extreme South American rainfall in 1960-2000 and links with sea surface temperature. J. Climate, 19, 1490-1512. Hoffmann, G., Ramirez, E., Taupin, J.D., Francou, B., Ribstein, P., Delmas, R., Duerr, H., Gallaire, R., Simoes, J., Schotterer, U., Stievenard, M., Werner, M., 2003: Coherent isotope history of Andean ice cores over the last century. Geophys. Res. Lett., 30(4), doi:10.1029/2002GL014870. Hoyos-Patino, F., 1998: Glaciers of Colombia. In: Williams, R.S., Ferrigno, J.G., (Eds.): Satellite Image atlas of the glaciers of the world- South America. USGS Professional paper 1386-I. Jordan, E., Brockmann, C., Fernandez, A., Alvarez, R., Jacobsen, K., 1980: The glacier inventory of Bolivia. World glacier inventory, Proc. Riederalp workshop, IAHS Publ., 126, 25-32. Jordan, E., 1983: The utility of a glacier inventory to developing countries such as Bolivia. In: The quaternary of South America and Antarctic Peninsula, 1, 125134. Jordan, E., 1985: Recent glacier distribution and present climate in the Central Andes of South America: Zeitschrift f端r Gletscherkunde und Glazialgeologie, 21, 213-224. Jordan, E., Geyer, K., Linder, W., Fernandez, B., Florez, A., Mojica, J., Ni単o, O., Torrez, C., Guarnizo, F., 1989: The recent glaciation of the Colombian Andes: Zentralblatt f端r Geologie und Pal辰ontologie, 1(5-6), 1113-1117. Jordan, E., Hastenrath, S., 1998: Glaciers of Ecuador. In: Williams, R.S., Ferrigno, J.G., (Eds.), Satellite image atlas of the glaciers of the world - South America. USGS Professional paper 1386-I, 31-50. Jordan, E., 1998: Glaciers of Bolivia. In: Williams, R.S., Ferrigno, J.G., (Eds.), Satellite image atlas of the world - South America. USGS Professional paper 1386-I, 81108. Jordan, E., Ungerechts, L., Caceres, B., Penafiel, A., Francou, B., 2005: Estimation by photogrammetry of the glacier recession on the Cotopaxi volcano (Ecuador) between 1956 and 1997. Hydrol. Sci. J., 50(6), 949-961.

Juen, I., 2006: Glacier mass balance and runoff in the Cordillera Blanca, Perú. Ph.D. thesis, University of Innsbruck, 173 p. Juen, I., Kaser, G., Georges, C., 2007: Modeling observed and future runoff from a glacierized tropical catchment (Cordillera Blanca, Perú). Global Planet. Change, in press. Kaser, G., Ames, A., Zamora, M., 1990: Glacier fluctuations and climate in the Cordillera Blanca, Peru. Ann. Glaciol., 14, 136-140. Kaser, G., 1995: Some notes on the behavior of tropical glaciers. Bull. Inst. fr. études andines, 24(3), 671-681. Kaser, G., Georges, C., Ames, A., 1996a: Modern glacier fluctuations in the HuascaranChopicalqui-Massif of the Cordillera Blanca, Peru. Zeitschrift fuer Gletscherkunde und Glazialgeologie, 32, 91-99. Kaser, G., Hastenrath, S., Ames, A., 1996b: Mass balance profiles on tropical glaciers. Zeitschrift fuer Gletscherkunde und Glazialgeologie, 32, 75-81. Kaser, G., Georges, C., 1997: Changes of the equilibrium-line altitude in the tropical Cordillera Blanca, Peru, 1930-1950, and their spatial variations. Ann. Glaciol., 24, 344-349. Kaser, G., Georges, C., 1999: On the mass balance of low latitude glaciers with particular consideration of the Peruvian Cordillera Blanca. Geograf. Ann., 81A(4), 643-651. Kaser, G., 1999: A review of the modern fluctuations of tropical glaciers. Global Planet. Change, 22, 93-103. Kaser, G., 2001: Glacier-climate interaction at low latitudes. J. Glaciol., 47(157), 195204. Kaser, G., 2002: Glacier mass balance and climate in the South American Andes. In: Casassa, G., Sepulveda, F.V., Sinclair, R.M., (Eds.): The Patagonian Icefields, 89-100. Kaser, G., Osmaston, H., 2002: Tropical Glaciers. Cambridge University Press, Cambridge. Kaser, G., Juen, I., Georges, C., Gomez, J., Tamayo, W., 2003: The impact of glaciers on the runoff and the reconstruction of mass balance history from hydrological data in the tropical Cordillera Blanca, Peru. J. Hydrol., 282(1-4), 130-144. Kaser, G., Georges, C., Juen, I., Mölg, T., 2004: High-altitude glaciers at low latitudes (in german). Innsbrucker Jahresberichte, 16, 82-103. Kaser, G., Georges, C., Juen, I., Mölg, T., 2005: Low-latitude glaciers: Unique global climate indicators and essential contributors to regional fresh water supply. A conceptual approach. In: Huber, U., Bugmann, H. K. M, Reasoner, M.A., (Eds.): Global Change and Mountain Regions: An overview of current knowledge, 23, 185-196, Springer, Dordrecht. Kaser, G., Cogley, J.G., Dyurgerov, M.B., Meier, M.F., Ohmura, A., 2006: Mass balance of glaciers and ice caps: Consensus estimates for 1961-2004. Geophys. Res. Lett., 33, L19501, doi:10.1029/2006GL027511. Klein, A.G., Masek, J.G., Isacks, B.L., 1996: Shortwave radiation geometry and glacier distribution in the Cordillera Real, Bolivia. In Colbeck, Samuel (Ed.): Glaciers, Ice Sheets and Volcanoes: A tribute to Mark F. Meier. CRREL Special Report 9627, 57-61.

Kousky, V. E. Kayano, M. T., 1994: Principal modes of outgoing longwave radiation and 250 mb circulation for the South American sector. J. Climate, 7, 1131–1143. Liebmann, B., Marengo, J. A., Glick, J. D., Kousky, V. E., Wainer, I. C., Massambani, O., 1998: A comparison of rainfall, outgoing longwave radiation and divergence over the Amazon basin. J. Climate, 11, 2898–2909. Mark, B.G., 2002: Hot ice: glaciers in the tropics are making the press. Hydrol. Proc., 16, 3297-3302. Mark, B.G., Seltzer, G.O., 2003: Tropical glacier meltwater contribution to stream discharge: a case study in the Cordillera Blanca, Peru. J. Glaciol., 49, 165, 271281. Mark, B.G., Seltzer, G.O., 2005a: Evaluation of recent glacier recession in the Cordillera Blanca, Peru (AD 1962-1999): spatial distribution of mass loss and climatic forcing. Quatern. Sci. Rev., 24, 2265-2280. Mark, B.G., Seltzer, G.O., 2005b: Deglaciation in the Peruvian Andes: Climatic forcing, hydrologic impact and comparative rates over time. In: Huber, U., Bugmann, H. K. M., Reasoner, M.A., (Eds.): Global Change and Mountain Regions: An overview of current knowledge, 23, 205-214, Springer, Dordrecht. Mark, B.G., McKenzie, J.M., Gomez, J., 2005: Hydrochemical evaluation of changing glacier meltwater contribution to stream discharge: Callejon de Huaylas, Peru. Hydrol. Sci. J., 50(6), 975-987. Mark, B.G., 2007: Tracing tropical Andean glaciers over space and time: Some lessons and transdiciplinary implications Global Planet. Change, in press. Messerli, B., Grosjean, M, Bonani, G., Bürgi, A., Geyh, M.A., Graf, K., Ramseyer, K., Romero, H., Schotterer, U., Schreier, H., Vuille, M., 1993: Climate Change and natural resource dynamics of the Atacama Altiplano during the last 18'000 years: a preliminary synthesis. Mountain Res. Devel., 13(2), 117-127. Messerli, B., 2001: The International Year of Mountains (IYM), the Mountain Research Initiative (MRI) and PAGES. Editorial, Pages News, 9(3), 2. Morales-Arnao, B., 1998: Glaciers of Peru. In: Williams, R.S., Ferrigno, J.G., (Eds.), Satellite image atlas of the world - South America. USGS Professional paper 1386-I, Nakicenovic, N., Swart, R., (Eds.) 2000: Special Report on Emissions Scenarios. Cambridge Univ. Press, Cambridge,U.K. New, M., Hulme, M., Jones, P., 2000: Representing twentieth-century space-time climate variability. Part II: Development of 1901-1996 monthly grids of terrestrial surface climate. J. Climate, 13, 2217-2238. Patzelt, G., 1983: Die Berg- und Eisstuerze von Huascaran, Cordillera Blanca, Peru. Hochgebirgsforschung – High Mountain Research, 6. Arbeitsgemeinschaft fuer vergleichende Hochgebirgsforschung. Portocarrero, C., 1995: Glacier retreat in Peru: Consequences on water resources and geodynamic hazards. Bull. Inst. fr. etudes andines, 24(3), 697-706. Pouyaud, B., Francou, B., Ribstein, P., 1995: A glacier monitoring network in the tropical Andes. Bull. Inst. fr. etudes andines, 24(3), 707-714. Pouyaud, B., Zapata, M., Yerren, J., Gomez, J., Rosas, G., Suarez, W., Ribstein, P., 2005: Avenir des resources en eau glaciare de la Cordillere Blanche. Hydrol. Sci. J., 50(6), 999-1022.

Quintana-Gomez, R.A., 1997: Changes in maximum and minimum temperatures at high elevation stations in the central Andes of South America. Proc. AMS Meeting on Climate Variations, J112-J113. Quintana-Gomez, R.A., 1999: Trends of maximum and minimum temperatures in northern South America, J. Climate, 12, 2104-2112. Quintana-Gomez, R.A., 2000: Trends of maximum and minimum temperatures in Ecuador and homogeneity evaluation during 1961-90. Preprints 6th Int. Conf. Southern Hemisphere Meteorology and Oceanography, Santiago Chile, 292-293. Rabatel, A., Jomelli, V., Naveau, P., Francou, B., Grancher, D., 2005: Dating of Little Ice Age glacier fluctuations in the tropical Andes: Charquini glaciers, Bolivia, 16S. Comptes Rendus Geoscience, 337, 1311-1322. Rabatel, A., Machaca, A., Francou, B., Jomelli, V., 2006: Glacier recession on Cerro Charquini (16°S), Bolivia since the maximum of the Little Ice Age (17th century). J. Glaciol., 52(176), 110-118. Racoviteanu, A.E., Manley, W., Arnaud, Y., Williams, M.W., 2007: Evaluating digital elevation models for glaciologic applications: An example from Nevado Coropuna, Peruvian Andes. Global Planet. Change, in press. Ramirez, E., Francou, B., Ribstein, P., Desclitres, M., Guerin, R., Mendoza, J., Gallaire, R., Pouyaud, B., Jordan, E., 2001: Small glaciers disappearing in the tropical Andes: a case study in Bolivia: Glaciar Chacaltaya (16°S). J. Glaciol., 47(157), 187-194. Raup, B., Racoviteanu, A., Khalsa, S.J.S., Helm, C., Armstrong, R., Arnaud, Y., 2006: The GLIMS geospatial glacier database: A new tool for studying glacier change. Global Planet. Change, in press. Regalado, A., 2005: Ukukus worry about glacier. The Wall Street Journal, June 17th, 2005. Ribstein, P., Tiriau, E., Francou, B., Saravia, R., 1995a: Tropical climate and glacier hydrology: a case study in Bolivia. J. Hydrol., 165, 221-234. Ribstein, P., Francou, B., Rigaudiere, P., Saravia, R., 1995b: Climatic variability and hydrologic modeling on Zongo glacier, Bolivia. Bull. Inst. fr etudes andines, 24(3), 639-649. Ronchail, J., 1995: Variabilidad interanual de las precipitaciones en Bolivia’, Bull. Inst. fr. etudes andines, 24, 369–378. Rossow, W. B., Schiffer, R.A., 1999: Advances in understanding Clouds from ISCCP. Bull. Amer. Meteorol. Soc., 80(11), 2261-2287. Santer, B.D., Wehner, M.F., Wigley, T.M.L., Sausen, R., Meehl, G.A., Taylor, K.E., Ammann, C., Arblaster, J., Washington, W.M., Boyle, J.S., Brueggemann, W., 2003: Contributions of anthropogenic and natural forcings to recent tropopause height changes. Science, 301, 479-483. Santer, B.D., Wigley, T.M.L., Mears, C., Wentz, F.J., Klein, S.A., Seidel, D.J., Taylor, K.E., Thorne, P.W., Wehner, M.F., Gleckler, P.J., Boyle, J.S., Collins, W.D., Dixon, K.W., Doutriaux, C., Free, M., Fu, Q., Hansen, J.E., Jones, G.S., Ruedy, R., Karl, T.R., Lanzante, J.R., Meehl, G.A., Ramaswamy, V., Russell, G., Schmidt, G.A., 2005: Amplification of surface temperature trends and variability in the tropical atmosphere. Science, 309, 1551-1556.

Schubert, C., 1992: The glaciers of the Sierra Nevada de Mérida (Venezuela), a photographic comparison of recent deglaciation. Erdkunde, 46, 58-64. Schubert, C., 1998: Glaciers of Venezuela. In: Williams, R.S., Ferrigno, J.G., (Eds.), Satellite image atlas of the glaciers of the world - South America. USGS Professional paper 1386-I, 81-108. Seidel, D.J., Free, M., 2003: Comparison of lower – tropospheric temperature climatologies and trends at low and high elevation radiosonde sites. Climatic Change, 59, 53-74. Semiond, H., Francou, B., Ayabaca, E., de la Cruz, A., Chango, R., 1997: El Glaciar 15 del Antizana (Ecuador). Investigaciones glaciologicas 1994-1997. 88p. Semiond, H., Francou, B., Ayabaca, E., de la Cruz, A., Gomez, G., Chango, R., Galárraga, R., 1998: La desglaciacion del Antizana (Ecuador) al partir del monitoreo del glaciar 15ª (1994-1997). Bull. Inst. fr. etudes andines, 27(3), 605610). Sicart, J. E., Ribstein, P., Francou, B., Gallaire, R., 2003: Etude des précipitations et de la fonte sur un glacier tropical: le glacier du Zongo, Bolivie, 16°S. Hydrol. Sci. J., 48(5), 799-808. Sicart, J.E., Wagnon, P., Ribstein, P., 2005: Atmospheric controls of the heat balance of Zongo Glacier (16°S, Bolivia). J. Geophys. Res., 110, D12106, doi:10.1029/2004JD005732. Sicart, J.E., Ribstein, P., Francou, B., Pouyaud, B., Condom, T., 2006: Glacier mass balance of tropical Zongo glacier, Bolivia, comparing hydrological and glaciological methods. Global Planet. Change, in press. Silverio, W., Jaquet, J.-M., 2005: Glacial cover mapping (1987-1996) of the Cordillera Blanca (Peru) using satellite imagery. Rem. Sens. Environ., 95, 342-350. Solomina, O., Jomelli, V., Kaser, G., Ames, A., Berger, B., Pouyaud, B., 2007: Lichenometry in the Cordillera Blanca, Peru: ‘Little Ice Age” moraine chronology. Global Planet. Change, in press. Thompson, L.G., Mosley-Thompson, E., Davis, M., Lin, P. N. Yao, T., Dyurgerov, M., Dai, J., 1993: ‘Recent warming’: Ice core evidence from tropical ice cores with emphasis on Central Asia. Global Planet. Change, 7, 145-156. Thompson, L.G., 2000: Ice core evidence for climate change in the Tropics: implications for our future. Quatern. Sci. Rev., 19(1-5), 19-35. Thompson, L.G., Mosley-Thompson, E., Davis, M.E., Lin, P.-N., Henderson, K., Mashiotta, T.A., 2003: Tropical glacier and ice core evidence of climate change on annual to millennial time scales. Climatic Change, 59, 137-155. Thompson, L.G., Mosley-Thompson, E., Brecher, H., Davis, M., Leon, B., Les, D., Lin, P.-N., Mashiotta, T., Mountain, K., 2006: Abrupt tropical climate change: Past and present. Proc. National Academy of Science, doi:10.1073/pnas.0603900103. Thouret, J.-C., 1996: Palaeoenvironmental changes and glacial stades of the last 50,000 years in the Cordillera Central, Colombia: Quatern. Res., 46(1), 1-18. Toumi, R., Hartell, N., Bignell, K., 1999: Mountain station pressure as an indicator of climate change. Geophys. Res. Lett., 26(12), 1751-1754. Trenberth, K. E., Guillemot, C.J., 1998: Evaluation of the atmospheric moisture and hydrological cycle in the NCEP/NCAR reanalyses. Climate Dyn., 14, 213-231.

Trenberth, K.E., Stepaniak, D.P., Hurrell, J.W., Fiorino, M., 2001: Quality of reanalyses in the Tropics. J. Climate, 14, 1499-1510. Trenberth, K. E., 2002: Changes in tropical clouds and radiation. Science, 296, 2095a. Vecchi, G.A., Soden, B., Wittenberg, A.T., Held, I.M., Leetmaa, A., Harrison, M.J., 2006: Weakening of tropical Pacific atmospheric circulation due to anthropogenic forcing. Nature, 441, 73-76. Vera, C., Silvestri, G., Liebmann, B., Gonzalez, P., 2006: Climate change scenarios for seasonal precipitation in South America from IPCC-AR4 models. Geophys. Res. Lett., 33, L13707, doi:10.1029/2006GL025759. Vergara, W., 2005: Adapting to Climate Change. Latin America and Caribbean Region Sustainable Development Working Paper, 25. The World Bank, Washington D.C. Villalba, R., Grau, H. R., Boninsegna, J. A., Jacoby, G. C., Ripalta, A., 1998: Tree-ring evidence for long-term precipitation changes in subtropical South America. Int. J. Clim., 18, 1463–1478. Vimeux, F., Gallaire, R., Bony, S., Hoffmann, G., Chiang, J.C.H., 2005: What are the climate controls on dD in precipitation in the Zongo Valley (Bolivia)? Implications for the Illimani ice core interpretation. Earth Planet. Sci. Lett., 240, 205-220. Vincent, L.A., Peterson, T.C., Barros, V.R., Marino, M.B., Rusticucci, M., Carrasco, G., Ramirez, E., Alves, L.M., Ambrizzi, T., Berlato, M.A., Grimm, A.M., Marengo, J.A., Molion, L., Moncunill, D.F., Rebello, E., Anunciaciao, Y.M.T., Quintana, J., Santos, J.L., Baez, J., Coronel, G., Garcia, J., Trebejo, I., Bidegain, M., Haylock, M.R., Karoly, D., 2005: Observed trends in indices of daily temperature extremes in South America, 1960-2000. J. Climate, 18, 5011-5023. Vörösmarty, C.J., Green, P., Salisbury, J., Lammers, R.B., 2000: Global water resources: Vulnerability from climate change and population growth. Science, 289, 284-288. Vuille, M., 1999: Atmospheric circulation over the Bolivian Altiplano during dry and wet periods and extreme phases of the Southern Oscillation. Int. J. Climatol., 19, 1579-1600. Vuille, M., Bradley, R. S., Keimig, F., 2000a: Interannual climate variability in the Central Andes and its relation to tropical Pacific and Atlantic forcing. J. Geophys. Res., 105,12,447-12,460. PDF-file Vuille, M., Bradley, R. S., Keimig, F., 2000b: Climate variability in the Andes of Ecuador and its relation to tropical Pacific and Atlantic sea surface temperatures anomalies. J. Climate, 13, 2520-2535. Vuille, M., Bradley, R. S., 2000: Mean annual temperature trends and their vertical structure in the tropical Andes. Geophys. Res. Lett., 27, 3885-3888. Vuille, M., Bradley, R.S., Werner, M., Keimig, F., 2003a: 20th century climate change in the tropical Andes: observations and model results. Climatic Change, 59(1-2), 7599. Vuille, M., Bradley, R.S., Werner, M., Healy, R., Keimig, F., 2003b: Modeling d18O in precipitation over the tropical Americas: 1. Interannual variability and climatic controls. J. Geophys. Res., 108, D6, 4174, doi:10.1029/2001JD002038 Vuille, M., Bradley, R.S., Healy, R., Werner, M., Hardy D. R., Thompson, L. G., Keimig, F., 2003c: Modeling d18O in precipitation over the tropical Americas: 2.

Simulation of the stable isotope signal in Andean ice cores. J. Geophys. Res., 108, D6, 4175, doi:10.1029/2001JD002039 Vuille, M., Keimig, F., 2004: Interannual variability of summertime convective cloudiness and precipitation in the central Andes derived from ISCCP-B3 data. J. Climate, 17, 3334-3348. Vuille, M., 2006: Climate Change in the tropical Andes – Observations, models, and simulated future impacts on glaciers and streamflow. Mountain Res. Dev., 26, 3. Vuille, M., Bradley, R.S., 2007: The vulnerable Andes – The impacts of climate change in the Cordillera. Climate Impacts and priority responses - World Bank Sustainable Development Series, in press. Vuille, M., Kaser, G., Juen, I., 2007: Multi-scale analysis of climate - glacier interactions in the Cordillera Blanca, Peru. J. Geophys. Res., (in prep). Wagnon, P., Ribstein, P., Francou, B., Pouyaud, B., 1999a: Annual cycle of energy balance of Zongo glacier, Cordillera Real, Bolivia. J. Geophys. Res., 104(D4), 3907-3923. Wagnon, P., Ribstein, P., Kaser, G., Berton, P., 1999b: Energy balance and runoff seasonality of a Bolivian glacier. Global Planet. Change, 22, 49-58. Wagnon, P., Ribstein, P., Francou, B., Sicart, J.E., 2001: Anomalous heat and mass budget of Glaciar Zongo, Bolivia, during the 1997-98 El Niño year. J. Glaciol., 47, 21-28. Wagnon, P., Sicart, J.-E., Berthier, E., Chazarin, J.-P., 2003: Wintertime high-altitude surface energy balance of a Bolivian glacier, Illimani, 6340 m above sea level. J. Geophys. Res., 108, 6, doi:10.1029/2002JD002088. Wentz, F.J., Schable, M, 2000: Precise climate monitoring using complementary satellite data sets. Nature, 403, 414-416 Wielicki, B. A., Wong, T., Allan, R. P., Slingo, A., Kiehl, J. T., Soden, B. J., Gordon, C. T., Miller, A. J., Yang, S. K., Randall, D. A., Robertson, F., Susskind, J., Jacobowitz, H., 2002: Evidence for large decadal variability in the tropical mean radiative energy budget. Science, 295, 841–844. Wigley, T.M.L., 2005: The climate change commitment. Science, 307, 1766-1769. Young, K.R., Lipton, J.K., 2006: Adaptive governance and climate change in the tropical highlands of western South America. Climatic Change, 78, 63-102.