Exploring hydrologic connections between tropical mountain wetlands and glacier recession in Peru's Cordillera Blanca
Graphical abstract
This research proposes relationships among peak water, glacier recession, changing glacier discharge, wetland area, and wetland ecosystem services with the approximate period of this study indicated by the grey shaded area.
Introduction
Landscape change can result from shifting hydrological processes and, in turn, feedbacks through altered evapotranspiration, runoff, and water storage can further influence landscape arrangement and composition. Landscape changes detected by satellite observations provide insights into the hydrologic connectivity that characterizes these processes, although quantifying the effects is complicated by high natural variability, typically unmeasured groundwater interconnections, and incomplete data archives (DeFries and Eshleman, 2004, Wohl et al., 2012). In this research, we adapt an econometric model to examine the influence of glacier retreat on downslope wetland ecosystems. To do this, we utilized an historic stream discharge dataset for selected Santa River watersheds in the Cordillera Blanca of northern Peru and compared it to land cover change derived from satellite imagery. The general approach can be adapted to study many other situations where the relationship between hydrologic connectivity and environmental drivers of change is in need of assessment (e.g. Ponette-González et al., 2014).
In joining econometric modeling with spatial categorical data, the explanatory variables can be conceived of as spatial driving forces, with ancillary datasets representing variables such as precipitation, evapotranspiration, or topography that can further specify the model (Chowdhury, 2006). Response variables can similarly be extracted from spatial categorical data derived from remotely-sensed imagery. Multiple logistic regression is often used (Cui et al., 2014, Mertens and Lambin, 2000), which provides a probability estimate of the effect of the explanatory variables on the response variable (Wooldridge, 2009). Another technique involves panel data and fixed effects regression (Andreß, Golsch, & Schmidt, 2012), which relies on a simple design: make each observation its own control. In this case, for each watershed we can estimate the effects of explanatory variables and control for unobserved variables. This is useful in studies where a non-experimental design is the only available strategy to evaluate interconnections. The response variable must be measured on two occasions at a minimum and the measurements must be directly comparable over time; additionally, the explanatory variables must exhibit variation across time (Allison, 2009). In contrast to multiple logistic regression that measures a probability response for change from one land cover to another, fixed effects regression generates an estimate of the relative effect of an explanatory variable over time, thus explicitly estimating how changes in the explanatory variables affect the response variable temporally (Wooldridge, 2009); doing so eliminates time-constant effects, both observed and unobserved. When two time periods are used, the technique is referred to as first differencing (Wooldridge, 2009).
Here we utilize the changes in wetland area as a way to evaluate hydrologic connectivity as affected by the glaciers themselves, and by other hydrologic and climatologic variables. Like many glaciated mountains nowadays, ice cover in the Cordillera Blanca is decreasing (Kaser, 1999, Rabatel et al., 2013, Thompson et al., 2006). Permanent ice cover declined from 723 km2 in 1970 (Kaser, Ames, & Zamora, 1990) to 482 km2 by 2010 (Burns & Nolin, 2014). Ice loss in the high Andes is associated with multiple social and biophysical consequences including changing water supplies for human uses as well as ecosystem integrity (Bradley, Vuille, Diaz, & Vergara, 2006), a coupled relationship that has been observed globally (Viviroli, Dürr, Messerli, Meybeck, & Weingartner, 2007). The effects on stream discharge are often region-specific based on climatological differences. There is, however, a widely accepted conceptual model that broadly describes the impact of glacier recession on stream discharge. Introduced by Braun, Weber, and Schulz (2000), the model has been reproduced in a number of studies with each contributing slight variations (Hock et al., 2005, Jansson et al., 2003, Moore et al., 2009, Nolin et al., 2010, Rees and Collins, 2006). More recently conceptualized as “peak water” by Carey et al. (2014), the model characterizes the impact of glacier retreat on surface hydrology by an initial increase in annual stream discharge during the early stages of deglaciation, followed by a decline in stream discharge as the glacier volume becomes too small to sustain flows. In addition to change in mean annual discharge, the hydrological transformation is characterized by an increase in discharge variability (Collins, 2008).
Recently, Baraer et al. (2012) have explored the peak water concept for the specific case of the Cordillera Blanca. Their study suggests that most of the streams in the Cordillera Blanca have undergone a decline in annual and dry season stream discharge since the 1980s that is related to glacier retreat. In addition to decreasing mean annual discharge, an increase in the annual variability of flows was observed. The researchers predicted that when glaciers lose their hydrological influence, yearly stream discharge should stabilize at a slightly lower level than earlier and dry season discharge is anticipated to be lower. Baraer et al. (2012) corroborated conclusions from previous studies focusing on the Cordillera Blanca: Juen, Kaser, and Georges (2007) and Vuille, Kaser, and Juen (2008) showed that reductions in glacier area may result in no major net change in annual total runoff, but seasonal variation should increase. In 2005, Pouyaud et al. (2005) projected a similar stream discharge evolution, but in a different timeframe.
Much of the high mountain landscapes within the Cordillera Blanca are within the boundary of Huascarán National Park where valley peatlands and wet meadows (together referred to herein as wetlands) form patches that are changing spatially. A case study from one valley, Quilcayhuanca, evaluated changes from 2000 to 2011. Over the 11-year interval, wetlands in Quilcayhuanca became smaller and more fragmented (Bury et al., 2013). Other research performed at the scale of the park (all valleys inside the park) found that wetland spatiotemporal change trends are nonlinear with wetlands increasing in area from 1987 to 1999, then decreasing in area from 1999 to 2010 (Polk, 2016). Relevant to this study is that in 1987, wetland occupied 12,402 ha and by 1995, wetland had increased to 13,858 ha, an 11.7% increase over 8 years (Polk, 2016). The wetlands receive input from groundwater, precipitation, and surface water flows, while they also contribute to aquifers and streamflow (Gordon et al., 2015, Maldonado Fonkén, 2015). Outflows from peatlands are seasonally variable, increasing during rainy periods and decreasing in the dry season when evapotranspiration is higher and inputs are lower (Rydin & Jeglum, 2013).
We explored the possible statistical relationships between area of wetlands and shifting hydrologic processes by testing seven hypotheses (Table 1). Because wetlands are spatially proximal to glaciers in the study area (Fig. 1), change in glacier area should be associated with change in wetland area (H1). In most Cordillera Blanca watersheds, proglacial lakes dammed by moraines (or engineered impoundments) store glacier meltwater. Ice loss is associated with increasing lake surface area and volume (Emmer, Vilímek, Klimeš, & Cochachin, 2014), so wetlands could be changing as a statistical function of both glaciers and lakes (H2). Baraer et al. (2012) showed that the Santa River discharge is decreasing while variability is increasing as glaciers recede, so changes in discharge and glacier area could be used jointly to estimate wetland area change (H3). By combining those three hypotheses, changes in glacier area, lake area, and discharge can be used as explanatory variables (H4). Some wetlands in the Peruvian Andes are thought to be groundwater-supported fens (e.g. Cooper et al., 2010), but in the study area wetlands may be maintained only by precipitation (bogs). As such, H5 introduces a lagged precipitation variable in place of lake area to address this possibility, and maintains the relationship between glacier area and discharge; a lagged variable is used because some water held in wetlands may be stored from prior rainy seasons, although the residence time is unknown (Gordon et al., 2015). In H6, we include four variables, changes in glacier area, discharge, lake area, and precipitation, to estimate wetland change. Finally, it may be that changing discharge and precipitation may alone predict wetland change (H7), as glaciers recede, assuming that the influence of meltwater will diminish and precipitation will be the key driver of streamflow (Baraer et al., 2012, Mark and Seltzer, 2003).
Section snippets
Study area
The Cordillera Blanca mountain range is located between 8.5 - 10° S and 77° 00′ - 77° 50′ W and is oriented on a northwest-southeast axis. It is comprised of a series of parallel valleys, each its own watershed in the Santa River basin (Fig. 2). The scale of analysis was the watershed level and we evaluated seven watersheds within the Santa River basin. All sample watersheds are located on the west side of the Cordillera Blanca and drain into the Santa River, which flows to the Pacific Ocean.
Results
Wetland change could be best predicted from changes in glacier area and mean annual discharge, although prior precipitation was also statistically significant (Table 4, Table 5). First, the bivariate regressions were computed to test H1 and to determine whether or not any of the explanatory variables might individually explain change in wetland area. One variable, prior precipitation (PrePrior) was highly statistically significant (p < 0.001, two-tailed) and the within R2 was 0.5571 (Table 4).
Discussion
Connectivity between land change and hydrologic processes are not well understood in tropical mountain regions, but the methods and results presented here show that a spatially explicit econometric analysis demonstrates a promising approach that can be further developed and used by land change scientists and land managers preparing for change associated with glacier recession. We are aware of only one other study located in the tropics that investigated the connections between landscape and
Conclusion
Decadal evaluations of other similar high mountain ecosystems show that wetland area may increase (H. Li et al., 2014, Postigo et al., 2008), decrease (Bury et al., 2013; Z. W.; Li et al., 2015, Nie and Li, 2011), exhibit stasis (Gao, Li, & Brierley, 2012), or both increase and decrease (Yao, Zhao, Gao, Sun, & Li, 2011). More could be done to evaluate effects of hydrologic connectivity (DeFries and Eshleman, 2004, Wohl et al., 2012), especially in light of the possible nonlinear relationships
Acknowledgments
This work was funded by the National Science Foundation under the Coupled Natural and Human Systems Program (“Hydrologic Transformation and Human Resilience to Climate Change in the Peruvian Andes” Award Number 1010381); the National Science Foundation under the Doctoral Dissertation Research Improvement Grant (“Glacier Recession and Implications for High-Elevation Peatlands” Award Number 1333141); the University of Texas Department of Geography and the Graduate School; and the American
References (71)
- et al.
Using atmospherically-corrected Landsat imagery to measure glacier area change in the Cordillera Blanca, Peru from 1987 to 2010
Remote Sensing of Environment
(2014) - et al.
Toward hydro-social modeling: Merging human variables and the social sciences with climate-glacier runoff models (Santa River, Peru)
Journal of Hydrology
(2014) - et al.
Sources and pathways of stream generation in tropical proglacial valleys of the Cordillera Blanca, Peru
Journal of Hydrology
(2015) - et al.
A review of large area monitoring of land cover change using Landsat data
Remote Sensing of Environment
(2012) - et al.
The concept of glacier storage: A review
Journal of Hydrology
(2003) - et al.
Societies consuming nature: A panel study of the ecological footprints of nations, 1960-2003
Social Science Research
(2011) - et al.
Modelling observed and future runoff from a glacierized tropical catchment (Cordillera Blanca, Perú)
Global and Planetary Change
(2007) A review of the modern fluctuations of tropical glaciers
Global and Planetary Change
(1999)- et al.
Shrinkage of the Ruoergai Swamp and changes to landscape connectivity, Qinghai-Tibet Plateau
CATENA
(2015) - et al.
Runoff from tropical alpine grasslands increases with areal extent of wetlands
CATENA
(2015)
Classification and change detection using Landsat TM data: When and how to correct atmospheric effects?
Remote Sensing of Environment
Glacier mass balance variability in the Cordillera Blanca, Peru and its relationship with climate and the large-scale circulation
Global and Planetary Change
Fixed effects regression models
Applied panel data analysis for economic and social surveys
Glacier recession and water resources in Peru's Cordillera Blanca
Journal of Glaciology
Potential impacts of a warming climate on water availability in snow-dominated regions
Nature
Threats to water supplies in the tropical Andes
Science
Consequences of climate change for runoff from Alpine regions
Annals of Glaciology
New geographies of water and climate change in Peru: Coupled natural and social transformations in the Santa River watershed
Annals of the Association of American Geographers
Peatlands and environmental change
Driving forces of tropical deforestation: The role of remote sensing and spatial models
Singapore Journal of Tropical Geography
The environmental consequences of rural and urban population change: An exploratory spatial panel study of forest cover in the southern United States, 2001–2006
Rural Sociology
Intensifying the countryside: A sociological study of cropland lost to the built environment in the United States, 2001–2006
Social Forces
Climatic warming, glacier recession and runoff from Alpine basins after the Little Ice Age maximum
Annals of Glaciology
Assessing the accuracy of remotely sensed data: Principles and practices
Alpine peatlands of the Andes, Cajamarca, Peru
Arctic, Antarctic, and Alpine Research
Quantitative analysis of the driving forces causing declines in marsh wetland landscapes in the Honghe region, northeast China, from 1975 to 2006
Environmental Earth Sciences
Land-use change and hydrologic processes: A major focus for the future
Hydrological Processes
Peatland response to global change
Science
Glacier retreat, lakes development and associated natural hazards in Cordilera Blanca, Peru
Topographic influence on wetland distribution and change in Maduo county, Qinghai-Tibet plateau, China
Journal of Mountain Science
Updated high-resolution grids of monthly climatic observations – the CRU TS3.10 Dataset
International Journal of Climatology
Modelling the response of mountain glacier discharge to climate warming
Cited by (52)
Ecohydrology and ecosystem services of a natural and an artificial bofedal wetland in the central Andes
2022, Science of the Total EnvironmentCitation Excerpt :This is particularly relevant as bofedales have suffered from rapid and dramatic degradation in the last decades, especially as a result of mining, peat harvesting, road construction and overgrazing (Maldonado-Fonkén, 2014; Salvador et al., 2014). Alterations to their hydrological connectivity due to on-going climate change has also been observed (Polk et al., 2017). To address these knowledge gaps and support bofedales management, this study aims to (1) identify the ecohydrological interactions (hydrological processes, soil characteristics, and vegetation) in a bofedal of southern Peru; (2) understand the consequences of these interactions for carbon storage and fodder provision, and (3) compare these characteristics between a natural and an artificial bofedal.
Payment for ecosystem services in Peru: Assessing the socio-ecological dimension of water services in the upper Santa River basin
2022, Ecosystem ServicesCitation Excerpt :As glaciers retreat, they expose metal-rich rock to the atmosphere, accelerating the detachment and entrainment of metals into water bodies (Drenkhan et al., 2015; Grande et al., 2019). Therefore, the capacity for provision and regulation of water services in these ecosystems has also decreased (Polk et al., 2017) and will continue to do so under future warming. At the social level, the Santa River basin is a territory characterised by divisive political dynamics and high levels of competition for the use and control of water resources among its various stakeholders (French, 2015).
Comparing model complexity for glacio-hydrological simulation in the data-scarce Peruvian Andes
2021, Journal of Hydrology: Regional StudiesLearning from indigenous and local knowledge: The deep history of nature-based solutions
2021, Nature-Based Solutions and Water Security: An Action Agenda for the 21st Century