MULTI-ANNUAL CHANGES OF GLACIAL LAKES IN PERU AND THEIR LINKS TO CLIMATE CHANGE AND GLOFS

Abstract:

The effect of climate change is causing a profound reduction in glacier’s coverage in the Cordilleras of the Peruvian Andes, giving rise to the development and growth of proglacial lakes. These glacial lakes can represent a significant hazard in the form of Glacier Lake Outburst Floods (GLOF) that have the potential to impact vulnerable populations and infrastructure located downstream of glaciated areas. In this study, we present a long-term spatio-temporal analysis of glacial lakes across the Peruvian Andes and address possible connections with GLOF events and climate change. Accessed and processed using Google Earth Engine and QGIS, this analysis was conducted using Landsat 5, 7 and 8 satellite data (Surface Reflectance Tier 1) acquired between 1984 to 2019. We detected and estimated glacier lake surface area changes using a robust semi-automated method which coupled empirical decision trees, sub-pixel analysis, spectral and topographic indices with many temporal and spatial filters to improve the quality of the glacial lake classifications. Using this classification as a result, we were able to categorize the glacial lakes based on their surface change dynamics over the past three decades. Results highlight a distinct spatial variability in glacial lake changes across the 20 Cordilleras of the Peruvian Andes. Overall, 12 Cordilleras experienced an increase in glacial lake area during the observation period (ranging from increases of 1% to 38%), whilst the remaining 8 experienced a reduction in area (ranging from reductions of -2% to -57%). A multi-temporal validation of the semi-automated glacier lake classification method presented, utilising manually digitized data, revealed an overall glacial lake mapping accuracy of >80% across the study area.