Characterizing regional precipitation-driven lake area change in Mongolia

Lake area is an important indicator for climate change and its relationship with climatic factors is critical for understanding the mechanisms that control lake level changes. In this study, lake area changes and their relations to precipitation were investigated using multi-temporal Landsat Thermatic Mapper (TM) and Enhanced Thermatic Mapper plus (ETM+) images collected from 10 different regions of Mongolia since the late 1980s. A linear-regression analysis was applied to examine the relationship between precipitation and lake area change for each region and across different regions of Mongolia. The relationships were interpreted in terms of regional climate regime and hydromorphological characteristics. A total of 165 lakes with areas greater than 10 hm(2) were identified from the Landsat images, which were aggregated for each region to estimate the regional lake area. Temporal lake area variability was larger in the Gobi regions, where small lakes are densely distributed. The regression analyses indicated that the regional patterns of precipitation-driven lake area changes varied considerably (R-2=0.028-0.950), depending on regional climate regime and hydromorphological characteristics. Generally, the lake area change in the hot-and-dry Gobi regions showed higher correlations with precipitation change. The precedent two-month precipitation was the best determining factor of lake area change across Mongolia. Our results indicate the usefulness of regression analysis based on satellite-derived multi-temporal lake area data to identify regions where factors other than precipitation might play important roles in determining lake area change.