Estimating lake temperature profile and evaporation losses by leveraging MODIS LST data

Global lake evaporation is a critical component of the terrestrial water cycle. Accurate quantification of lake evaporation dynamics is of high importance for understanding lake energy budgets, land-atmosphere interactions, as well as regional water availability. However, the accurate quantification of lake evaporation has been hindered by the complexity involved with addressing the heat storage of water bodies. In this study, a new model—the Lake Temperature and Evaporation Model (LTEM)—was developed to simulate lake water temperature profiles, which were then used to calculate heat storage changes and evaporation rates. Inputs for the LTEM include the meteorological and bathymetric data, as well as the Moderate Resolution Imaging Spectroradiometer (MODIS) water surface temperature (WST)—which is the land surface temperature (LST) over water. The MODIS WST was leveraged to constrain the hydrodynamic simulations. Model results over 11 lakes around the world show robust performance of LTEM. The long term average temperature biases range from -0.5 °C to 0.5 °C, and the evaporation rate biases range from -0.19 mm/day to 0.28 mm/day. In particular, it is found that LTEM significantly improves the simulation of the seasonality of lake evaporation rates. The validation results suggest that the averaged coefficient of determination (R²) for the evaporation rate is 0.84, which is 0.28 higher than that obtained when the conventional Penman equation (without heat storage) is used. The volumetric evaporation time series was then calculated as a product of the monthly evaporation rate and lake surface area (derived from MODIS near-infrared image classifications). This study provides an end-to-end framework for quantifying volumetric evaporation for the world's lakes and reservoirs. It also provides the capability to investigate the thermal dynamics of lake systems, and thus can benefit the various water resources applications across scales.