Validating and Mapping Surface Water Temperatures in Lake Taihu: Results From MODIS Land Surface Temperature Products

Abstract

Data from four in situ high-frequency monitoring sites in Lake Taihu, China were used to 1) assess the accuracy of MODIS land surface temperature (LST) products; 2) characterize the spatial and temporal variability in lake surface water temperature (LSWT) with Aqua MODIS thermal-infrared imagery; and 3) explore the causes of these variations. The validation showed that MODIS-derived LSWTs and in situ water temperatures were significantly correlated, with a coefficient of determination higher than 0.96 and a root mean square error between 1.2°C and 1.8°C. These results indicate MODIS LST products can be used to assess the spatial and temporal thermal behavior of Lake Taihu. A spatial analysis of daytime LSWT showed different thermal zones along the lake due to the differential heat storage capacity between the deep and shallow regions of the lake, such as the littoral and East Lake Taihu. In contrast, the nighttime LSWT showed little variation, and there was a uniform surface water temperature. Following a change in solar radiation, the annual cycle of the average LSWT (with a nearly 24°C range) started to increase in January and peaked in July. In Lake Taihu, the mean nighttime LSWT from July to September rapidly warmed over the period of 2002-2013 at an average rate of 0.059 ± 0.053°C/year. Our study promotes the use of MODIS LST products in limnology through a validation using a large amount of real-time synchronous water temperature data based on high-frequency observations, which were seldom used in previous studies.