Evaluation of DisTRAD and TsHARP Downscaling Methods to Increase the Spatial Resolution of MODIS Thermal Images

Land surface temperature (LST) is a key boundary condition in many ground-based modeling schemes based on remote sensing. Previous literature has shown that LST products from satellite imagery can be used to detect land surface changes, including urbanization, deforestation and desertification, which can improve our ability to monitor surface changes continuously. The objective of the present study was to evaluate the results of DisTRAD and TsHARP thermal sharpening methods to downscale the spatial resolution of MODIS LST from 1000 m to 250 m.

The research method in the present article is applied in terms of purpose and based on correlation relations in terms of method of work.

The performance of DisTRAD and TsHARP thermal downscaling methods were evaluated by the Root Mean Square Error (RMSE) and the Mean Bias Error (MBE) between the downscaled and original LSTs. Statistical analysis showed that the RMSE between the downscaled images of DisTRAD and TsHARP methods with the original LST (1000 m (terra)) for 3 May 2019 were found to be 1.77 ° C and 1.7 ° C, respectively, whereas the R2 were found to be about 53% and for 17 October 2019, the RMSE were found to be 2.44 ° C and 2.38 ° C respectively, whereas the R2 were found to be about 85%.

The study of the results of Terra and Aqua satellites generally shows the superiority of Terra satellite results over Aqua. The main reason could be the different passage times of the satellites from the study area. Since that changes in soil moisture and water body such as the Karun River are common sources of error, so the use of these methods is recommended only in areas without excessive changes in moisture.