Temperature and Emissivity Separation from MODIS Multispectral TIR Data Using the UVEG-TES Algorithm: Development and Validation within ESA's Climate Change Initiative

Extended abstract / summary of an ongoing research article prepared for submission to a peer-reviewed scientific journal. Land Surface Temperature (LST) and Land Surface Emissivity (LSE) are fundamental to terrestrial energy balance and climate change studies. Within the European Space Agency’s Climate Change Initiative (ESA CCI), this study presents an alternative dataset to MODIS products (MOD21/MYD21) designed to provide consistent Essential Climate Variables (ECVs) across global and local scales, including urban and agricultural settings. We implemented UVEG-TES, a workflow that applies the Temperature and Emissivity Separation (TES) algorithm alongside radiative transfer modelling. Atmospheric effects were corrected using ECMWF profiles and the RTTOV model. The framework combines standard TES modules (NEM and MMD) with the NDVI Thresholds Method (NDVI-THM) to refine emissivity over vegetated surfaces, complemented by a Split-Window approach to constrain band 32 under noisy atmospheric conditions. Data were processed on the JASMIN supercomputing facility (UK) via CEDA. Validation comprised global cross-comparisons against NASA’s MOD21/MYD21 and in situ assessments at Cabo de Gata and Fuente Duque (Spain). Globally, UVEG-TES showed close agreement with operational products (LST RMSE: 1.8 K by day, 1.0 K by night; LSE RMSE: 0.012–0.016). In situ results showed comparable daytime performance to MODIS (RMSE: 2.2–2.7 K), whilst reducing nocturnal bias and improving nighttime accuracy to an RMSE of 1.1–1.5 K.