Land Surface Temperature retrieval from Sentinel 2 and 3 data: the SEN4LST project
Sobrino, Jose Antonio1; Jiménez-Muñoz, Juan Carlos1; Ruescas, A. B.2; Danne, Olaf2; Brockmann, Carsten2; Ghent, D.3; Remedios, J3; North, P.R.J.4; Davies, W. H.4; Mitraka, Z5; Mathieu, Pierre Philippe5; Merchant, C6
1Universidad de Valencia, SPAIN; 2Brockmann Consult, GERMANY; 3University of Leicester, UNITED KINGDOM; 4University of Swansea, UNITED KINGDOM; 5ESA, ITALY; 6University of Edinbourgh, UNITED KINGDOM
Land surface Temperature (LST) is one of the key parameters in the physics of land-surface processes on regional and global scales, combining the results of all surface-atmosphere interactions and energy fluxes between the surface and the atmosphere. An adequate characterization of LST distribution and its temporal evolution requires measurements with detailed spatial and temporal frequencies. With the advent of the Sentinel 2 (S2) and 3 (S3) series of satellites a unique opportunity exists to go beyond the current state of the art of single instrument algorithms. The Synergistic Use of The Sentinel Missions For Estimating And Monitoring Land Surface Temperature (SEN4LST) project aims at developing techniques to fully utilize synergy between S2 and S3 instruments in order to improve LST retrievals. In the framework of the SEN4LST project, three LST retrieval algorithms were proposed using the thermal infrared bands of the Sea and Land Surface Temperature Retrieval (SLSTR) instrument on board the S3 platform: split-window (SW), dual-angle (DA) and a combined algorithm using both split-window and dual-angle techniques (SW-DA). One of the objectives of the project is to select the best algorithm to generate LST products from the synergy between S2/S3 instruments. In this sense, validation is a critical step in the selection process for the best performing candidate algorithm. Accurate atmospheric correction and emissivity estimation is also a crucial step in the LST retrieval. During the SEN4LST project, a simulated dataset was constructed using synthetic images under different atmospheric conditions, as well as a match-up database of in situ observations and MERIS/AATSR data. Atmospheric correction (AOD and at-surface reflectance retrieval) and LST algorithms were tested using both the simulated datasets and the real pairs of MERIS/AATSR imagery. The results show an accurate retrieval of both AOD and at-surface reflectance, and a better performance of the SW algorithm when compared to the DA or SW-DA algorithm. However, over homogeneous areas, the DA algorithm provided the lowest errors on the LST retrieval. Un-mixing techniques were also developed to explore the synergy between S2 and S3 data in order to improve the emissivity estimation.