An All-Weather Land Surface Temperature Product Based on MSG/SEVIRI Observations

João P A Martins,Isabel F Trigo,Carlos Jimenez,Alirio Arboleda,Françoise Gellens-Meulenberghs,Sofia L Ermida,Nicolas Ghilain,Frank-M Göttsche,Folke-S Olesen

doi:10.3390/rs11243044

Abstract

A new all-weather land surface temperature (LST) product derived at the Satellite Application Facility on Land Surface Analysis (LSA-SAF) is presented. It is the first all-weather LST product based on visible and infrared observations combining clear-sky LST retrieved from the Spinning Enhanced Visible and Infrared Imager on Meteosat Second Generation (MSG/SEVIRI) infrared (IR) measurements with LST estimated with a land surface energy balance (EB) model to fill gaps caused by clouds. The EB model solves the surface energy balance mostly using products derived at LSA-SAF. The new product is compared with in situ observations made at 3 dedicated validation stations, and with a microwave (MW)-based LST product derived from Advanced Microwave Scanning Radiometer-Earth Observing System (AMSR-E) measurements. The validation against in-situ LST indicates an accuracy of the new product between -0.8 K and 1.1 K and a precision between 1.0 K and 1.4 K, generally showing a better performance than the MW product. The EB model shows some limitations concerning the representation of the LST diurnal cycle. Comparisons with MW LST generally show higher LST of the new product over desert areas, and lower LST over tropical regions. Several other imagers provide suitable measurements for implementing the proposed methodology, which offers the potential to obtain a global, nearly gap-free LST product.

Highlights

Land surface temperature (LST) translates the response of land surface to environmental factors and constrains the energy exchanges at the land–atmosphere interface [1,2,3]
The all-weather land surface temperature (LST) dataset consists of a combination of LSA-SAF standard LST product derived from thermal IR brightness temperatures, and cloudy sky estimates obtained with the surface energy balance (EB) model
This is the case for most imagers on current geostationary satellites (e.g. the Geostationary Operational Environmental Satellite (GOES)-R operated by the National Oceanic and Atmospheric Administration (NOAA), Himawari operated by the Japan Meteorological Agency, or MSG/SEVIRI Indian Ocean Coverage, operated by the European Organisation for the Exploitation of Meteorological Satellites – EUMETSAT) and polar orbiters (e.g. the Visible Infrared Imaging Radiometer Suite – VIIRS – operated by National Aeronautics and Space Administration (NASA)/NOAA, Moderate Resolution Imaging Spectroradiometer (MODIS) operated by NASA, MetOp operated by EUMETSAT, Sentinel-3 operated by the European Space Agency)

Summary

Introduction

Land surface temperature (LST) translates the response of land surface to environmental factors and constrains the energy exchanges at the land–atmosphere interface [1,2,3] It is an essential variable for computing longwave surface-emitted radiation, as well as for estimating turbulent fluxes of latent and sensible heat; in some regions, the amplitude of the diurnal cycle of LST is strongly correlated to soil moisture [4,5]. The availability of long-term satellite data records is providing unique opportunities to derive climate-related information from satellite-derived LST, its diurnal cycle, and in regions with sparse surface station data coverage. This motivated the inclusion of LST in the list of essential climate variables [7], which underlines its relevance for climate applications. Alternative causes include: 1) a misrepresentation of surface processes and land-atmosphere interactions in surface schemes and 2) inaccuracies in the main forcing variables and surface parameters (e.g. radiative fluxes and soil moisture)

Methods

Results

Discussion

Conclusion