Search

Thermal sensors onboard Landsat satellites have been underutilized due to the lack of consistent and accurate methodologies for retrieving the land surface temperature (LST) at global scales over all land cover types. We present an operational algorithm for generating Landsat LST consistently for all sensors that will be implemented by the United States Geological Survey/The National Aeronautics and Space Administration and made available at the Land Processes Distributed Active Archive Center. The LST algorithm involves three steps. The observed thermal radiance is atmospherically corrected using a radiative transfer model and reanalysis data. The Advanced Spaceborne Thermal Emission and Reflection Radiometer Global Emissivity Data Set version 3 is spectrally adjusted and then modified to account for vegetation phenology and snow cover using Landsat visible-shortwave infrared data. The LST is retrieved by inverting the atmospherically and emissivity corrected Landsat radiances with a lookup-table approach. Landsat-derived emissivities were validated at two pseudoinvariant sand dune sites within an average absolute error of 0.54% when compared with laboratory measurements. The Landsat LST retrievals were validated with in situ observations from four surface radiation budget network (SURFRAD) sites, and two inland water bodies (Salton Sea and Lake Tahoe) in the USA. The LST retrievals for Landsat 5 and 7 had a mean bias (root mean square error) of 0.7 K (2.2 K) and 0.9 K (2.3 K) for the SURFRAD sites, and −0.3 K (0.6 K) and 0.4 K (0.7 K) for the inland water bodies, respectively. The operational algorithm will provide a consistent LST record from four decades of historical Landsat thermal data enabling the long-term monitoring of temperature and trends, land cover and land use changes, and improved utilization in models. [ABSTRACT FROM AUTHOR]

Land surface temperature (LST) is a key climate variable for studying the energy and water balance of the earth surface and monitoring the effects of climate change. This paper presents a physics-based temperature emissivity separation (TES) algorithm for the simultaneous retrieval of LST and emissivity (LST&E) from the thermal infrared bands of the Suomi National Polar-Orbiting Partnership’s Visible Infrared Imaging Radiometer Suite (VIIRS) payload. The new VIIRS LST&E product (VNP21) was developed to provide continuity with the Moderate-Resolution Imaging Spectroradiometer (MODIS) equivalent LST&E product (MxD21) product, which is available in Collection 6, and to address inconsistencies between the current MODIS and VIIRS split-window LST products. The TES algorithm uses full radiative transfer simulations to isolate the surface emitted radiance, and an emissivity calibration curve based on the variability in the surface radiance data to dynamically retrieve both LST and spectral emissivity. Furthermore, an improved water vapor scaling model was implemented to improve the accuracy and stability of the atmospheric correction for conditions with high atmospheric water vapor content. An independent assessment of the VIIRS LST retrievals was performed against in situ LST measurements over two dedicated validation sites at Lake Tahoe and Salton Sea in the Southwestern USA, while the VIIRS emissivity retrievals were evaluated with the latest ASTER Global Emissivity Dataset Version 3 (GEDv3). The bias and root-mean-square error (RMSE) in retrieved VIIRS LST were 0.50 and 1.40 K, respectively for the two sites combined, while mean emissivity differences between VIIRS and ASTER GEDv3 were 0.2%, 0.1%, and 0.3% for bands M14 ( 8.55~\mu \textm ), M15 ( 10.76~\mu \textm ), and M16 ( 12.01~\mu \textm ), respectively, with an RMSE of 1%. We further demonstrate close agreement between the MODIS and VIIRS TES algorithm LST products to within ~0.3 K difference, as opposed to the current MODIS and VIIRS split window products, which had an average difference of 3 K. [ABSTRACT FROM PUBLISHER]