Enterprise LST Algorithm Development and Its Evaluation on NOAA-20 VIIRS Data

Land Surface Temperature (LST) is one of Essential Climate Variables, which has been widely used in varieties of fields such as evapotranspiration, irrigation and hydrological cycle particularly agricultural drought monitoring and urban heat island monitoring etc. Satellite LSTs have been routinely produced for decades from a variety of polar-orbiting and geostationary satellites. For producing seamless LST climate data record from these programs, an enterprise LST algorithm is developed for consistency of the LST products from different meteorological satellite missions including geostationary orbit (e.g. ABI on GOES-16) and Low Earth orbit missions. Consistent LST products from different satellite missions is beneficial to obtain high quality LST product validation.

The enterprise LST algorithm employs the split window technique and uses the emissivity explicitly in its formula. In this study, focuses are on the enterprise LST algorithm design, development and its evaluations on NOAA-20 VIIRS data since January 5th, 2018. A theoretical analysis was conducted for the algorithm uncertainty attributed to the regression model, sensor noise, emissivity uncertainty and water vapor uncertainty. A preliminary validation is performed through the comparison with in-situ LST observations from Surface Radiation Budget Network, Baseline Surface Radiation Network and Global Monitoring Division stations. Further, a cross satellite data comparison analysis has been performed for evaluating NOAA-20 LST products using the MODIS LST data, SEVIRI LST data and GOES-16 LST data.