This dataset contains land surface temperatures (LSTs) and their uncertainty estimates from the Advanced Very High Resolution Radiometer-3 (AVHRR-3) on NOAA-19. Satellite land surface temperatures are skin temperatures, which means, for example, the temperature of the ground surface in bare soil areas, the temperature of the canopy over forests, and a mix of the soil and leaf temperature over sparse vegetation. The skin temperature is an important variable when considering surface fluxes of, for instance, heat and water.

Separate files are provided for temperatures retrieved from descending and ascending orbits. Descending and ascending equator crossing times were originally 02:00 and 14:00 local solar time but have been allowed to drift.
This product includes a time correction which can be added to the observed LST to correct to a consistent overpass time.

Per pixel uncertainty estimates are given in two forms, first, an estimate of the total uncertainty for the pixel and second, a breakdown of the uncertainty into components by correlation length. Also provided in the files, on a per pixel basis, are the observation time, the satellite viewing and solar geometry angles, a quality flag, and land cover class.

The dataset coverage is global over the land surface. LSTs are provided on a global equal angle grid at a resolution of 0.05° longitude and 0.05° latitude. AVHRR achieves full Earth coverage twice per day. LSTs are not produced where clouds are present since under these circumstances the IR radiometer observes the cloud top which is usually much colder than the surface.

Dataset coverage starts on 22nd February 2009 and continues to 17th September 2020. There are minor interruptions (1-10 days) during satellite/instrument maintenance periods or instrument anomalies.

The dataset was produced by the University of Leicester (UoL) and LSTs were retrieved using a two channel Generalised Split Window retrieval algorithm and data were processed in the UoL processing chain.