Correction of AVHRR Pathfinder SST data for volcanic aerosol effects using ATSR SSTs and TOMS aerosol optical depth

The Pathfinder sea surface temperature (SST) data set, from the Advanced Very High Resolution Radiometer (AVHRR), is an important climate data record for SST due to the longevity of the sensor series of potentially 38 years (from 1981 to 2020). However high amounts of aerosol in the atmosphere, such as after a major volcanic eruption, lead to a cool bias on the sea surface temperatures retrieved from single view sensors such as the AVHRR. The Along Track Scanning Radiometer (ATSR) series of instruments provide a better climate-quality SST data record but are only available from 1991. The purpose of the work presented in this paper was to develop and test a method of reducing the bias in AVHRR Pathfinder SST during periods of high stratospheric volcanic aerosol. In this paper we take the ATSR record to be the reference SST for calculating the bias on the AVHRR Pathfinder SSTs during the Pinatubo eruption aerosol period (1991–1992). The analysis gives a bias of approximately − 0.1 K for AVHRR Pathfinder SSTs in periods with low atmospheric aerosol and a much larger bias (globally − 0.44 K in September to November 1991) in periods affected by volcanic aerosol. This paper shows the bias can largely be corrected using a multi-linear regression with aerosol optical depths from the Total Ozone Mapping Spectrometer (TOMS) during periods of high volcanic aerosol. The AVHRR Pathfinder minus ATSR SST bias, for grid-boxes where the correction with TOMS can be applied, is reduced from − 0.53 K to − 0.12 K in a three month period heavily affected by the stratospheric aerosol. This bias corrected value is close to the bias during low atmospheric aerosol loading periods. A three-way error analysis technique was used to evaluate the random retrieval uncertainty as measured by standard deviation on the SST datasets before and after bias correction. AVHRR Pathfinder SST error in the months after the Pinatubo eruption fell from 0.53 K uncorrected to 0.20 K after correction which shows the correction technique has value.