UCL Discovery
UCL home » Library Services » Electronic resources » UCL Discovery

Eliminating bias in satellite retrievals of sea surface temperature

Merchant, Christopher John; (1998) Eliminating bias in satellite retrievals of sea surface temperature. Doctoral thesis (Ph.D), UCL (University College London). Green open access

[thumbnail of Eliminating_bias_in_satellite_.pdf]
Preview
Text
Eliminating_bias_in_satellite_.pdf

Download (10MB) | Preview

Abstract

Sea surface temperature (SST) is a critical parameter for climate research, and needs to be measured with an absolute accuracy of ~0.3 K (average over ~100 km scale on a weekly to monthly time scale) and with a long term stability of 0.1 K per decade. These stringent requirements present a formidable challenge to satellite based SST measurement. The most promising satellite radiometer is the ATSR (and successors), but bias and spurious trends have arisen in the ATSR SST retrieval process. Eliminating such retrieval bias is the focus of this thesis. SSTs derived from the ATSR using the prelaunch retrieval scheme are biased by up to -1.5 K by stratospheric aerosol from the eruption of Mount Pinatubo shortly before launch. An "aerosol-robust" retrieval scheme is derived which has no detectable aerosol- related bias. Another bias of up to 0.5 K arising from a deficiency of the radiative transfer model used to develop the prelaunch retrieval scheme is resolved by implementing an updated parameterisation of water vapour continuum absorption. The new SSTs are shown to have an accuracy better than 0.3 K (error in a single retrieval over a -20 km spatial scale) and to be robust to aerosol effects, by a validation exercise against buoys measuring SST in situ. The validation data consist of 620 satellite-buoy coincidences in the tropical Pacific between September 1991 and May 1992, a region and period associated with high loadings of stratospheric aerosol and tropospheric water vapour. This is the first validation exercise to correct for the effects of the difference between bulk SSTs (measured by buoys) and skin SSTs (measured radiometrically). The factor now limiting accuracy is residual cloud contamination. The new retrieval scheme has been adopted for the reprocessing of all archived ATSR data to SST.

Type: Thesis (Doctoral)
Qualification: Ph.D
Title: Eliminating bias in satellite retrievals of sea surface temperature
Open access status: An open access version is available from UCL Discovery
Language: English
Additional information: Thesis digitised by ProQuest.
Keywords: Earth sciences; Sea surface temperature
URI: https://discovery.ucl.ac.uk/id/eprint/10097999
Downloads since deposit
33Downloads
Download activity - last month
Download activity - last 12 months
Downloads by country - last 12 months

Archive Staff Only

View Item View Item