Evaluation of the multichannel sea surface temperature algorithm along the Brazilian coast

摘要

The accuracy to which sea surface temperature (SST) can be estimated using satellite infrared radiometers is limited mainly by atmospheric contamination. This work reports an investigation on the global coefficients of the Multichannel Sea Surface Temperature (MCSST) algorithm along the Brazilian coast (8/spl deg/N-36/spl deg/S). MCSST data were evaluated for the period November 1981-December 1986 in conjunction with temperature data from the Comprehensive Ocean-Atmosphere Data Set (COADS). Monthly statistics of 2/spl deg/ bin between satellite and ship SST differences ranged from -3.44/spl deg/ to 2.84/spl deg/C and mean differences from -0.18/spl deg/ to 9.47/spl deg/C (standard deviation from 0.52/spl deg/ to 0.65/spl deg/C). Interannual oscillations of the SST differences were observed along the western Atlantic Ocean and can be partly explained by the 1982-1983 El Nino event.