Toward Estimating Climatic Trends in SST. Part III: Systematic Biases

摘要: A method is developed to quantify systematic errors in two types of sea surface temperature (SST) observations: bucket and engine-intake measurements. simple linear model proposed where the SST measured using a cooled or warmed by fraction air–sea difference an engine intake has constant bias. The applied collocated nighttime observations made at moderate wind speeds, allowing effects solar radiation strong vertical gradients upper ocean be neglected. analysis complicated large random all variables used. To estimate coefficients this model, novel type regression, are correlated with each other, introduced. Because uncertainty priori estimates error covariance matrix, Bayesian regression problem developed, maximum likelihood approximations posterior distributions parameters obtained. Results show that change resulting from can detected. suggests may 0.12° 0.02° 0.16° 0.02°C difference. When accounted for, warm bias mid- late 1970s 1980s was found smaller than suggested previous studies, ranging between 0.09° 0.06° 0.18° 0.05°C. For early 1990s SSTs have cold 0.13° 0.07°C.