A simple method of interpolating sea-surface temperature (SST) and its measurement method is applied to western North Pacific ship and buoy observations taken between 1970 and 2000. Comparisons are made between the resulting quasi-daily in situ analyses and more inclusive analyses based mainly on satellite infrared observations. In terms of analysis differences, spatial correlation, and temporal spectra, the in situ analyses are found to be of moderate quality and are taken to be appropriate for composite diagnoses. This use is illustrated in a novel comparison of short-time-scale SST changes observed during the passage of two groups of midlatitude cold-season cyclones.
Transient cooling is found to be preferentially associated with the stronger of the two cyclone groups. The significance of this cooling is quantified and the possibility of oceanic mixed-layer cooling and heat loss from bucket observations is explored. Although the bias in bucket observations is confirmed using composites constructed without such observations, the signature of transient cooling is still apparent. (Independent bucket observations also permit composite analysis errors to be estimated.) Engine intake observations are then identified as a primary source of both SST information in general and of the transient cooling signal in particular. A related bias in this type of observation is proposed. Greater sophistication in the analysis of in situ observations is also briefly discussed. Copyright © 2010 Royal Meteorological Society