Biases in dew point temperature are generally associated with contamination of the wet bulb thermometer by salt or insufficient moistening of the wet bulb thermometer wick. These instrumentation problems lead to a systematically higher value of the specific humidity. On average, the dew point temperature is biased about 0.5° C (Isemer and Hasse 1987 and references within). Since this bias is due mainly to human errors for which no record is available, it is very difficult to devise a correction.
Kent et al. (1993a) have found that dew point temperature measured by psychrometers were on average lower than those reported from ships using screens. By considering the many factors that might have contributed to this discrepancy, Kent et al. (1993a) concluded that psychrometer observations were more reliable and a simple regression equation was devised to correct screen-measured dew point temperatures. Because of inadequate metadata in COADS/CMR-5 (there is no indicator of the method of measurement for dew point temperature) this correction could not be implemented at the present time.