Biases in MOS Forecasts of Maximum and Minimum Temperatures at Phoenix, Arizona

Abstract

The biases of objective forecasts of maximum and minimum temperature for Phoenix are evaluated relative to the observed temperatures. The temperature forecasts were calculated from regression equations that had been derived from model output statistics (MOS). During the analysis period, from October 1982 through September 1984, significant cold biases of ?1.5?2.5°F (0.8?1.4°C), were determined for the MOS minimum temperature forecast at 24, 36 and 48 h. The maximum temperature forecasts had warm biases <1.0°F (0.6°C) that were significant only at 24 h. The minimum and maximum temperatures from the most recent 30-year normals (1951?80) were, respectively, ?5°F (3°C) and 1°F (0.6°C) colder than the observed temperatures during the analysis period. The difference between the climatological and observed minimum temperatures is significant at the 1% level and suggests that a local or regional change in weather conditions may be an important factor in explaining the MOS biases.