Spatial and Elevational Variations of Summer Rainfall in the Southwestern United States

Abstract

This study examines the spatial variability of mean monthly summer rainfall in the southwestern United States, with special attention given to the effect elevation. Rain gauge data from a consistent 60-yr period show that mean rainfall increases linearly with elevation within a local area. A simple model (rain = normalized rainfall as a function of latitude and longitude + elevation coefficient ? elevation) explains a large part of the spatial variability of mean rainfall. The rainfall model (the MSWR model) and digital elevation data were used to produce a 1° ? 1° gridded rainfall climatology for July, August, and September. Regional rainfall estimated with this model is 9.3% higher than an estimate based on arithmetic averaging of gauge data over 2° ? 2° areas. For individual 2° ? 2° cells, the difference between model rainfall and the arithmetic mean of gauge rainfall ranged from ?250% to +41%. The MSWR model was used to remove orographic effects from regional rainfall fields. When rainfall is normalized to sea level, two rainfall maximums emerge: one in south-central Arizona associated with the Mexican monsoon maximum and one in southeastern New Mexico associated with the Gulf of Mexico. Detrended block kriging (using the MSWR model as an estimate of the long-term trend) and monthly rain gauge data were used to produce unbiased areal rainfall estimates that were compared to 1° ? 1° satellite-based rainfall estimates. On a month-by-month basis, there were large differences between the two estimates, although the comparison improved after temporal averaging.