Using synoptic surface and geopotential height fields for generating grid-scale precipitation

Abstract

Grid-scale precipitation data were derived from synoptic-scale fields of 50 kPa geopotentials by a simple method that relates synoptic-scale circulation to local climate. The study addressed the concerns that: (i) different sizes of grids might affect the results; (ii) different gridpoint spacing might affect the results; (iii) days that remain unclassified might be important to precipitation delivery. Three different grids were used. Grid A is 45 points with a 3.7°×3.7° spacing with limited spatial coverage; Grid B is 117 points on a 5°×5° grid which covered most of North America and the north-eastern Pacific Ocean; and Grid C is 35 points on a 10°×10° grid with the same area as Grid B. The Kirchhofer technique was applied, and generated 20 (Grid A), 18 (Grid B), and 14 (Grid C) different synoptic types from a 10-year sample data set (1960–1969). Comparisons between simulated and observed precipitation for 1970–1989 showed that the method produced compatible precipitation totals regardless of which grid was employed, but the simulated variability was too low. Grid B gave the closest agreement between simulated and observed precipitation, with Grids C and A being ranked second and third, respectively. The spatial coverage of the grid was therefore deemed to be more important than the number of grid points within it. Similar results were found when comparing observed precipitation with that generated using general circulation model (GCM) geopotential height data. Precipitation regimes generated from the GCM suggest that little change in the synoptic control of precipitation would arise from a doubling of atmospheric CO2. Copyright © 1999 Royal Meteorological Society