Page Banner

United States Department of Agriculture

Agricultural Research Service

Research Project: INTEGRATION OF CLIMATE VARIABILITY AND FORECASTS INTO RISK-BASED MANAGEMENT TOOLS FOR AGRICULTURE PRODUCTION AND RESOURCE CONSERVATION

Location: Great Plains Agroclimate and Natural Resources Research Unit

Title: Evaluating PRISM precipitation grid data as possible surrogates for station data at four sites in Oklahoma

Authors
item Schneider, Jeanne
item Ford Jr, Donald

Submitted to: Oklahoma Academy of Science Proceedings
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: September 29, 2010
Publication Date: May 1, 2011
Citation: Schneider, J.M., Ford Jr, D.L. 2011. Evaluating PRISM precipitation grid data as possible surrogates for station data at four sites in Oklahoma. Oklahoma Academy of Science Proceedings. 90:77-88.

Interpretive Summary: The development of climate-sensitive decision support for agriculture or water resource management requires long time series of monthly precipitation for specific locations. Archived station data for many locations is available, but time continuity, quality, and spatial coverage of station data remain significant issues. One possible alternative for station data are continuous, gridded monthly data produced by the PRISM Climate Group, with each grid cell roughly 4 km per side. The PRISM monthly precipitation data is evaluated against station data for four sites in Oklahoma, for possible use whenever station data is unavailable or insufficient. The station and PRISM data are found to be very similar in two key respects (specifically 30-year monthly means and certain characteristics of probability density functions), but significantly different in variance. The difference in variance is attributed to situations where precipitation varied by significant amounts over short distances, with the gridded PRISM data “smearing” the heavy rainfall over locations that received lower amounts. As a result, PRISM precipitation data would be suitable for downscaling seasonal climate forecasts or other analyses that require knowledge of the mean and central shape of the local probability density function, but not for specifying variance as input for weather generators.

Technical Abstract: The development of climate-sensitive decision support for agriculture or water resource management requires long time series of monthly precipitation for specific locations. Archived station data for many locations is available, but time continuity, quality, and spatial coverage of station data remain significant issues. One possible alternative for station data are continuous, gridded monthly data produced by the PRISM Climate Group, with each grid cell roughly 4 km per side. The PRISM monthly precipitation data is evaluated against station data for four sites in Oklahoma, for possible use whenever station data is unavailable or insufficient. The station and PRISM data are found to be very similar in two key respects (specifically 30-year monthly means and certain characteristics of probability density functions), but significantly different in variance. The difference in variance is attributed to situations where precipitation varied by significant amounts over short distances, with the gridded PRISM data “smearing” the heavy rainfall over locations that received lower amounts. As a result, PRISM precipitation data would be suitable for downscaling seasonal climate forecasts or other analyses that require knowledge of the mean and central shape of the local probability density function, but not for specifying variance as input for weather generators.

Last Modified: 9/22/2014
Footer Content Back to Top of Page