Gridded Percent of Normal Precipitation Mapping for the Arkansas and Red River Basins
John Schmidt and Bill Lawrence
The Arkansas-Red Basin River Forecast Center (ABRFC) is tasked with monitoring the hydrometeorological state of the Arkansas and Red rivers from their headwaters to Pine Bluff, Arkansas and Fulton, Arkansas, respectively. These 200,000 plus square miles of terrain range from the 14,000 ft peaks of the Continental Divide in Colorado through the Southern Plains toward the Mississippi River Valley. The climate across this region is as diverse as the topography. Specifically, annual precipitation values range from near 11 inches in southeastern Colorado to over 60 inches in the higher elevations of western Arkansas, just 550 miles away.
Since July 1994, the ABRFC has utilized Weather Surveillance Radar-88 Doppler (WSR-88D) and observed precipitation amounts to create its gridded precipitation estimates. These estimates are meticulously quality-controlled and are in turn used as input to the ABRFC's operational hydrologic forecast model. In the summer of 1999, the ABRFC received gridded Parameter-Elevation Regressions on Independent Slopes Model (PRISM) climate data from a cooperative venture between Oregon State University and the United States Department of Agriculture-Natural Resources Conservation Service (NRCS). These detailed climate data were compared to accumulated operational precipitation maps to generate monthly percent-of-normal precipitation maps on a 4x4-km grid.
PRISM Climate Data
Many of the current climatology maps in use are derived solely from point data, with some sort of interpolation scheme used between observation stations. Unfortunately, the network of observing stations is not dense enough to accurately capture the localized variability of precipitation distribution, especially in mountainous areas. PRISM data uses National Oceanic and Atmospheric Administration Cooperative Stations and NRCS SNOTEL station data from 1961-1990 and a digital elevation model (DEM) along with a set of rules, decisions and calculations to simulate the thought process of a climatologist in generating its spatial climate maps. These rules account for increased precipitation with elevation, rain shadows and coastal locations. The farther a given location is from an observation site, the more these rules influence a given climate parameter.
ABRFC Precipitation Fields
In 1993, the ABRFC began using Stage III, a National Weather Service (NWS), Office of Hydrology (OH) software package, to merge WSR-88D precipitation estimates with observed data on an hourly basis. When accumulated, the resultant fields were often too low when compared to 24-hour cooperative rainfall observations. Additionally, the graphical user interface had some shortcomings that made operational use cumbersome. In 1996, the ABRFC, in cooperation with the Corps of Engineers, Tulsa District, developed Process 1 (P1). Since December 1996, the ABRFC has used P1 almost exclusively for its operational precipitation processing. This geographic method optimizes the accuracy of physical measurements while maintaining the variability of intensity that the WSR-88D estimates provide. At gage locations, the gridded value is set to the observed value and a bias is calculated. Between gage locations, a double linear interpolation technique is applied to a triangulated irregular network (TIN) to vary the bias for each grid. The result is a unique bias for each 4x4-km grid. West of the front range of Colorado's Rocky Mountains, beam blockage of the radar forces precipitation estimates to be gage only, losing the spatial variability of precipitation that the radar offers. Precipitation fields created after December 1996 are considered to be more accurate, as they employ a full coverage from the WSR-88D network and are processed using P1.
Calculation of Percent of Normal Maps
The most important factor to account for when dealing with spatial data is projection. PRISM data are delivered in the Albers Equal Area projection, whereas ABRFC operational precipitation fields are created in the Hydrologic Rainfall Analysis Project (HRAP) projection, a type of polar stereographic projection. Unfortunately, the HRAP isn't a widely used projection and therefore conversion routines do not exist in ArcView, a geographic information system (GIS) software package. The NWS-OH supplied the ABRFC with an Avenue script to execute the conversion. A center point file of each HRAP grid was created for the ABRFC area in a geographic projection and then overlaid on top of the gridded PRISM data. Each of the HRAP grids was then assigned a value based on the PRISM grid with which it coincided. Using this procedure, monthly precipitation normal maps were created. Having the PRISM data in HRAP coordinates allows easy calculations of percent-of-normal precipitation values, derived from the monthly P1 precipitation fields and the corresponding monthly mean precipitation data. Images of these percents-of-normal are available in durations of the last 10 days, last 30 days, last 60 days, last 90 days, last 180 days, current month and current year, as well as every month and year since July 1994. These images can be accessed at http://info.abrfc.noaa.gov/longterm.html.
Results and Conclusions
High-resolution PRISM data are the most detailed, highest-quality spatial climate datasets currently available. Additionally, P1 offers the best estimates of gridded precipitation fields currently available in an operational environment. These two enhanced products, when compared to one another, offer detailed insight into the distribution of rainfall across the ABRFC on a 4x4-km scale. Using these products, it is possible to capture localized departures that are often lost when using regionalized climate and drought maps. (See attached: fig. 1,fig. 2 and fig. 3 for examples of ninety-day accumulated precipitation, ninety-day accumulated PRISM normals and the ninety-day percent of normal maps, respectively).
Future uses of PRISM data and these percent of normal precipitation products at ABRFC might include the ability to adjust the state variables of the ABRFC's operational hydrologic forecast model in real-time. Additionally, a volumetric analysis of varying precipitation processing methods is in progress. This analysis could lead to improvements in the calibration of the ABRFC's operational hydrologic forecast model.
The authors would like to thank James Paul (senior hydrologist, ABRFC) for his assistance in converting precipitation fields into tabular data. Thanks also to Seann Reed (research hydrologist, NWS-OH, Hydrologic Research Lab) for writing and supplying the authors with the ArcView Avenue scripts and programs necessary to perform the HRAP to geographic projection conversion.
Hudlow, M.D., "Modern Era of Rainfall Estimation," Presented at the International Symposium on Remote Sensing and Water Resources, Enschede, Netherlands, August 1990, 11 pp.
Johnson, G., "The NRCS PRISM Climate Mapping Project," available online at http://www.ocs.orst.edu/prism/.
McCormick, B.S., 1995: "ViewRain and Associated Utilities, cal_rad." US Army Corps of Engineers, Tulsa Division.
Figure 1: Ninety days of accumulated P1 precipitation processing ending February 10, 2000.
Figure 2: Ninety days of accumulated PRISM normal precipitation ending February 10, 2000.
Figure 3: Percent of normal precipitation over ninety days ending February 10, 2000.