This Resource serves to explain and contain the methodology, R codes, and results of the PRISM freshwater supply key indicator analysis for my thesis. For more information, see my thesis at the USU Digital Commons.

Freshwater availability in the state can be summarized using streamflow, reservoir level, precipitation, and temperature data. Climate data for this study have a period of record greater than 30 years, preferably extending beyond 1950, and are representative of natural conditions at the county-level.

Oregon State University, Northwest Alliance for Computational Science and Engineering PRISM precipitation and temperature gridded data are representative of statewide, to county-level, from 1895-2015. These data are available online from the PRISM Climate Group. Using the R ‘prism’ package, monthly PRISM 4km raster grids were downloaded. Boundary shapefiles of Utah state, and each county, were obtained online from the Utah Geospatial Resource Center webpage. Using the R ‘rgdal’ and ‘sp’ packages, these shapefiles were transformed from their native World Geodetic System 1984 coordinate system to match the PRISM BIL raster’s native North American Datum 1983 coordinate system. Using the R ‘raster’ package, medians of PRISM precipitation grids at each spatial area of interest were calculated and summed for water years and seasons. Medians were also calculated for PRISM temperature grids and averaged over water years and seasons. For analysis of single months, the median results were used for all PRISM indicators. Seasons were analyzed for the calendar year which they are in, Winter being the first season of each year. Freshwater availability key indicators were non-parametrically separated per temporal/spatial delineation into quintiles representing Very Wet/Very High/Hot (top 20% of values), Wet/High/Hot (60-80%), Moderate/Mid-level (40-60%), Dry/Low/Cool (20-40%), to Very Dry/Very Low/Cool (bottom 20%). Each quintile bin was assigned a rank value 1-5, with ‘5’ being the value of the top quintile, in preparation for the Kendall Tau-b correlation analysis. These results, along with USGS irrigation withdrawal and acreage data, were loaded into R. State-level quintile results were matched according to USGS report year. County quintile results were matched with corresponding USGS irrigation withdrawal and acreage county-level data per report year for all other areas of interest. Using the base R function cor(), with the “kendall” method selected (which is, by default, the Kendall Tau-b calculation), relationship correlation matrices were produced for all areas of interest. The USGS irrigation withdrawal and acreage data correlation analysis matrices were created using the R ‘corrplot’ package for all areas of interest.

See Word file for an Example PRISM Analysis, made by Alan Butler at the United States Bureau of Reclamation, which was used as a guide for this analysis.