Datasets

As in previous volumes in this atlas, to compare the modern distributions of plant taxa with climatic parameters, we used an equal-area grid for North America; each point of which was approximately 25 km from adjacent points ("25-km grid" in this report). A map of the number of taxa present at each grid point is shown in figure 1. For this report, we updated the modern climatic data underlying our analyses. In previous volumes, we employed a North American dataset constructed of thirty-year climatic "normals" for the period 1951 to 1980 based on a diverse array of data sources (see Thompson and others, 1999a for further description). In this report, this dataset is replaced by one based on averages for the period 1961–1990 from New and others (1999, 2000, 2002) that were interpolated onto the 25-km grid using local trend-surface regression with elevation as a covariate. The updated climatic data are the same employed in Thompson and others (2008, 2012) and in our usage include both observed climatic parameters (mean annual temperature [ANNT], mean temperature of the coldest and warmest months [MTCO and MTWA], and mean January, July, and annual precipitation [JANP, JULP, and ANNP]), as well as the calculated variables of growing degree days (Newman, 1980) on a 5°C base (GDD5) and a moisture index (AE/PE) that expresses the ratio of actual to potential evapotranspiration (following the equations of Priestley and Taylor, 1972). Maps of the modern distributions of these climatic parameters are shown in figures 2, 3, 4, and 5.

We digitized the present-day distributions of 530 woody plant species based largely on maps in "Atlas of United States Trees" compiled by the U.S. Forest Service (Little 1971, 1976, 1977, 1978, 1981; Critchfield and Little, 1966; Viereck and Little, 1975). Our dataset includes a few additional taxa from arid and semiarid settings in the Western United States from other sources (Bailey, 1970; Benson and Darrow, 1981; Yang, 1970). For each taxon, the digitized distribution maps were used to assign its presence (or absence) to each point on the 25-km grid. Since the MCR approach requires two or more taxa, we excluded grid points with only a single taxon present, and a total of 25,043 grid points provided information for the analyses in this report. In addition, nine taxa used in Thompson and others (2008) were excluded because their modern geographic and climatic distributions are extremely limited. Figure 1 illustrates the resulting taxonomic richness across the North American dataset. We followed the original taxonomic usages of the source publications, and table 1 provides the currently accepted taxonomy. Currently accepted botanical names were verified in Tropicos (Missouri Botanical Garden, 2012a), eFloras (Missouri Botanical Garden, 2012b), and ITIS (USGS Center for Biological Informatics, 2012) nomenclatural databases.