Dataset of forb composition in a Pacific Northwest Bunchgrass Prairie

This dataset supports the research article "Forb composition gradients and intra-annual variation in a threatened Pacific Northwest Bunchgrass Prairie; Averett and Endress. In Print.  Ecology and Evolution".  This data includes: (1) perennial forb species composition data from Pacific Northwest Bunchgrass Pairie habitat in the Starkey Experimental Forest and Range, northeastern Oregon; (2) environmental, abiotic, and species trait variables measured from each sampling site; (3) plant species list; (4) densities of culturally important forb species; and (5) long-term sample dates for vegetation plots from the Starkey Experimental Forest and surrounding National Forest lands.  Forb composition data was collected from 29 plots in the Starkey Experimental Forest and Range, northeastern Oregon, at three different times during 2016 (April; May; July).

Methods
We used a stratified random design to sample vegetation composition from 29 (154 m2) plots in Pacific Northwest Bunchgrass (PNB) habitat within the SEFR.  Areas within the Main and Campbell study areas of the Starkey Experimental Forest and Range were stratified by percent tree cover, then 30 plots were randomly located within areas with < 5% tree cover (grassland plots).  One plot was located within an open forest when visited, and therefore, excluded, resulting in a total of 29 plots.  We sampled each plot at three different times during the growing season in 2016: April (April 18th – May 2nd); May (May 23rd – June 1st); and July (July 11th – 18th).  The first two sample periods coincided with growth of spring ephemerals.  The third was within the traditional vegetation sampling window for PNB.

One circular plot (radius = 7m) was established at each sampling site and 12 quadrats (1m2) were systematically located within each plot (Appendix A; Fig A1).  Two transect lines were laid out perpendicular to each other and intersecting at the center of the plot, resulting in one line running 14 m in length from north (0 m) to south (14 m), and the other 14 m in length from west (0 m) to east (14 m).  Four quadrats were centered at 0.5 m, 4 m, 10 m, and 13.5 m along each of the two transects for a total of eight quadrats along the north/south and west/east lines.  Four additional quadrats were located 4 m from the plot center along each of the NE, SE, SW, and NW cardinal directions for a total of 12 quadrats per plot. 

Within each quadrat, presence/absence was recorded and canopy cover was estimated for all forb species during the early (April and May) sampling periods, and all vascular plant species during the late (July) sampling period.  Canopy cover was classified into one of eight cover categories (<1%; >1-5%; >5-25%; >25-50%; >50-75%; >75-95%; >95-99%; >99-100%).  Plot level abundance for each species was calculated as the frequency of quadrats occupied per plot. Plot-level cover was calculated as the average arithmetic midpoint of the cover classes.  Graminoids (grasses, sedges, and rushes) were not identified to species during April and May due to their early phenological development. Plot-level cover of the soil surface, i.e., litter, rock, biotic crust (moss or lichen), and bare ground along with cover of total vegetation and each functional group (i.e., perennial forbs, annual forbs, perennial graminoids, annual graminoids, and shrubs) were estimated the same as species cover.

A tile probe was used to measure depth to soil restrictive layer (average of nine samples per plot) within 80 cm of the mineral soil surface during April at the center of the plot and at 1.75 m and 12.25 m along the North/South and West/East lines as well as at 6.5 m from the plot center along each of the NE, SE, SW, and NW lines.  Nine soil cores (< 25 cm deep depending on soil depth) were collected offset (0.25 m towards the plot center) from the tile probe measurement locations.  The nine soil cores were then mixed for each plot, ground, and dried at 60 ˚C for 48 hours.  Soil chemical and textural analyses were performed at AgSource Laboratory (Umatilla, OR, USA) for pH, cation exchange capacity (CEC), organic matter (%), Phosphorous (P; Olsen); Potassium (K; Ammonium Acetate), Magnesium (Mg; Ammonium Acetate), Calcium (Ca; Ammonium Acetate), Sodium (Na; Ammonium Acetate), and percent Sand, Silt, and Clay. 

Elevation, slope, and aspect were extracted from 30-m resolution digital elevation models (U.S. Geological Survey 2006) using ArcGIS 9.1.  We transformed aspect by folding the aspect about the NE – SW lines to align with an expected heat load gradient (SW = maximum heat load and NE = minimum heat load). 

For more detail regarding methods please refer to Averett and Endress. In Print. Ecology and Evolution "Forb composition gradients and intra-annual variation in a threatened Pacific Northwest Bunchgrass Prairie".