Zumwalt Prairie Fire Grazing Study Summarized Cover Data
Description
This dataset was used to analyze the effects of prescribed fire, livestock grazing and ventenata cover on vegetation and soil surface characteristics in an experiment with four treatments: prescribed fire (BURN), livestock grazing (GRAZ), combined fire and grazing (BOTH), and no fire or grazing (CNTL). This experiment took place on a bunchgrass grassland in northeast Oregon, USA and spanned 14 years (2008 - 2021), with data collected in 5 of those years. Treatments were replicated four times. Burn treatments were applied on the landscape within blocks and grazing treatments were nested within the burned and unburned sections of each block. Prescribed fire was applied in two years (2006, 2016) and livestock grazing occurred annually. The dataset contains 15 response values, summarized by treatment and year: 1) cover data for some individual species and several guilds (groupings of plant species); 2) percent bare ground 3) calculated diversity and richness indices 3) Normalized Difference Perennial Dominance Index calculated from annual and perennial values and 4) estimated biomass derived from a model developed for this particular grassland that correlates Landsat satellite imagery to measured biomass. Raw data used for all variables except biomass were collected with line-point intercept methods.
Files
Steps to reproduce
Raw data were collected using line-point intercept methods (see Bonham, C.D., 2013 Measurements for Terrestrial Vegetation, 2nd edition. UK: John Wiley & Sons.). Plants were identified to species level or, in some cases, only to genus level. The percent cover of each species (by plot) was calculated as the number of points with that species divided by the total number of points per plot. To produce summary data by guild, first each species observation was assigned to the appropriate guild (e.g. "Perennial Native Grass"). LPI is a frequency measure (presence/absence at each point). Assigning individual species to guilds often resulted in the same guild being represented more than once at each point. Therefore, when summarizing cover values by guild, repeat values for the same guild at the same point were deleted so that the percent cover for each guild in each plot was the number of points where that guild was present divided by the total number of points per plot.