Legacy effects of livestock grazing on vegetation of an arid landscape exist more than 30 years after livestock were excluded
Data files
Mar 21, 2026 version files 401.67 KB
-
Fowlers_data_dryad.xlsx
387.01 KB
-
README.md
14.66 KB
Abstract
Introduction: Globally, livestock grazing is key driver of vegetation degradation in arid rangelands. This degradation tends to be most severe near artificial water points which are focal points for livestock activity. Conservation reserves have been established on former rangelands, however, the timeframes for vegetation recovery following the cessation of livestock grazing are poorly understood.
Objectives: To assess changes in rangeland vegetation structure and composition following long term herbivore exclusion to inform understanding of the timeframes required for rangeland vegetation recovery.
Methods: We quantified recovery of vegetation following the removal of livestock in an Australian arid rangeland by comparing vegetation structure and composition inside and outside of herbivore exclosures that had been established for 36 years and were situated along gradients of distance from water.
Results: Inside of long-term exclosures, vegetation cover was higher and taller and forb density was higher than in unfenced control plots. Inside exclosures, annual species richness was higher but did not respond to distance from water. Perennial species richness did not respond to herbivore exclusion but increased with distance from water both inside and outside of exclosures.
Conclusions: After 36 years of livestock exclusion, vegetation recovery occurred structurally but floristically, recovery was mainly represented by short-lived annual species. That the richness of perennial species inside of herbivore exclusion increased with increasing distance from water suggests that the composition of perennial plant species was a legacy of the effects that water-focused livestock grazing had on vegetation when the plots were established 36 years previously.
Implications for practice: Following livestock removal, timeframes for vegetation recovery may vary depending on the indicator used. Changes in ground layer structure, community composition and abundance of annuals may be observed within a decade, however, recovery of slow growing perennials may occur over multiple decades. A key implication of this study is that the progress of full passive recovery in arid plant communities following livestock removal should, at the minimum, be expressed in the order of decades. Active recovery approaches such as direct seeding or out-planting may be required to enhance the rate of recovery of perennial plants following the livestock removal.
https://doi.org/10.5061/dryad.jh9w0vtnk
Description of the data and file structure
This data were collected to examine recovery of vegetation inside and outside of herbivore exclosures established 36 years prior that were situated along gradients of distance from water in arid Australia (Fowlers Gap, Nov 2021). We predicted that species richness and the cover, height of vegetation and the abundance of individual species would increase with increasing distance from water outside of exclosures and be greater inside of exclosures. Our criterion for vegetation recovery to have occurred was that the height, density, cover and species richness of vegetation inside of the exclosures would not vary with distance from water and be at similar levels to plots outside of exclosures that were situated distant from water. Our specific aims were to:
1. Use Landsat imagery to quantify the effects of distance from water on bare ground.
2. Conduct vegetation surveys to quantify bare ground cover, the cover of vegetation elements, vegetation height, species composition and species richness of vegetation inside and outside exclosures.
3. Determine the responses of annual and perennial plant species to grazing exclusion and distance from water.
4. Examine the response of plant species to herbivore exclusion and distance from water and categorize them as grazing increasers or decreasers.
The study used paired fenced and unfenced vegetation plots (20 x 20 m) forming 16 blocks of 32 plots. Measurements either sampled over the entire 20 x 20 m area or within 4 x 4m x 4 m subplots within the plot area.
Files and variables
File: Fowlers_data_dryad.xlsx
Data is organised by sheets within the .xlsx workbook.
-
Plot height contains data for the height of vegetative material above ground, at 10 random points within the 20 x 20 m plot areas. Each point measurement is numbered under column "Sample".
Variable Units Comment Block - 16 blocks across 4 paddocks. Block names are coded {Paddock name}_{Block number}. Plot - Plots within fenced exclosures that exclude sheep and kangaroos (Fence) and plots adjacent to and outside of exclosures that are accessible to grazers (Unfenced). Sample - A number to denote an individual measurement within a plot (10 measurements per plot). Distance from water Metres The euclidean distance a plot is from the paddock water point. Height (cm) Centimetres The height of vegetative material above ground as measured with a ruler to the nearest cm. -
Growth form cover contains data for the cover of plant growth forms as a percentage of 100 regular points, spread over 5 random transects, sampled within the 20 x 20 m plot areas. Bare ground, forbs (dead + alive), grasses (dead + alive), shrubs (dead + alive) and subshrubs (dead + alive) are distinguished.
Variable Units Comment Block - 16 blocks across 4 paddocks. Block names are coded {Paddock name}_{Block number}. Plot - Plots within fenced exclosures that exclude sheep and kangaroos (Fence) and plots adjacent to and outside of exclosures that are accessible to grazers (Unfenced). Distance from water Metres The euclidean distance a plot is from the paddock water point. Category - All bare = bare ground, All forbs = forbs: dead + alive, All grass = tufts and tussocks, dead + alive, All shrub = shrubs: dead + alive and All subshrub = subshrubs: dead + alive. % cover Percent The estimated ground cover of each category feature -
Species density contains data for the density of individual plant species encountered within 4 x 4m x 4m concentric nested quadrats placed randomly (without replacement) within the 20 x 20 m plots. Density is measured as individuals/m sq. derived from the survey that use the importance score method (Morrison et al. 1995).
Variable Units Comment Block - 16 blocks across 4 paddocks. Block names are coded {Paddock name}_{Block number}. Plot - Plots within fenced exclosures that exclude sheep and kangaroos (Fence) and plots adjacent to and outside of exclosures that are accessible to grazers (Unfenced). Sample - A number to denote an individual measurement within a plot (4 sub-plots per plot). Distance from water Metres The euclidean distance a plot is from the paddock water point. Species - The name of the plant species sampled Density (ind/m2) Individuals of a species / metres2 The estimated density of plant species within the 4 x 4 m sub-plot sampled -
Species richness contains data for the species richness of each 20 x 20 m plot. Species richness is categorised under "Group" column to include all species, species classified as annuals only or species classified as perennials only.
Variable Units Comment Block - 16 blocks across 4 paddocks. Block names are coded {Paddock name}_{Block number}. Plot - Plots within fenced exclosures that exclude sheep and kangaroos (Fence) and plots adjacent to and outside of exclosures that are accessible to grazers (Unfenced). Distance from water Metres The euclidean distance a plot is from the paddock water point. Group - All plant species, annuals only and perennials only Counts - The number of species counted in a 20 x 20 m plot. -
Seasonal FVC samples contains data for the values of seasonal fractional vegetation cover at 2000 random sample points, taken over the broader study area. SFVC values were taken from a mean Landsat SFVC raster of images collected from December 1987 to December 2021. Response values are band 1 (bare ground), band 2 (green vegetation) and band 3 (non-green vegetation). In this study, bands 2 and 3 are added.
Variable Units Comment Sample point - A number to denote each sample taken from the mean raster image Band - Band 1 = bare ground, Band 2 = green vegetation, Band 3 = non-green vegetation Value Percent The percent cover of each band category Distance from water Metres The euclidean distance a sample point is from a paddock water point. -
Plot geographic data contains geographic data associated with each plot.
Variable Units Comment Block - 16 blocks across 4 paddocks. Block names are coded {Paddock name}_{Block number}. Plot - Plots within fenced exclosures that exclude sheep and kangaroos (Fence) and plots adjacent to and outside of exclosures that are accessible to grazers (Unfenced). Plot size (m) Metres Plot dimensions Latitude Decimal degrees Geographic latitude Longitude Decimal degrees Geographic longitude Distance from water Metres The euclidean distance a plot is from the paddock water point. -
Plant traits contains plant traits, relevant to the study, associated with the plant species sampled. Traits include "Growth form" (forb, tuft, shrub, subshrub, trailer, tussock), "Geography" (native or introduced species), "Life history" (annual or perennial) and "Physical defence" (spiny or non-spiny).
Variable Units Comment Species - Scientific name of plant species encountered Family - Family of plant species encountered Growth form - Growth form category assigned to plant species Geography - A category to assign whether the species is native or introduced to study area Life history - Life history type assigned to species (annual or perennial) Physical defence - A category to group species with spines and those without. -
Sheep numbers contains counts of sheep at Fowlers Gap from 1890 - 2021 with a 20 year gap of missing data from 1944. Sheep counts are reported as sheep per unit area and sheep per water points in the study area over time. Prior to 1945, sheep counts were estimated using knowledge of paddock sizes at Fowlers Gap over time and historic sheep density data of the Fowlers Gap region as reported in Beadle (1948). Sheep counts from 1966 were gathered from station annual reports.
Variable Units Comment Year - Paddock names - Name of paddock within Fowlers Gap Station historic records Paddock area (km2) km2 Area of paddock Sheep/km2 Sheep / km2 Density of sheep Sheep/ha Sheep / ha Density of sheep Sheep/100 acres Sheep / 100 acres Density of sheep No. water points - Number of known water points available in paddock Sheep/water point Sheep / water point Number of sheep per known water point -
Rain contains annual rainfall recorded at Fowlers Gap from 1966 to 2021 as reported by the local weather station and as recorded in Fowlers Gap annual reports.
Variable Units Comment Year - Rainfall mm Annual rainfall total Source - Data source
References
Beadle NCW. (1948) The Vegetation and Pastures of Western New South Wales with Special reference to Soil Erosion. Thomas Henry Tennant, Government Printer, Sydney
Morrison DA, Le Brocque AF, Clarke PJ (1995) An assessment of some improved techniques for estimating the abundance (frequency) of sedentary organisms. Vegetatio 120:131-145
Code/software
Data can be read into R for analysis and plotting. Data can also be read using spreadsheet software such as Microsoft Excel.
Access information
Other publicly accessible locations of the data:
- N/A
Data was derived from the following sources:
- N/A
Study location:
This study was undertaken at Fowlers Gap Research Station, NSW, Australia. Data were collected from 20/11/2021 to 3/12/2021 in Chenopod shrubland within four long term sheep paddocks. Each paddock had a single water point and four 20 x 20 m sheep and kangaroo exclosure fences located at increasing distances from the water point. These fences were built for a previous study in 1986.
Ground layer vegetation structure and floristics:
Vegetation was measured inside 16 pairs of 20 x 20m plots that had excluded livestock for 36 years or been continuously exposed to livestock in that time. Height (10 point samples) and growth form cover (100 point samples) were measured at regular points along five random transects within plots. The density of plant species was measured in four sub-plots using nine concentric nested quadrats ranging from 1/6m2 to 16m2.
Landsat seasonal fractional vegetation cover (SFVC):
The seasonal fractional vegetation cover Image series of the study area from 1987 to 2021 were collected and a mean raster created. The SFVC data was regressed against a Euclidean distance from water point raster generated, constrained by paddock boundaries with origins from known paddock water points established in 1971. Band 1 (Bare ground) was treated as a separate regression from Bands 2 and 3. Band 2 values were added to Band 3 values to include photosyntheic and non-photosynthtic vegetation in one regression.