The proliferation of woody plants in grasslands has been occurring worldwide for over two centuries. Encroachment is typically due to increasing abundance of native woody species within their historic range. Livestock grazing is commonly proposed as the primary driver of woody plant encroachment. However, this assumption has not been rigorously evaluated; and qualitative evidence in support of it is equivocal. We used a meta-analytic approach to determine the extent to which livestock grazing is a primary driver of woody plant encroachment.
We compiled a literature database of 109 peer-reviewed studies that empirically measured woody plant encroachment into current or historic grasslands under grazed conditions. Only 22 of these studies experimentally addressed the role of grazing on treatment and control sites, reported woody plant encroachment rates, and specified both error and sample size. These 22 studies measured different metrics of woody plant encroachment (e.g., density, canopy cover, basal area), so a meta-analysis was conducted using the 13 studies reporting canopy cover. An effect size (ES) for woody plant cover was calculated using a standardized mean difference (difference of control and treatment means divided by a pooled variance). ES values were tested for deviation from zero (indicating no difference between treatment and control sites) with a z-test. The grain size of measured woody plant cover, data source (field or aerial photo), grazing duration, geographic location, and mean annual rainfall were also recorded. A mixed model with study specified as a random effect and ES weighted by inverse variance was used to assess the importance of these variables.
The average grazing ES was 0.0115 ± 0.36 S.E and did not differ significantly from zero, suggesting that grazing does not predictably lead to an increase or a decrease of woody plant cover. Data source was the only variable which significantly (p < 0.01) predicted grazing ES. Studies using aerial photos were more likely to conclude that grazing promotes woody plant encroachment than were studies based on field assessments. Our conclusions are constrained by the small sample size of the meta-analysis. Surprisingly few studies provided all the information needed to fully utilize them in a meta-analytic context, and many failed to report basic data (e.g., error and sample size). To improve future synthesis, it is imperative that researchers more thoroughly and consistently report their results. Synthesis of woody plant encroachment research will enable improved management of vegetation in the world’s drylands.