The effects of livestock grazing on vegetation in a semiarid grassland: a test of three hypotheses.

Questions: Livestock grazing is an important driver of plant diversity in grasslands. Different mechanisms summarized in three main hypotheses have been proposed to explain its varying effects depending on productivity. The insufficient compensation hypothesis (ICH) states that livestock causes the extinction of species that cannot compensate for biomass removal in unproductive sites, reducing species richness. The competitive release hypothesis (CRH) states that livestock increases plant diversity by attenuating competition, especially in productive environments. In the convergent evolution hypothesis (CEH), livestock has minor effects on diversity in unproductive sites because plants that endure stress are pre-adapted to grazing. These hypotheses have rarely been compared formally, and have mostly been tested on large spatial scales. Here we simultaneously test these hypotheses on a local scale. Location: Natural semi-arid grassland, Oaxaca, Mexico. Methods: We performed a long-term exclosure experiment, recording plant diversity annually over 17 years. In 2017 we measured species richness over a grazing intensity gradient. We used linear mixed-effects models, nestedness and beta-diversity analyses to evaluate diversity and composition in relation to grazing and productivity. Results: Plant diversity generally increased with grazing intensity. In line with CRH, species richness increased as grazing became more intensive due to a progressive incorporation of increasingly poor competitors to the community, with a seemingly stronger effect under high productivity. This suggests that herbivores limit competitive species and prevent competitive exclusion. ICH may play a secondary role because the effect was reversed in low productivity inside exclosures. Conclusions: Our finding implies that in productive sites, the exclusion of livestock can be detrimental to plant diversity as a lack of grazing accelerates competitive exclusion. By contrast, in limiting environments, grazing can preserve diversity if livestock are kept at low stocking densities. Our results show that small-scale variations in productivity need to be considered when designing livestock management. [ABSTRACT FROM AUTHOR]