1. The current human-induced intensification of grazing pressure and the increase of aridity as a result of climate alterations are unprecedented and have been identified as the main drivers that cause desertification in rangelands worldwide. In these ecosystems, human well-being mostly depends on plant species that provide forage for domestic herbivores. However, scarce evidence exists about the interaction between regional aridity level and human-induced disturbances as determinants of forage plant populations’ structure and dynamics.

2. We studied the effects of domestic grazing intensification on the population structure of dominant native grasses, in three rangeland sites located across a regional aridity gradient: a semi-desert (high-aridity site), a shrub-grass steppe (intermediate-aridity site) and a grass steppe (low-aridity site). We also studied the effect of two-year grazing exclusion on plant growth of a key native forage grass species common to the three sites.

3. Grazing decreased total grass density and increased the frequency of small plants in all sites, particularly for forage species. However, the size of the grazing intensification effect was the greatest in the high-aridity site, where intensive grazing produced a ten-fold reduction of grass density. Moreover, plant recovery (growth) after grazing exclusion was lower as aridity increased.

4. Synthesis and applications. Our study provides evidence of a negative synergistic effect of grazing pressure and aridity that may lead to the collapse of grass populations. Long-term grazing intensification degrades the population structure of grasses, particularly in high-aridity sites, where the forage provision is substantially reduced. These results refute the hypothesis that plant-traits of dominant species adapted to high-aridity allow them to resist herbivory. Besides, high-aridity delays plant recovery after defoliation (low resilience). The management of both the grazing pressure and the length of grazing-rest according to the ecological-site aridity are key aspects for maintaining the forage provision of rangelands. Monitoring plant populations’ structure through time and space strengthens inferences about responses of forage species to ongoing changes in disturbance and stress regimes. This knowledge is complementary to regional and worldwide monitoring endeavors based on land cover, and it contributes to the robust design of sustainable management of global rangelands.