Short- and long-term vegetation change related to grazing systems, precipitation, and mesquite cover

Rangeland scientists struggle with how long rangeland experiments must continue in order to detect treatment effects, particularly in semiarid ecosystems characterized by slow responses and high spatiotemporal variability. We compared changes in eight grass and three shrub categories to grazing systems (yearlong vs. seasonal rotation with equivalent long-term stocking rates), and covariates (precipitation and mesquite [Prosopis velutina] gradients) over 12 yr (1972-1984) and 34 yr (1972-2006) on the Santa Rita Experimental Range, Arizona. We used split-plot analysis of variance, with year as the split, to make these comparisons. Grazing systems did not influence plant dynamics as shown by the lack of grazing system by year effect on all response variables in either time period. The absence of a detectable grazing effect on vegetation changes may be due to overriding influences of grazing intensity, pasture size, precipitation variability, and few replicates. Also, more time may be needed to detect the small accumulating and potentially temporary effects from grazing systems. The grazing system main effects present at the beginning and throughout the study suggest that pastures assigned to each grazing system had different potentials to support vegetation. Nearly twice the number of response variables were related to the precipitation covariate than to mesquite cover, but only about half of all the relationships were consistent between time periods. The struggle to know how long to observe before detecting a grazing system effect was not resolved with the additional 22 yr of observation because we cannot definitively reject that either more time is needed to detect small but cumulative effects or that the two grazing systems are not different.