Protection from livestock fails to deter shrub proliferation in a desert landscape with a history of heavy grazing.

Desertification is often characterized by the replacement of mesophytic grasses with xerophytic shrubs. Livestock grazing is considered a key driver of shrub encroachment, although most evidence is anecdotal or confounded by other factors. Mapping of velvet mesquite (Prosopis velutina) shrubs in and out of exclosures in 1932, 1948, and 2006 in semiarid grasslands of southeastern Arizona, USA, afforded the opportunity to quantify livestock grazing effects on mesquite proliferation over 74 years in the absence of fire to test the widespread assumption that livestock grazing promotes shrub proliferation. In 1932, shrub cover, density, and aboveground biomass were compared on grazed (12%, 173 plants/ha, 4182 kg/ha) and newly protected areas (8%, 203 plants/ha, 3119 kg/ha). By 1948, cover on both areas increased to 18%; yet, density on the protected area increased 300% (to 620 plants/ha), nearly twice that of the grazed area (325 plants/ha). From 1932 to 1948, differences in recruitment of new plants and growth of existing plants were reflected in biomass, which was higher on the protected area (415 plants/ha, 8788 kg/ha) relative to the grazed area (155 plants/ha, 7085 kg/ha), although mortality was equally low ( 0.06%). In 2006, 42 years after an herbicide application reset mesquite cover to 10% on both areas, aboveground mesquite mass was comparable on both areas ( 4700 kg/ha), but cover and density on the protected area (22%, 960 plants/ha) exceeded that on the grazed area (15%, 433 plants/ha). Mesquite mass in 2006 was substantially below 1948 levels, so continued accrual is likely. That shrub recovery from herbicides on a biomass basis was much less than recovery on a cover basis suggests that remotely sensed biomass estimates should integrate land management history. Contrary to widely held assumptions, protection from livestock since 1932 not only failed to deter woody-plant proliferation, but actually promoted it relative to grazed areas. Results suggest (1) that thresholds for grassland resistance to shrub encroachment had been crossed by the 1930s, and (2) fire management rather than grazing management may be key to maintaining grassland physiognomy in this bioclimatic region.