干旱区生态环境知识资源中心

The effects of cattle grazing were evaluated on range dynamics of the Caatinga which is a deciduous dry woodland, covering most of the semi-arid Brazilian Northeast. Three stocking rates (SR) were studied (heavy, 1 steer 6.7 ha(-1); moderate 1 steer 10 ha(-1); light, 1 steer 13.3 ha(-1)), in addition to an ungrazed exclosure (zero stocking). In the first phase (1978-81) each stocking rate was tested under continuous and deferred grazing. In the second phase (1981-84), deferred grazing was eliminated, so that pastures became replications of continous grazing. Six steers per pasture were used, and pasture size was used to vary stocking rate. There was no effect of stocking rate or grazing system period on the frequency of the herbaceous species. They were, however, influenced by rainfall in the period, and could be divided into 3 groups. Sixteen species increased with increasing rainfall during the last months of the rainy season, and reached the highest frequency in 1984. Eleven species also increased with increasing rainfall but reached the highest frequency in 1983. Rainfall had no effect on the frequency of 2 important species, Herissantia crispa (L.) Brit. and Selaginella convoluta Spring. Death rate of 5 shrubs (Lippia microphylla Cham., Croton rhamnifolius (Kunth em.) Mull. Arg. Calliandra depauperata Benth, Cordia leucocephala Moric., and Bauhinia cheilantha (Bong.) Steud.) decreased with decreasing stocking rate, 11.7, 9.3, 7.7, and 4.5%, respectively on heavy, moderate, light, and zero stocking. Death rates were higher in easily broken shrub species, L. microphylla and C. leucocephala. Stocking rate also influenced the height growth rate of the tagged shrubs, being respectively -2.7 and 9.8% for heavy and zero stocking. Mean density of shrubs and trees, determined by the Point-Centered Quarter Method, was respectively 21,109, and 447 plants ha(-1) in 1982, and 13,230 and 401 plants in 1984; the main cause of the high shrub death (37.3%) was probably the 1982 drought. Density was not affected by stocking rate. Considering the 7 experimental areas separately, there was no regression between 1982 and 1984 shrub densities. There was, however, regression between 1982 density and the difference between 1982 and 1984 densities.