A dynamic programming approach to optimal grazing strategies using a succession model for a tropical grassland.
Abstract
Plant cover data collected at Varanasi, India, for 2 successive years in pastures at different grazing intensities were used to formulate a simulated succession computer program. A transfer matrix was used giving yearly changes in plant cover between species; transfer functions were then developed for each of 4 intensities of grazing from ungrazed to heavily grazed and these were used to develop a long-term optimum grazing pattern using dynamic programming, taking into account such factors as species diversity, legumes, bare ground and density of the the dominant species Dichanthium annulatum in addition to forage utilization. Results indicated that a combination of moderate grazing with an occasional year of light grazing was the best strategy over a 12-year period.