STATE AND TRANSITION MODELS FOR RANGELANDS .5. THE USE OF STATE AND TRANSITION MODELS FOR PREDICTING VEGETATION CHANGE IN RANGELANDS

State and transition models are similar in nature to Markov models whi ch have been applied in the field of ecology since the 1960s. Variatio ns to true Markov processes that have been used in ecology include sec ond-order, discrete Markov and semi-Markov processes and continuous-ti me Markov processes. The uses of discrete-time and continuous-time Mar kov models are discussed. Three examples of how vegetation dynamics ca n be simulated by Markov models are presented. The way in which altere d climate may alter the course of vegetation change is described for P rosopis savanna in south Texas (USA). Chemical control strategies for Acacia nilotica management in north-western Queensland were compared b y incorporating a Markov model into a simulation model which included the effect of woody vegetation on pasture growth, as well as predictio n of liveweight gain of cattle. The impact of altering grazing pressur e on pasture composition change is presented using a continuous-time M arkov model of pastures in tropical woodlands of northern Australia. T hese examples indicate how state and transition models can be used. St ate and transition models may be used for prediction and analysis in a ddition to aiding communication. The integration of Markov models into process models shows promise for devising complex management models f or rangelands.