Markov Model for Predicting the Land Cover Changes in Shimla District

Abstract

Forests have been the key element in maintaining sustainability of many global phenomena. Human dependency on forests is both necessary and unavoidable and hence degradation of this natural resource is inevitable. The study aims to understand the change dynamics over past few decades in the Shimla district, using remote sensing and GIS based techniques. The tree cover area estimated during 1970s, 80s and 90s were 50.65%, 48.30% and 52.31% respectively. The classified images were analysed for changes and found that 2.35% of net tree cover changed into non-tree cover during 1972 to 1989 but during 1989 to 1999 the trend changed into a net positive one with the increase of tree cover by 4.01%. Transition probabilities of each land cover features were calculated for the three-time periods (72-89,89-99 and 72-99) and then analysed for their statistical significance using Markov chain model. Based on the findings, a non-spatial temporal Markov prediction was made for the year 2009. The predicted forest area in 2009 is 55.49% with the 5% error under Markovian assumption of stationarity.