A comparison of Markov model-based methods for predicting the ecosystem service value of land use in Wuhan, central China

Aggressive human activity and limited natural resources cause complex land use changes that significantly affect the ecosystem service of land types. In this paper, we used the Markov model to predict future changes in the ecosystem service of each land type. We used remote sensing to evaluate the changes in five land use categories, and previously published value coefficients to calculate the ecosystem service value of each land type. Two methods were applied to acquire useful results. The first was called the A-E method. It involved predicting the future changes of areas in land use using the Markov model and multiplying the value coefficients of the ecosystem service. The second was called the B-E method. It involved calculating the initial ecosystem service value of the land type, and then directly predicting these values. From comparison the actual values in 2011 and the stationarity of the values, we determined that the predicted ecosystem service values of the five land types by the A-E method was better than those by the B-E method. Despite considering the annual changes in the coefficients of ecosystem services, our results can reflect the changes in ecosystem services with land use transformation.