Modeling the effect of land use and climate change scenarios on future soil loss rate in Kasilian watershed of northern Iran

Accelerated erosion processes caused by global climate and land use changes in many regions of the world constitute a major restrictive factor in their sustainability. This study proposes a method to estimate soil loss rate under changes in future land use and climate in Kasilian watershed of northern Iran within two periods. The first period is related to current climate and land use (1991-2010), and the second concerns climate and land use scenarios (2011-2030). Downscaling global climate model projections of future climate was applied at the regional scale. A statistical downscaling model was then used to downscale precipitation for three scenarios, i.e., 10% increase in rainfall, 10% decrease in rainfall, and unchanged rainfall. Next, cellular automata-Markov model was used for characterization based on two scenarios of land use future that were designed using suitability maps. The soil loss mean for the current period was found to be 6.3 t ha(-1) year(-1), thereby indicating a low sustainability of soils. The results of simulated soil loss maps indicate a similar pattern in spatial distribution of loss rates compared with those of current periods, but the amount of risk has increased such that simulated erosion mean was 31-58% higher than the current period in all scenarios. Soil loss is thoroughly influenced by climate and land cover patterns in future. In other words, rainfall erosivity has increased by 20 MJ mm ha(-1) h(-1) year(-1), based on unchanged rainfall scenario and National Centers for Environmental Prediction data, simulated that cover management factor has increased by 35% compared with the current period. However, simulations indicated that land use changes may potentially induce much larger changes in erosion. The results also showed that soil loss is closely related to land use change and various scenarios of climate change and that revised universal soil loss equation is suitable model to investigate these relationships.