Soil erosion is known as a complex and hazardous geomorphological phenomenon that affects land management quality. Researchers developed various models to measure soil erosion because it has been identified as one of the main reasons for land degradation and desertification all around the world. Since it is difficult to measure the amount of soil erosion directly, varoius models has been introduced to estimate it indirectly. There is generally two defeciencies with these models: firstly they measure the amount of soil erosion deterministically and the likelihood of soil erosion is also expressed equal to its risk without considering land use type and asset value. Land use impact model (LUIM) has been invented in Australia for evaluating the risk of soil erosion in different types of land uses in order to identify areas with high risk of soil erosion. In this model several maps including management quality, land susceptibility and sensitivity to soil erosion for all land use types are developed using Bayesian belief networks. This model provides a basis for comparing the influences of different land use types on soil degradation processes. CORINE is another appropriate model for mapping soil erosion based on USLE model. This model estimates erosibility, erodibility and potential erosion in the study area. Finally, CORINE measures the real soil erosion by considering land cover types and potential erosion map. This study aimed to develop a desertification early warning system for Hanna region based on soil physical, chemical and biological characteristics. According to the LUIM analysis, 41.5 and 58.5 percent of the studied area is exposed to moderate and high risk of soil erosion, respectively. According to CORINE model, 49.2, 8.3 and 42.5 percent of the area had high, moderate and low potential soil erosion, respectively. According to this model 47.2, 4.4 and 48.4 percents of the study area were exposed to high, medium and low real soil erosion, respectively. Key words : erosion, land use impact model, bayesian belief network, CORINE
