Prediction of Soil Hydrological Responses under Land/use/Cover Changes using Markov Chains in Jiroft Watershed, Iran

Abstract

Aims: This study aims to evaluate the Soil Hydrological Response (SHR) under land-use/ land-cover (LU/LC) (current management/current land cover) using a field-oriented and remote sensing database in the Jiroft watershed, Iran. Materials & Methods: Land-use maps were extracted from Landsat images using the supervised classification method for 1987-2017. The results were validated against the field data from 100 points, where we found the Kapp index to be greater than 80%, indicating an acceptable accuracy in land-use classification. The LU/LC map was then projected for 2047 using the CA-Markov model. The Curve Number (CN) for each land-use was determined from superimposing LU/LC and the soil hydrological group map. The Soil Conservation ServicesCurve Number method (SCS-CN) was employed to estimate runoff. Findings: Rangelands with good (densely vegetated) and moderate vegetation conditions had a decreasing trend in terms of area (i.e., -2.94% and -3.64% in 1987- 2017), while croplands, orchards, residential, and saline areas had an increasing trend (by 1.46%, 0.88%, 0.33%, and 7.21%). We found that agricultural lands, saline lands, and residential areas will increase by 0.75, 5.5, and 0.13% by 2047, respectively. The highest Curve Number (CN) values were detected in desert lands, residential areas, and poor-condition rangelands with the respective values ​​of 88, 88.67, and 84.67. As for 2047, residential areas, deserts, and rangelands with poor conditions with the respective CN values of 89, 88, and 84.67 would contribute substantially to runoff generation. Conclusion: We found a considerable increase (up to 6 mm increase) in runoff depth in some land-uses and more than 3.4% increase in the area of the high runoff producing class (IV). We believe that higher runoff production potential and more intense and short rain showers should be considered seriously in terms of possible flash floods in the future.