Introduction Soil degradation is a phenomenon which damages the soil structure and reduces its capacity for production. Soil erosion, as one of the most common forms of soil degradation, leads to loss of soil surface including on-site and off-site effects. Although soil erosion is a natural process on the earth, some of the human activities such as burning agriculture residues, deforestation, overgrazing, and lack of proper soil conservation practices accelerate soil erosion and enhance the negative consequences of erosion. Selecting and implementing of management scenarios requires assessment of soil losses from different management operations. Generally, management practices consist of structural and non-structural methods used to mitigate erosion, prevent nutrient removal, and increase soil infiltration capacity. Application of simulation models is an appropriate technique to evaluate erosional conditions. GeoWEPP is a process-based, distributed-parameter, and continuous simulation model of water erosion in watersheds with the possibility to simulate hillslopes and hydrographical network. Identifying problems in the real world usually produces large amounts of information and decision space, which requires optimization using evolutionary algorithms due to the variety of aims considered. Considering diversity of evolutionary algorithms, NSGA-II is one of the most common and effective multiobjective evolutionary algorithms (MOEA) and a very powerful tool for solving problems with conflicting objectives. Development of simulation models with optimization algorithms that are capable of analyzing very complex systems, has been found to be very efficient in real world problems. Simulation-optimization models are powerful tools for solving problems for least cost and best performance. Materials and Methods In this study, to predict sediment yield and runoff using GeoWEPP model, the integration of WEPP, TOPAZ, (Topography Parameterization), CLIGEN (Climate Generation) and GIS tool (ArcGIS) were used. The GeoWEPP model provides the processing of digital data including DEM, soil and landcover (The format of inputs was ASCII file). To generate climate file, the CLIGEN module which is a stochastic weather generation model was utilized. Furthermore, in TOPAZ part the CSA (critical source area) and MSCL (minimum source channel length) to delineate streams and also the outlet point of studied watershed were defined using GeoWEPP linked to ArcGIS. Using the basic maps including DEM, slope, soil great groups and soil database the GeoWEPP model simulates and generates the hillslopes automatically; therefore, this is an important advantage of GeoWEPP compared to WEPP model, which is capable of performing the simulation of watershed components spontaneously. In this study. in order to optimize the placement of gabions, 118 channels and 5110 candidate sites for gabion construction were simulated and evaluated. For optimization process of the number of objectives the AHP technique was initially used to prioritize the effective factors on the placement of gabions. Analytical hierarchy process is a structured technique for organizing and analyzing complicated decisions based on mathematical calculations. The AHP depicts the accurate approach for quantifying the weights of criteria and estimates the relative magnitudes of factors through pair-wise comparisons. The AHP technique includes creating hierarchical structure, prioritizing and calculating relative weights of the criteria, calculating the final weights and system results compatibility. The main criteria (objectives) for our study were minimum distance from road, minimum distance from residential area, maximum length of main channel, maximum sediment yield, maximum discharge volume and maximum volume structure. The AHP technique made it possible to restrict the decision making space and the number of possible options,, therefore simplify the optimization process. Then, NSGA-II (Non-dominated Sorting Genetic Algorithm) was applied in order to find the best solutions, i.e., the Pareto front, of alternatives for optimal location of structures based on the two objectives with higher priority and distance constraint.
Results and discussion The results of paired comparison matrix and prioritizing showed that the length of main channel in the watershed is the main effective criterion in locating gabion structures. The first priority is the most critical channel which produces the highest sediment yield; therefore, the most expensive structure is established on that channel. After channel length, the volume discharge is the second priority of effective factors for gabion placement. Using the results of AHP, based on channel length and discharge volume, the non-dominated sorting genetic algorithm (NSGA-II) was performed and the priority of critical channels and the specific position was determined from 1 to 35 among 5110 candidate sites for gabion construction. Using the ArcGIS, slope map and the lowest width of the critical channels, the best place for gabion construction was determined. Moreover, the main output of GeoWEPP is the spatial distribution of sediment yield and based on this map the sediment yield was classified in the watershed. Based on this map, the red color was the highest amount of sediment yield (more than 4 ton) in the watershed.
Conclusion Results confirmed that application of simulation-optimization techniques helps to select the best sites to construct gabion as the best management practice in the watershed.
Key words: GeoWEPP model, Optimization, Multi-objective decision making, NSGA-II algorithm, Analytic hierarchy process