HYDROLOGICAL MODELLING AND EVALUATION OF BEST MANAGEMENT PRACTICES IN A WATERSHED

Abstract

In this study, a spatially distributed sediment yield model was developed in GIS environment which is capable to identify the sediment source and sink areas within the catchment. Transport capacity and gross soil erosion maps were generated by overlaying different thematic maps. Employing transport capacity maps along with gross soil erosion (GSE), a programme was developed in Interactive Data Language (IDL) to route the sediment form ridge pixel to the outlet of the Pokhariya watershed. Lumped as well as spatial accuracy of the model was checked by comparing the model output with the observed sediment data at the outlet. By overlaying the gross soil erosion and deposition maps, net erosion/deposition maps were generated. A distributed parameter model, Soil and Water Assessment Tool (SWAT) was tested on monthly and seasonal basis and used for developing management scenarios for the critical sub-watersheds of Pokhariya watershed. Sensitivity analysis was performed. It was found that the parameters namely Soil Conservation Service (SCS) curve number (CN) followed by hydraulic conductivity for tributary (CH_K1), Manning roughness coefficient for overland flow (OV_N) are more sensitive. SWAT model was calibrated and validated for simulation of runoff and sediment yield of the Pokhariya watershed for monsoon season (June to September) of the years 2000 to 2005 and 2006 to 2008 respectively. For runoff simulation, the values of R2, CC, NSE, PBIAS and RSR were found to be 0.88, 0.94, 0.70, -5.43 and 0.52, respectively, during calibration and 0.93, 0.96, 0.78, -24.78 and 0.41 respectively, during validation period. Similarly for sediment yield simulation, the value of R2, CC, NSE, PBIAS and RSR were found to be 0.82, 0.91, 0.91, 9.25 and 0.25, respectively, during calibration and 0.84, 0.92, 0.97, 5.12 and 0.13 during validation period respectively. The critical sub-watersheds were identified on the basis of average annual sediment yield obtained during the years 2000 - 2008. The calibrated and validated model was used for planning and management of critical sub-watersheds. The ranking of different critical sub-watersheds was carried out as per their sediment yield. The sub-watershed numbers SW-6, SW-4, SW-2, SW-5, SW-10, SW-7, SW-9, SW-8, SW-1 and SW-3 were vi considered in order of priority for developing the best management practices for conservation. Sub watershed number six was found to be most critical. The total twelve combinations of treatments were selected for employing the Best Management Practices in the Pokhariya watershed. The treatment included four crops (i.e. Maize, groundnut, sorghum and soybean), five tillage practices (i.e. zero tillage, conservation tillage, field cultivator, Harrow spike tooth and disc plough) and three levels of fertilizer (Low, medium and high). The results showed that crops like maize, groundnut and soybean cannot replace the existing rice crop on the basis of sediment yield reduction criteria. The field cultivator is recommended to replace the conventional tillage (country plough/Mouldboard plough) because it reduces the sediment yield by 15 percent as compared to conventional tillage. Zero and conservation tillage practices can be recommended because these tillage reduces sediment yield by 25 and 23 percent respectively as compared to existing tillage. Soil and water conservation measures like nala bund, check dam and percolation tanks are recommended in the study watershed. Comparison of sub-watershed prioritization using SYI, Morphometric and AHP techniques was also carried out. It was found SWAT based prioritization is found to be better with AHP based prioritization followed by SYI and morphometric analysis derived prioritization. The poor performance of the morphometric analysis for the prioritization of sub-watersheds is may be due to the non-involvement of land use and land cover data of the study area. Since the data requirement for the SWAT is excessive. Therefore, in the absence of field data multi-Criteria Decision Making based on Analytical Hierarchy Process (AHP) has been found to be better than Silt Yield Index (SYI) and morphometric analysis. The performance evaluation of five machine learning or data-driven models viz. ANNLM, ANNSCG, LS-SVR, REPTree and M5 model was carried out for predicting runoff and sediment yield. The input variable were selected using the trial-and-error procedure which represents the hydrological process in the watershed. The Seven input variables to all the models were comprised of a combination of rainfall, average temperature, relative humidity, pan evaporation, sunshine duration, solar radiation and Wind speed. Evaluation of model’s performances were carried out using the five statistical indices. Comparative analysis showed that ANNLM model marginally outperformed LSvii SVR model and all the other models investigated during calibration and validation for runoff modelling whereas LS-SVR model surpassed ANN model and other models for sediment yield modelling. Moreover, M5 model tree is better in simulating sediment yield and runoff than its near counterpart, the REPTree model and marginally inferior when compared to LS-SVR and ANN models.