Evaluation of static and dynamic land use data for watershed hydrologic process simulation: A case study in Gummara watershed, Ethiopia

摘要

Land Use Land Cover (LULC) change significantly affects hydrological processes. Several studies attempted to understand the effect of LULC change on biophysical processes; however, limited studies accounted dynamic nature of land use change. In this study, Soil and Water Assessment Tool (SWAT 2012) hydrological model and statistical analysis were applied to assess the impacts of land use change on hydrological responses such as surface runoff, evapotranspiration, and peak flow in Gummara watershed, Ethiopia. Moreover, the effects of static and dynamic land use data application on the SWAT model performance were evaluated. Two model setups, Static Land Use (SLU) and Dynamic Land Use (DLU), were studied to investigate the effects of accounting dynamic land use on hydrological responses. Both SLU and DLU model setups used the same meteorological, soil, and DEM data, but different land use. The SLU setup used the 1985 land use layer, whereas the DLU setup used 1985, 1995, 2005, and 2015 land use data. The calibration (validation) results showed that the model satisfactorily predicts temporal variation and peak streamflow with Nash Sutcliffe Efficiency (NSE) of 0.75 (0.71) and 0.73 (0.71) in the DLU and SLU setups, respectively. However, the DLU model setup simulated the detailed biophysical processes better during the calibration period. Both model setups equally predicted daily streamflow during the validation period. Better performance was obtained while applying the DLU model setup because of improved representation of the dynamic watershed characteristics such as curve number (CN2), overland Manning's (OV_N), and canopy storage (CANMX). Expansion of agricultural land use by 11.1% and the reduction of forest cover by 2.3% during the period from 1985 to 2015 increased the average annual surface runoff and peak flow by 11.6 mm and 2.4 m(3)/s, respectively and decreased the evapotranspiration by 5.3 mm. On the other hand, expansion of shrubland by 1% decreased the surface runoff by 1.2 mm and increased the evapotranspiration by 1.1 mm. The results showed that accounting DLU into the SWAT model simulation leads to a more realistic representation of temporal land use changes, thereby improving the accuracy of temporal and spatial hydrological processes estimation.