IMPACT OF DIFFERENT SPATIAL AND PRECIPITATION INPUT DATA ON SWAT-DERIVED CATCHMENT FEATURES

Abstract

Computer-based mathematical models are used in water management research to represent ecological problems, simulate different processes, deal with and support solving such problems. Such studies require an interdisciplinary approach that is both user-friendly and comprehensive to integrate all of the processes that occur in nature. The use of available hydrological models requires a model setup with a reasonable level of data quality and quantity to represent catchment features to emphasize the distinctive local character of ecosystems. One of the models that comply with such requirements is Soil and Water Assessment Tool (SWAT). This paper aims to address the possible impact of different spatial and precipitation input data on the catchment features by using SWAT for the case study of the floodplain located within the Danube River Basin in Serbia. The objective was to evaluate the impact of digital elevation models, land use, and soil types with different resolutions, respectively (1) SRTM and ASTER (30 m), TanDEM-X (12,5 m) resolution, (2) CORINE and GlobCover land use databases, and (3) FAO/UNESCO world soil map and digitized soil map of Vojvodina Province, on catchment delineation. The research was conducted alongside the analysis of precipitation using data from CFSR, CarpatClim, and national yearbooks. Regarding spatial data, the results indicate that high-resolution data for this area needs to be adjusted, while the ASTER layer is suitable at an acceptable level for further modelling in SWAT. Interpolated precipitation data are better to use due to their higher resolution (10km) and heterogeneous distributions of rain gauge stations.