Prediction of mean runoff in a small ungauged basin using raster climatological data

An analysis of mean annual runoff was conducted for a small ungauged catchment of the Tulovska River in the South Morava River Basin. Langbein?s method, Turk?s method and regional regression models were applied for obtaining the mean runoff. Long-term mean temperature, annual precipitation and the mean catchment elevation are used as input data. The estimations were conducted using various sources of climatological input data and the results were then compared. Precipitation and temperature data were derived, in a 10 km x 10 km resolution raster data format, from the digital data repository of the CarpatClim project. Point measurements of temperature and precipitation by the Republic Hydrometeorological Service of Serbia were also used as input data and compared with the raster data. The difference between the raster and point meteorological data were attributed to the effects of terrain elevation spatial variability, not captured in the raster data. The analyses showed that the predicted mean runoff can differ up to 50%, depending on the chosen method and the input data.