Prediction of low-flow indices in ungauged basins through physiographical space-based interpolation

Low-flow estimation is of great importance to assess the availability of water resources. This study investigates the applicability of physiographical space-based interpolation techniques for the prediction of low-flow indices in ungauged basins (basins for which discharge observations are sparse or unavailable). The study considers 51 catchments located in a wide region of central Italy, for which several geomorphological and climatic descriptors are available. The analysis applies both deterministic and geostatistical techniques for interpolating low-flow indices in the physiographical space. A jack-knife cross-validation procedure is applied in order to quantify the accuracy of each technique when it is applied to ungauged basins. The results of the study show that physiographical space-based interpolation is a viable approach for estimating low-flow indices in ungauged basins and geostatistical techniques outperform deterministic techniques.