Abstract
This study attempted to improve a methodology for estimating watershed-scale storage changes from hourly discharge data and to verify its effect in the upper Abukuma River watershed in Japan. The previous methodology separate hydrographs into several discharge sub-components by a filter-separation method, and then it explored relationships between discharge sub-component Q and watershed-scale storage S, assuming power-law relationships between derivative of discharge sub-components dQ/dt and Q. The present study employed linear relationships between Q and S to be theoretically consistent with the filter-separation method in which linear relationships are assumed between Q and S. Based on this theoretical revision, we re-estimated watershed-scale storages to compare with those estimated by the previous methodology. As the result, we found instantaneous increases of storage after rainfall events become smaller and storage change become smoother. In addition, we confirmed the present methodology can estimates more realistic storage than previous one in terms of event-scale water balance.
