Performance Evaluation of TANK Model Parameter Optimization According to Data Intervals

Abstract. For the establishment of water quality management plans or design of hydraulic structures, long-term runoff data of watersheds is required. However, runoff data is insufficient in some watersheds, and the runoff is estimated through empirical formulas, specific discharge measurement or simulation using hydrologic models in these watersheds. Specific discharge measurement with extrapolation is not logical in estimating the runoff of small watersheds, and empirical formulas is also inappropriate for the estimation of daily runoff data. Among the runoff estimation methods, simulation using hydrologic model such as tank model is the most reasonable method for estimating daily runoff. In the case of the tank model, the parameter can be estimated through the parameter estimation formula or optimization methods. If the parameters of a hydrologic model are calibrated by optimization, the simulation result is similar with actual survey runoff, but the survey runoff data is necessary for optimization. In the measurement network for water quality planning, it‘s common to measure at intervals of 8 days to 1 month instead of daily measurements. There is not sufficient research on whether accurate data can be estimated by applying the runoff data with interval to the calibration of the long-term runoff model. Therefore, the study carried out the evaluation of the applicability of tank model calibration with periodic data by calibrating modified tank model with periodic survey runoff data and analyzing simulated result. The results of this study can be used to establish the hydrological monitoring plans and to utilize the runoff data with periodicity.