ABSTRACT

The storage function model (SFM) is a rainfall–runoff model widely used in Korea for flood forecasting and warning systems. The traditional equations for parameter estimation are based on observed data under limited conditions, and so it is difficult to sensitively consider the factors affecting actual runoff. The aim of this study is to develop synthetic equations for SFM parameter estimation that are useful for ungauged watersheds, through 35 000 applications of the distributed kinematic wave model (KWM) for virtual watersheds. The parameter estimation equation developed in this study was applied to 101 observed events of 16 basins in Korea. We obtained satisfactory results from the application of a synthetic equation to actual basins. That is to say, the peak flood volume simulated by the equation was comparable with the observed volume. The relative error (RE), mean absolute percent error (MAPE), and percent bias (PBIAS) were used as criteria to evaluate the performance of the equation. For the results, RE and MAPE were within 10%, and PBIAS was within 15%.

摘要

存储功能模型（SFM）是一种降雨径流模型，在韩国广泛用于洪水预报和预警系统。传统的参数估算公式基于有限条件下的观测数据，因此很难敏感地考虑影响实际径流的因素。这项研究的目的是通过35,000个虚拟流域的分布式运动波模型（KWM）的应用，开发出用于SFM参数估计的合成方程，这些方程对于未流域的流域是有用的。本研究开发的参数估计方程被应用于韩国16个盆地的101个观测事件。通过将合成方程应用于实际盆地，我们获得了令人满意的结果。也就是说，由方程式模拟的洪峰峰值与观测到的洪峰相当。相对误差（RE），平均绝对百分比误差（MAPE）和百分比偏差（PBIAS）被用作评估该方程式性能的标准。对于结果，RE和MAPE在10％之内，而PBIAS在15％之内。