Parameter interdependence and uncertainty induced by lumping in a hydrologic model

摘要: [1] Throughout the world, watershed modeling is undertaken using lumped parameter hydrologic models that represent real-world processes in a manner at once abstract, but nevertheless relies on algorithms reflect and parameters hydraulic properties. In most cases, values are assigned to of such through calibration against flows outlets. One criterion by which utility model success process judged realistic this process. This study employs regularization theory examine relationship between corresponding It demonstrates any kind lumping or averaging can induce substantial amount "structural noise,'' devices as Box-Cox transformation autoregressive moving average (ARMA) residuals unlikely render homoscedastic uncorrelated. Furthermore, estimated for properties after they named often contaminated greater lesser degree do not purport all. As result, question how rigidly should be bounded during estimation still an open one.