水流
地表径流
环境科学
蒸散量
概念模型
校准
水文学(农业)
水文模型
流量(数学)
前期水分
径流曲线数
流域
气候学
计算机科学
数学
地质学
统计
岩土工程
几何学
地理
生物
数据库
地图学
生态学
摘要
A set of tests is proposed to determine the reliability of soil moisture accounting rainfall-runoff models for prediction beyond the calibration experience and for extreme event forecasts under dry and wet antecedent conditions. The tests are developed for small (less than 0.5 km2) hypothetical hillslope catchments whose hydrologic response to rainfall and evapotranspiration (ET) fluxes is modeled with an accurate physically based model (S-H) developed by Smith and Hebbert (1983). The tests are based on the assumption that the climatic and resulting S-;H streamflow time series are error free. The conceptual model explored is a modified version of the Sacramento model (used by the U.S. National Weather Service) reprogrammed for inputs at various time increments as small as 1 min. A Nelder and Mead (1965) direct search optimization scheme is also included in this modified model (denoted SMA) to assist in model calibration. The set of tests is designed to illuminate differences in modeled ET fluxes, streamflow time series, and the relative amounts of Hortonian and saturated overland flow and subsurface flow under a wide variety of climate and hillslope combinations. The test scheme ensures that the conceptual model is tested for a number of hydrodynamic conditions where flow responses range from primarily surface, to primarily subsurface, to relatively even amounts of surface and subsurface flow for humid and subhumid environments. It also provides direct comparisons of the physical (hillslope catchments) and conceptual model storages, and explicit evaluation of the combined influence of conceptual model structure and calibration errors. Specific model structure differences between S-H and SMA are detailed to elucidate any differences between model responses reported in the companion paper. Hydrologic responses are evaluated with statistical measures and graphical comparisons for time increments ranging from minutes to years.
科研通智能强力驱动
Strongly Powered by AbleSci AI