Daily water balance modeling is important for both hydrological science and water resources management. The majority of daily water balance models express the output fluxes (including runoff and evapotranspiration) as independent functions of soil moisture store(s), and as such they involve a large number of parameters, the estimation of which is a major challenge. In this paper, following a downward approach, we develop a parsimonious daily water balance model, DPWM, based on the Proportionality Hypothesis, which assumes that a proportional relation exists among different hydrological components, which contributes to model parsimony. DPWM includes just 5 parameters, 3 of which are taken up to capture water balance partitioning and 2 for runoff routing. With daily precipitation and potential evapotranspiration as inputs, DPWM invokes Proportionality Hypothesis to generate four main daily hydrological components, namely evapotranspiration, storage change, fast runoff and slow runoff. DPWM was applied to simulate daily streamflow (Q) and evapotranspiration (ET) in 671 US catchments covering a wide range of climates and landscapes, and showed catchment-average KGE(Q), KGE(1/Q) and KGE(ET) of 0.64, 0.68, 0.75 respectively. DPWM is seen to capture temporal variability of streamflow in rainfall-dominated humid catchments particularly well. DPWM shows comparable performance with the MARRMoT version of the XINANJIANG model, a traditional 12-parameter rainfall-runoff model that shows top performance amongst over 30 conceptual rainfall-runoff models. DPWM’s performance on KGE(1/Q) is even higher than that of XINANJIANG model. These results provide evidence that the Proportionality Hypothesis is useful for modeling daily water balance across a wide range of climates and landscape characteristics.

中文翻译：

---

基于比例性假设的简约每日水平衡模型


日常水平衡建模对于水文科学和水资源管理都非常重要。大多数日水分平衡模型将输出通量（包括径流和蒸散）表示为土壤储水量的独立函数，因此它们涉及大量参数，估计这些参数是一个重大挑战。在本文中，我们遵循向下的方法，基于比例假设开发了一个简洁的每日水平衡模型 DPWM，该假设不同水文成分之间存在比例关系，这有助于模型简洁。DPWM 仅包含 5 个参数，其中 3 个用于捕获水平衡分区，2 个用于径流路由。以日降雨量和潜在蒸散量为输入，DPWM 调用比例假设来生成四个主要的日水文组成部分，即蒸散、储存变化、快速径流和慢速径流。DPWM 用于模拟 671 个美国集水区的日径流 （Q） 和蒸散量 （ET），涵盖广泛的气候和景观，并显示流域平均 KGE（Q）、KGE（1/Q） 和 KGE（ET） 分别为 0.64、0.68、0.75。DPWM 被认为可以特别好地捕捉以降雨为主的潮湿集水区中径流的时间变化。DPWM 的性能与西南江模型的 MARRMoT 版本相当，后者是一种传统的 12 参数降雨径流模型，在 30 多个概念性降雨径流模型中表现出最佳性能。DPWM 在 KGE（1/Q） 上的性能甚至高于湘南江模型。 这些结果提供了证据，证明比例性假说可用于在各种气候和景观特征下对日常水平衡进行建模。