Abstract. Understanding the propagation of prolonged meteorological drought helps solve the problem of intensified water scarcity around the world. Most of the existing literature studied the propagation of drought from one type to another (e.g., from meteorological to hydrological drought) with statistical approaches; there remains difficulty in revealing the causality between meteorological drought and potential changes in the catchment water storage capacity (CWSC). This study aims to identify the response of the CWSC to the meteorological drought by examining the changes of hydrological-model parameters after drought events. Firstly, the temporal variation of a model parameter that denotes that the CWSC is estimated to reflect the potential changes in the real CWSC. Next, the change points of the CWSC parameter were determined based on the Bayesian change point analysis. Finally, the possible association and linkage between the shift in the CWSC and the time lag of the catchment (i.e., time lag between the onset of the drought and the change point) with multiple catchment properties and climate characteristics were identified. A total of 83 catchments from southeastern Australia were selected as the study areas. Results indicated that (1) significant shifts in the CWSC can be observed in 62.7 % of the catchments, which can be divided into two subgroups with the opposite response, i.e., 48.2 % of catchments had lower runoff generation rates, while 14.5 % of catchments had higher runoff generation rate; (2) the increase in the CWSC during a chronic drought can be observed in smaller catchments with lower elevation, slope and forest coverage of evergreen broadleaf forest, while the decrease in the CWSC can be observed in larger catchments with higher elevation and larger coverage of evergreen broadleaf forest; (3) catchments with a lower proportion of evergreen broadleaf forest usually have a longer time lag and are more resilient. This study improves our understanding of possible changes in the CWSC induced by a prolonged meteorological drought, which will help improve our ability to simulate the hydrological system under climate change.

This study presents a new attempt of applying the hydrological model SWAT to the Three Gorges watershed in China for addressing its non-point source (NPS) pollution control issues. The model was calibrated and validated using the monitoring data collected during 2002-2008, and satisfactory values of R 2 and E NS (Nash-Suttclife Efficiency) were obtained. The calibrated SWAT model was then used to simulate 6 different land use scenarios for investigating the effects of each scenario on the non-point source (NPS) pollution control in the watershed. Six scenarios were designed with distinct land use focuses and include five newly-designed scenarios (Q1-Q5) representing 5 different land use alternatives and a baseline scenario (Q6) representing the land use pattern the watershed had in 2005. It was identified that the farmland is the dominant contributor to the NPS pollution in the watershed in terms of yields of sediment, TN and TP. If the farmland is changed to the woodland, grassland or shrubland, a better control and reduction over the NPS pollution could be achieved. This study provides a good understanding of the interactions between different land use patterns and the NPS pollution control for decision-makers to make sound decisions. Changing the land use pattern and implementing alternative management practices could help reduce the non-point source pollution effectively and thus play a significant role in improving reservoir water quality of the watershed.

Reliable partitioning of precipitation (P) into runoff (Q) and evapotranspiration (E a ) is crucial for hydrological research and application, especially for regions with scarce data (Yang et al., 2007;Zhang et al., 2018). Understanding the controls of catchment properties on hydrological partitioning helps achieve reliable hydrology partitioning, but remains a challenging task (Sinha et al., 2020). The Budyko framework has been widely used to establish the relationship between evaporative ratio (i.e., E a /P) and relative availability of water and energy (i.e., aridity index (AI), the ratio between long-term E P and P, AI = E p /P) (Budyko, 1974;Cheng et al., 2011). In the widely used Budyko framework proposed by Fu (1981) (i.e., Fu's equation), the controls of other secondary factors on partitioning are lumped into a landscape parameter (ω) that includes intra-annual climate variability, soil, vegetation, and topography (Cheng et al., 2021;Fu, 1981;Zhang et al., 2004). Note that Fu's equation is further explained in Section 2.1. Parameterizing ω with catchment properties not only improves the simulation accuracy of Fu's equation (Greve et al., 2015), but also reveals the controls of climate, physiography, and vegetation on hydrological partitioning (Abatzoglou & Ficklin, 2017). However, current understanding of controls on hydrological partitioning are still very limited, and building a physically-based relationship between ω and the control factors is difficult due to the complex (nonlinear) interactions between climate and catchment processes (

No abstract