Hydrological Processes最新文献

This study investigates canopy storage capacity (S)—the maximum amount of rainfall stored in a forest canopy—using field data from 648 storm events across 15 diversely vegetated sites of the United States National Ecological Observatory Network (NEON). We compared these S values to remotely sensed and ground-based vegetation metrics of leaf area index (LAI), mean canopy height, and biomass density. We found that S values were generally constrained within a narrow range (site-level medians averaged 0.9 mm and ranged from 0.5 to 1.5 mm) relative to larger ranges in biomass, height, and LAI. Vegetation height partially explained this variation; other tested vegetation-structure variables did not significantly explain S. Significant seasonal variations in S were also observed in a subset of sites. Within sites, S often varied with vegetation traits, but those relationships were inconsistent across sites. We also found that specific storage capacity (S_spec; mm of S per unit LAI, as inferred from ground-based LAI measurements), with site-level median values averaging 0.6 and ranging from 0.3 to 1.7, was not effectively explained by variations in vegetation structure. The limited ability for vegetation metrics to explain S or S_spec may be due to challenges in a single vegetation metric capturing the diverse traits that explain canopy storage, or due to scale mismatches in throughfall measurements relative to available vegetation-structure data. Regardless of reason, our findings contrast with common assumptions and modelling conventions and do not suggest that using vegetation-structure metrics to scale S is more appropriate than assuming S to be constant across forested sites.

Rainfall-runoff modelling is pivotal in hydrology, yet reliance on natural rainfall poses significant challenges since it is so erratic. Rainfall simulators provide a controlled environment; however, replicating natural rainfall accurately in simulators remains challenging due to factors like raindrop velocity, terminal speed, spray dynamics, kinetic energy and droplet size. In this study, these parameters were comprehensively evaluated to ensure the rainfall simulator effectively replicates natural rainfall conditions. A series of laboratory experiments were conducted using this simulator, involving 12 distinct rainfall intensities and three rainfall durations. Importantly, the study extends beyond constant rainfall events by examining moving rainfall patterns, categorised into upstream (u/s), downstream (d/s), extreme upstream (Ex u/s) and extreme downstream (Ex d/s) rainfall. The influence of these dynamic patterns on surface flow and hydrograph components was systematically analysed. Special emphasis was placed on the rising limb of the hydrograph, which captures the initial response of runoff to rainfall within a catchment and its relationship with rainfall characteristics. The regression-based model effectively captures rainfall-runoff interactions, with Nash-Sutcliffe Efficiency (NSE) and coefficient of determination (R²) exceeding 0.8 and Percent Bias (PBIAS) below 10% in both calibration and validation. Nonlinear regression proved to be the most reliable prediction approach. The study elucidates the relationship between rainfall characteristics and runoff dynamics under moving rainfall conditions. It provides critical insights into the relationship between rainfall patterns and runoff behaviour, contributing to advancements in hydrological modelling and water resource management.

Adequate quantification of precipitation and its spatiotemporal variability is crucial for understanding the physical and biological processes within a watershed. Mountain watersheds pose particular challenges due to strong spatial heterogeneity in precipitation and typically sparse in situ monitoring networks. The increasing availability of gridded precipitation products can help address these limitations, but their reliability at local or sub-regional scales remains difficult to assess. This study proposes a novel, process-based approach that incorporates daily streamflow data to evaluate the performance of four widely used gridded precipitation datasets (CHIRPS-v2, MSWEP-v2.8, GPCC-v2022 and TerraClimate). Five key watersheds in central-western Argentina (ca. 30°–37° S) serve as case studies. The evaluation framework is based on five process-informed expectations derived from the region's climate and topography: (a) most annual precipitation should fall as snow during winter (April–September); (b) a strong positive relationship should exist between winter precipitation and summer streamflow; (c) interannual variability of precipitation should exceed that of streamflow due to basin-scale damping effects; (d) runoff coefficients should be statistically lower than unity, reflecting mass conservation; and (e) winter precipitation should be concentrated at higher elevations. We apply simple non-parametric statistical tests to evaluate how well each dataset meets these expectations. A comparative assessment identifies the most reliable product for each watershed. Our findings show that MSWEP and TerraClimate perform best overall, particularly in capturing total precipitation and its seasonality. Other datasets fail to reproduce key hydrological signals, likely due to a lack of physically based inputs (e.g., reanalysis). Overall, this process-based, catchment-integrative evaluation offers a promising framework for assessing precipitation products in other snow-dominated mountain regions with limited ground observations, provided that the dominant hydroclimatic processes are well understood.

Adequate quantification of precipitation and its spatiotemporal variability is crucial for understanding the physical and biological processes within a watershed. Mountain watersheds pose particular challenges due to strong spatial heterogeneity in precipitation and typically sparse in situ monitoring networks. The increasing availability of gridded precipitation products can help address these limitations, but their reliability at local or sub-regional scales remains difficult to assess. This study proposes a novel, process-based approach that incorporates daily streamflow data to evaluate the performance of four widely used gridded precipitation datasets (CHIRPS-v2, MSWEP-v2.8, GPCC-v2022 and TerraClimate). Five key watersheds in central-western Argentina (ca. 30°–37° S) serve as case studies. The evaluation framework is based on five process-informed expectations derived from the region's climate and topography: (a) most annual precipitation should fall as snow during winter (April–September); (b) a strong positive relationship should exist between winter precipitation and summer streamflow; (c) interannual variability of precipitation should exceed that of streamflow due to basin-scale damping effects; (d) runoff coefficients should be statistically lower than unity, reflecting mass conservation; and (e) winter precipitation should be concentrated at higher elevations. We apply simple non-parametric statistical tests to evaluate how well each dataset meets these expectations. A comparative assessment identifies the most reliable product for each watershed. Our findings show that MSWEP and TerraClimate perform best overall, particularly in capturing total precipitation and its seasonality. Other datasets fail to reproduce key hydrological signals, likely due to a lack of physically based inputs (e.g., reanalysis). Overall, this process-based, catchment-integrative evaluation offers a promising framework for assessing precipitation products in other snow-dominated mountain regions with limited ground observations, provided that the dominant hydroclimatic processes are well understood.

Water scarcity and uneven rainfall continue to affect many parts of Yunnan Province, despite its generally high rainfall. This study used Geographic Information Systems (GIS) and three Multi-Criteria Decision-Making (MCDM) models: Multi-Influence Factor (MIF), Analytical Hierarchy Process (AHP) and Weighted Linear Combination (WLC) to map rainwater harvesting (RWH) suitability. Eleven parameters were included, covering topographic, hydrological, environmental, and accessibility factors: rainfall, slope, elevation, soil type, land use/land cover, drainage density, lineament density, TWI, distance to rivers, distance to roads, and distance to built-up areas. Across all three models, the northern and north-western mountainous regions were consistently mapped as very low to low suitability, dominated by steep slopes and fast runoff. Under the MIF model, very low (15.42%) and low suitability (33.89%) together covered nearly half the province, while moderate suitability was highest at 34.23%, followed by high (11.95%) and very high (4.50%). The AHP-based WLC model classified 12.78% as very low, 33.79% as low, 30.24% as moderate, 20.83% as high and 2.36% as very high. The standalone WLC model showed 25.16% very low, 29.40% low, 25.47% moderate, 15.56% high and 4.41% very high suitability. Despite numerical differences, all models consistently identified the central, southern and southeastern regions as the main RWH hotspots, driven by gentler topography and favourable moisture conditions. These results show that a significant portion of Yunnan, especially the moderate- and high-suitability zones, can support small-scale RWH structures such as check dams, contour trenches and percolation tanks. The strong agreement among the three models increases confidence in the mapped zones.