ABSTRACT Numerous rainfall‐runoff event studies exist and focus on collecting different catchment response variables, from streamflow to catchment storage and solute concentrations. However, few to no studies provide multi‐faceted classifications and typologies of events using multiple types of response metrics. The goal of this study was to use rainfall event data, antecedent condition proxies and response metrics related to streamflow, groundwater and water quality dynamics to classify rainfall‐runoff events. A small, forested catchment was monitored and 30 rainfall‐runoff events were delineated. Individual storm hydrographs were used to calculate streamflow response metrics, water table data were used to estimate the magnitude and timing of groundwater responses and water quality statistics were analyzed for the rising and falling limbs of individual event hydrographs. Spatial variability in groundwater dynamics was observed, with the water table responding differently in different parts of the catchment. Nutrients typically displayed higher concentrations on storm hydrograph rising limbs, which is consistent with the flushing of shallow organic soils. Runoff generation for most events was primarily through shallow subsurface flow, with larger magnitude events also involving overland flow contributions. Classifications of rainfall‐runoff events, using a clustering algorithm, revealed very different clusters depending on the catchment response metrics considered. Classifications incorporating groundwater data notably led to more homogeneous classifications (with two clusters) compared to classifications using other catchment response metrics (i.e., rainfall event and antecedent conditions, streamflow, water quality) that yielded four clusters. Multi‐faceted classifications are therefore recommended whenever possible, as they can help infer the dominant drivers of variability of catchment response to rainfall.
ABSTRACT Operating riverbed infiltration galleries requires balancing water supply against groundwater‐level limits under seasonally varying hydrological conditions. This study develops a reusable monthly pumping rule and examines how seasonal groundwater responses help explain its structure. A calibrated HydroGeoSphere (HGS) model simulated the coupled river–aquifer–gallery system on the Hutuo River, northern China, under sampled inflow, rainfall and pumping conditions. An Extra Trees surrogate was coupled with regularized multi‐start simulated annealing to optimize monthly inflow thresholds, maximizing pumping volume over 30 historical annual sequences subject to a prescribed month‐end control‐well level. Application requires only the calendar month and anticipated monthly mean river inflow. HGS back‐testing over the same sequences met this target in 350 of 360 months, with a mean annual pumping volume of 20.57 × 10 6 m 3 . The optimized rule generally allowed higher pumping rates at lower inflows in summer than in winter. Monthly regression analysis of the HGS records estimated drawdown associated with pumping after adjustment for inflow, rainfall and initial control‐well level. At 1.2 m 3 /s, estimated drawdown ranged from 1.01 m in August to 2.75 m in January. Within‐month interaction analyses associated higher rainfall with smaller pumping‐induced drawdown, although statistical support varied among months. These model‐based findings suggest that seasonal pumping opportunities reflect differences in both water availability and the groundwater response to abstraction. Relating these responses to operating thresholds provides a hydrological basis for interpreting the rule and a framework for deriving site‐specific rules for comparable river‐connected intakes.
ABSTRACT Traditional snow depletion curves (SDC) fail to relate aggregated snow covered area (SCA) with the specific location of the snowline on a daily time step. This spatial uncertainty is compounded with temporal date uncertainty in cloud‐gap filling (CGF)—the process of infilling cloud covered pixels in daily remotely sensed snow cover—and limits their utility for long term forecasting. This study formalizes the spatial Snow Depletion Curve (sSDC): a dual‐model framework that anchors the temporal regression based depletion forecasts in the dynamic seasonally recurrent snow depletion pattern which is a stationary watershed characteristic. We present a dual stage architecture where Model 1 uses dynamic seasonally recurrent snow depletion patterns to resolve data uncertainty in CGF, which then serves as the foundation for Model 2—the sSDC—to convert SCA to τ * that creates a linear relationship between snow depletion and time. The ability of Model 2 to enable reliable forecasts results from the main finding of this work: that there is a linear relationship between day of year of melt out and the pixel relative timing variable ( τ *) of the dynamic seasonally recurrent snow depletion pattern. Validation across a 21‐year record demonstrates the framework's ability to track snow depletion ( R 2 = 0.97) and crucially predict final melt timing within ±0.6 days more than 2 months in advance. Testing over the Upper Snake River Basin—a large and critical water tower in the Western US that experienced a range of climactic variability—confirms the framework's robustness. Our dual‐model framework provides the high fidelity snowline on a daily timestep required for snowmelt runoff models, effectively translating static watershed patterns into dynamic, predictive temporal forecasts. The framework extracts inherent watershed characteristics and provides a promising method for application in mountainous, seasonal snow cover/snowmelt basins worldwide.
ABSTRACT The hydrological effects of Eucalyptus plantations on landscapes remain controversial. Previous studies have identified these plantations as a major contributor to soil‐water depletion, reduced groundwater recharge and lower baseflow relative to native forests. In Brazil, the area under Eucalyptus cultivation has expanded substantially in recent decades, raising concerns about potential ecohydrological impacts. Direct comparisons of the hydrological behaviour of Eucalyptus plantations and Atlantic Forest remnants under similar climatic and topographic conditions, and on soils within the same taxonomic class, are needed to clarify these effects, because comparable water balance totals may conceal distinct hydrological pathways in the two forest types. This paper compares water balance components between a mature Eucalyptus plantation and an adjacent semi‐deciduous Atlantic Forest remnant over four hydrological years (2013–2017). Both study sites share the same soil type (Dystroferric Red Latosol), climate and physiographic conditions. We monitored throughfall (TF), stemflow (SF), soil moisture to 1.0 m depth and meteorological data, enabling calculation of water balance, focusing on evapotranspiration (ET), canopy interception (CI), percolation (Pp) and soil water storage (SWS). The observation period comprised 1172 days for Eucalyptus and 1295 days for Atlantic Forest, with gross precipitation (GP) totalling 3659.4 and 4665.7 mm, respectively. ET accounted for 86.21% of GP in Eucalyptus versus 79.44% in Atlantic Forest. TF represented 71.52% (Eucalyptus) and 77.01% (Atlantic Forest) of GP, while SF displayed the greatest relative difference: 3.49% for Eucalyptus compared to 1.57% for Atlantic Forest. Pp from the 1.0 m soil layer was 12.72% (Eucalyptus) and 14.73% (Atlantic Forest) of GP. The Atlantic Forest exhibited higher water input to the forest floor but also greater percolation, leading to a more pronounced cumulative soil water deficit (−942 mm) by the end of the study period compared to the Eucalyptus site (−551 mm). This is likely due to the higher organic matter and biological activity in the Atlantic Forest, which enhance infiltration, accelerating deep drainage. Despite similar evapotranspiration, the Eucalyptus plantation exhibited lower soil water depletion than the adjacent native forest, which has been more severely affected by droughts, with implications for water resources management and land‐use planning in tropical regions.
ABSTRACT Xylem water isotope values are widely used to trace plant water sources. However, the sampling decisions that precede extraction are rarely fully reported in isotope studies, limiting reproducibility and cross comparisons. A survey of 42 researchers across five continents, complemented by a virtual community workshop, revealed substantial methodological diversity: only 51% maintain standard operating procedures, with variation in organ selection among respondents (branches 73.8%, stems 64.3%, roots 35.7%), sapwood‐heartwood separation (63.2% separate), bark removal (59% always) and transport conditions (56% ice packs vs. 39% ambient). From the surveyed researchers, 77.5% adapt sampling strategies to their extraction system, but these adaptations are rarely documented. This means that reported methods systematically underrepresent the process chain influencing isotopic outcomes. Given that varying circumstances and research objectives might preclude universal standardization, we introduce a Xylem Sampling Form: a structured metadata tool that makes sampling details explicit and comparable across studies without prescribing specific protocols, laying groundwork for future data‐driven evaluation of how upstream sampling choices propagate into isotopic uncertainty.
