Chris Soulsby’s research while affiliated with University of Aberdeen and other places

During the last decade, tracer-aided hydrological models (TAMs) have been applied in numerous studies and have successfully evolved for different purposes. Such studies confirmed the value of tracer data in hydrological modeling, offering insights into internal storages, water sources, flow pathways, mixing processes, and water ages, which cannot be derived from hydrometric data alone. The direct coupling of tracers into flux tracking and the water balance of hydrological models can reduce model uncertainty through increased hydrological and biogeochemical process knowledge. More specifically, such models can simultaneously capture the celerity of hydrological responses with the velocities (and age) of water particles. As a result of the increased availability of high-resolution tracer data characterizing hydrological functioning across the Critical Zone and entire landscapes, together with the rapid improvement of computing capacity, four major advances reshaped the capability of TAMs, which we review in this paper: (1) enhanced representation of spatial heterogeneity, (2) more explicit conceptualization of ecohydrological partitioning, (3) application to larger catchment scales, and (4) incorporation of non-conservative tracers in coupled water quality modeling. However, persistent challenges have also emerged, particularly in relation to data acquisition, mismatches between the information content of data and scale of application, uncertainties in model structures, as well as adaptation of multi-criteria calibration. In this review, recent advances and remaining challenges of TAMs have been summarized and discussed with a particular focus on conservative tracers and flux tracking models.

The Limits of Acceptability approach has been demonstrated to be an effective conditioning tool due to its capacity to consider epistemic uncertainty. However, its application faces two challenges—the low efficiency when random sampling is used and the difficulty in setting limits prior to calibration. Here an algorithm DREAM(LoAX) was developed and added to GLUE framework. As an extension of DREAM(LoA) of Vrugt and Beven (2018), https://doi.org/10.1016/j.jhydrol.2018.02.026, it evaluates model performance based on limit boundaries, thus inherits the merits of the GLUE framework (explicit consideration of epistemic errors). Moreover, the importance of initial choice of limits is strongly reduced by allowing iterative evolution of limits based on historical model performance. By testing a series of examples (including a high‐dimensional numeric example, a single‐objective hydrological example, and a multi‐objective hydrological example) with or without error‐free assumption using synthetic or real observations, the search capacity of DREAM(LoAX) to locate acceptable models is demonstrated. The algorithm also shows comparable efficiency to DREAM and DREAM(LoA). More importantly, it provides real‐time diagnostic information regarding when (at which timestep), where (for which objective), and how (to which direction and to which extent) the model fails when uncertainty is pronounced, allowing potential uncertainty sources in the data or flaws in the model structure to be identified. In this context, DREAM(LoAX) is not only a useful conditioning tool, but also a diagnostic and learning tool for development of improved modeling.

Ecohydrological partitioning of rainfall into different sources of evaporated and transpired water is crucial to quantify water balance impacts from land cover change. However, resolving ecohydrological partitioning into component fluxes can be ambiguous and uncertain, even where detailed, small-scale measurements are available. To constrain ecohydrological fluxes at the scale of an individual tree in an urban setting, we combined hydro-meteorological, sap flow, soil water and high-resolution in situ plant xylem and atmospheric vapor stable isotope measurements over the growing season from April to October 2022. These data were integrated with parsimonious tracer-aided conceptual modeling. The data helped isolate temporal patterns of shifting preferential fractionation in xylem and atmospheric vapor from δ 18 O to δ 2 H mainly depending on air temperature and relative humidity. Modeling high-resolution in situ isotope data revealed the dominant local influence of interception, soil evaporation and transpired water sources on atmospheric vapor particularly during dry periods , whereas wet periods were driven by more variable non-local moisture sources. Additionally, modeling tree water storage did not explain the highly variable and more depleted xylem isotope data compared to enriched and fractionated soil water. Despite volumetrically constrained (within transpiration measurement uncertainty bounds) ecohydrological partitioning, the atmospheric vapor isotope data showed that fine-scale variations of interception and soil evaporation vapor sources can have nuanced impacts on the atmospheric vapor mixture. The comparison of a more complex conceptualization of modeled soil storages (three soil storages) with a minimalist two-storage model indicated the notoriously difficult isotopic discrimination of root water uptake depths. Nonetheless, the combination of soil moisture, transpiration and high-resolution in situ isotope measurements with modeling helped enhance our understanding of plot-scale vegetation-mediated urban hydrological processes.

Urban green spaces (UGS) provide essential ecosystem services (ES), for example, precipitation infiltration for flood mitigation, transpiration (Tr) for local atmosphere cooling and groundwater recharge (Gr) for drinking water provision. However, vegetation type impacts the ecohydrological partitioning of incoming precipitation and therefore ES provision, whilst flux rate potential is different in disparate hydroclimates. Consequently, paired studies in different hydroclimates are useful to understand similarities and differences in vegetation controlled ecohydrological partitioning to effectively guide UGS management. We simultaneously undertook sub‐daily soil moisture measurements beneath three contrasting urban vegetation types (grass, shrub, mature tree) between 01/01/2021 and 31/12/2023 for an inter‐comparison of an energy‐limited Scottish and a moisture‐limited region of Germany. These data were integrated with hydroclimatic and sapflux data in the EcoHydroPlot model to constrain estimates of ecohydrological fluxes. Soil moisture data showed clear effects of the contrasting hydroclimates, with high and low VWC values in Scotland and Germany, respectively, whilst evapotranspiration potential was ~50% greater in Germany. Consequently, ecohydrological functioning and flux rates were fundamentally different, with Tr dominant in Germany and Gr dominant in Scotland. However, vegetation cover was shown in both countries to be a key control on urban ecohydrological partitioning with grass encouraging Gr, contrasting to evergreen shrubs in Scotland and mature trees in Germany elevating Tr. In Germany, impacts to hydrological functioning due to low soil VWC were marked with the mature trees high Tr rate shutting down Gr for the majority of the study period. The German site also showed greater hydrological functioning susceptibility to inter‐annual hydroclimatic variability with all fluxes heavily suppressed during the 2022 drought. In contrast, the high VWC in Scotland provided some buffer against ongoing negative rainfall anomalies. Overall, the study indicated the importance of diverse UGS vegetation cover to encourage contrasting ecohydrological fluxes.

Increasing drought frequency and severity from climate change are causing streamflow to become increasingly intermittent in many areas. This has implications for the spatio-temporal characteristics of water quality regimes which need to be understood in terms of risks to the provision of clean water for public supplies and instream habitats. Recent advances in sensor technology allow reliable and accurate high-resolution monitoring of a growing number of water quality parameters. Here, we continuously monitored a suite of water quality parameters over 3 years in an intermittent stream network in the eutrophic, lowland Demnitzer Millcreek catchment, Germany. We focused on the effects of wetland systems impacted by beaver dams on the diurnal, seasonal and inter-annual variation in water quality dynamics at two sites, upstream and downstream of these wetlands. We then used the data to model stream metabolism. Dissolved oxygen and pH were higher upstream of the wetlands, while conductivity, turbidity, chlorophyll a and phosphorous concentrations were higher downstream. We found clear diurnal cycling of dissolved oxygen and pH at both sites. These dynamics were correlated with seasonal hydroclimatic changes and stream metabolism, becoming increasingly pronounced as temperatures increased and flows decreased in spring and summer. Upstream of the wetlands this corresponded to the stream rapidly becoming increasingly heterotrophic as modelled Gross Primary Production (GPP) was exceeded by Ecosystem Respiration (ER). Downstream, where GPP was lower, the stream was usually strongly heterotrophic and prone to increasingly hypoxic conditions (i.e., insufficient oxygen) before streamflow ceased in summer. This coincided with lower velocities and deeper channels in beaver impacted areas. Seasonal and inter-annual variations in water quality were found to mainly correlate with hydroclimatic factors (particularly temperature) and their influence on streamflow. This study highlights that heterotrophy and hypoxia in lowland rivers in central Europe is an important seasonal feature of intermittent streams where agricultural landscapes continue leaching nutrients. These insights contribute to an evidence base for understanding how climate change will affect the quantity and quality of rural water resources in intermittent lowland streams with wetlands where the presence of beavers requires management responses. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Over the last 20 years, we have dramatically improved hydrometeorological data including isotopes, but are we making the most of this data? Stable isotopes of oxygen and hydrogen in the water molecule (stable water isotopes – SWI) are well known tracers of the global hydrological cycle producing critical climate science. Despite this, stable water isotopes are not explicitly included in influential climate reports (e.g. Intergovernmental Panel on Climate Change, IPCC) except for paleoclimate reconstructions. Continuous developments in modelling approaches have now made isotope-enabled modelling of climate and hydrology more powerful and easier to perform, reducing prediction uncertainty and providing more robust simulations. We argue that it is time to incorporate stable water isotopes and isotope-enabled modelling into mainstream hydroclimatic forecasting with the prospect of vastly improving climate change predictions and evidence.

The dimensionality of parameters and objectives has been increasing due to the accelerating development of models and monitoring network, which brings potential challenges for calibration. In this study, two common philosophies for multi‐objective optimisation in hydrology (the use of aggregated scalar criterion or vector functions) were revisited with different sampling strategies: (a) random sampling, (b) DiffeRential Evolution Adaptive Metropolis (DREAM as an example of an aggregated scalar function), and (c) Non‐Dominated Sorting Genetic Algorithm II (NSGA‐II as Pareto‐based multi‐objective optimisation). By testing the ability of algorithms to simultaneously capture soil moisture and soil water isotopes at three depths under four vegetation covers, we found random sampling performed poorly in matching observations due to its inability to explore high‐dimensional parameter space. DREAM, in contrast, could provide efficient parameter convergence with informal likelihood functions, but the choice of formal likelihood function is difficult due to the lack of knowledge about model residuals, leading to poor performance. NSGA‐II is effective and efficient after aggregating objectives to ≤4, but failed when calibrating against all 24 objectives. Overall, both philosophies and all three approaches are challenged by increasing dimensionality, and it generally requires a degree of trial‐and‐error before achieving a successful calibration. This suggests the potential to explore a more flexible way to describe model residuals (e.g., by defining limits of acceptability). Alternatively, improvements could be made by using an ensemble of models to represent the system (instead of “best” model) given the average of a calibrated ensemble usually performed better than any individual model.

Tropical cyclones (TCs) are one of the major natural hazards to island and coastal communities and ecosystems. However, isotopic compositions of TC‐derived precipitation (P) in surface water (SW) and groundwater (GW) reservoirs are still lacking. We tested the three main assumptions of the isotope storm “spike” hypothesis (sudden spikes in isotopic ratios). Our database covers 40 TCs and is divided into recent (N = 778; 2012–2023) and archived (N = 236; 1984–1995) rainfall isotope observations and SW/GW isotope monitoring (N = 6013; 2014–2023). Seasonal rainfall contribution from TCs ranged from less than 1% to over 54% (4% on average) between 1984 and 2023. Mean δ¹⁸O compositions across TCs domains were significantly lower than the regional (noncyclonic) δ¹⁸O mean (−5.24 ± 4.27‰): maritime (−6.29 ± 3.28‰), coastal (−7.78 ± 4.28‰), and inland (−9.80 ± 5.18‰) values. Coastal and maritime TC convection resulted in large rainfall amounts with high isotope compositions. This could bias past climate reconstructions toward unrealistic drier conditions. Significant δ¹⁸O and d‐excess differences were found between storm intensities. P/SW and P/GW isotope ratios revealed the rapid propagation of TC excursions in freshwater systems. Our findings highlight the potential of TC isotope observations for diagnosing intensity and frequency in paleoproxies beyond idealized TC models.