Science, despite its status as objective and searching for truth, is inherently a social activity. Research is conducted by scientists that collaborate, work in teams, get advised by their supervisor, get funding to study particular questions, know one another from earlier projects, and so on. In these social interactions, we together define what we consider important to study, or what we deem unimportant. This occurs at multiple levels: Funding agencies, for example, have the power to determine which research questions should be addressed. As hydrological modelling community, we have implicitly agreed that discharge is the main variable of interest - focusing on other fluxes or states is often presented as an advancement.  And at the modelling team level, we often (implicitly) agree on a modelling vision. From interviews with modellers from different teams, it for example became apparent that one team had the modelling vision to `keep things as simple as possible’. Given this vision, the modeller was inclined to choose the simpler parameterization over a more complex one to describe the same process. In another team, ‘scale invariance’ was considered more important, and therefore process representations were selected based on their scalability. Therefore, if we want to “advance” process-representation in models across spatial and temporal scales, the theme of this session, we should acknowledge that different researchers have different perceptions of what advancing comprehends, that there is no objective measure to define advancement, and that the first step probably is, that we have to clarify and express our modelling vision.

How to cite: Melsen, L.: The sociology of modelling: how we shape a perception together, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-21063, https://doi.org/10.5194/egusphere-egu24-21063, 2024.

The North American continent is home to a wide range of different hydro-climates. A key research gap is that there is currently limited understanding on the spatial variability of dominant hydrologic processes across these different hydro-climates. This limited understanding makes it difficult to select computational models that faithfully represent the hydrologic processes across such large domains, yet faithful representation of the different hydro-climatic behaviors is critical for accurate numerical prediction.

Here we present progress on a synthesis of dominant hydrologic processes under different combinations of climate-terrain-human forcings, engaging the broader community of catchment and Critical Zone scientists. The product from this research will be a continental “Hydrologic Mosaic”, with each landscape in the mosaic described by a set of perceptual and conceptual models. In this first step, we produce a continental map of hydrologic landscapes defined through the juxtaposition of hydroclimate, terrain and geology, and vegetation, land use, and management. We will define hydrologically meaningful indicators of terrestrial hydrology that concisely describe a location’s (i) hydroclimate (e.g., aridity, snow fraction, energy/water seasonality), (ii) topography and geology (e.g. depth to bedrock, soil porosity, topographic slope), and (iii) vegetation, land use and management (e.g., vegetation type, agricultural drainage, reservoir size), and calculate values for these indicators for each location on the continent. We then use clustering analysis to create a manageable number of representative hydrologic landscapes.

This work functions as a starting point in a wider project, where these initial hydrologic landscapes will be refined through interactions with regional experts. Together, we will develop perceptual (sketches and descriptions) and conceptual (box-and-arrow diagrams) of the dominant processes in each hydrologic landscape. These conceptual diagrams will contribute to large-domain modeling efforts by allowing targeted model selection and comparison efforts for each hydrologic landscape.

How to cite: Knoben, W., Clark, M., Fan, Y., McMillan, H., Read, J., and van Werkhoven, K.: Developing perceptual models of hydrologic behavior across the North American continent, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16034, https://doi.org/10.5194/egusphere-egu24-16034, 2024.

Brazil faced a severe water crisis during the mid-2010s, resulting in water scarcity and water rationing in various cities. The Belo Horizonte Metropolitan Region was seriously affected. It is located in the southeastern part of the country and home to roughly 5 million people (Costa et al., 2015) and mining industry. The region’s water supply relies on a complex and integrated system, which combines a water abstraction at the Velhas river and three reservoirs. One of these reservoirs, named Serra Azul, reached a minimum of only 5,4 % of its total capacity during the crisis. Here, we demonstrate tools for improving the water management in this area, by developing a hydrological model suitable for mountainous regions with tropical climates. Our case study was the Serra Azul reservoir’s well-gauged catchment. We selected 12 gauges that cover several head waters and rivers section in the 260 km² area.  We used these discharge data (3-5 years), and available static catchments' attributes (e.g. subsurface properties), to adapt a flexible framework for conceptual hydrological modeling. Hence, we identified a suitable model structure using SUPERFLEX (Fenicia et al., 2014). The findings show that by including soil type, lithology and land cover as explanatory variables in the model, we obtained significant improvements in performance, e.g. the correlation between the base flow index estimated for observed and simulated time-series increased from 0.40 to 0.76. We also accounted for groundwater contributions to the streamflow, modelling the relation between the percentage of porous aquifer within each catchment and its flow magnitude. Thereby, we improved the average NSE and timeseries correlation considerably.  Overall, we successfully set up a parsimonious hydrologic model for water resources management in a region that is notoriously difficult to predict, where anthropic activities such as mining and agriculture have a decisive impact on the water cycle.

References:

Costa et al. Caracterização e Quadros de Análise Comparativa da Governança Metropolitana no Brasil: análise comparativa das funções públicas de interesse comum (Componente 2)-RM do Rio de Janeiro (Relatório de Pesquisa). (2015). Rio de Janeiro: Institute for Applied Economic Research–Ipea.

Fenicia et al. "Catchment properties, function, and conceptual model representation: is there a correspondence?." Hydrological Processes 28.4 (2014): 2451-2467.

How to cite: de Sousa Matos, A. C., Höge, M., Medeiros do Nascimento, T. V., de Oliveira Corrêa, G., Oliveira e Silva, F. E., and Fenicia, F.: Model development for a water supply catchment in southeast Brazil, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-973, https://doi.org/10.5194/egusphere-egu24-973, 2024.

Hydrological modelling of ungauged catchments, which lack observed streamflow data, is an important practical goal in hydrology. A major challenge is to identify a model structure that reflects the hydrological processes relevant to the catchment of interest. Paraphrasing a well-known adage, “all models are wrong, but some model-mechanisms (process representations) might be useful.”

We extend a method previously introduced for mechanism identification in gauged basins, by formulating the Bayesian inference equations in the space of (regionalized) flow indices principal components and by accounting for posterior parameter uncertainty. We use a flexible hydrological model to generate candidate mechanisms and model structures. Then, we use statistical hypothesis testing to identify the "dominant" (more a posteriori probable) hydrological mechanism. We assume that the error in the regionalization of flow indices principal components dominates the error of the hydrological model structure.

The method is illustrated in 92 catchments from northern Spain. We treat 16 out of the 92 catchments as ungauged. We use 624 model-structures from FUSE (flexible hydrological model framework). The case study includes real data and synthetic experiments.

The findings show that routing is among the most identifiable processes, whereas percolation and unsaturated zone processes are the least identifiable. The probability of making an identification (correct or wrong), remains stable at ~25%, both in the real and in the synthetic experiments. In the synthetic experiments, where the “true” mechanism is known, we can evaluate the reliability, i.e., the probability of identifying the true mechanism when the method makes an identification. Reliability varies between 60%-95% depending on the magnitude of the combined regionalization and hydrological error. The study contributes perspectives on hydrological mechanism identification under data-scarce conditions.

Prieto et al. (2022) An Exploration of Bayesian Identification of Dominant Hydrological Mechanisms in Ungauged Catchments, WRR, 58(3), doi:10.1029/2021WR030705.

How to cite: Prieto, C., Le Vine, N., Kavetski, D., Fenicia, F., Scheidegger, A., and Vitolo, C.: Can we identify dominant hydrological mechanisms in ungauged catchments?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4385, https://doi.org/10.5194/egusphere-egu24-4385, 2024.

Abstract: Subsurface stormflow is the dominant runoff generation mechanism during certain flash flood events. The collapse characteristics and threshold behavior are critical to the transition from the slow runoff stage to the rapid runoff stage and the relevant study is not only essential for hydrology theory but also for flash flood disaster prevention. In recent years, efforts have been made to explore the common principles of hydrological processes under strong spatial heterogeneity, but findings from field experiments and numerical studies were difficult to apply to the modeling process. Furthermore, we have yet to develop a deeper understanding of the mechanisms for the impact of complex factors on subsurface stormflow and lack a comprehensive understanding of the formation mechanism of threshold.
This presentation discusses how we plan to address this research gap. Firstly, around the phenomenon of “burst-block-burst” in the subsurface stormflow runoff generation process, rainfall-runoff simulation experiments were carried out and factor analysis was conducted to determine the main influencing factors of subsurface stormflow runoff generation. The main influencing factors include soil texture factors, collapse state factors, initial state factors, and other factors, and the influence of these four types of factors decreases in turn. In the second step, we constructed a field hydrological station in the Huanggou Watershed located in Hubei Province, China, collected the rainfall-runoff data, and found that the subsurface stormflow process shows a three-stage-double-threshold behavior: the water storage stage, the initial flow stage, and the rapid flow stage. In the third step, synthesizing the main influencing factors, the three-stage double-threshold process was quantified. Further, the three-stage subsurface stormflow-based model (TSSM) was developed and applied to the Huanggou Hillslope and the Huanggou Watershed. The results show that TSSM performed well, with NSEs of 0.82 and 0.67 in the calibration and verification periods of the Huanggou slope, and NSEs of 0.76 and 0.74 in the calibration and verification periods of the Huanggou Watershed, respectively.
This study elucidated the collapse characteristics and threshold behavior of subsurface stormflow and developed an effective simulation model, which contributes to increasing our understanding of three-stage subsurface stormflow and is beneficial for hydrologists to develop more realistic hydrological models.

Keywords: Subsurface stormflow; Collapse characteristics; Threshold behavior; Three-stage subsurface stormflow mechanism; TSSM

How to cite: Song, Y., Zhang, Y., Li, S., and Yang, J.: Study on the collapse characteristics and threshold behavior of the subsurface stormflow mechanism, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2729, https://doi.org/10.5194/egusphere-egu24-2729, 2024.

Hydrologic modelling in the low-lying, flat prairie or arctic pothole regions is challenging because of variable contributing areas that modify the transformation of local runoff into streamflow. Most hydrological and land surface models fail in predicting prairie hydrology due to overlooking or inadequately representing the variable contributing area dynamics. In this study, we develop an open-source, model-agnostic version of a revised formulation of the recently developed Hysteretic Depressional Storage (HDS) model. This revised formulation accounts for the hysteretic relationship of pothole depressions and its effects on streamflow generation. The revised HDS model is implemented and tested with two different hydrological models of varying complexity (MESH and HYPE). The modified hydrological models are tested on a number of prairie pothole basins in Canada. Results show improved simulations of the streamflow response in the tested basins. Importantly, the modified models replicate the known hysteretic relationships between depressional storage and contributing areas in that region. The open-source HDS implementation approach is designed for use in hydrologic or land surface modelling systems, enabling improvements in simulating the complex hydrology and streamflow regimes globally.

How to cite: Stadnyk, T., Ahmed, M. I., Clark, M., and Pietroniro, A.: An Improved Representation of The Variable Contributing Area Concept in Hydrologic and Land Surface Models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13230, https://doi.org/10.5194/egusphere-egu24-13230, 2024.

Hydrological losses play a significant role in determining the runoff coefficient, influencing the amount of precipitation that ultimately contributes to surface runoff. These losses are influenced by various characteristics, including rainfall, the physical and geographical attributes of the watershed (such as slope and land use), and the soil moisture content within the watershed.

Here, we evaluate the effect of the variability of the loss function on the amount of simulated runoff and the runoff coefficient in specific watersheds in Iran. The investigation entails the assessment of two distinct conditions. First, the runoff coefficient is calculated under the assumption of a constant loss, utilizing the φ index. Second, a variable loss function, derived from a soil moisture algorithm, is employed to determine the runoff coefficient.

Our analysis shows that the assumption of a variable loss function yields more realistic results. When the variable losses are considered, the simulated runoff coefficient is closer to the observed values and determines the runoff coefficient for all months, including those characterized by low rainfall. The constant loss φ index exhibits two significant practical limitations: the overestimation of runoff coefficient values, and an inability to estimate runoff coefficient during months with low rainfall. The study emphasizes the need for a variable loss function to provide more realistic results. Our findings suggest that utilizing the variable loss function within the soil moisture algorithm produces more accurate results. Thus, the application for improved forecasting of rainfall and runoff processes is recommended.

How to cite: Eslami, Z., Abdollahi, K., and Kirchner, J.: Analyzing the fixed or variable effect of considering hydrological loss functions on the runoff coefficient in continuous modeling, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5982, https://doi.org/10.5194/egusphere-egu24-5982, 2024.

While there is little opposition to the idea that groundwater can play a central role in rainfall-runoff response, there is little consensus on how this should be modelled. Here, we present modelling supporting a recently advanced hypothesis that links groundwater-surface water interactions with observed shifts in the relationship between rainfall and runoff in south-east Australia.  While many approaches assume a direct and simplistic relationship between groundwater head and baseflow, evidence arising from a multi-year drought in Australia challenges traditional notions. GRACE and bore data during the Millennium Drought (1997-2010) show multi-year declines in groundwater storage, of such severity that we might expect the baseflow to cease, giving a flashier regime.  In reality, the shape of the hydrograph is mostly unchanged, but other changes abound: a year of given rainfall generates less runoff today than it did pre-drought (ie. shift in rainfall-runoff relationship). In other words, during and after the drought we see a hydrograph of similar shape to before, but diminished. While many Australian hydrologists are convinced that groundwater played a key role in this behaviour, it is unclear how these observations can be explained by existing hypotheses or modelling methods for groundwater surface-water interaction, and new paradigms are required.

The hypothesis explored here is that these observations can be explained by leaky bedrock in headwater catchments, which facilitates gradual groundwater export from upslope areas to downslope areas (within the same catchment).  Upslope areas subject to groundwater decline then see groundwater-surface water decoupling and reduced runoff. The hypothesised leakage is slow enough to go unnoticed during wetter periods, but in drier periods recharge may be too low to balance the export, leading to reduced groundwater levels and groundwater surface-water decoupling. When wetter conditions resume, the groundwater deficits may take a while to be replenished, delaying recovery of rainfall-runoff relationships (as observed in Australia).  In downslope areas, the drained water may contribute to streamflow, but may also be lost to evaporation and transpiration, particularly in drier catchments with flatter valley bottoms of alluvium or colluvium. In such catchments, the net effect of these processes is to allow groundwater originating from upslope to supplement evaporative budgets downslope rather than increasing streamflow.

We advance this hypothesis, firstly by presenting evidence of its applicability in south-east Australia; and secondly by building and testing improved numerical models that incorporate a simplified representation of these processes. Modelling results show improved performance when tested across several catchments affected by rainfall-runoff relationship changes, and improved realism such as multi-year declines in simulated groundwater storage, consistent with observations.  These results suggest a promising avenue for further research relevant to a variety of water resource applications including climate change impact assessment.

How to cite: Fowler, K., Regan-Beasley, D., Nixon, M., and Walker, G.: New modelling paradigm linking groundwater, surface water and rainfall-runoff relationship shifts under multi-year drought, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19135, https://doi.org/10.5194/egusphere-egu24-19135, 2024.

The occurrence of agricultural catchments covered by plastic greenhouses is growing worldwide. Horticultural greenhouse production allows for saving irrigation water at the farm scale, but also alters the natural hydrological cycle. Currently, these alterations are not accounted for in physics-based hydrological models. In this study, we aim to couple the greenhouse climate model KASPRO¹, to estimate indoor crop transpiration from outdoor meteorological variables, with the integrated surface-subsurface hydrological model CATchment HYdrology (CATHY²) to simulate the stream discharge as well as the shallow groundwater depth in an agricultural catchment (11 km²) covered by plastic greenhouses in South Italy. The dynamic presence of greenhouses, along with bare soils and vegetated lands, is mapped with the Open field and Protected Agriculture land cover Classifier (OPAC³).

We first compare our simulations against indoor measurements of water use (drip- and sub-irrigation) and soil moisture dynamics at different depths at the plot scale. Next, we run CATHY at the catchment scale and compare the output against measured stream water level.

The aim of our study is to validate the capabilities of KASPRO and CATHY to provide high-fidelity spatially distributed dynamic simulations of evapotranspiration and irrigation fluxes, as well as soil moisture and groundwater flows. Such capabilities are essentials to understand the implications of plastic greenhouse districts on the hydrological cycle and thus making these models useful tools for a more sustainable management of agricultural catchments.

References

¹ De Zwart, H.F., 1996. Analyzing Energy-Saving Options in Greenhouse Cultivation Using a Simulation Model. Landbouwuniversiteit, Wageningen.

² Camporese, M., Paniconi, C., Putti, M., & Orlandini, S. (2010). Surface--subsurface flow modeling with path-based runoff routing, boundary condition-based coupling, and assimilation of multisource observation data. Water Resources Research, 46, W02512.

³ la Cecilia, D., Tom, M., Stamm, C., Odermatt, D., 2023. Pixel-based mapping of open field and protected agriculture using constrained Sentinel-2 data. ISPRS Open Journal of Photogrammetry and Remote Sensing 8. https://doi.org/10.1016/j.ophoto.2023.100033.

Acknowledgements: We thank the Consorzio di Bonifica in Destra del Fiume Sele for the continuous support in the MSCA-PF REWATERING project.

Funding: This project has received funding from the European Union’s Horizon Europe research and innovation under the Marie Skłodowska-Curie grant agreement No. 101062255

How to cite: la Cecilia, D., Venezia, A., Maino, D., and Camporese, M.: Towards an understanding of the hydrological processes of greenhouse horticulture districts, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9140, https://doi.org/10.5194/egusphere-egu24-9140, 2024.

While natural flood management (NFM) as a resilience flood mitigation strategy is widely used in the UK and Europe, there remains a lack of scientific evidence regarding its effectiveness. The primary uncertainties stem from two aspects: the determination of NFM effectiveness on flood mitigation is limited by the scale of impact assessment; and the combination of multiple NFM interventions implemented within a catchment which may result in flood synchronicity. We argue that the effectiveness of combined scenarios involving multiple NFM interventions within a catchment can vary.  We utilize a hydrological model that simulates both instream and terrestrial interventions at a large catchment scale. To demonstrate how scale and interventions interact to determine flood peaks, we integrated various NFM interventions and land cover changes within the upstream catchment into a model, including afforestation, soil aeration, catchment/floodplain restoration and hedge planting. We modelled existing and planned scenarios using a spatially distributed hydrological model, Spatially Distributed TOPMODEL (SD-TOPMODEL). In comparison to previous versions of TOPMODEL, we have improved the simulation efficiency to allow for the simulation of up to a 200-year return period flood event at a larger catchment scale (~84 km²); and simplified the model parameters which are not related to the effects of NFM interventions and retained three key parameters which are physically significant. Following extensive parameter calibration and validation, the model is stable, providing a reliable fit for flood peaks, with the Nash-Sutcliffe coefficient (NS) between modelled and observed discharge reaching up to 0.905. The modelling results illustrated the effectiveness of NFM interventions in reducing flood peaks at a large catchment scale. Further refinements will involve incorporating additional types of NFM interventions into our next coupled model. 

How to cite: Zhu, Q., Klaar, M., Willis, T., and Holden, J.: Modelling the effectiveness of multiple natural flood management interventions at a large catchment scale, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12534, https://doi.org/10.5194/egusphere-egu24-12534, 2024.

The ability of contemporary hydrological models to serve as a basis for credible prediction and decision making is increasingly challenged – especially as hydrological systems are pushed outside the envelope of historical experience. Conceptual models are the most common type of surface water hydrological model used for decision support due to reasonable performance in the absence of change, ease of use and computational speed that facilitate scenario, sensitivity and uncertainty analysis. Hence, conceptual models arguably represent the current "shopfront" of hydrological science as seen by practitioners. However, these models have notable limitations in their ability to resolve internal catchment processes and subsequently capture hydrological change. New thinking is needed to confront the challenges faced by the current generation of conceptual models in dealing with a changing environment. We argue that the next generation of conceptual models should combine the parsimony of conceptual models with our best available scientific understanding. We propose a strategy to develop such models using multiple hydrological lines of evidence. This strategy includes using appropriately selected physically-resolved models as "Virtual Hydrological Laboratories" to test and refine the simpler models' ability to predict future hydrological changes. This approach moves beyond the sole focus on "predictive skill" measured using metrics of historical performance, facilitating the development of the next generation of conceptual models with hydrological fidelity - i.e., that "get the right answers for the right reasons". This quest is more than a scientific curiosity – it is expected by environmental policy makers and broader stakeholders.

How to cite: Kavetski, D., Thyer, M., McInerney, D., Gupta, H., Westra, S., Maier, H., Jakeman, A., Croke, B., Partington, D., Shanafield, M., Simmons, C., and Tague, C.: Virtual Hydrological Laboratories to develop the next generation of conceptual models and support decision-making under change, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9070, https://doi.org/10.5194/egusphere-egu24-9070, 2024.

SWAT+ is a completely restructured version of the SWAT model. It includes new capabilities, and the possibility of modelling water resources management is particularly relevant. A new water allocation module allows to allocate water for different purposes inside and outside a basin. Reservoir management can be modelled using decision tables that define which actions occur under different scenarios. Despite the relevance of these novelties, there is a lack of studies demonstrating accurate simulations of reservoir outflows using decision tables in SWAT+, and there are to date no publications regarding the water allocation module.

The Tagus River basin (Spain) is the most populated (11 million inhabitants) basin on the Iberian Peninsula and its water resources management is highly controversial. This basin is a clear example of the importance of including anthropogenic water management in the modelling process, since it is intensively regulated by more than 50 reservoirs, several water transfers, and irrigation. Therefore, the water allocation module (for simulating water transfers and irrigation) and decision tables (for reservoir management and irrigation) were used in a detailed model of the Upper Tagus River Basin (UTRB), where most of the water demands of the basin are located.

Firstly, more than 30 reservoirs were introduced to the model and their management was analyzed using observed data. Different decision table structures were created considering the properties of the reservoirs (purpose, storage, etc.) and then adapted to each of the reservoirs. A satisfactory simulation of reservoir storage and outflow was achieved in most of the cases, demonstrating the reliability of the model and the adequacy of this approach.

There are numerous water transfers in the UTRB, of which the Tagus-Segura water transfer (TSWT) is the most relevant one. Some transfer water from one reservoir to another, while two of them divert water outside the modelled basin. In addition, water is transferred from reservoirs to water treatment plants and subsequently released to selected receiving channels. All these transfers were modelled using the SWAT+ water allocation module and for most of them the modelled volumes matched the observed ones well.

The agricultural water demand was estimated from the River Basin Management Plan. To simulate the irrigation, all demand objects within the UTRB (irrigated agricultural lands, 282 objects) and their respective sources (closest channel to those objects, 118 sources) were identified. An irrigation decision table was developed for the basin, allowing to simulate a demand close to the calculated and to supply enough water to meet more than 80% of this demand.

This works presents a novel approach to simulating water resources management in a highly regulated river basin using SWAT+. Results shows a satisfactory simulation of different management actions (reservoirs, irrigation, water transfers inside and outside the basin, wastewater discharges). Further work on the water allocation module will boost even more the application of SWAT+.

How to cite: Sánchez-Gómez, A., Arnold, J., Bieger, K., Sammons, N., Martínez-Pérez, S., and Molina-Navarro, E.: Application of the new SWAT+ water allocation module in the Tagus River basin, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19765, https://doi.org/10.5194/egusphere-egu24-19765, 2024.

One the most difficult challenges in applied hydrology is predicting runoff in ungauged or poorly gauged watersheds. Thus, simple approaches for runoff estimation are especially useful in hydrologic applications. A simple, well established, and widely used technique for predicting the direct runoff depths of rainfall events is the Soil Conservation Service - Curve Number (SCS-CN) method. Due to its straightforward but well-proven approach, readily available and well documented environmental inputs, and incorporation of numerous variables influencing runoff generation into a single CN parameter, it quickly rose to prominence among engineers and practitioners. Tables can be used to identify the CN parameter values corresponding to prevailing soil, land cover and land management conditions. However, it is always better to estimate the CN value using observed rainfall-runoff (P-Q) data when available. Estimating appropriate CN values for additional soil – land cover conditions and additional regions is also critical for extending and updating the method’s documentation given that the SCS-CN approach is extremely sensitive to variations in the CN values.

However, even when the CN value is determined from measured P-Q data, the estimated CN values vary substantially from storm to storm on any given watershed. For this reason, various methods to estimate the CN value characterizing each watershed have been proposed up to know, and many theories on the reasons behind the observed relationships between CN and P for each watershed have been stated. Though, after many years of research, there isn’t still a unique agreed method to estimate the CN values characterizing a watershed or a soil-land cover complex, while the proposed methods lead to different CN values and in many cases neglect spatial variability. Further, an increasing number of modified SCS-CN versions are continuously developed, and new parameters are introduced complicating the situation even more.

Accordingly, this study attempts to collect, categorize, and systematically analyze the huge number of studies on SCS-CN method published in the last 30 years. We selected this period as 30 years ago, in 1993, R.H. Hawkins published his emblematic study on the “Asymptotic determination of runoff curve numbers from data” (J. Irrigat. Drain. Div. ASCE, 119(2): 334–345). In this review study, specific attention is given to the methods focusing on CN value determination from measured P-Q data. The advantages and limitations of the various approaches are investigated, as well as trends and gaps in existing literature. The analysed methods are classified and the main paths are identified. Based on the obtained results, conclusions on the current status are being made, and the more promising approaches are highlighted.  Then, ideas on future research pathways towards the target of a unified CN values determination approach are discussed.

How to cite: Soulis, K., Palli Gravani, S., and Kalivas, D.: SCS-CN parameter determination from observed rainfall runoff data. A critical review., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4058, https://doi.org/10.5194/egusphere-egu24-4058, 2024.

Due to the difficulties of gathering relevant data of groundwater systems and the lack of fundamental physically-based understanding on the processes involved, the representation of groundwater flow heterogeneity in catchment- to regional-scale hydrological models is often overlooked. We often limit the representation of groundwater with simplified homogeneous and shallow aquifers where effective hydraulic properties are derived from global-scale database. This raises questions regarding the validity of such models to quantify the potential impacts of climate change, where subsurface heterogeneity is expected to play a major role in their short- to long- term regulation.

We will present the results of a numerical modelling experiment designed to explore the role of the vertical compartmentalization of hillslopes on groundwater flow and recession discharge. We found that, when hydraulic properties are vertically compartmentalized, streamflow recession behaviour may strongly deviate from what is predicted by groundwater theory that considers the drainage of shallow reservoirs with homogeneous properties. We further identified the hillslope configurations for which the homogeneous theory derived from the Boussinesq solution approximately holds and, conversely, for those for which it does not. By comparing the modelled streamflow recession discharge and the groundwater table dynamics, we identify the critical hydrogeological conditions responsible for the emergence of strong deviations. We further present new solutions to better represent subsurface heterogeneity in catchment-scale models and calibrate hydraulic parameters that properly capture the groundwater and streamflow dynamics.

How to cite: Roques, C., Abhervé, R., Marti, E., Cornette, N., de Dreuzy, J.-R., Rupp, D., Boisson, A., Leray, S., Brunner, P., and Selker, J.: Recession discharge from compartmentalized bedrock hillslopes: hydrogeological processes and solutions for model calibration , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18302, https://doi.org/10.5194/egusphere-egu24-18302, 2024.

Flash floods, characterized by rapid streamflow response to rainfall, pose a significant natural hazard, particularly in small tropical watersheds (< 40 km²). Understanding the role of rainfall event characteristics, including amount, intensity, and spatial structure, is crucial for addressing and predicting flash floods. This study employs the WRF-Hydro® model with high-resolution (250-m, hourly) rainfall data and the Random Balance Designs – Fourier Amplitude Sensitivity Testing (RBD-FAST) method to investigate how rainfall impacts streamflow, specifically peak flow events, in seven watersheds on Oʻahu, USA.

Analyzing storm events from 2015 to 2020, we examined peak flow responses to corresponding rainfall event characteristics and estimated their contributions to model efficiency. In addition, (1) random redistribution of rainfall and (2) spatial shifting of rainfall were experimented with to assess the sensitivity of peak flow to rainfall event characteristics. Not only the rainfall amount and intensity but heavy rainfall areas (>= 25 mm) within an event also exerted a significant impact on peak flow, while other spatial features contributed varying degrees of influence. Notably, spatially shifting rainfall for at least 250-m in any direction highly affected event peak streamflow, emphasizing the importance of rainfall amount, intensity, heavy rainfall areas, total rainfall areas, and connectivity among rainfall areas.

Given the significance of rainfall's spatial heterogeneity, these findings underscore the benefits of incorporating rainfall spatial characteristics in probabilistic flood forecasting and the mitigation of flood risks. This research contributes valuable insights for enhancing flood prediction strategies in small tropical watersheds, providing a basis for informed decision-making and risk management.

How to cite: Huang, Y.-F. and Tsang, Y.: Sensitivity analysis of streamflow responses to varied rainfall spatial patterns in small tropical watersheds, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7129, https://doi.org/10.5194/egusphere-egu24-7129, 2024.

Efficiently modeling water and energy fluxes across spatial scales has historically involved grouping landscapes based on hydrological similarities. The HydroBlocks (HB) modeling framework, using unsupervised machine learning of high-resolution environmental datasets, emerges as a robust tool for representing heterogeneity in Land Surface Models (LSMs). This framework effectively discretizes complex gridded LSMs, such as Noah-MP, into spatially unstructured Hydrological Response Units (HRUs), facilitating the modeling of hydrological processes at hyper-resolution (10-100 m) with computational efficiency suitable for continental and global simulations. However, extending process-based hydrological models to such scales does not inherently ensure heightened simulation accuracy. For operational purposes, especially in flood warning systems, calibrating new LSMs remains imperative. Therefore, this study proposes a spatial parameter sensitivity methodology based on the pyVISCOUS algorithm, with the potential to facilitate HRU-level parameter calibration and enhance the application of hyper-resolution resolving LSMs for real-time streamflow prediction.

Our investigation delves into the relationship between spatial parameter sensitivity and model discretization across the Contiguous United States (CONUS), mainly focusing on surface and subsurface runoff states. Two clustering architectures were used to generate HB HRUs for an ensemble of simulations varying Noah-MP LSM parameters. The simplified HB configuration clusters HRUs based on terrain and hillslope variations, while the formal HB incorporates finer-scale land heterogeneity from high-resolution land cover and soil properties maps. Results reveal that saturated hydraulic conductivity was considered the most sensitive parameter for runoff production independent of the HRU grid configuration. The infiltration controlling parameter REFDK was ranked as the second most important in first-order sensitivity and had a higher spatial impact (% of HRUs) over the experiment with a higher level of clustering small-scale heterogeneity. Lower sensitivities were found in HRUs classified as urban areas, while soil properties parameters demonstrate reduced sensitivity near streams, where the floodplain remains closer to saturation. We intend to demonstrate that excluding the least sensitive HRU groups within a defined parameter range from calibration could potentially minimize computational costs while preserving physically realistic spatial patterns of LSM fluxes and states at field-scale resolutions, mitigating artifacts introduced by conventional methods (e.g. constant parameter multiplier over subbasins).

How to cite: Bacelar, L., Liu, H., Tang, G., Vergopolan, N., Mizukami, N., Wood, A., and Chaney, N.: How can spatial parameter sensitivity analysis enhance streamflow calibration routines in hyper-resolution hydrological models?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13626, https://doi.org/10.5194/egusphere-egu24-13626, 2024.

Land surface processes exert a significant impact on local, regional, and global climate through intricate physical exchanges, including energy, water cycle dynamics, vegetation response, soil moisture variations, and heat fluxes between land and atmosphere. A comprehensive understanding of these processes necessitates the analysis of land surface states (e.g., soil moisture, temperature) and fluxes (e.g., evapotranspiration, runoff) over an extended period for various research fields such as hydrological process modeling, weather and climate forecast, drought/flood monitoring, and water resource conservation. However, the accuracy of analyses is hindered by the sparse and uneven distribution of in-situ measurements. To overcome this limitation, satellite-based data and land surface models are employed. While satellites provide continuous global data, they only capture surface-level conditions and have limited daily spatial coverage. Daily, multi-depth soil profile information is essential for understanding land condition dynamics and their impact on the water cycle and agriculture. The Community Land Model (CLM), specifically its latest version CLM5, stands as a pivotal tool for simulating biophysical and biogeochemical processes, including interactions with the atmosphere. Nevertheless, its efficacy in accurately simulating water and energy cycles over India, where local land surface changes are particularly pertinent due to sparse in-situ data remains to be evaluated. To address this gap, our study employs CLM5 to simulate the land surface process at a 0.1° resolution from 1980 to 2020 over India. The evaluation process is comprehensive, involving comparisons with diverse land surface datasets, encompassing in-situ, remotely sensed, and reanalysis measurements. For soil moisture, CLM5 demonstrates good agreement with in-situ data (correlation: 0.66 to 0.67) but exhibits wet biases when compared to in-situ and GLEAM. In the case of evapotranspiration and runoff, CLM5SP closely matches the patterns observed in GLEAM and GRUN datasets (correlation: 0.89 to 0.95 for evapotranspiration and 0.77 to 0.96 for runoff). However, it is noteworthy that CLM5SP tends to overestimate both evapotranspiration and runoff when compared to the reference datasets. The anticipated outcome of this study provides valuable insights into the capabilities of CLM5 simulations over India, offering applications and references for enhancing the model's characterization of water and energy fluxes in the future.

How to cite: Devavat, C. N. and Chandrika Thulaseedharan, D.: How well does CLM5 simulate water and energy cycles over India? - A performance evaluation , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12287, https://doi.org/10.5194/egusphere-egu24-12287, 2024.

Recent advances in the application of machine learning techniques to estimate soil hydraulic properties using soil datasets have shown promising results. PedoTransfer Functions (PTFs) can facilitate the mapping of the complex relationship between soil properties and soil hydraulic properties, e.g., lateral hydraulic conductivity—a necessity for estimating lateral subsurface flow in spatially distributed hydrological models like wflow_sbm. The vertical-to-horizontal saturated hydraulic conductivity ratio (f_Kh0) is crucial for model calibration, but an established PTF is currently lacking. Our objective is to investigate the potential of ML algorithms in estimating PTFs for f_Kh0 prediction. First, optimized f_Kh0 across Great Britain (GB) resulting from a sensitivity analysis of the wflow_sbm model (Weerts et al., 2024) were used to train two ML algorithms; Random Forest (RF) and Boosted Regression Trees (BRT), using seven soil parameters from SoilGrids v1.0. Both algorithms effectively predicted f_Kh0 of 92 subbasins (i.e., test set of 25%) with high performance as compared against the optimized values, and RF slightly outperformed BRT. As a next step, we compared wflow_sbm simulated discharge results using uncalibrated f_Kh0 (default value of 100) and predicted values. The predictions notably improved wflow_sbm predictive accuracy by rising the median KGE from 0.55 (using uncalibrated f_Kh0) to 0.75 (using predicted f_Kh0). Following, we generated two globally distributed f_Kh0 maps, allowing us to further investigate the transferability of the ML-based PTFs. Therefore, we tested the predicted f_Kh0 across 559 gauge stations within the Loire basin in France. The utilization of either RF or BRT improved performance in around 75% of these subbasins with a KGE that was, on average, 0.06 higher. Furthermore, f_Kh0 prediction uncertainty and the impact of model spatial resolution were further analyzed. In conclusion, our study demonstrates the potential of ML methods to find relationships between soil properties and (model) soil hydraulic properties, which assists in parameter estimates for distributed hydrological models in gauged and ungauged basins.

How to cite: Ali, A. M., Imhoff, R. O., and Weerts, A. H.: Machine learning for predicting spatially variable lateral hydraulic conductivity: a step towards efficient hydrological model calibration and global applicability, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8347, https://doi.org/10.5194/egusphere-egu24-8347, 2024.

Accurately and efficiently estimating parameters for spatially distributed environmental models is impossible without proper regularization of the parameter space. The Multiscale Parameter Regionalization (MPR, Samaniego et al. 2010) makes use of high-resolution physiographic data (i.e., physiographic data such as soil maps and land cover information) to translate local land surface properties into model parameters. MPR consists of two steps: first, the high-resolution model parameters are derived from physiographic data via transfer functions at the native resolution. Second, the model parameters are upscaled to the target resolution the environmental model is applied on. MPR has already been successfully applied to the mesoscale hydrologic model (mHM, Samaniego et al. 2010, Kumar et al. 2013). An agnostic, stand-alone version implementation of MPR (Schweppe et al., 2022) allows applying this technique to any land-surface model or hydrological model.

In this study, we apply MPR to optimize parameters for the land-surface model ECLand (Boussetta et al. 2021) of the ECMWF Integrated Forecasting System. ECLand is calibrated at multiple locations simultaneously to provide an improved representation of river discharge at a global scale. We demonstrate the flexibility of the MPR approach by optimizing different transfer functions including the default one used in ECLand. In particular, we will discuss how specific choices in the calibration setting (i.e., chosen model parameters and ranges, basin locations, transfer function) affect the obtained ECLand model performance.

References:

Samaniego L., Kumar, R., and Attinger, S.: “Multiscale parameter regionalization of a grid-based hydrologic model at the mesoscale”, Water Resour. Res., 46, 2010.

Kumar, R., Samaniego, L., and Attinger, S.: “Implications of distributed hydrologic model parameterization on water fluxes at multiple scales and locations”, Water Resources Res, 2013

Schweppe, R., Thober, S., Müller, S., Kelbling, M., Kumar, R., Attinger, S., and Samaniego, L.: MPR 1.0: a stand-alone multiscale parameter regionalization tool for improved parameter estimation of land surface models, Geosci. Model Dev., 15, 859–882, https://doi.org/10.5194/gmd-15-859-2022, 2022

Boussetta S, Balsamo G, Arduini G, Dutra E, McNorton J, Choulga M, Agustí-Panareda A, Beljaars A, Wedi N, Munõz-Sabater J, de Rosnay P, Sandu I, Hadade I, Carver G, Mazzetti C, Prudhomme C, Yamazaki D, Zsoter E. ECLand: The ECMWF Land Surface Modelling System. Atmosphere. 2021; 12(6):723. https://doi.org/10.3390/atmos12060723

How to cite: Thober, S., Mai, J., Mazzetti, C., Balsamo, G., Prudhomme, C., Schweppe, R., Kelbling, M., Müller, S., and Samaniego, L.: Multi-basin calibration of the ECMWF land-surface model ECLand, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13129, https://doi.org/10.5194/egusphere-egu24-13129, 2024.

ParFlow CONCN model is the first integrated groundwater-surface water modeling platform over the mainland of China with a resolution of 30-arcsec and a depth of 492 m. With the flexibility of reconstruction and prediction of groundwater and surface water states and fluxes, it is undoubtedly an efficient tool for scientific understanding on water cycle and decision making on water resources and thus to tackle China’s water crisis in the changing world. Nonetheless, the CONCN model may have broader significances to the hydrologic community. Model evaluation and comparison is a common practice in the geoscientific modeling communities, such as those of land surface and earth system models. Due to the challenges in aquifer parameterization and the expensive computing requirement, high-resolution, large-scale, 3D groundwater modeling or integrated hydrologic modeling with 3D groundwater component is still under development though becoming more active in the past decade. Several national-scale such integrated hydrologic models, for example, ParFlow models over CONUS, west Africa, and Germany, have been built at 1 km resolution or higher with satisfying performances. However, the wide extension of modeling in this category to other places worldwide or to global scale is limitedly explored, preventing the evaluation of modeling workflows at different places and comparison with models using other parameterization schemes. Here, we demonstrate the construction and the first-phase evaluation of CONCN model by leveraging global datasets. Global permeability (GLHYMPS 1.0) and hydrography (MERIT Hydro) products were helpful to build the model while global water table depth (Fan et al., Science, 2013 and Zeng et al., JAMES, 2018) and streamflow (GRADES-HYDRDL and CNRD v1.0) products were adopted to preliminarily evaluate the simulation results. In this data-poor modeling area, both the construction and evaluation of the CONCN model are impossible about five years earlier as most of these global datasets did not exist. Therefore, the CONCN model can be one of the pioneers to evaluate and then to improve the current workflow of the existing models and address the challenges in new modeling areas with hydrogeology, hydrography, and climatology unseen in existing models. We also expect our dilemma caused by lacking observations as many other modelers in China can push the data-sharing to constrain hydrologic models and to motivate the collaboration such as model intercomparison in the Chinese hydrologic modeling community, which are well developed in the global community.

How to cite: Yang, C., Condon, L., and Maxwell, R.: Building and evaluating the high-resolution, integrated groundwater-surface water ParFlow modeling platform of continental China (CONCN): leveraging global datasets in a data-poor region   , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2600, https://doi.org/10.5194/egusphere-egu24-2600, 2024.

To represent the physical processes at hillslope scales for hyper-resolution land surface modeling, we propose a hierarchical, catchment-based spatial tessellation method. The land surface is divided into a hierarchical structure: catchments, height bands along hillslopes within a catchment, and land cover patches within a height band. This catchment-based structure explicitly represents hillslope drainage networks and can be applied at various resolutions determined by a pre-defined maximum height band size. The proposed tessellation method is superior to the conventional grid-based structure in representing land surface heterogeneity, resulting in a higher aggregation skill through the height band representation. The spatial variations in air temperature, leaf area index, saturated soil hydraulic conductivity, and soil porosity are generally lower within a height band than those in a conventional rectangular grid, reflecting the nature of topographic control on climate, vegetation, and soil distribution. The improvement in aggregation skill depends on resolutions and terrain slope angle, more pronounced at 1/6° model resolution and over steeper terrains. Finally, we demonstrate that our proposed catchment-based structure performs better than the grid-based structure through modeling tests over the Columbia River basin at resolutions of 1/2°, 1/6°, and 1/20° and a global test at 1/2° using the ILAMB model evaluation metrics.

How to cite: Lina, H. and Shupeng, Z.: A Catchment-Based Hierarchical Spatial Tessellation Approach to a Better Representation of Land Heterogeneity for Hyper-Resolution Land Surface Modeling , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3343, https://doi.org/10.5194/egusphere-egu24-3343, 2024.

Conducting highly standardized model intercomparison studies of hydrologic models across large scales is beneficial in various aspects such as improving model accuracy and robustness, informing decision making, addressing uncertainties, enhancing educational and outreach opportunities and facilitating model benchmarking among others. However, looking beyond streamflow for hydrologic models is required to ensure that models simulate the right results for the right reasons. Continental scale analyses provide further insights into which systematic limitations a model has.

In this study, seven models (GR4J-CemaNeige, HMETS, Blended-v1, Blended-v2, HBV-EC, HYPR, SAC-SMA) have been setup for more than 2500 watersheds across Canada and the US using the RAVEN modeling framework. The models are setup using a standardized set of meteorologic and geophysical datasets to inform the model regarding forcings, soil, landcover, and terrain. All models are calibrated with respect to daily streamflow (2001-2015) and are subsequently validated on an independent time period (1986-2000). Calibration was performed using 10 independent trials of the Dynamically Dimensioned Search algorithm each using a budget of 2000 model evaluations and Kling-Gupta Efficiency (KGE) as the objective function. Additional variables such as actual evapotranspiration (AET), surface soil moisture (SSM), and snow water equivalent (SWE) for the calibrated model setups were recorded and compared against independent gridded reference datasets (AET and SSM from GLEAM, SWE from ERA5-Land). 

The results (surprisingly) show that all tested models perform equally well for streamflow prediction (range of median KGE values across all sites during calibration period is [0.83, 0.87] and validation period is [0.46, 0.54]). 

Differences between models are most apparent for the auxiliary variables analyzed, i.e. AET, SSM, and SWE. The most interesting differences between the models lie in their abilities to predict AET, with median KGE being the highest for SAC-SMA (0.71), followed by GR4J-CemaNeige (0.65), while the lowest values were observed for HMETS (0.37) and HBC-EC (0.17). Indeed SAC-SMA showed highest performances across 51% of locations while the second-best model is GR4J-CemaNeige with best performance at 13% of locations. 

The SSM, evaluated using the Pearson correlation (r) coefficient, was predicted relatively well by all models (r ranging between 0.62 and 0.72); however, while most models had poorer predictions in the Rocky mountains and at higher latitudes, the SAC-SMA was definitely a better predictor of the temporal dynamics in SSM in these regions.

While the median performance for SWE prediction was relatively low across all models (median KGE between 0.23 and 0.40), poorer predictions mostly occurred in regions with low annual SWE, and predictions improved with increasing annual snow amounts. 

The study reveals novel insights regarding the consistent ability of a suite of models to predict streamflow, while clear ranking of models was apparent based on their ability to simulate spatially distributed variables like AET. Such differences likely arise due to model equifinality highlighting the value of model evaluation against multiple spatially distributed and lumped metrics, generating the correct streamflow for the right reasons.

How to cite: Mai, J. and Basu, N. B.: Beyond streamflow predictions: A continental scale hydrologic model intercomparison experiment, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12485, https://doi.org/10.5194/egusphere-egu24-12485, 2024.

The calibration and validation of hydrological models often involve a suite of established statistical metrics, which may not always match the needs of local stakeholders, thereby constraining the evaluative scope, particularly in the context of global climate services. This study introduces an alternative, complementary evaluation approach by formulating two types of application-based evaluation metrics (Du et al., 2024), representing model performance in terms of (i) temporal matching of the extreme quantiles and (ii) reproduction of the maximized split-sample difference in flow signatures. The introduced metrics are compared to conventional statistical metrics, at seven case study areas across the world, with three model settings representing different datasets and calibrations, generated from the global hydrological model World-Wide HYPE (WWH; Arheimer et al., 2020). The different performances found using application-based and conventional metrics, respectively, reveal their ability to uncover the models' capability in various aspects. Ultimately, the comprehensive analysis of conventional and application-based metrics allows us to delineate two scenarios for model application: generally applicable models, and conditionally applicable models. For example, in some areas the WWH model, when applied with global dataset and local calibration, is well capable of producing predictions for the timing of extreme quantiles and the relative difference in flow signatures, even though it may not excel according to conventional evaluation metrics. Consequently, this model can be classified as conditionally applicable, suitable for areas where local data is scarce, yet providing reliable information that can aid decision-makers in developing strategies for water resources management.

Arheimer, B., Pimentel, R., Isberg, K., Crochemore, L., Andersson, J.C.M., Hasan, A.,  Pineda, L. (2020). Global catchment modelling using World-Wide HYPE (WWH), open data, and stepwise parameter estimation. Hydrology and Earth System Sciences, 24, 535-559.

Du, Y., Olsson, J., Isberg, K., Strömqvist, J., Hundecha., Y., Silva, B.C., Rafee, S.A.A., Fragoso Jr., C.R., Beldring, S., Hansen, A., Uvo, C.B., Sörensen, J. (2024). Application-based evaluation of multi-basin hydrological models. Journal of Hydrology, under revision.

How to cite: Olsson, J., Du, Y., Isberg, K., Strömqvist, J., and Hundecha, Y.: Assessment of a global hydrological service by application-based metrics, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9082, https://doi.org/10.5194/egusphere-egu24-9082, 2024.