Search

Estuarine salt intrusion causes problems with freshwater availability in many deltas. Water managers require timely and accurate forecasts to be able to mitigate and adapt to salt intrusion. Data-driven models derived with machine learning are ideally suited for this, as they can mimic complex non-linear systems and are computationally efficient. We set up a long short-term memory (LSTM) model to forecast salt intrusion in the Rhine-Meuse delta, the Netherlands. Inputs for this model are chloride concentrations, water levels, discharges and wind speed, measured at nine locations. It forecasts daily minimum, mean and maximum chloride concentrations up to 7 d ahead at Krimpen aan den IJssel, an important location for freshwater provision. The model forecasts baseline concentrations and peak timing well but peak height is underestimated, a problem that becomes worse with increasing lead time. Between lead times of 1 and 7 d, forecast precision declines from 0.9 to 0.7 and forecast recall declines from 0.7 to 0.5 on average. Given these results, we aim to extend the model to other locations in the delta. We expect that a similar setup can work in other deltas, especially those with a similar or simpler channel network., Rivers, Ports, Waterways and Dredging Engineering

Estuarine salt intrusion causes problems with freshwater availability in many deltas. Water managers require timely and accurate forecasts to be able to mitigate and adapt to salt intrusion. Data-driven models derived with machine learning are ideally suited for this, as they can mimic complex non-linear systems and are computationally efficient. We set up a long short-term memory (LSTM) model to forecast salt intrusion in the Rhine–Meuse delta, the Netherlands. Inputs for this model are chloride concentrations, water levels, discharges and wind speed, measured at nine locations. It forecasts daily minimum, mean and maximum chloride concentrations up to 7 d ahead at Krimpen aan den IJssel, an important location for freshwater provision. The model forecasts baseline concentrations and peak timing well but peak height is underestimated, a problem that becomes worse with increasing lead time. Between lead times of 1 and 7 d, forecast precision declines from 0.9 to 0.7 and forecast recall declines from 0.7 to 0.5 on average. Given these results, we aim to extend the model to other locations in the delta. We expect that a similar setup can work in other deltas, especially those with a similar or simpler channel network.

Flash flood early warning requires accurate rainfall forecasts with a high spatial and temporal resolution. As the first few hours ahead are already not sufficiently well captured by the rainfall forecasts of numerical weather prediction (NWP) models, radar rainfall nowcasting can provide an alternative. Because this observation-based method quickly loses skill after the first 2 hr of the forecast, it needs to be combined with NWP forecasts to extend the skillful lead time of short-term rainfall forecasts, which should increase decision-making times. We implemented an adaptive scale-dependent ensemble blending method in the open-source pysteps library, based on the Short-Term Ensemble Prediction System scheme. In this implementation, the extrapolation (ensemble) nowcast, (ensemble) NWP, and noise components are combined with skill-dependent weights that vary per spatial scale level. To constrain the (dis)appearance of rain in the ensemble members to regions around the rainy areas, we have developed a Lagrangian blended probability matching scheme and incremental masking strategy. We describe the implementation details and evaluate the method using three heavy and extreme (July 2021) rainfall events in four Belgian and Dutch catchments. We benchmark the results of the 48-member blended forecasts against the Belgian NWP forecast, a 48-member nowcast, and a simple 48-member linear blending approach. Both on the radar domain and catchment scale, the introduced blending approach predominantly performs similarly or better than only nowcasting (in terms of event-averaged continuous ranked probability score and critical success index values) and adds value compared with NWP for the first hours of the forecast, although the difference, particularly with the linear blending method, reduces when we focus on catchment-average cumulative rainfall sums instead of instantaneous rainfall rates. By properly combining observations and NWP forecasts, blending methods such as these are

Flash flood early warning requires accurate rainfall forecasts with a high spatial and temporal resolution. As the first few hours ahead are already not sufficiently well captured by the rainfall forecasts of numerical weather prediction (NWP) models, radar rainfall nowcasting can provide an alternative. Because this observation-based method quickly loses skill after the first 2 hr of the forecast, it needs to be combined with NWP forecasts to extend the skillful lead time of short-term rainfall forecasts, which should increase decision-making times. We implemented an adaptive scale-dependent ensemble blending method in the open-source pysteps library, based on the Short-Term Ensemble Prediction System scheme. In this implementation, the extrapolation (ensemble) nowcast, (ensemble) NWP, and noise components are combined with skill-dependent weights that vary per spatial scale level. To constrain the (dis)appearance of rain in the ensemble members to regions around the rainy areas, we have developed a Lagrangian blended probability matching scheme and incremental masking strategy. We describe the implementation details and evaluate the method using three heavy and extreme (July 2021) rainfall events in four Belgian and Dutch catchments. We benchmark the results of the 48-member blended forecasts against the Belgian NWP forecast, a 48-member nowcast, and a simple 48-member linear blending approach. Both on the radar domain and catchment scale, the introduced blending approach predominantly performs similarly or better than only nowcasting (in terms of event-averaged continuous ranked probability score and critical success index values) and adds value compared with NWP for the first hours of the forecast, although the difference, particularly with the linear blending method, reduces when we focus on catchment-average cumulative rainfall sums instead of instantaneous rainfall rates. By properly combining observations and NWP forecasts, blending methods such as these, Water Resources

Although shallow ((Formula presented.) 1.5 m) soil water storage has been extensively studied, the significance of deeper unsaturated zone water storage to flow generation is poorly documented. However, a limited but growing body of empirical work shows that the weathered bedrock vadose zone, not soil, stores the majority of plant available water in many seasonally dry and semi-arid landscapes. Moreover, this storage dynamic mediates recharge to hillslope groundwater systems that generate stream discharge and support ecologically significant baseflows. Explicit representations of bedrock vadose zone processes are rarely incorporated into runoff models, due in part to a paucity of observations that can constrain simulations. Here, we develop a simple representation of the weathered bedrock vadose zone that is guided by in situ field observations. We incorporate this representation into a rainfall-runoff model, and calibrate it on streamflow alone, on rock moisture (i.e., weathered bedrock vadose zone moisture) alone, and on both using the concept of Pareto optimality. We find that the model is capable of accurately simultaneously simulating dynamics in rock moisture and streamflow, in terms of Kling-Gupta Efficiency, when using Pareto optimal parameter sets. Calibration on streamflow alone, however, is insufficient to accurately simulate rock moisture dynamics. We further find that the posterior distributions of some model parameters are sensitive to choice of calibration scenario. The posterior distribution of high-performing model parameters resulting from the streamflow only calibration scenario include physically unrealistic values that are not yielded by the rock moisture only or Pareto calibration strategies. These results suggest that the accuracy of some model results can be increased and parameter uncertainty decreased via incorporation of rock moisture data in calibration, without sacrificing streamflow simulation quality. Emerging recognition of the global

Streamflow is often the only variable used to evaluate hydrological models. In a previous international comparison study, eight research groups followed an identical protocol to calibrate 12 hydrological models using observed streamflow of catchments within the Meuse basin. In the current study, we quantify the differences in five states and fluxes of these 12 process-based models with similar streamflow performance, in a systematic and comprehensive way. Next, we assess model behavior plausibility by ranking the models for a set of criteria using streamflow and remote-sensing data of evaporation, snow cover, soil moisture and total storage anomalies. We found substantial dissimilarities between models for annual interception and seasonal evaporation rates, the annual number of days with water stored as snow, the mean annual maximum snow storage and the size of the root-zone storage capacity. These differences in internal process representation imply that these models cannot all simultaneously be close to reality. Modeled annual evaporation rates are consistent with Global Land Evaporation Amsterdam Model (GLEAM) estimates. However, there is a large uncertainty in modeled and remote-sensing annual interception. Substantial differences are also found between Moderate Resolution Imaging Spectroradiometer (MODIS) and modeled number of days with snow storage. Models with relatively small root-zone storage capacities and without root water uptake reduction under dry conditions tend to have an empty root-zone storage for several days each summer, while this is not suggested by remote-sensing data of evaporation, soil moisture and vegetation indices. On the other hand, models with relatively large root-zone storage capacities tend to overestimate very dry total storage anomalies of the Gravity Recovery and Climate Experiment (GRACE). None of the models is systematically consistent with the information available from all different (remote-sensing) data sources. Yet we di, Water Resources

The data are the hydrological model results of 12 models calibrated to the Ourthe catchment at Tabreux in the Belgian Ardennes. It provides hourly modeled streamflow, states and fluxes results for a selection of feasible parameter sets applied to 5 catchments (Ourthe at Tabreux, Ourthe Orientale at Mabompré, Ourthe Occidentale at Ortho, Semois at Membre-Pont, Lesse at Gendron). The models were calibrated by several institutes and universities working on the Meuse basin and gathering at the Meuse International Symposium in Liège. The data were used for the study Behind the scenes of streamflow performance by Bouaziz et al., 2021, HESS., The data are the hydrological model results of 12 models calibrated to the Ourthe catchment at Tabreux in the Belgian Ardennes. It provides hourly modeled streamflow, states and fluxes results for a selection of feasible parameter sets applied to 5 catchments (Ourthe at Tabreux, Ourthe Orientale at Mabompré, Ourthe Occidentale at Ortho, Semois at Membre-Pont, Lesse at Gendron). The models were calibrated by several institutes and universities working on the Meuse basin and gathering at the Meuse International Symposium in Liège. The data were used for the study Behind the scenes of streamflow performance by Bouaziz et al., 2021, HESS.

Streamflow is often the only variable used to evaluate hydrological models. In a previous international comparison study, eight research groups followed an identical protocol to calibrate 12 hydrological models using observed streamflow of catchments within the Meuse basin. In the current study, we quantify the differences in five states and fluxes of these 12 process-based models with similar streamflow performance, in a systematic and comprehensive way. Next, we assess model behavior plausibility by ranking the models for a set of criteria using streamflow and remote-sensing data of evaporation, snow cover, soil moisture and total storage anomalies. We found substantial dissimilarities between models for annual interception and seasonal evaporation rates, the annual number of days with water stored as snow, the mean annual maximum snow storage and the size of the root-zone storage capacity. These differences in internal process representation imply that these models cannot all simultaneously be close to reality. Modeled annual evaporation rates are consistent with Global Land Evaporation Amsterdam Model (GLEAM) estimates. However, there is a large uncertainty in modeled and remote-sensing annual interception. Substantial differences are also found between Moderate Resolution Imaging Spectroradiometer (MODIS) and modeled number of days with snow storage. Models with relatively small root-zone storage capacities and without root water uptake reduction under dry conditions tend to have an empty root-zone storage for several days each summer, while this is not suggested by remote-sensing data of evaporation, soil moisture and vegetation indices. On the other hand, models with relatively large root-zone storage capacities tend to overestimate very dry total storage anomalies of the Gravity Recovery and Climate Experiment (GRACE). None of the models is systematically consistent with the information available from all different (remote-sensing) data sources. Yet we did n

Following the rise of R as a scientific programming language, the increasing requirement for more transferable research and the growth of data availability in hydrology, R packages containing hydrological models are becoming more and more available as an open-source resource to hydrologists. Corresponding to the core of the hydrological studies workflow, their value is increasingly meaningful regarding the reliability of methods and results. Despite package and model distinctiveness, no study has ever provided a comparison of R packages for conceptual rainfall-runoff modelling from a user perspective by contrasting their philosophy, model characteristics and ease of use. We have selected eight packages based on our ability to consistently run their models on simple hydrology modelling examples. We have uniformly analysed the exact structure of seven of the hydrological models integrated into these R packages in terms of conceptual storages and fluxes, spatial discretisation, data requirements and output provided. The analysis showed that very different modelling choices are associated with these packages, which emphasises various hydrological concepts. These specificities are not always sufficiently well explained by the package documentation. Therefore a synthesis of the package functionalities was performed from a user perspective. This synthesis helps to inform the selection of which packages could/should be used depending on the problem at hand. In this regard, the technical features, documentation, R implementations and computational times were investigated. Moreover, by providing a framework for package comparison, this study is a step forward towards supporting more transferable and reusable methods and results for hydrological modelling in R.

Accurate and robust short-term rainfall forecasts (nowcasts) are useful in operational flood forecasting. However, the high temporal and spatial variability of rainfall fields make rainfall nowcasting a challenging endeavour. To cope with this variability, nowcasting techniques based on weather radar imagery have been proposed. Here, we employ radar rainfall nowcasting for discharge predictions in three lowland catchments in the Netherlands, with surface areas ranging from 6.5 to 957 km2. Deterministic (Lagrangian persistence) and probabilistic (SBMcast) nowcasting techniques are used to produce short-term rainfall forecasts (up to a few hours ahead), which are used as input for the hydrological model WALRUS. Rainfall forecasts were found to deteriorate with increasing lead time, often due to underestimation. Discharge could be forecasted 25–170 min earlier than without rainfall nowcasting, with the best performance for the largest catchment. When accounting for catchment response time, the best (but most variable) relative performance was found for the smallest catchment. Probabilistic nowcasting effectively accounted for the uncertainty associated with rainfall and discharge forecasts. The uncertainty in rainfall forecasts was found to be largest for the smaller catchments. The uncertainty in how much earlier the discharge could be forecasted (the gain in lead time) ranged from 15 to 50 min.