Search

Variably saturated subsurface flow models require knowledge of the soil hydraulic parameters. However, the determination of these parameters in heterogeneous soils is not easily feasible and subject to large uncertainties. As the modeled soil moisture is very sensitive to these parameters, especially the saturated hydraulic conductivity, porosity, and the parameters describing the retention and relative permeability functions, it is likewise highly uncertain. Data assimilation can be used to handle and reduce both the state and parameter uncertainty. In this work, we apply the ensemble Kalman filter (EnKF) to a three-dimensional heterogeneous hillslope model and investigate the influence of updating the different soil hydraulic parameters on the accuracy of the estimated soil moisture. We further examine the usage of a simplified layered soil structure instead of the fully resolved heterogeneous soil structure in the ensemble. It is shown that the best estimates are obtained when performing a joint update of porosity and the van Genuchten parameters and (optionally) the saturated hydraulic conductivity. The usage of a simplified soil structure gave decent estimates of spatially averaged soil moisture in combination with parameter updates but led to a failure of the EnKF and very poor soil moisture estimates at non-observed locations.

Reliable estimates of soil water content and groundwater levels are essential in evaluating water availability for plants and as drinking water and thus both subsurface components (vadose zone and groundwater) are commonly monitored. Such measurements can be used for data assimilation in order to improve predictions of numerical subsurface flow models. Within this work, we investigate to what extent measurements from one subsurface component are able to improve predictions in the other one.For this purpose, we utilize idealized test cases at a subcatchment scale using a Localized Ensemble Kalman Filter to update the water table height and soil moisture at certain depths with measurements taken from a numerical reference model. We do joint, as well as single component updates. We test strongly coupled data assimilation, which implies utilizing correlations between the subsurface components for updating the ensemble and compare it to weakly coupled data assimilation. We also update soil hydraulic parameters and examine the role of their heterogeneity with respect to data assimilation. We run simulations with both a complex 3D model (using TSMP-PDAF) as well as a more simplified and computationally efficient 2.5D model, which consists of multiple 1D vadose-zone columns coupled iteratively with a 2D groundwater-flow model. In idealized settings, such as homogeneous subsurface structures, we find that predictions in one component consistently benefit from updating the other component.

Pluvial urban flood events are prone to cause huge damages to infrastructures and can also endanger human lives. A strategy for dealing with natural disasters like urban flood events is to build up detailed models to predict potential implications of an event. These models are commonly physically based hydrodynamic models. Using such models for gaining better understanding of historical and possible future events can be beneficial. For damage mitigation during a storm event, the computational demand of these models is, however, too high. Therefore, substitute models have been developed in recent years, which are fast enough to allow for real time prediction. We present a machine learning model for real-time urban flood prediction with spatial and temporal resolution. The model was tested with promising results for a flat urban catchment. The model is based on a combination of autoencoders and a NARX neural network structure. The spatial resolution is 6 x 6 meters and the temporal resolution is 5 minutes. During the present research we applied the model to a steep urban catchment. Database for training the model was generated with the 1D/2D bidirectional coupled hydrodynamic model Hystem Extran 2D. As input we used design storm events with return periods of up to 100 years.

Convective rainfall cells have a maximum prediction time of 30 to 120 minutes, which means that physically based models are usually too slow to calculate the expected flooding depths within the available time. For this reason, novel models for short-term flood and flash flood prediction are needed.The work is being carried out as part of the WaX program and its aim is to develop a fast flood prediction model for temporal and spatial precipitation data (for example high-resolution radar forecast) and runoff-relevant soil parameters with machine learning methods for catchments with strong topography.The developed AI model aims to predict maximum flood depths for an urban catchment (Emmendingen in Baden-Württemberg, Germany) and temporally resolved flood depths in predefined neuralgic areas. It is based on supervised learning and therefore requires a database of input and associated output for training and validation.The effective rainfall, which is based on spatially and temporally distributed rainfall (from radar hindcast) and runoff-relevant parameters such as land use, slope, soil type and especially soil moisture, serves as the training input. The associated results for training and validation are the spatially distributed flood depth within the catchment. To build up the database both, the flood depths and effective rainfall rates, were precalculated using established hydrological respectively hydrodynamic models. To predict flood depths during real heavy rainfall events, a number of different high-resolution radar forecasts are used and combined with a range of soil moisture assumptions.The model is a combined method, in which the hydrological processes are done by a fast and calibrated physically based model, while the time-consuming hydrodynamic calculation is replaced by machine learning methods. This leads to a utilization of the low calculation time in the range of seconds with promising accuracy compared to the results from hydrodynamic simulations. Due to the fast prognosis time it is possible to calculate a series of ensembles for different soil moisture conditions and precipitation loads as a result of the uncertain soil moisture conditions and the uncertain radar forecast. The contribution will compare preliminary AI model predictions to the physically based model results to assess the potential and limitations of the model.

Multi-phase multi-component flow and transport models are a key instrument for analyzing and predicting gas transport inside the vadose zone. The soil moisture distribution within the vadose zone varies in time and space. Thus, accurate gas transport prediction relies on the precise knowledge of the saturation-dependency of the transport parameters such as the effective gas diffusion coefficient Deff. Although recent advances from typical small scale experiments (diffusion apparatus with typical soil core size of 100 cm3) show that Deff-saturation(S)-relationships are not only dependent on general soil characteristics such as air-filled porosity and total porosity, but can also be derived from pore network characteristics, such as pore connectivity and geometry, most model frameworks rely on simple empirical formulations such as Millington & Quirk (1961), which find wide acceptance, but have been found to not be universally applicable.The current state of research lacks extensive performance tests for the application of Deff-S-relationships beyond the small scale and especially for realistic natural conditions, where soil moisture changes with depth and gas transport processes may be more complex than in standard laboratory setups used for the experimental determination of Deff, which leads to unknown errors in the numerical prediction of sub-surface gas transport processes.We test different Deff-S-functions within a multi-phase-multi-component flow and transport model by reproducing a laboratory gas transport experiment, where a tracer gas is injected into a quasi-2D Darcy-scale sand tank with a soil moisture distribution that covers the full range from wet to dry and comparing simulated and measured gas concentrations at several locations within the tank over time. The systematic evaluation of different functions leads to the conclusion that the saturation-dependency of Deff in the tested sand follows power law-scaling at low gas phase saturation and linear scaling above, in line with the physically based concepts of percolation theory and effective medium theory and with recent experimental results (Ghanbarian et al., 2018). Other approaches such as (Buckingham, 1904; Penman, 1940; Millington & Quirk, 1961; Moldrup et al., 2000) lead to large errors between numerical and experimental results. We demonstrate that the use of an inaccurate Deff-S-function can lead to a misrepresentation of diffusion coefficients by a factor of up to 105, which underlines the need for a correct representation of the saturation-dependency of Deff in numerical modeling of sub-surface gas transport.

Models of variably saturated subsurface flow require knowledge of the soil hydraulic parameters. However, the determination of these parameters in heterogeneous soils is not easily feasible and subject to large uncertainties. As the modeled soil moisture is very sensitive to these parameters, especially the saturated hydraulic conductivity, porosity and the parameters describing the retention and relative permeability functions, it is likewise highly uncertain. Data assimilation can be used to handle and reduce both, state and parameter uncertainty. In this work, we apply the ensemble Kalman filter (EnKF) to a three-dimensional heterogeneous hillslope model and investigate the influence of updating the different soil hydraulic parameters on the accuracy of the estimated soil moisture. We further examine the usage of a simplified layered soil structure instead of the fully resolved heterogeneous soil structure in the ensemble. It is shown that the best estimates are obtained when performing a joint update of porosity and van Genuchten parameters and optionally the saturated hydraulic conductivity. The usage of a simplified soil structure gave decent estimates of spatially averaged soil moisture in combination with parameter updates but led to a failure of the EnKF and very poor soil moisture estimates at non-observed locations.

We study the mechanisms of advective trapping in composite porous media that consist of circular inclusions of distributed permeability embedded in a high conductivity matrix. Advective trapping occurs when solute enters low velocity regions in the media. Transport is analyzed in terms of breakthrough curves measured at the outlet of the system. The curve's peak behavior depends on the volume fraction occupied by the inclusions, while the tail behavior depends on the distribution of permeability values. In order to quantify the observed behaviors we derive two equivalent upscaled transport models. First, we derive a Lagrangian trapping model using the continuous time random walk framework that is parameterized in terms of volume fraction and the distribution of conductivites in the inclusions. Second, we establish a non-local partial differential equation for the mobile solute concentration by volume averaging of the microscale transport equation. We show the equivalence between the two models as well as (first-order) multirate mass transfer models. The upscaled approach, parameterized by medium and flow properties captures all features of the observed solute breakthrough curves, and sheds new light on the modeling of advective trapping in heterogeneous media.

We test the improvement of flux predictions with data assimilation (DA) in a coupled land surface/subsurface model. We present results of DA experiments in an idealized testcase with an extent of 1km x 5km x 50m. Our model considers multiple heterogeneous soil units, different plant functional types and a sophisticated topographical design chosen to induce lateral flow and rivers at specific areas. We use TSMP-PDAF to couple the land-surface model CLM and the subsurface/surface flow model ParFlow with the DA framework PDAF. We use a Localized Ensemble Kalman Filter (LEnKF) with an ensemble of 93 members. We consider uncertainty in the atmosphere, soil properties and initial conditions by different atmospheric forcings, distinct heterogeneous soil parameter distributions and an individual spinup for each ensemble member. The ensemble, which has a horizontal grid resolution of 40m, is updated with virtual measurements from a high resolution (10m) reference model.In the scope of this work, we address the impact of updating different state variables (soil moisture and pressure head) on groundwater recharge, lateral subsurface flow, surface runoff, and evapotranspiration. While surface runoff and evapotranspiration directly depend on pressure head and soil moisture, subsurface flow depends on pressure head gradients. For groundwater recharge, our estimate depends on groundwater storage changes (which can directly be enforced by the updates during DA) as well as subsurface flow. To investigate if DA can directly improve these fluxes, we run multiple experiments with different observation frequencies and localization radii. Further, we investigate if there are improvements in the fluxes during open loop forecasting periods subsequent to DA.

Soil hydrological processes (SHP) support ecosystems, modulate the impact of climate change on terrestrial systems and control feedback mechanisms between water, energy and biogeochemical cycles. However, land-use changes and extreme events are increasingly impacting these processes. In this Review, we describe SHP across scales and examine their links with soil properties, ecosystem processes and climate. Soil structure influences SHP such as infiltration, soil water redistribution and root water uptake on small scales. On local scales, SHP are driven by root water uptake, vegetation and groundwater dynamics. Regionally, SHP are impacted by extreme events such as droughts, floods, heatwaves and land-use change; however, antecedent and current SHP partially determine the broader effects of extreme events. Emerging technologies such as wireless and automated sensing, soil moisture observation through novel synthetic aperture radars satellites, big data analysis and machine learning approaches offer unique opportunities to advance soil hydrology. These advances, in tandem with the inclusion of more key soil types and properties in models, will be pivotal in predicting the role of SHP during global change.

The vadose zone is subject to dynamic boundary conditions in the form of infiltration and evaporation. A better understanding of implications for flow and solute transport, arising from these dynamic boundary conditions in combination with heterogeneous structure, will help to improve the prediction of the fate of solutes. We present laboratory experiments and numerical simulations of heterogeneous porous media under unsaturated conditions where controlled, temporally varying precipitation and evaporation are applied to study the effect of dynamic boundary conditions on solute transport in the presence of material interfaces. Dye tracers Eosine Y and Brilliant Blue FCF are utilized to visualize solute transport and analyze redistribution processes in a flow cell. Water and solute fluxes in and out of the flow cell are quantified. While in dynamic experiments application of small infiltration rates (significantly below the saturated hydraulic conductivities of the materials) led to a reversal of transport paths between infiltration and succeeding evaporation, larger infiltration rates altered downward transport such that flow and transport paths differed from those observed during evaporation. Differences in transport paths ultimately led to a redistribution and trapping of solute in one material which manifested as pronounced tailing in breakthrough curves. Trapping was induced not by the formation of a stagnant zone as result of large parameter contrast but by an interplay of dynamic boundary conditions and material heterogeneity. This study thereby highlights the importance to consider dynamic boundary conditions in predictions of solute leaching. © 2019. The Authors.

We present a multi-dimensional continuum mathematical model for modeling the growth of a symbiotic biofilm system. We take a dual-species namely, the Streptococcus - Veillonella sp. biofilm system as an example for numerical investigations. The presented model describes both the cooperation and competition between these species of bacteria. The coupled partial differential equations are solved by using an integrative finite element numerical strategy. Numerical examples are carried out for studying the evolution and distribution of the bio-components. The results demonstrate that the presented model is capable of describing the symbiotic behavior of the biofilm system. However, homogenized numerical solutions are observed locally. To study the homogenization behavior of the model, numerical investigations regarding on how random initial biomass distribution influences the homogenization process are carried out. We found that a smaller correlation length of the initial biomass distribution leads to faster homogenization of the solution globally, however, shows more fluctuated biomass profiles along the biofilm thickness direction. More realistic scenarios with bacteria in patches are also investigated numerically in this study.

Soil moisture is an important variable for land surface processes. To make good model predictions of soil moisture, the flow processes in the subsurface need to be captured well. Flow in the subsurface strongly depends on the soil hydraulic parameters. Information about model parameters is often not available, at least not for the entire domain of interest. The resulting parameter uncertainty needs to be accounted for in the applied model. Data assimilation can account for parameter and model errors as well as for all other possible sources of uncertainty if observations are available that can be used to condition the model states. Thus, the parameter uncertainty might be reduced and model predictions improved. However, including the parameters increases the size of the state vector and thus the computational burden. Especially for large models, this can be a problem. Furthermore, the updates can produce unphysical parameter combinations which in unsaturated zone models often lead to numerical problems.In this work, we test the effect of updating the soil hydraulic parameters along with soil moisture in a 3D subsurface hillslope model. We use the ensemble Kalman filter for data assimilation and synthetic observations of soil moisture. In a similar study using a 1D unsaturated flow model, parameter updates were found to be the best way to handle parameter uncertainty. Updating parameters resulted in improved predictions of soil moisture, although not necessarily in more realistic model parameters. The parameter updates should rather be considered a method of treating parameter uncertainty than a method for parameter identification. In the 1D settings, updating all uncertain parameters led to the best results. Whether this still holds and is feasible for a more complex 3D model is the question addressed in this presentation.

Estimating states and fluxes of the water cycle with terrestrial system models needs a large amount of input data, including soil and vegetation parameters, resulting in large uncertainties in model predictions. Assimilation of pressure head and/or soil moisture data can better constrain states and parameters of a terrestrial system model. Here we assimilate pressure head data and soil moisture data in a terrestrial system model over the Neckar catchment (13928 km2) with a spatial horizontal resolution of 800 m. We use the Terrestrial System Modeling Platform (TSMP), which consists of an atmospheric model component (not used in this work), the Community Land Model version 3.5 (CLM3.5), and the subsurface hydrological model Parflow, coupled by OASIS. TSMP is coupled to the Parallel Data Assimilation Framework (PDAF), which allows the assimilation of land surface and subsurface observations to estimate the model states and parameters. In this work the localized Ensemble Kalman Filter (LEnKF) was used to update hydraulic head, soil moisture and/or saturated hydraulic conductivity by assimilating hydraulic head or in situ soil moisture observations for a period of one year. Ensembles of soil properties, leaf area index and atmospheric forcings were generated. The ensemble of atmospheric forcings considered correlations among four variables, and spatio-temporal correlations of the atmospheric variables using a geostatistical procedure. The characterization of the water table depth and river discharge without data assimilation and for different scenarios of pressure head and soil moisture data assimilation were compared.

Pluvial floods in urban areas are caused by local, fast storm events with very high rainfall rates, which lead to inundation of streets and buildings before the storm water reaches a watercourse. An increase in frequency and intensity of heavy rainfall events and an ongoing urbanization may further increase the risk of pluvial flooding in many urban areas. Currently, warnings for pluvial floods are mostly limited to information on rainfall intensities and durations over larger areas, which is often not detailed enough to effectively protect people and goods. We present a proof‐of‐concept for an impact‐based forecasting system for pluvial floods. Using a model chain consisting of a rainfall forecast, an inundation, a contaminant transport and a damage model, we are able to provide predictions for the expected rainfall, the inundated areas, spreading of potential contamination and the expected damage to residential buildings. We use a neural network‐based inundation model, which significantly reduces the computation time of the model chain. To demonstrate the feasibility, we perform a hindcast of a recent pluvial flood event in an urban area in Germany. The required spatio‐temporal accuracy of rainfall forecasts is still a major challenge, but our results show that reliable impact‐based warnings can be forecasts are available up to 5 min before the peak of an extreme rainfall event. Based on our results, we discuss how the outputs of the impact‐based forecast could be used to disseminate impact‐based early warnings., Plain Language Summary: Pluvial floods are caused by local rain storms with extreme rainfall rates, which leads to immediate flooding of streets and buildings in urban areas. These events are expected to increase in the future due to climate change and growing urban areas. Pluvial floods are directly caused by a rainstorm, which gives citizens and emergency responders usually only a few minutes to act. Existing forecasting systems for pluvial floods are limited to rainfall forecasts that neither provide information about where a flood might occur nor how severe the impacts will be. Here, the main challenge is that current computer models that predict inundation take too long to run to release flood forecasts early enough. We present a new inundation model that can predict inundation for an upcoming flood event in a fraction of the time of existing models. We combine this model with models that predict the spreading of contamination (e.g., from a car accident) and the damage to residential buildings. For a real flood event we can show that this information can be released up to 5 min before the rainfall peak, which gives citizens and emergency responders the opportunities to safe lives and protect important valuables., Key Points: First impact‐based forecasting for pluvial foods. Artificial neural network inundation model significantly cuts calculation time to 0.1% of a physically based model with comparable accuracy. Forecast with estimates for inundated areas, spreading of contaminants and expected damage could be released 5 min before peak rainfall., Bundesministerium für Bildung und Forschung (BMBF), Z Zurich Foundation, Grantham Foundation for the Protection of the Environment, ESRC Centre for Climate Change Economics and Policy: ES/R009708/1

Fully integrated three-dimensional (3D) physically based hydrologic models usually require many computational resources. For many applications, simplified models can be a cost-effective alternative. The 3D models of subsurface flow are often simplified by coupling a 2D groundwater model with multiple 1D models for the unsaturated zone. The crucial part of such models is the coupling between the two model compartments. In this work we compare two approaches for the coupling. One is iterative where the 1D unsaturated zone models go down to the impervious bottom of the aquifer, and the other one is non-iterative and uses a moving lower boundary for the unsaturated zone. In this context we also propose a new way of treating the specific yield, which plays a crucial role in linking the unsaturated and the groundwater model. Both models are applied to three test cases with increasing complexity and analyzed in terms of accuracy and speed compared to fully integrated model runs. The non-iterative approach is faster but does not yield a good accuracy for the model parameters in all applied test cases, whereas the iterative one gives good results in all cases. Which strategy is applied depends on the requirements: computational speed vs. model accuracy.

Advective trapping occurs when solute enters low velocity zones in heterogeneous porous media. Classical local modelling approaches combine the impact of slow advection and diffusion into a hydrodynamic dispersion coefficient and many temporally non-local approaches lump these mechanisms into a single memory function. This joint treatment makes parameterization difficult and thus prediction of large-scale transport a challenge. Here, we investigate the mechanisms of advective trapping and their impact on transport in media composed of a high conductivity background and isolated low permeability inclusions. Breakthrough curves show that effective transport changes from a streamtube-like behaviour to genuine random trapping as the degree of disorder of the inclusion arrangement increases. We upscale this behaviour using a Lagrangian view point, in which idealized solute particles transition over a fixed distance at random advection times combined with Poissonian advective trapping events. We discuss the mathematical formulation of the upscaled model in the continuous time random walk and mobile–immobile mass transfer frameworks, and derive a model for large-scale solute non-Fickian dispersion. These findings give new insight into transport in highly heterogeneous media., Data used for producing the figures can be downloaded from digital.csic.es (https://digital.csic.es/handle/10261/216991) and by solving the respective equations. The authors thank Professor T.R. Ginn and two anonymous reviewers for their comments on the paper. J.J.H. and M.D. acknowledge the support of the Spanish Ministry of Science and Innovation (project CEX2018-000794-S and project HydroPore PID2019-106887GB-C31). J.J.H. acknowledges the support of the European Social Fund and Spanish Ministry of Ministry of Science, Innovation and Universities through the ‘Ramón y Cajal’ fellowship (RYC-2017-22300). The authors thank P. Uszes for providing the code to measure the average distance between inclusions.

The DLVO (Derjaguin-Landau-Verwey-Overbeek) approach to predict colloid mobility in porous media is centered on solution ionic strength and physicochemical surface properties of colloids and solid matrix. However, several colloid retention mechanisms are not related to such interfacial properties, but instead to hydraulic features like flow regime and pore structure. We aimed to determine the relative importance of DLVO-related and non-DLVO-related retention mechanisms, which remains poorly understood. For that, we developed a conceptual approach based on previous research on organic matter-coated goethite (OMCG) colloid mobility in quartz sand. OMCG colloid retention by DLVO mechanisms was negligible at 0.0 mM ionic strength. Therefore, any retention at 0.0 mM can be assigned to non-DLVO retention. At increasing ionic strength, the amount of DLVO retention is rising, while the amount of non-DLVO retention is independent from ionic strength and thus remains constant. This allows for a differentiation between the two types of retention mechanisms. To test this conceptual approach, we conducted OMCG colloid breakthrough experiments at varying interfacial conditions (ionic strength: 0.0–5.53 mM) and hydraulic conditions (flow rate: 0.11 – 5.02 cm min−1). From sessile drop contact angles and zeta potentials, DLVO and extended DLVO (XDLVO) interactions including Lewis acid-base interactions were approximated. The results show that colloid retention was almost exclusively related to DLVO retention mechanisms, while retention by hydraulic factors was practically irrelevant. We conclude that our conceptual approach can be applied to determine the relative importance of colloid retention caused by DLVO and non-DLVO mechanisms for further colloid-solid matrix systems.

A Lagrangean particle based transport model is introduced to calculate the potential contamination paths of solutes in drainage water in an urban area during a pluvial flood event. The necessity to capture the complexity of a surface runoff model and the coupling of pipe and surface are investigated. Results show that fully coupled hydrodynamic models are needed for a good representation of transport paths and breakthrough curves of contaminant concentration in drainage water. An analysis for time variation of a single location spill event is performed that shows a high variability in solute spreading depending on the spill relative to the rainfall characteristics.

Transport of gas components in the unsaturated zone and across the soil surface plays a role for transport of volatile contaminants, gases from pipe leaks or greenhouse gases. When estimating flow rates from the soil into the atmosphere, a good understanding of the transport processes is important. In general, component transport in the gas phase is considered to be mainly due to diffusion. However, the wind field above the soil surface can induce flow into the subsurface and influence transport and mass fluxes.We present a study on gas component transport through dry and partially saturated soil into a free air flow above the soil surface, considering gas components of different density. Laboratory experiments in a quasi-2d sand tank were carried out. The tank was placed underneath a wind tunnel, and different wind velocities were used. Gases with different densities were injected with constant rate at an inlet port. Concentration distributions were measured continuously with sensors that were installed inside of the tank. After establishing a steady state concentration distribution, the gas injection was stopped and the decrease of gas concentrations inside the tank was monitored.The experiments show that the concentration profiles under steady state gas injection depend on gas density and the different diffusion coefficients. They depend only slightly on the velocity of the overlaying wind field and the influence is mainly seen very close to the soil surface. The transient gas transport out of the soil, however, did not only depend on the different diffusion coefficients, but was clearly influenced by the wind field. The transient 2d concentration distribution fields illustrate that the wind field induced a flow field inside the tank that depends on the wind velocity and the component density and influences the gas component transport. The influence increases under partly saturated conditions.To reproduce the transport correctly, it is necessary to capture the coupling between free flow and porous medium flow and the transport in the coupled flow. To do so, we use a fully coupled flow and transport model implemented into the environment DuMux ((Dune for Multi-(Phase,Component, Scale, ...) flow and transport in porous media). It can be shown that including the coupling concept, the main features of the concentration distributions can be reproduced for both the steady state and the transient case. With the model it is also demonstrated, that although advective fluxes inside the porous medium introduced by the wind field (horizontal and lateral) are relatively small in comparison to the diffusive fluxes, they cause relevant changes in the concentration distribution and thus indirectly influence the mass fluxes inside the porous medium and across the soil-atmosphere interface.

Pluvial floods in urban areas are caused by local, fast storm events with very high rainfall rates, which lead to inundation of streets and buildings before the storm water reaches a watercourse. An increase in frequency and intensity of heavy rainfall events and an on-going urbanization may further increase the risk of pluvial flooding in many urban areas. Current early warning systems for pluvial floods are limited to rainfall predictions with fixed thresholds for rainfall duration and intensity and often do not provide the necessary information to effectively protect people and goods. We present a proof-of-concept for an impact-based early warning system for pluvial floods. Using a model chain consisting of a rainfall forecast, an inundation, a contaminant transport and a damage model, we are able to provide predictions for the expected rainfall, the inundated areas, spreading of potential contamination and the expected damage to residential buildings. We use a neural network-based inundation model, which significantly reduces the computation time of the model chain. To demonstrate the feasibility, we perform a hindcast of a recent pluvial flood event in an urban area in Germany. The required spatio-temporal accuracy of rainfall forecasts is still a major challenge, but our results show that reliable impact-based warnings can be issued up to 5 minutes before the peak of an extreme rainfall event. To effectively disseminate the warnings issued by the model chain we propose a two-way mobile warning application that allows for the collection of real-time validation data.

Early warning systems for floods in urban areas should forecast water levels and damage estimation to protect vulnerable regions. To estimate the danger of a flood for buildings and people, the energy of the flood has to be taken into account additionally to the water level. The energy is related to the flow velocity. For directing rescue workers or trace spreading of contaminants through flooded streets, a high resolution of the water’s energy in space and time is required. Direct numerical run-off calculation is too slow for a flood forecast in time. Therefore a database with pre-calculated events is needed and a method to select the water levels and velocity fields that are similar to a forecasted rain event. We present a method, how to create a real-time forecast based on pre-calculated data. The selection and weighting of the pre-calculated data is based on the precipitation pattern in the flood region. A nearest neighbor approach is applied to find water levels and velocity fields from a database that are similar to the forecasting event. For the ranking of similarity, different new metrics are compared against each other. The quality of the metrics is tested with a new approach of comparing velocity fields on the surface and in the pipe system. Considering both domains is crucial for understanding the complex dynamic flow paths on the surface. An urban catchment of 5 km² with high resolution (~3 m³) triangular surface mesh and connected drainage system is used for a hydrodynamic run-off simulation. The 1D-2D coupled software HYSTEM EXTRAN is used to generate the water levels and velocity fields for strong rainfall events of the past 20 years. More than 900 events with a duration between 15 minutes and 24 hours and return periods between 10 and 100 years were calculated and stored as the “pre-calculated” dataset.For comparing two events, the mean square error is calculated between the precipitation patterns with different approaches to select the start index and number of intervals. This number depends on the hydraulic response time, the temporal resolution and the length of the reference pattern. The quality of the nearest neighbor selection is quantified using the Nash–Sutcliffe model efficiency coefficient of pipe flow and the root mean square error of water level and velocity in significant surface cells. Additionally, the transport paths of artificial contamination spills are compared between the events to show the reproducibility of velocity fields for each metric. Results show that the reaction time and the wetting state of the surface is very important. Single cell values correspond well between a forecasted and a dataset event. However, complex transport paths have a very high variability that is not reproducible with similar events. Further research is required to clarify if this is a result of the random walk approach or of the injection time of the particles.

Urban flood warning systems need fast response times between the rainfall forecast and the flood alarm. Flood forecasts from physically based numerical models usually need much computation time. Flood forecasts based on databases from previous events or pre-simulated events can speed up the process of decision making. This work introduces and compares four distance metrics for temporal rainfall patterns used in a nearest neighbour based forecast system for dynamic water levels and velocities during pluvial floods. The system uses a database of 960 pre-calculated flood events. The performance of each metric is evaluated by analysis of time-series of flow variables. For the error quantification,a procedure to find a small number of representative locations in an urban catchment is described . A new approach to quantify the similarity of dynamic flow fields is introduced, which makes use of particle transport tracking. The four distance metrics are tested on forecast for four exemplary pluvial flood events resulting from rainfall of durations between 15 and 50 minutes and return-periods between 10 and 100 years. The best metric is based on the temporal precipitation pattern and takes the response time of the drainage system into account. If the pattern has a very short duration, a simpler characterisation-metric can be used, which takes the total volume, peak intensity and peak position into account.

For predicting flow in the unsaturated zone, an adequate choice of the model parameters, especially the soil hydraulic parameters, is essential. It is difficult to determine these parameters, as the parameter estimation problem easily becomes ill-posed, e.g. due to pseudo-correlations among two or more of the unknown parameters. In the field, this problem is strongly related to the available observations which, in monitoring networks, are not optimized to be used for parameter estimation. In this paper, we investigate the potential of data assimilation using the ensemble Kalman filter (EnKF) with unsaturated zone models under conditions where model parameters are highly uncertain and not identifiable. Different ways of dealing with the parameter uncertainty, such as parameter updates and bias correction, are discussed and compared. It is shown that jointly updating all uncertain parameters and states is the best method to account for the error induced by parameter uncertainty.

In urban water pipe networks, pipe leakage may lead to subsurface contamination or to reduced waste water treatment efficiency. The quantification of pipe leakage is challenging due to inaccessibility and unknown hydraulic properties of the soil. A novel physically-based model for three-dimensional numerical simulation of pipe leakage in variably saturated soil is presented. We describe the newly implemented coupling between the pipe flow simulator HYSTEM-EXTRAN and the groundwater flow simulator OpenGeoSys and its validation. We further describe a novel upscaling of leakage using transfer functions derived from numerical simulations. This upscaling enables the simulation of numerous pipe defects with the benefit of reduced computation times. Finally, we investigate the response of leakage to different time-dependent pipe flow events and conclude that larger pipe flow volume and duration lead to larger leakage while the peak position in time has a small effect on leakage.

Due to the complexity of flow conditions as well as solid matrix and colloid surface properties, solid knowledge about the transport of iron oxide colloids in soils remains scarce. In order to analyze the influence of flow conditions on iron oxide colloid transport and retention, breakthrough behavior of negatively charged, organic matter-coated goethite (OMCG) colloids in saturated quartz sand columns was investigated under continuous and stagnant flow conditions. Classic DLVO and extended DLVO (XDLVO) interaction energies including Lewis acid/base parameters were evaluated using measurements of sessile drop contact angles and zeta potentials of OMCG colloids and quartz. Results elucidated that under continuous flow conditions, OMCG colloids were highly mobile, which was in agreement with calculated unfavorable attachment conditions revealed by predictions of both DLVO approaches. In contrast, during intervals of flow interruption, significant amounts of OMCG colloids were retained in the solid matrix and could not be remobilized via re-establishment of flow. The magnitude of colloid retention increased with the duration of flow interruption; OMCG colloids were almost completely immobilized after 112 h. Further experiments were conducted in order to determine possible colloid retention mechanisms. Results indicated that the major cause for retention during flow stagnation was OMCG colloid capture at locations with attractive DLVO/XDLVO interactions, promoted by fast gravitational settling of colloids onto the solid matrix. We compared breakthrough curves to model predictions, where we demonstrated that an attachment term with a stagnant fluid switch was required in the mass balance in order to reproduce the measurements. We conclude that high mobility of OMCG colloids and prediction of that transport behavior with the applied DLVO approaches were only valid under continuous flow conditions. Under more discontinuous hydraulic conditions relevant in natural soils, such as flow interruption, OMCG colloid transport behavior was modified significantly.

Models for flow in environmental systems are subject to uncertainty. Models can thus be interpreted as hypotheses on the validity of the underlying model assumptions. One important source of uncertainty in models for flow and transport processes in the subsurface is the model concept. While uncertain model parameters or forcing terms can be captured as random processes and random fields, this type of uncertainty cannot be included into a model in a straightforward manner. This is particularly true if established model descriptions of a given process are not known or are still being debated. In this contribution, we outline several examples of subsurface flow and transport modeling where uncertainty of the model concept plays an important role. We discuss the need for the development of methods and standards to deal with this type of uncertainty in model hypothesis testing. This article is protected by copyright. All rights reserved.

The paper presents a 2D upward infiltration experiment performed on a model porous medium consisting of fine sand background with two inclusions made of coarser sands. The purpose of the experiment was to investigate the effects of structural air trapping, which occurs during infiltration as a result of heterogeneous material structure. The experiment shows that a significant amount of air becomes trapped in each of the inclusions. Numerical simulations were carried out using the two-phase water-air flow model and the Richards equation. The experimental results can be reproduced with good accuracy only using a two-phase flow model, which accounts for both structural and pore-scale trapping. On the other hand, the Richards equation was not able to represent the structural trapping caused by material heterogeneity.

Editors of several journals in the field of hydrology met during the Assembly of the International Association of Hydrological Sciences - IAHS (within the Assembly of the International Union of Geodesy and Geophysics - IUGG) in Prague in June 2015. This event was a follow-up of a similar meeting held in July 2013 in Gothenburg (as reported by Bloschl et al. [2014]). These meetings enable the group of editors to review the current status of the journals and the publication process, and share thoughts on future strategies. Journals were represented in the 2015 meeting through their editors, as shown in the list of authors. The main points on fostering innovation and improving impact assessment in journal publications in hydrology are communicated in this joint editorial published in the above journals.

International audience; Editors of several journals in the field of hydrology met during the Assembly of the International Association of Hydrological Sciences—IAHS (within the Assembly of the International Union of Geodesy and Geophysics—IUGG) in Prague in June 2015. This event was a follow-up of a similar meeting held in July 2013 in Gothenburg (as reported by Blöschl et al. 2014). These meetings enable the group of editors to review the current status of the journals and the publication process, and share thoughts on future strategies. Journals were represented in the 2015 meeting through their editors, as shown in the list of authors. The main points on fostering innovation and improving impact assessment in journal publications in hydrology are communicated in this joint editorial published in the above journals.

The real-time forecasting of urban flooding is a challenging task for the following two reasons: (1) urban flooding is often characterized by short lead times, (2) the uncertainty in precipitation forecasting is usually high. Standard physically based numerical models are often too slow for the use in real-time forecasting systems. Data driven models have small computational costs and fast computation times and may be useful to overcome this problem. The present study presents an artificial neural network based model for the prediction of maximum water levels during a flash flood event. The challenge of finding a suitable structure for the neural network was solved with a new growing algorithm. The model is successfully tested for spatially uniformly distributed synthetic rain events in two real but slightly modified urban catchments with different surface slopes. The computation time of the model in the order of seconds and the accuracy of the results are convincing, which suggest that the method may be useful for real-time forecasts. Bundesministerium für Bildung und Forschung/Sonderprogramm GEOTECHNOLOGIEN/03G0846A/EU

International audience; We develop an efficient computational model for simulating fluid invasion patterns emerging in variable aperture fractures. This two-dimensional model takes into account the effect of capillary force on the fluid-fluid interfaces and viscous pressure drop in both fluid phases. The pressure distribution is solved at each time step based on mass balance and local cubic law, considering an imposed pressure jump condition at the fluid-fluid interface. This pressure jump corresponds to the Laplace pressure which includes both terms related to the out-of-plane (aperture-spanning) curvature and to the in-plane curvature. Simulating a configuration that emulates viscous fingering in two-dimensional random porous media confirms that the model accounts properly for the role of viscous forces. Furthermore, direct comparison with previously obtained experimental results shows that the model reproduces the observed drainage patterns in a rough fracture reasonably well. The evolutions of tip location, the inlet pressures, and the invading phase fractal dimensions are analyzed to characterize the transition from capillary fingering to viscous fingering regimes. A radial injection scenario of immiscible invasion is also studied with varying modified capillary number and viscosity ratio, showing displacement patterns ranging from capillary fingering to viscous fingering to stable displacement. Such simulations using two contact angles show that the invading phase becomes more compact when the wetting condition changes from strong to weak drainage, as already observed in 2-D porous media. The model can be used to bridge the gap in spatial scales of two-phase flow between pore-scale modeling approaches and the continuum Darcy-scale models.

In this paper a tool to automatically design drainage systems for urban districts is provided. Since pipe networks typically follow public streets, the drainage network is placed below the existing streets. The presented approach supplies a complete workflow in R, including an interface to hydraulic solvers implemented in the simulation software Stormwater Management Model. Algorithms to compute (1) subcatchments from land use data and (2) stormwater drainage networks from street polylines and topography are implemented. A test case, focussing on the stormwater drainage network on district scale, is setup. The stormwater network generator performs well for the small test case, however, some improvements on deriving the percentage of impervious surface cover have to be implemented. Further work includes the application to a more complex urban area, with pumps, combined sewer systems and multiple outfalls.

The ensemble Kalman filter is, due to its computational efficiency, becoming more and more popular as a method for estimating both model states and parameters in hydrologic modeling, also for nonlinear state propagation models. In the ensemble Kalman filter the calculation of the error correlations, and hence the filter update, is done based on the ensemble of model evaluations and can therefore be strongly influenced by a few ensemble members with extreme values. With nonlinear state propagation models, extreme values can be a common phenomenon that can be, especially if there are nonlinearities between the observed variable and the modeled states, problematic during the filter update. An illustrative example of this problem is shown using an unsaturated flow model where the modeled states are pressure heads and observations are water content. It is demonstrated that the ensemble Kalman filter can in this case yield a deterioration of state predictions. We discuss the normal score transform and the transform with the retention function applied to the model states in order to mitigate this problem. It is shown that both transforms improve the estimation of the model states and parameters.

Streptococcus gordonii is one of the first colonizing bacteria on tooth or dental implant materials forming so-called biofilm. These biofilms cause severe inflammation that, as a consequence, lead to tooth or implant failure. A promising way to study the S. gordonii biofilm formation is by combining experimental investigations and numerical simulation. Our previous research has shown the potential to model the growth process of the S. gordonii biofilm in a mimic human oral environment by using a mathematical model developed by Alpkvist and Klapper (the A-K model). The parameters used for the simulation were calibrated by the experimental results. However, what are the crucial parameters that have a strong influence on the biofilm growth behavior and thus need to be determined accurately is an open question. In this paper, parameter studies on four independent parameters are carried out.

Porous transport layers (PTL) are key components of proton exchange membrane water electrolysis (PEMWE) cells controlling species transport. Further optimization requires better understanding of how PTLs influence overpotentials. In this work, the data from an electrochemical overpotential breakdown is compared to a state-of-the-art model, which includes a Nernstian overpotential description, two-phase Darcian flow and advective-diffusive mass transport. Model parameters are derived from X-ray tomographic measurements, pore-scale calculations, standard models for porous materials and by transferring ex situ measurements from other materials. If the parameter set is available, model results and experimental data match well concerning PTL-related overpotentials at different current densities and operating pressures. Both experimental and modeling results suggest that mass transport through PTLs does not affect a considerable, pressure-independent share of mass transport overpotentials. Both model results and experimental findings conclude that mass transport through the cathode PTL causes overpotentials more than twice as high as through its anode counterpart. Further research opportunities regarding the relationship between PTL bulk properties and experimentally determined mass transport overpotentials are identified. © 2020 The Author(s). Published on behalf of The Electrochemical Society by IOP Publishing Limited.