Causal prior-embedded physics-informed neural networks and a case study on metformin transport in porous media.

This study introduces a novel approach to transport modelling by integrating experimentally derived causal priors into neural networks. We illustrate this paradigm using a case study of metformin, a ubiquitous pharmaceutical emerging pollutant, and its transport behaviour in sandy media. Specifically, data from metformin's sandy column transport experiment was used to estimate unobservable parameters through a physics-based model Hydrus-1D, followed by a data augmentation to produce a more comprehensive dataset. A causal graph incorporating key variables was constructed, aiding in identifying impactful variables and estimating their causal dynamics or "causal prior." The causal priors extracted from the augmented dataset included underexplored system parameters such as the type-1 sorption fraction F, first-order reaction rate coefficient α, and transport system scale. Their moderate impact on the transport process has been quantitatively evaluated (normalized causal effect 0.0423, -0.1447 and -0.0351, respectively) with adequate confounders considered for the first time. The prior was later embedded into multilayer neural networks via two methods: causal weight initialization and causal prior regularization. Based on the results from AutoML hyperparameter tuning experiments, using two embedding methods simultaneously emerged as a more advantageous practice since our proposed causal weight initialization technique can enhance model stability, particularly when used in conjunction with causal prior regularization. amongst those experiments utilizing both techniques, the R-squared values peaked at 0.881. This study demonstrates a balanced approach between expert knowledge and data-driven methods, providing enhanced interpretability in black-box models such as neural networks for environmental modelling.

Modelling groundwater pollutant transfer mineral micropollutants in a multi-layered aquifer in Burkina Faso (West African Sahel).

In Burkina Faso, human activities around water points in rural areas affect groundwater resources, which become unfit for consumption. Nearly 33.5% of boreholes are subject to point source pollution. The assessment of the evolution of such pollution should be monitored to assess groundwater quality. In addition, withdrawals for irrigation alone are estimated at 85%, i.e. 46% of the water demand, heightening the deterioration in quality while creating depression zones further leading to an increase in recharge. It is therefore critical to understand the evolution and fate of the transfer of pollutants in such environments. In this study, we aimed to model the transfer of pollutant and predict the future state of pollution using the MT3D-USGS Groundwater Solute Transport Simulator code through the Groundwater Modelling Software (GMS) over the period 2012-2062 (50 years). A mathematical model is further developed through inferential statistics and used as a surrogate model for comparison. The results showed that deterioration in water quality was more attributable to withdrawals, especially for Cyanide (Cn) and Arsenic (As). A rather slow degradation is reported for Lead (Pb), which extends over 22 km, and Fluoride (F), which extends from 4 to 10 km due to localized recharge. A faster degradation for Cn over a distance of 2-16 km and as from 3 to 11 km is also observed because of the geological setting of the subsoil. These results might assist decision-makers for the quantitative and qualitative management of groundwater resources, and the management of the basement aquifer in the area through the establishment of protection zones.

Management of saltwater intrusion using 3D numerical modelling: a first for Pacific Island country of Vanuatu.

Small island countries like Vanuatu are facing the brunt of climate change, sea level rise (SLR), tropical cyclones, and limited or declining access to freshwater. The Tagabe coastal aquifer in Port Vila (the capital of Vanuatu) shows the presence of salinity, indicating saltwater intrusion (SWI). This study aims to develop and evaluate effective SWI management strategies for Tagabe coastal aquifer. To manage SWI, the numerical simulation model for the study area was developed using the SEAWAT code. The flow model was developed using MODFLOW and the transport model was developed using MT3DMS. Whereby SEAWAT solved flow and transport equations simultaneously. The model was calibrated, and different scenarios were evaluated for the management of SWI. The SLR was also considered in the model simulations. The results indicated that increased population, pumping rates, and SLR affect the SWI rates. To manage the SWI, we introduced hydraulic barriers like barrier wells and injection wells which effectively managed SWI in Tagabe coastal aquifer. The results from this study are significantly important whereby, the water managers, site owners, and governing bodies can use the management strategies presented in this study to create policies and regulations for managing SWI rates in Port Vila. Additionally, the water industry, private businesses, and investors who wish to extract groundwater from the Tagabe can use this study as a reference for daily or yearly freshwater production rates without the risk of SWI.

Data scarce modelling the impact of present and future groundwater development on Jordan multiaquifer groundwater resources.

Rapidly growing demands and climate change stresses water resources worldwide and leads to highly competitive situations between the environment and socio-economic development of a region, calling for a smart and modelling driven water resources management. However, data scarcity often prevents the realisation of a comprehensive, nation-wide resources model, which provides reliable and spatially discretized results of water resources development. We present a workflow approach to set up a large-scale multi-aquifer model, overcoming data shortage by stepwise calibration and integrating hydrological and numerical groundwater flow modelling into a coupled system. The study aims to develop such a system to assess how groundwater resources react on anthropogenic impacts on the example of the Kingdom of Jordan, one of the water poorest countries on globe. Simulated heads reliably resembled the monitored ones in >70 % of the observation wells. That makes us confident, the model represents all the states well from 1970, prior to the intense development of the country until 2015. The water balance shows an annual deficit of 1.16 million cubic meter (MCM) due to an overdraft. The discharge to the Dead Sea increased from 564 MCM/yr to 696 MCM/yr over the time period. Regional drawdowns of >250 m and groundwater depression with an extension of approx.100 km are observable in both large aquifer complexes. Most severe areas in the upper calcareous aquifer are located in the north of Amman and practically in all urban and agricultural agglomerations across the country. Groundwater tables in the deeper sandstone aquifer are particularly affected in the south as well as in the wider vicinity of the Dead Sea as consequence of its continuous dropping. Simulations of the future development of the groundwater tables indicate a severe deterioration of the situation with further declines in groundwater levels of up to 70 m.

Do transient hydrological processes explain the variability of strontium-90 activity in groundwater downstream of a radioactive trench near Chernobyl?

The Chernobyl Pilot Site (CPS) was created in 2000 in order to study radionuclide migration processes to the geosphere from radioactive material of the Red forest buried in a trench. In this article, the data collected in the CPS up to 2015 are analyzed to identify the links between hydrological conditions and release of strontium-90 (90Sr) from the trench. Then, a flow-and-transport model is used for simulating distribution of 90Sr both in the unsaturated and saturated zones downstream of the trench. The results show that the 90Sr activity in groundwater is strongly transient in time, due to the high inter-annual variability of both the recharge rate and the groundwater level (some particularly wet winters resulted in saturation of the bottom part of the trench). In addition, the parameters that govern the sorption of 90Sr in trench material appear to vary significantly in space (the retardation factor ranges from 10 to 50 depending on the location). This spatiotemporal variability could hide some critical processes, e.g., related to a long-term trend, and needs to be characterized through an appropriate sampling frequency.

Importance of hydraulic travel time for the evaluation of organic compounds removal in bank filtration.

The growing global demand for drinking water is driving both the diversification of water supply sources and their sustainability. River bank filtration (RBF) is an excellent option since it strongly reduces the extent of treatment steps compared to direct usage of surface water. Organic micropollutants (e.g. pharmaceuticals) are widely recognized as a hazard in drinking water production from surface water. Due to their potentially high mobility, stability, bioaccumulation and persistency, these substances can pass through RBF-systems. Scientific studies on compound removal and attenuation efficiency of RBF rely on the knowledge of travel time to compare concentrations in the river to the ones in the bank filtrate since water quality in rivers can change rapidly. However, bank filtrate samples represent a mixture of water with different travel times as the flow paths vary. This has not yet been considered in studies of bank filtration removal efficiency for organic micro pollutants. Here we present a method that considers the residence-time distribution of the bank filtrate sample obtained by groundwater modelling to evaluate the removal efficiency of RBF for organic micropollutants. The method was tested in a comprehensive study with 50 samples taken over a one-year-period at a river bank filtration site in Vienna (Austria). Our findings revealed that better coverage of varying river water quality (higher sampling frequency during the period of infiltration) resulted not only in a higher number of compounds considered as removed but also significantly reduced the number of compounds considered to have formed during the RBF process. The application of the presented method indicated that RBF is very effective in removing organic micropollutants. Considering different travel times will provide better models and a better understanding of the potential of RBF for pollutant removal and thus supports its safe application as a solution to the growing demand for drinking water.

Projecting the hydrochemical trajectory of a constructed fen watershed: Implications for long-term wetland function.

Surface mining operations for bitumen have fundamentally altered large areas of boreal forest and fen peatland in the Athabasca Oil Sands Region (AOSR) of Alberta, Canada. Pilot projects intended to assess the feasibility of fen construction as a reclamation option have been designed, built, and are currently undergoing monitoring. Initial assessments of ecohydrologic function have been conducted for these systems but offer limited insight into their evolution and likely successional pathway. Thus, this study projects the hydrologic and geochemical behaviour of a constructed fen watershed to understand whether the system will be capable of supporting peatland processes into the future. A numerical groundwater flow and sodium transport model was calibrated and validated with 7 years of hydraulic head, water flux, and water chemistry data. Based on Monte Carlo simulations, the projected fen water table would be stable and remain close to the surface (<15 cm), indicating that the design of the system can generate sufficient water quantity to meet evaporative demand and maintain surface water discharge. However, water quality was more sensitive to climatic variability, which induced a large range in potential sodium concentrations at the fen surface (450-850 mg L-1). Evapoconcentration of salts across the surface of the fen will likely limit moss establishment for decades following construction. Yet stress-thresholds of salt-tolerant vegetation like sedges will not be exceeded. Ultimately, these projections support the original design principles and philosophy that guided the creation of the watershed. Nonetheless, this work indicates that increasing the area of the fen relative to the upland would not have a detrimental impact on the ability of the system to maintain a high water table. This could allow for the proportion of peatlands on the reclamation landscape to reflect the pre-disturbance environment more faithfully.

Beneficial effects of dynamic groundwater flow and redox conditions on Natural Attenuation of mono-, poly-, and NSO-heterocyclic hydrocarbons.

Natural Attenuation (NA) processes have been demonstrated to reduce pollutant loads at different contaminated groundwater sites world-wide and are increasingly considered in contaminated site management concepts. However, data are mainly available for steady state groundwater flow and stable redox conditions as well as pollutants listed in standard regulatory schemes. In this study, the influence of transient groundwater flow and redox conditions on NA was examined at a former gas works site near the river Rhine in Germany. The investigated 78 pollutants included 40 mono- and polyaromatic hydrocarbons (MAHs, PAHs) and 38 NSO-heterocyclic aromatic hydrocarbons (NSO-HET). In the highly polluted areas, the MAHs benzene, indene and indane, the PAHs naphthalene, acenaphthene, 1- and 2-methylnaphthalene and the NSO-HET 2-methylquinoline, carbazole, benzothiophene, dibenzofuran and benzofuran were predominant. Pollutant concentrations decreased with increasing distance from the sources of contamination. At the plume fringes, the MAHs benzene and indane, the PAH acenaphthene, the NSO-HET carbazole, 5-methylbenzothiophene, 2- and 3-methylbenzofuran and 2-methyldibenzofuran were predominant, indicating low retention and slow intrinsic biodegradation of these compounds. The influence of surface water on groundwater level, pollutant concentrations, and redox conditions in the monitoring wells was observed with a permanently installed groundwater sensor. The temporary availability of oxygen was observed at the plume fringes, resulting in aerobic and ferric iron reducing biodegradation processes. Field and laboratory data were used to set-up a groundwater flow and reactive transport model used for quantification of the field mass transfer rates. In conclusion, the study demonstrates that NA is effective under transient flow and redox conditions. A conceptual model and reactive transport simulation can facilitate the interpretation of pronounced fluctuations of pollutant concentration in monitoring wells. Based on the analysis of 78 pollutants, indane, indene and several NSO-HET like carbazole, benzothiophene and 2-methyldibenzofuran are recommended for monitoring at tar oil polluted sites, besides EPA-PAHs and BTEX.

Numerical assessment of the effect of water-saving irrigation on the water cycle at the Manas River Basin oasis, China.

Mulch drip irrigation is widely used in the arid areas of Northwest China. Consequently, the Manas River Basin has developed into the fourth largest irrigated agricultural area in China. In this study, a groundwater model of the regional water cycle was developed to quantitatively assess the groundwater balance in response to different irrigation schemes, including traditional irrigation, conventional water-saving irrigation, and high-efficiency water-saving irrigation schemes. Our results reveal that 1) The water-saving irrigation technology has affected the water cycle process in farmlands. The higher the degree of water conservation, the lower the infiltration into groundwater, the higher the deficit of the groundwater balance, and the more significant the decline of the groundwater level. 2) The groundwater at the Manas River Basin remains in a negative equilibrium state. To achieve an equilibrium state of the groundwater at the Manas River Basin, the catchment management agencies should restrict the scale of oasis development and the utilization of groundwater.

Topsoil structure stability in a restored floodplain: Impacts of fluctuating water levels, soil parameters and ecosystem engineers.

Ecosystem services provided by floodplains are strongly controlled by the structural stability of soils. The development of a stable structure in floodplain soils is affected by a complex and poorly understood interplay of hydrological, physico-chemical and biological processes. This paper aims at analysing relations between fluctuating groundwater levels, soil physico-chemical and biological parameters on soil structure stability in a restored floodplain. Water level fluctuations in the soil are modelled using a numerical surface-water-groundwater flow model and correlated to soil physico-chemical parameters and abundances of plants and earthworms. Causal relations and multiple interactions between the investigated parameters are tested through structural equation modelling (SEM). Fluctuating water levels in the soil did not directly affect the topsoil structure stability, but indirectly through affecting plant roots and soil parameters that in turn determine topsoil structure stability. These relations remain significant for mean annual days of complete and partial (>25%) water saturation. Ecosystem functioning of a restored floodplain might already be affected by the fluctuation of groundwater levels alone, and not only through complete flooding by surface water during a flood period. Surprisingly, abundances of earthworms did not show any relation to other variables in the SEM. These findings emphasise that earthworms have efficiently adapted to periodic stress and harsh environmental conditions. Variability of the topsoil structure stability is thus stronger driven by the influence of fluctuating water levels on plants than by the abundance of earthworms. This knowledge about the functional network of soil engineering organisms, soil parameters and fluctuating water levels and how they affect soil structural stability is of fundamental importance to define management strategies of near-natural or restored floodplains in the future.

Probabilistic risk assessment of nitrate groundwater contamination from greenhouses in Albenga plain (Liguria, Italy) using lysimeters.

The use of fertilizers in greenhouse-grown crops can pose a threat to groundwater quality and, consequently, to human beings and subterranean ecosystem, where intensive farming produces pollutants leaching. Albenga plain (Liguria, Italy) is an alluvial area of about 45km2 historically devoted to farming. Recently the crops have evolved to greenhouses horticulture and floriculture production. In the area high levels of nitrates in groundwater have been detected. Lysimeters with three types of reconstituted soils (loamy sand, sandy clay loam and sandy loam) collected from different areas of Albenga plain were used in this study to evaluate the leaching loss of nitrate (NO3-) over a period of 12weeks. Leaf lettuce (Lactuca sativa L.) was selected as a representative green-grown crop. Each of the soil samples was treated with a slow release fertilizer, simulating the real fertilizing strategy of the tillage. In order to estimate the potential risk for aquifers as well as for organisms exposed via pore water, nitrate concentrations in groundwater were evaluated by applying a simplified attenuation model to the experimental data. Results were refined and extended from comparison of single effects and exposure values (Tier I level) up to the evaluation of probabilistic distributions of exposure and related effects (Tier II, III IV levels). HHRA suggested HI >1 and about 20% probability of exceeding RfD for all the greenhouses, regardless of the soil. ERA suggested HQ>100 for all the greenhouses; 93% probability of PNEC exceedance for greenhouses containing sand clay loam. The probability of exceeding LC50 for 5% of the species was about 40% and the probability corresponding to DBQ of DEC/EC50>0.001 was >90% for all the greenhouses. The significantly high risk, related to the detected nitrate leaching loss, can be attributed to excessive and inappropriate fertigation strategies.

The hydrological functioning of a constructed fen wetland watershed.

Mine reclamation requires the reconstruction of entire landforms and drainage systems. The hydrological regime of reclaimed landscapes will be a manifestation of the processes operating within the individual landforms that comprise it. Hydrology is the most important process regulating wetland function and development, via strong controls on chemical and biotic processes. Accordingly, this research addresses the growing and immediate need to understand the hydrological processes that operate within reconstructed landscapes following resource extraction. In this study, the function of a constructed fen watershed (the Nikanotee Fen watershed) is evaluated for the first two years following construction (2013-2014) and is assessed and discussed within the context of the construction-level design. The system design was capable of sustaining wet conditions within the Nikanotee Fen during the snow-free period in 2013 and 2014, with persistent ponded water in some areas. Evapotranspiration dominated the water fluxes from the system. These losses were partially offset by groundwater discharge from the upland aquifer, which demonstrated strong hydrologic connectivity with the fen in spite of most construction materials having lower than targeted saturated hydraulic conductivities. However, the variable surface infiltration rates and thick placement of a soil-capping layer constrained recharge to the upland aquifer, which remained below designed water contents in much of the upland. These findings indicate that it is possible to engineer the landscape to accommodate the hydrological functions of a fen peatland following surface oil sands extraction. Future research priorities should include understanding the storage and release of water within coarse-grained reclaimed landforms as well as evaluating the relative importance of external water sources and internal water conservation mechanisms for the viability of fen ecosystems over the longer-term.

Shallow aquifer response to climate change scenarios in a small catchment in the Guarani Aquifer outcrop zone

ABSTRACT Water availability restrictions are already a reality in several countries. This issue is likely to worsen due to climate change, predicted for the upcoming decades. This study aims to estimate the impacts of climate change on groundwater system in the Guarani Aquifer outcrop zone. Global Climate Models (GCM) outputs were used as inputs to a water balance model, which produced recharge estimates for the groundwater model. Recharge was estimated across different land use types considering a control period from 2004 to 2014, and a future period from 2081 to 2099. Major changes in monthly rainfall means are expected to take place in dry seasons. Most of the analysed scenarios predict increase of more than 2 ºC in monthly mean temperatures. Comparing the control and future runs, our results showed a mean recharge change among scenarios that ranged from ~-80 to ~+60%, depending on the land use type. As a result of such decrease in recharge rates, the response given by the groundwater model indicates a lowering of the water table under most scenarios.

Understanding changes in the hydrological behaviour within a karst aquifer (Lurbach system, Austria).

A thorough data analysis combined with groundwater modelling was conducted in an Austrian binary karst aquifer to better understand changes in the hydrological behaviour observed at a karst spring. During a period of 4 years after a major flood event the spring hydrograph appears to be more damped with lower peak flow and higher baseflow than in the years before. The analysis of the hydrograph recession suggests that the observed hydrological change is caused by changes within the karst system rather than by varying hydro-meteorological conditions. The functioning of the aquifer and potential causes of the observed changes are further examined using the groundwater flow model MODFLOW. The simulation results suggest that a modification of hydraulic conductivity and storage within the conduit network, e.g. due to the plugging of the drainage conduits with sediments, may be the cause of the different behaviour. MODFLOW was able to reproduce the observed dynamics of spring flow, although it does not account for turbulent flow within karst conduits. Using a simplified model scenario it is demonstrated that the damping of the hydrograph is much stronger if turbulent conduit flow is taken into account. Thus, a turbulent flow model is needed to assess potential changes in the storage properties quantitatively.

Coastal hydrogeological system of Mar Piccolo (Taranto, Italy).

The Mar Piccolo basin is an internal sea basin located along the Ionian coast (Southern Italy), and it is surrounded primarily by fractured carbonate karstic environment. Because of the karstic features, the main continental water inflow is from groundwater discharge. The Mar Piccolo basin represents a peculiar and sensitive environment and a social emergency because of sea water and sediment pollution. This pollution appears to be caused by the overlapping effects of dangerous anthropogenic activities, including heavy industries and commercial and navy dockyards. The paper aims to define the contribution of subaerial and submarine coastal springs to the hydrological dynamic equilibrium of this internal sea basin. A general approach was defined, including a hydrogeological basin border assessment to detect inflowing springs, detailed geological and hydrogeological conceptualisation, in situ submarine and subaerial spring measurements, and flow numerical modelling. Multiple sources of data were obtained to define a relevant geodatabase, and it contained information on approximately 2000 wells, located in the study area (1600 km(2)). The conceptualisation of the hydrogeological basin, which is 978 km(2) wide, was supported by a 3D geological model that interpolated 716 stratigraphic logs. The variability in hydraulic conductivity was determined using hundreds of pumping tests. Five surveys were performed to acquire hydro-geochemical data and spring flow-yield measurements; the isotope groundwater age was assessed and used for model validation. The mean annual volume exchanged by the hydrogeological basin was assessed equal to 106.93 10(6) m(3). The numerical modelling permitted an assessment of the mean monthly yield of each spring outflow (surveyed or not), travel time, and main path flow.

Bioremediation: how to deal with removal efficiency uncertainty? An economic application.

Bioremediation is a remediation strategy, which has considerable strength but also certain limitations. Complex and uncertain relationships among biomass, contaminants, and nutrients lead to an uncertain level of removal efficiency. The uncertainty inherent to a bioremediation strategy should be addressed in the remediation selection process. In order to evaluate the bioremediation strategy economically, this study takes into account the reversibility of a decision. A decision tree structures the different remediation strategies, thus giving the possible courses of action open to the decision maker. The option value indicates the importance of having the possibility to reverse a previously made decision. Compared with conventional economic evaluation tools, more information to ground the selection made is revealed.