Search

Groundwater resources play an important role for irrigation, particularly in arid and semi-arid regions, where groundwater depletion poses a critical threat to agricultural production and associated local livelihoods. However, the relationship between groundwater use, farming, and poverty, particularly with regards to informal mechanisms of resources management, remains poorly understood. Here, we assess this relationship by developing a behavioural model of groundwater user groups, empirically grounded in the politically fragile context of Tunisia. The model integrates biophysical aquifer dynamics, institutional governance, and farmer decision-making, all of which are co-occurring under conditions of aquifer depletion and illicit groundwater extraction. The paper examines how community-level norms drive distributional outcomes of farmer behaviours and traces pathways of local system collapse - whether hydrogeological or financial. Through this model, we explore how varying levels of trust and leadership, ecological conditions, and agricultural strategies can delay or avoid collapse of the social-ecological system. Results indicate limits to collective action under path-dependent aquifer depletion, which ultimately leads to the hydrogeological collapse of groundwater user groups independent of social and institutional norms. Despite this inevitable hydrogeological collapse of user groups, the most common cause of water user group failure is bankruptcy, which is linked to the erosion of social norms regarding fee payment. Social and institutional norms, however, can serve to delay the financial collapse of user groups. In the politically fragile system of Tunisia, low levels of trust in government result in low social penalties for illicit water withdrawals. In the absence of alternative irrigation sources, this serves as a temporary buffer against income-poverty. These results highlight the need for polycentric coordination at the aquifer-level as well as income div

To develop appropriate climate change adaptation plans, evidence of the effectiveness of adaptation measures is required. At a regional scale, however, this information is usually lacking. The region of Seewinkel in Austria was taken as a case study because of its extensive agricultural industry and its unique ecosystem of saline lakes. The goal of the study was to provide stakeholders with evidence to support their climate change adaptation process. Adaptation measures discussed by local stakeholders were analyzed to determine their efficacy. A system dynamics (SD) based model was developed to serve as a tool for the water policy analysis and to be used in place of advanced hydrological models. The model was calibrated using observational data and forced with bias-adjusted EURO-CORDEX climate data for three representative concentration pathways (RCPs) (2010–2100). Three parameters in the model were changed to simulate adaptation measures. The results showed that combined measures, increasing irrigation efficiency and changing crops could reduce water demand by an average of 40 %, 23 % and 23 %, respectively, for all RCPs. The local aquifer's level could be increased above the historical average by an average of 0.43 m by combined measures, 0.20 m by increasing irrigation efficiency, 0.20 m by changing crops and 0.06 m by artificially recharging the aquifer.

Hydro-economic modeling (HEM) addresses research and policy questions from socioeconomic and biophysical perspectives under a broad range of water-related topics. Applications of HEM include economic evaluations of existing and new water projects, alternative water management actions or policies, risk assessments from hydro-climatic uncertainty (e.g., climate change), and the costs and benefits of mitigation and/or adaptation to such events. This paper reviews applications of HEM in five different categories: (1) climate change impacts and adaptation, (2) water–food–energy–ecosystems nexus management, (3) capability to link to other models, (4) innovative water management options, and (5) the ability to address and integrate uncertainty. We find that (i) the increasing complexity and heterogeneity of water resource management problems due to the growing demand and competition for water across economic sectors, (ii) limited availability and high costs of developing additional supplies, and (iii) emerging recognition and consideration of environmental water demands and value, have inspired new integrated hydro-economic problems and models to address issues of water–food–energy nexus sustainability, resilience, reliability through water (re)allocation based on the relative “value” of water uses. In the past decade, the field of HEM has improved the integration of ecosystem needs, but their representation is still insufficient and mostly ineffective. HEM studies address how to sustainably manage water resources, including groundwater which has become an area of particular interest in climate change adaptation. The current most used spatial and temporal resolutions (basin-scale and yearly time-step) are appropriate for planning but not for operational decisions and could be underestimating impacts from extreme events (e.g., flood risk) captured only by sub-monthly time scales. In addition, HEM primarily focuses on biophysical and economic indicators but often overlooks p

Neural networks have been shown to be extremely effective rainfall-runoff models, where the river discharge is predicted from meteorological inputs. However, the question remains: what have these models learned? Is it possible to extract information about the learned relationships that map inputs to outputs, and do these mappings represent known hydrological concepts? Small-scale experiments have demonstrated that the internal states of long short-term memory networks (LSTMs), a particular neural network architecture predisposed to hydrological modelling, can be interpreted. By extracting the tensors which represent the learned translation from inputs (precipitation, temperature, and potential evapotranspiration) to outputs (discharge), this research seeks to understand what information the LSTM captures about the hydrological system. We assess the hypothesis that the LSTM replicates real-world processes and that we can extract information about these processes from the internal states of the LSTM. We examine the cell-state vector, which represents the memory of the LSTM, and explore the ways in which the LSTM learns to reproduce stores of water, such as soil moisture and snow cover. We use a simple regression approach to map the LSTM state vector to our target stores (soil moisture and snow). Good correlations (R2>0.8) between the probe outputs and the target variables of interest provide evidence that the LSTM contains information that reflects known hydrological processes comparable with the concept of variable-capacity soil moisture stores. The implications of this study are threefold: (1) LSTMs reproduce known hydrological processes. (2) While conceptual models have theoretical assumptions embedded in the model a priori, the LSTM derives these from the data. These learned representations are interpretable by scientists. (3) LSTMs can be used to gain an estimate of intermediate stores of water such as soil moisture. While machine learning interpretability is sti

This study focuses on the impact of direct assimilation of Tropical Rainfall Measuring Mission (TRMM) Microwave Imager (TMI) and Visible and Infrared Scanner (VIRS) channels radiances in the prediction of Tropical cyclones (TCs) in the Bay of Bengal (BOB) region. For this purpose, two TCs, viz., Jal and Thane are simulated by using the Weather Research and Forecasting (WRF) model. Artificial Neural Network (ANN) based fast forward radiative transfer codes are developed for both the TMI and VIRS channels to speed up the simulation of radiances from vertical profiles of the atmosphere. For the WRF model initialization, initial ensembles are generated by perturbing atmospheric variables such as temperature (T, K), pressure (P, hpa), relative humidity (RH, %), meridional (U, m/s) and zonal winds (V, m/s) using Empirical Orthogonal function (EOF) technique. Further, each ensemble member is integrated up to a time that is close to the subsequent overpass of TRMM. Simulated profiles are obtained from the assimilated ensemble members which are used to generate the brightness temperatures through the fast ANN based fast forward radiative transfer codes. A Bayesian-based ensemble data assimilation technique is then developed for assimilating both the rainy and clear sky radiances, wherein the likelihoods are used to determine the conditional probabilities of all the candidates in the ensemble by comparing the TRMM observed radiances with the simulated radiances. Based on the posterior probability densities of each member of the ensemble, the initial conditions (ICs) at 00 UTC are corrected using a linear weighted average of initial ensembles for the all atmospheric variables. With these weighted average ICs, the WRF model is then executed all the way up to the required forecast period. Simulation results thus obtained with the assimilation are compared with the observations provided by the Joint Typhoon Warning Center (JTWC) and also the control run (i.e., WRF simulations san

In Austria, increase in demand as well as climate change might create local and seasonal hot-spots of water stress. It is thus important to understand the status quo and future development of these phenomena to identify potential areas of tension. WaterStressAT assesses water availability and demand in two Austrian case studies under a set of regional development and climate change scenarios.

Unsustainable depletion of aquifer storage and diversion of groundwater from downstream users continue to be serious problems, even though their adverse effects are widely recognized. Groundwater depletion involves interactions between economically motivated pumping decisions and physical constraints. Here, we investigate these interactions by using optimal control techniques to describe the pumping decisions of economic agents who share an aquifer. Our approach relies on a multi-scale description of unconfined groundwater flow, applied to a computational experiment that illustrates some important impacts of aquifer development. We show that cooperative groundwater management can provide higher economic benefits, with less storage depletion, than an uncooperative alternative. However, demand for water still drives pumping decisions, even when users cooperate. In many aquifers the benefits of pumping are necessarily accompanied by a reduction in aquifer outflow and by depletion of the groundwater reserve, for both cooperative and uncooperative management. These benefits decrease significantly when pumping is limited by well yield constraints or when the diversion of aquifer outflow is restricted. Our overall analysis emphasizes the importance of viewing aquifer management as part of a larger resource allocation problem that considers the conflicting needs of well operators, downstream users, and future consumers of the resource.

Sustainable management of groundwater resources under changing climatic conditions require an application of reliable and accurate predictions of groundwater levels. Mechanistic multi-scale, multi-physics simulation models are often too hard to use for this purpose, especially for groundwater managers who do not have access to the complex compute resources and data. Therefore, we analyzed the applicability and performance of four modern deep learning computational models for predictions of groundwater levels. We compare three methods for optimizing the models’ hyperparameters, including two surrogate model-based algorithms and a random sampling method. The models were tested using predictions of the groundwater level in Butte County, California, USA, taking into account the temporal variability of streamflow, precipitation, and ambient temperature. Our numerical study shows that the optimization of the hyperparameters can lead to reasonably accurate performance of all models (root mean squared errors of groundwater predictions of 2 meters or less), but the “simplest” network, namely a multilayer perceptron (MLP) performs overall better for learning and predicting groundwater data than the more advanced long short-term memory or convolutional neural networks in terms of prediction accuracy and time-to-solution, making the MLP a suitable candidate for groundwater prediction.