In this study we investigate how climate change will directly influence the groundwater resources in Germany during the 21
century. We apply a machine learning groundwater level prediction approach ...based on convolutional neural networks to 118 sites well distributed over Germany to assess the groundwater level development under different RCP scenarios (2.6, 4.5, 8.5). We consider only direct meteorological inputs, while highly uncertain anthropogenic factors such as groundwater extractions are excluded. While less pronounced and fewer significant trends can be found under RCP2.6 and RCP4.5, we detect significantly declining trends of groundwater levels for most of the sites under RCP8.5, revealing a spatial pattern of stronger decreases, especially in the northern and eastern part of Germany, emphasizing already existing decreasing trends in these regions. We can further show an increased variability and longer periods of low groundwater levels during the annual cycle towards the end of the century.
Despite many existing approaches, modeling karst water resources remains challenging as conventional approaches usually heavily rely on distinct system knowledge. Artificial neural networks (ANNs), ...however, require only little prior knowledge to automatically establish an input–output relationship. For ANN modeling in karst, the temporal and spatial data availability is often an important constraint, as usually no or few climate stations are located within or near karst spring catchments. Hence, spatial coverage is often not satisfactory and can result in substantial uncertainties about the true conditions in the catchment, leading to lower model performance. To overcome these problems, we apply convolutional neural networks (CNNs) to simulate karst spring discharge and to directly learn from spatially distributed climate input data (combined 2D–1D CNNs). We investigate three karst spring catchments in the Alpine and Mediterranean region with different meteorological–hydrological characteristics and hydrodynamic system properties. We compare the proposed approach both to existing modeling studies in these regions and to our own 1D CNN models that are conventionally trained with climate station input data. Our results show that all the models are excellently suited to modeling karst spring discharge (NSE: 0.73–0.87, KGE: 0.63–0.86) and can compete with the simulation results of existing approaches in the respective areas. The 2D models show a better fit than the 1D models in two of three cases and automatically learn to focus on the relevant areas of the input domain. By performing a spatial input sensitivity analysis, we can further show their usefulness in localizing the position of karst catchments.
•NARX were applied to obtain groundwater level forecasts with lead times up to half a year.•Porous, fractured and karst aquifers with and without external influences on groundwater levels.•The ...developed approach is easily transferable on other wells.•Input- and feedback delays were determined by applying STL time series decomposition.•The results indicate an outstanding suitability of NARX for groundwater level predictions.
While the application of neural networks for groundwater level forecasting in general has been investigated by many authors, the use of nonlinear autoregressive networks with exogenous inputs (NARX) is relatively new. For this work NARX were applied to obtain groundwater level forecasts for several wells in southwest Germany. Wells in porous, fractured and karst aquifers were investigated and forecasts of lead times up to half a year were conducted for both influenced (e.g. nearby pumping) and uninfluenced wells. Precipitation and temperature were chosen as predictors, which makes the selected approach easily transferable, since both parameters are widely available and simple to measure. Input and feedback delays were determined by applying STL time series decomposition on the data and using auto- and cross-correlation functions on the remainders to determine significant time lags. Coefficient of determination, (relative) root mean squared error and Nash-Sutcliffe efficiency were used to evaluate forecasts, the model selection was based on an out-of-sample validation on rolling basis. The results are promising and indicate an outstanding suitability of NARX for groundwater level predictions with such a small set of inputs in all three aquifer types.
The prediction of groundwater nitrate concentration's response to geo-environmental and human-influenced factors is essential to better restore groundwater quality and improve land use management ...practices. In this paper, we regionalize groundwater nitrate concentration using different machine learning methods (Random forest (RF), unimodal 2D and 3D convolutional neural networks (CNN), and multi-stream early and late fusion 2D-CNNs) so that the nitrate situation in unobserved areas can be predicted. CNNs take into account not only the nitrate values of the grid cells of the observation wells but also the values around them. This has the added benefit of allowing them to learn directly about the influence of the surroundings. The predictive performance of the models was tested on a dataset from a pilot region in Germany, and the results show that, in general, all the machine learning models, after a Bayesian optimization hyperparameter search and training, achieve good spatial predictive performance compared to previous studies based on Kriging and numerical models. Based on the mean absolute error (MAE), the random forest model and the 2DCNN late fusion model performed best with an MAE (STD) of 9.55 (0.367) mg/l, R
2
= 0.43 and 10.32 (0.27) mg/l, R
2
= 0.27, respectively. The 3DCNN with an MAE (STD) of 11.66 (0.21) mg/l and largest resources consumption is the worst performing model. Feature importance learning from the models was used in conjunction with partial dependency analysis of the most important features to gain greater insight into the major factors explaining the nitrate spatial variability. Large uncertainties in nitrate prediction have been shown in previous studies. Therefore, the models were extended to quantify uncertainty using prediction intervals (PIs) derived from bootstrapping. Knowledge of uncertainty helps the water manager reduce risk and plan more reliably.
•Nine groundwater level time-series with records of more than 100 years were analyzed.•First study of aquifer responses to the teleconnections AMO and ENSO in Europe.•Results are a valuable ...contribution to longer-term forecasting of groundwater levels.
Groundwater is an important resource for drinking water supply, and is subject to natural variation as well as climate change effects. It has been shown that long-term natural variations of groundwater levels can often be attributed to climatic oscillations. Long-term groundwater level records are rare, but of special importance for the detection of longer, decadal to multi-decadal periodicities, which are vital for predictions of future development of groundwater levels and the distinction between natural variation and climate change effects. We have examined periodicities of nine groundwater level time-series with records of more than 100 years, as well as possible impacts of climatic teleconnections (NAO, AMO and ENSO) with wavelet analysis. The monitoring wells are located in Germany, Netherlands, UK and Denmark and cover different depths to groundwater, different aquifer types and hydraulic conditions. Our results show that all evaluated monitoring wells exhibit significant relations to long-term climatic periodicities of NAO, AMO and ENSO. Among the wells in phreatic porous aquifers, there is a signal damping, which can be related to the thickness of the unsaturated zone. Further, the damping is higher in the lower permeable aquifers and there is less damping in the karstic aquifers compared to the porous aquifers, in spite of a much thicker unsaturated zone.
It is now well established to use shallow artificial neural networks (ANNs) to obtain accurate and reliable groundwater level forecasts, which are an important tool for sustainable groundwater ...management. However, we observe an increasing shift from conventional shallow ANNs to state-of-the-art deep-learning (DL) techniques, but a direct comparison of the performance is often lacking. Although they have already clearly proven their suitability, shallow recurrent networks frequently seem to be excluded from the study design due to the euphoria about new DL techniques and its successes in various disciplines. Therefore, we aim to provide an overview on the predictive ability in terms of groundwater levels of shallow conventional recurrent ANNs, namely non-linear autoregressive networks with exogenous input (NARX) and popular state-of-the-art DL techniques such as long short-term memory (LSTM) and convolutional neural networks (CNNs). We compare the performance on both sequence-to-value (seq2val) and sequence-to-sequence (seq2seq) forecasting on a 4-year period while using only few, widely available and easy to measure meteorological input parameters, which makes our approach widely applicable. Further, we also investigate the data dependency in terms of time series length of the different ANN architectures. For seq2val forecasts, NARX models on average perform best; however, CNNs are much faster and only slightly worse in terms of accuracy. For seq2seq forecasts, mostly NARX outperform both DL models and even almost reach the speed of CNNs. However, NARX are the least robust against initialization effects, which nevertheless can be handled easily using ensemble forecasting. We showed that shallow neural networks, such as NARX, should not be neglected in comparison to DL techniques especially when only small amounts of training data are available, where they can clearly outperform LSTMs and CNNs; however, LSTMs and CNNs might perform substantially better with a larger dataset, where DL really can demonstrate its strengths, which is rarely available in the groundwater domain though.
The polymorphic major histocompatibility complex class I chain-related molecule A (MICA) and its soluble form (sMICA) interact with activating receptor natural-killer group 2 member D (NKG2D) on ...natural-killer (NK) and T cells, thereby modifying immune responses to transplantation and infectious agents (e.g., cytomegalovirus). Two single-nucleotide polymorphisms (SNPs), rs2596538GA in the
promoter and rs1051792AG in the coding region (
-129Val/Met), influence MICA expression or binding to NKG2D, with MICA-129Met molecules showing higher receptor affinity. To investigate the impact of these SNPs on the occurrence of cytomegalovirus infection or acute rejection (AR) in individuals who underwent simultaneous pancreas⁻kidney transplantation (SPKT), 50 recipient-donor pairs were genotyped, and sMICA levels were measured during the first year post-transplantation. Recipients with a Val-mismatch (recipient Met/Met and donor Val/Met or Val/Val) showed shorter cytomegalovirus infection-free and shorter kidney AR-free survival. Additionally, Val mismatch was an independent predictor of cytomegalovirus infection and kidney AR in the first year post-transplantation. Interestingly, sMICA levels were lower in rs2596538AA and MICA129Met/Met-homozygous recipients. These results provide further evidence that genetic variants of
influence sMICA levels, and that Val mismatch at position 129 increases cytomegalovirus infection and kidney AR risk during the first year post-SPKT.
Hydrograph clustering helps to identify dynamic patterns within aquifers systems, an important foundation of characterizing groundwater systems and their influences, which is necessary to effectively ...manage groundwater resources. We develope an unsupervised modeling approach to characterize and cluster hydrographs on regional scale according to their dynamics. We apply feature-based clustering to improve the exploitation of heterogeneous datasets, explore the usefulness of existing features and propose new features specifically useful to describe groundwater hydrographs. The clustering itself is based on a powerful combination of Self-Organizing Maps with a modified DS2L-Algorithm, which automatically derives the cluster number but also allows to influence the level of detail of the clustering. We further develop a framework that combines these methods with ensemble modeling, internal cluster validation indices, resampling and consensus voting to finally obtain a robust clustering result and remove arbitrariness from the feature selection process. Further we propose a measure to sort hydrographs within clusters, useful for both interpretability and visualization. We test the framework with weekly data from the Upper Rhine Graben System, using more than 1800 hydrographs from a period of 30 years (1986-2016). The results show that our approach is adaptively capable of identifying homogeneous groups of hydrograph dynamics. The resulting clusters show both spatially known and unknown patterns, some of which correspond clearly to external controlling factors, such as intensive groundwater management in the northern part of the test area. This framework is easily transferable to other regions and, by adapting the describing features, also to other time series-clustering applications.
It is now well established to use shallow artificial neural networks (ANNs) to obtain accurate and reliable groundwater level forecasts, which are an important tool for sustainable groundwater ...management. However, we observe an increasing shift from conventional shallow ANNs to state-of-the-art deep-learning (DL) techniques, but a direct comparison of the performance is often lacking. Although they have already clearly proven their suitability, shallow recurrent networks frequently seem to be excluded from the study design due to the euphoria about new DL techniques and its successes in various disciplines. Therefore, we aim to provide an overview on the predictive ability in terms of groundwater levels of shallow conventional recurrent ANNs, namely non-linear autoregressive networks with exogenous input (NARX) and popular state-of-the-art DL techniques such as long short-term memory (LSTM) and convolutional neural networks (CNNs). We compare the performance on both sequence-to-value (seq2val) and sequence-to-sequence (seq2seq) forecasting on a 4-year period while using only few, widely available and easy to measure meteorological input parameters, which makes our approach widely applicable. Further, we also investigate the data dependency in terms of time series length of the different ANN architectures. For seq2val forecasts, NARX models on average perform best; however, CNNs are much faster and only slightly worse in terms of accuracy. For seq2seq forecasts, mostly NARX outperform both DL models and even almost reach the speed of CNNs. However, NARX are the least robust against initialization effects, which nevertheless can be handled easily using ensemble forecasting. We showed that shallow neural networks, such as NARX, should not be neglected in comparison to DL techniques especially when only small amounts of training data are available, where they can clearly outperform LSTMs and CNNs; however, LSTMs and CNNs might perform substantially better with a larger dataset, where DL really can demonstrate its strengths, which is rarely available in the groundwater domain though.
Seasons are known to have a major influence on groundwater recharge and therefore groundwater levels; however, underlying relationships are complex and partly unknown. The goal of this study is to ...investigate the influence of the seasons on groundwater levels (GWLs), especially during low-water periods. For this purpose, we train artificial neural networks on data from 24 locations spread throughout Germany. We exclusively focus on precipitation and temperature as input data and apply layer-wise relevance propagation to understand the relationships learned by the models to simulate GWLs. We find that the learned relationships are plausible and thus consistent with our understanding of the major physical processes. Our results show that for the investigated locations, the models learn that summer is the key season for periods of low GWLs in fall, with a connection to the preceding winter usually only being subordinate. Specifically, dry summers exhibit a strong influence on low-water periods and generate a water deficit that (preceding) wet winters cannot compensate for. Temperature is thus an important proxy for evapotranspiration in summer and is generally identified as more important than precipitation, albeit only on average. Single precipitation events show by far the largest influences on GWLs, and summer precipitation seems to mainly control the severeness of low-GWL periods in fall, while higher summer temperatures do not systematically cause more severe low-water periods.