An Indirect Approach Based on Long Short-Term Memory Networks to Estimate Groundwater Table Depth Anomalies Across Europe With an Application for Drought Analysis

The lack of high-quality continental-scale groundwater table depth observations necessitates developing an indirect method to produce reliable estimation for water table depth anomalies (wtda) over Europe to facilitate European groundwater management under drought conditions. Long Short-Term Memory (LSTM) networks are a deep learning technology to exploit long-short-term dependencies in the input-output relationship, which have been observed in the response of groundwater dynamics to atmospheric and land surface processes. Here, we introduced different input variables including precipitation anomalies (pra), which is the most common proxy of wtda, for the networks to arrive at improved wtda estimates at individual pixels over Europe in various experiments. All input and target data involved in this study were obtained from the simulated TSMP-G2A data set. We performed wavelet coherence analysis to gain a comprehensive understanding of the contributions of different input variable combinations to wtda estimates. Based on the different experiments, we derived an indirect method utilizing LSTM networks with pra and soil moisture anomaly (θa) as input, which achieved the optimal network performance. The regional medians of test R2 scores and RMSEs obtained by the method in the areas with wtd ≤ 3.0 m were 76–95% and 0.17–0.30, respectively, constituting a 20–66% increase in median R2 and a 0.19–0.30 decrease in median RMSEs compared to the LSTM networks only with pra as input. Our results show that introducing θa significantly improved the performance of the trained networks to predict wtda, indicating the substantial contribution of θa to explain groundwater anomalies. Also, the European wtda map reproduced by the method had good agreement with that derived from the TSMP-G2A data set with respect to drought severity, successfully detecting ~41% of strong drought events (wtda ≥ 1.5) and ~29% of extreme drought events (wtda ≥ 2) in August 2015. The study emphasizes the importance to combine soil moisture information with precipitation information in quantifying or predicting groundwater anomalies. In the future, the indirect method derived in this study can be transferred to real-time monitoring of groundwater drought at the continental scale using remotely sensed soil moisture and precipitation observations or respective information from weather prediction models.

Knowledge transfer from simulation to reality via Long Short-Term Memory networks: Estimating groundwater table depth anomalies over Europe

<p>Near real-time groundwater table depth measurements are scarce over Europe, leading to challenges in monitoring groundwater resources at the continental scale. In this study, we leveraged knowledge learned from simulation results by Long Short-Term Memory (LSTM) networks to estimate monthly groundwater table depth anomaly (<em>wtd<sub>a</sub></em>) data over Europe. The LSTM networks were trained, validated, and tested at individual pixels on anomaly data derived from daily integrated hydrologic simulation results over Europe from 1996 to 2016, with a spatial resolution of 0.11° (Furusho-Percot et al., 2019), to predict monthly <em>wtd<sub>a</sub></em> based on monthly precipitation anomalies (<em>pr<sub>a</sub></em>) and soil moisture anomalies (<em>θ<sub>a</sub></em>). Without additional training, we directly fed the networks with averaged monthly <em>pr<sub>a</sub></em> and <em>θ<sub>a</sub></em> data from 1996 to 2016 obtained from commonly available observational datasets and reanalysis products, and compared the network outputs with available borehole <em>in situ</em> measured <em>wtd<sub>a</sub></em>. The LSTM network estimates show good agreement with the <em>in situ</em> observations, resulting in Pearson correlation coefficients of regional averaged <em>wtd<sub>a</sub></em> data in seven PRUDENCE regions ranging from 42% to 76%, which are ~ 10% higher than the original simulation results except for the Iberian Peninsula. Our study demonstrates the potential of LSTM networks to transfer knowledge from simulation to reality for the estimation of <em>wtd<sub>a</sub></em> over Europe. The proposed method can be used to provide spatiotemporally continuous information at large spatial scales in case of sparse ground-based observations, which is common for groundwater table depth measurements. Moreover, the results highlight the advantage of combining physically-based models with machine learning techniques in data processing.</p><p> </p><p>Reference:</p><p>Furusho-Percot, C., Goergen, K., Hartick, C., Kulkarni, K., Keune, J. and Kollet, S. (2019). Pan-European groundwater to atmosphere terrestrial systems climatology from a physically consistent simulation. Scientific Data, 6(1).</p>

Artificial Neural Network Optimized with a Genetic Algorithm for Seasonal Groundwater Table Depth Prediction in Uttar Pradesh, India

Accurate information about groundwater level prediction is crucial for effective planning and management of groundwater resources. In the present study, the Artificial Neural Network (ANN), optimized with a Genetic Algorithm (GA-ANN), was employed for seasonal groundwater table depth (GWTD) prediction in the area between the Ganga and Hindon rivers located in Uttar Pradesh State, India. A total of 18 models for both seasons (nine for the pre-monsoon and nine for the post-monsoon) have been formulated by using groundwater recharge (GWR), groundwater discharge (GWD), and previous groundwater level data from a 21-year period (1994–2014). The hybrid GA-ANN models’ predictive ability was evaluated against the traditional GA models based on statistical indicators and visual inspection. The results appraisal indicates that the hybrid GA-ANN models outperformed the GA models for predicting the seasonal GWTD in the study region. Overall, the hybrid GA-ANN-8 model with an 8-9-1 structure (i.e., 8: inputs, 9: neurons in the hidden layer, and 1: output) was nominated optimal for predicting the GWTD during pre- and post-monsoon seasons. Additionally, it was noted that the maximum number of input variables in the hybrid GA-ANN approach improved the prediction accuracy. In conclusion, the proposed hybrid GA-ANN model’s findings could be readily transferable or implemented in other parts of the world, specifically those with similar geology and hydrogeology conditions for sustainable planning and groundwater resources management.

Evaluation of Satellite-Based Optical and Thermal Trapezoid Methods for Groundwater Table Depth Monitoring in Estonian Bogs

<p>Groundwater table depth and peat moisture content are of crucial importance for many peatland processes, like for example their greenhouse gas budget. Thus, there is a strong need for remote sensing techniques that allow to spatially monitor these critical moisture conditions to quantify the hydrological responses to climate change and other anthropogenic disturbances. Previous studies have demonstrated the usefulness but also limitations of microwave observations for peatland moisture monitoring at the large scale. Here, we explore the potential of techniques based on optical and thermal imagery for smaller scale applications.</p><p>Satellite-derived land surface temperature (LST) as well as shortwave infrared transformed reflectance (STR) are sensitive to soil moisture conditions in mineral soils. Both data form, together with remotely sensed vegetation indices (VIs), trapezoids in the LST-VI and STR-VI space with the highest range of possible LST and STR for bare soil conditions. The lowest and highest LST and STR along the vegetation cover gradient define the wet and dry edge, respectively. In this study, we used Landsat 7 and Landsat 8 satellite data for the vegetation periods from 2008 through 2019 to calculate various VIs, LST and STR for hemiboreal raised bogs in Estonia. Two common approaches for the determination of wet and dry edges for the LST-based method were applied and compared. The first approach estimates the edges directly from the observed values of VIs and LST for each scene; while the second one relies on modelled theoretical edges for each scene. In contrast, the STR-VI trapezoid is derived from observed values from all scenes as proposed in literature. The trapezoids are used to calculate the dryness index of each Landsat pixel by linearly scaling between the wet and dry edge. These indices are evaluated with measured groundwater table depth time series. Preliminary results indicate that, for our study area, suitable LST-based trapezoids cannot be derived from satellite observations alone, indicated by the low dependency of the resulting dryness index on groundwater table depth. Evaluation of the theoretically-derived trapezoids and the STR-VI is ongoing and will be discussed.</p>

10 yrs of Improved Groundwater Table Estimates in Northern Peatlands Through Assimilation of Passive Microwave Observations into PEATCLSM

<p>Groundwater table depth and peat moisture, exert a first order control on a range of biogeochemical and -physical peatland processes, and the susceptibility to peat fires. Therefore, one of the first critical measures to identify “peatlands under pressure” is the change of hydrological conditions, e.g. due to changing climatic conditions or direct “hydraulic” human influence. In this presentation, we introduce a new opportunity for the global-scale monitoring of moisture conditions in peatlands. We assimilate L-band brightness temperature (Tb) data from the Soil Moisture Ocean Salinity (SMOS) into the Catchment land surface model (CLSM) to improve the simulation of Northern peatland hydrology from 2010 through 2019. We compare four simulation experiments: two open loop and two data assimilation simulations, either using the default CLSM or a recently-developed peatland-specific adaptation of it (PEATCLSM, Bechtold et al. 2019). The assimilation system uses a spatially distributed ensemble Kalman filter to update soil moisture and groundwater table depth. The simulation experiments are evaluated against an in-situ dataset of groundwater table depth in about 20 natural and semi-natural peatlands that are large enough to be dominant in the corresponding 81-km<sup>2</sup> model grid cells. For PEATCLSM, Tb data assimilation increases the temporal Pearson correlation (R) and anomaly correlation (aR) between simulated and measured groundwater table from 0.53 and 0.38 (open-loop) to 0.58 and 0.45 (analysis), respectively. Time series comparison at monitoring sites demonstrates how the assimilation effectively corrects for remaining deficiencies in model physics and/or errors of the global meteorological data forcing the model. The generally lower coefficients of 0.30 (R) and 0.09 (aR) for the default CLSM also improve after Tb data assimilation to values of 0.39 (R) and 0.28 (aR). However, even with Tb data assimilation, the skill of CLSM remains inferior to that of PEATCLSM. The more realistic model physics of PEATCLSM are also supported by a reduction of the Tb misfits (observed Tb – forecasted Tb) over 94 % of the Northern peatland area. The temporal variance of Tb misfits is reduced by 20 % on average and is largest over the large peatland areas of the Western Siberian (25 %) and Hudson Bay Lowlands (40 %). This study demonstrates, for the first time, an improved estimation of the peatland hydrological dynamics by the assimilation of SMOS L-band brightness data into a global land surface model and suggests a new route of research focusing on the incorporation of additional satellite observations into peatland-specific modeling schemes.</p><p>Bechtold, M., De Lannoy, G.J M., Koster, R.D., Reichle, R.H., et al. (2019). PEAT-CLSM: A Specific Treatment of Peatland Hydrology in the NASA Catchment Land Surface Model. JOURNAL OF ADVANCES IN MODELING EARTH SYSTEMS, 11 (7), 2130-2162. doi: 10.1029/2018MS001574.</p>