Groundwater Level Time Series Research Articles

Spatial-temporal graph neural networks for groundwater data.

This paper introduces a novel application of spatial-temporal graph neural networks (ST-GNNs) to predict groundwater levels. Groundwater level prediction is inherently complex, influenced by various hydrological, meteorological, and anthropogenic factors. Traditional prediction models often struggle with the nonlinearity and non-stationary characteristics of groundwater data. Our study leverages the capabilities of ST-GNNs to address these challenges in the Overbetuwe area, Netherlands. We utilize a comprehensive dataset encompassing 395 groundwater level time series and auxiliary data such as precipitation, evaporation, river stages, and pumping well data. The graph-based framework of our ST-GNN model facilitates the integration of spatial interconnectivity and temporal dynamics, capturing the complex interactions within the groundwater system. Our modified Multivariate Time Graph Neural Network model shows significant improvements over traditional methods, particularly in handling missing data and forecasting future groundwater levels with minimal bias. The model's performance is rigorously evaluated when trained and applied with both synthetic and measured data, demonstrating superior accuracy and robustness in comparison to traditional numerical models in long-term forecasting. The study's findings highlight the potential of ST-GNNs in environmental modeling, offering a significant step forward in predictive modeling of groundwater levels.

Assessing groundwater level modelling using a 1-D convolutional neural network (CNN): linking model performances to geospatial and time series features

Abstract. Groundwater level (GWL) forecasting with machine learning has been widely studied due to its generally accurate results and low input data requirements. Furthermore, machine learning models for this purpose can be set up and trained quickly compared to the effort required for process-based numerical models. Despite demonstrating high performance at specific locations, applying the same model architecture to multiple sites across a regional area can lead to varying accuracies. The reasons behind this discrepancy in model performance have been scarcely examined in previous studies. Here, we explore the relationship between model performance and the geospatial and time series features of the sites. Using precipitation (P) and temperature (T) as predictors, we model monthly groundwater levels at approximately 500 observation wells in Lower Saxony, Germany, applying a 1-D convolutional neural network (CNN) with a fixed architecture and hyperparameters tuned for each time series individually. The GWL observations range from 21 to 71 years, resulting in variable test and training dataset time ranges. The performances are evaluated against selected geospatial characteristics (e.g. land cover, distance to waterworks, and leaf area index) and time series features (e.g. autocorrelation, flat spots, and number of peaks) using Pearson correlation coefficients. Results indicate that model performance is negatively influenced at sites near waterworks and densely vegetated areas. Longer subsequences of GWL measurements above or below the mean negatively impact the model accuracy. Besides, GWL time series containing more irregular patterns and with a higher number of peaks might lead to higher model performances, possibly due to a closer link with precipitation dynamics. As deep learning models are known to be black-box models missing the understanding of physical processes, our work provides new insights into how geospatial and time series features link to the input–output relationship of a GWL forecasting model.

Filling the ponds of Hattuşa - a geohydrological approach to determine inflows by time and volume

In the ruins of Hattuşa, the capital of the Hittite Empire more than 3,000 years ago, large, sedimented water storage ponds have been uncovered since 1998. An investigation based on time series of measured groundwater levels, showed that the ponds were supplied from layered aquifers which were opened at their source horizons directly at the uphill edge of the pond. When the groundwater level at the pond exceeds a threshold given by the geology at that edge, water pours into the pond. The principles and the path of filling become clear, but not the volumes and time distribution of the transfers. These results of the previous studies were amply published and are shortly summarized. The present investigation is also based on the time series of groundwater levels and of the discharge of a modern fountain as well as the local hydrogeological properties. The approach is simple but shows the effect of storage in the aquifer which delays the outflow into the pond, so that water from the rainy season was available in the dry season. Also, the annual outflow values are determined. Besides gaining knowledge it is an aim of the article to encourage interdisciplinary and process-oriented thinking.

Bee-inspired insights: Unleashing the potential of artificial bee colony optimized hybrid neural networks for enhanced groundwater level time series prediction.

Analysis of the change in groundwater used as a drinking and irrigation water source is of critical importance in terms of monitoring aquifers, planning water resources, energy production, combating climate change, and agricultural production. Therefore, it is necessary to model groundwater level (GWL) fluctuations to monitor and predict groundwater storage. Artificial intelligence-based models in water resource management have become prevalent due to their proven success in hydrological studies. This study proposed a hybrid model that combines the artificial neural network (ANN) and the artificial bee colony optimization (ABC) algorithm, along with the ensemble empirical mode decomposition (EEMD) and the local mean decomposition (LMD) techniques, to model groundwater levels in Erzurum province, Türkiye. GWL estimation results were evaluated with mean square error (MSE), coefficient of determination (R2), and residual sum of squares (RSS) and visually with violin, scatter, and time series plot. The study results indicated that the EEMD-ABC-ANN hybrid model was superior to other models in estimating GWL, with R2 values ranging from 0.91 to 0.99 and MSE values ranging from 0.004 to 0.07. It has also been revealed that promising GWL predictions can be made with previous GWL data.

Deformation Monitoring Based on SBAS-InSAR and Leveling Measurement: A Case Study of the Jing-Mi Diversion Canal in China.

The Jing-Mi Diversion Canal is a large-scale water diversion project in Beijing. Routine monitoring is crucial for the reliability and stability of urban water supply. Compared with traditional monitoring methods, interferometric synthetic aperture radar (InSAR) has the advantages of large scale and high accuracy. Based on the small baseline subset InSAR, 187 ascending and 102 descending SAR images obtained from Sentinel-1 were used to detect the deformation along the diversion canal from 2017 to 2023. The results show that there was a sinking trend along the diversion canal. The subsidence was serious in the first half of the canal, and continued to sink from 2019 to 2020. The subsidence was alleviated in 2023. Combined with leveling measurements, the InSAR deformation monitoring results of important pumping station buildings were verified. The measurement accuracy of InSAR can reach the millimeter level. We extracted the groundwater level time series and subsidence for risky canal segments. Through pixel-by-pixel comparison, it was found that fluctuations in groundwater level would have some impact on surface deformation. Severe local subsidence or uplift deformation occasionally occurred. To ensure the safety of water diversion, the monitoring and maintenance of relevant pump station buildings in risky areas should be increased in the future.

Groundwater Recharge Assessment in Central Benin: The Case of the Collines Region (West Africa)

The objective of this study was to assess groundwater recharge in the hard-rock central region of Benin so as to compare it with the water needs of the local population. To reach this objective, we applied the Water Table Fluctuation (WTF) method, which requires long-term monitoring of groundwater level fluctuations. Groundwater level time series were used in combination with other data (including time series of surface water discharge and rainfall) to estimate groundwater recharge but also to shed further light on the relationship between surface water and groundwater. The results demonstrated that the minimum inter-annual groundwater recharge amount is about 1.09 × 109 m3, which is enough to cover the basic water needs of the local population. It should be highlighted that in sub-regions where the density of the population is high, water shortage can still occur with the above estimated groundwater recharge amount. This study has also illustrated that when applying the WTF method, sites with a highly uncertain specific yield can be detected.

Trend detection and depletion effects evidence in time series of groundwater levels in the southern sector of the left bank of the Tagus-Sado Basin (Portugal, Iberian Peninsula)

In the last 20 years in Portugal, water resources have been affected to the point that water storage has decreased by 20% since 2000. Creating strategies to manage water resources requires a comprehensive understanding of the factors influencing water storage and their effects over time. This study is focused on the evolution of Groundwater Deep Levels (GDL) by applying a two-phase trend analysis methodology to examine the dynamic changes in GDL within a series of monitoring wells located in the Central and Southern sectors of the Left Bank of the Tagus-Sado Cenozoic age Basin, situated in Portugal In the initial phase of trend analysis, Factorial Analysis of Mixed Data (FAMD) was employed and posteriorly the Hierarchical Classification Analysis (HCA). These techniques enabled us to identify distinct GDL trend profiles and generate interpretative maps illustrating their spatial distribution. In the second phase, the non-parametric Mann–Kendall Analysis (MKA) and Innovative Trend Analysis (ITA) were applied, allowing for a quantified confirmation of the different trend profiles previously detected. These techniques allowed the identification of positive and negative hydrodynamic trends in distinct sections of the Basin. In the SE sector they are characterized by a significative increase of GDL associated with overexploitation and in the Central sector with a decrease of GDL. Nevertheless, significant depletion effects can result from natural factors such as prolonged droughts, and in certain regions, changes in geological and hydrothermal dynamics, such as Alpine-age faults, graben, and horst structures, may account for these alterations.

A maximal overlap discrete wavelet packet transform coupled with an LSTM deep learning model for improving multilevel groundwater level forecasts

Developing precise groundwater level (GWL) forecast models is essential for the optimal usage of limited groundwater resources and sustainable planning and management of water resources. In this study, an improved forecasting accuracy for up to 3 weeks ahead of GWLs in Bangladesh was achieved by employing a coupled Long Short Term Memory (LSTM) network-based deep learning algorithm and Maximal Overlap Discrete Wavelet Packet Transform (MODWPT) data preprocessing. The coupled LSTM-MODWPT model’s performance was compared with that of the LSTM model. For both standalone LSTM and LSTM-MODWPT models, the Random Forest feature selection approach was employed to select the ideal inputs from the candidate GWL lags. In the LSTM-MODWPT model, input GWL time series were decomposed using MODWPT. The ‘Fejér-Korovkin’ mother wavelet with a filter length of 18 was used to obtain a collection of scaling coefficients and wavelets for every single input time series. Model performance was assessed using five performance indices: Root Mean Squared Error; Scatter Index; Maximum Absolute Error; Median Absolute Deviation; and an a-20 index. The LSTM-MODWPT model outperformed standalone LSTM models for all time horizons in GWL forecasting. The percentage improvements in the forecasting accuracies were 36.28%, 32.97%, and 30.77%, respectively, for 1-, 2-, and 3-weeks ahead forecasts at the observation well GT3330001. Accordingly, the coupled LSTM-MODWPT model could potentially be used to enhance multiscale GWL forecasts. This research demonstrates that the coupled LSTM-MODWPT model could generate more precise GWL forecasts at the Bangladesh study site, with potential applications in other geographic locations globally.

Investigating climatic and human effects on the decline in the groundwater level (a case study of the Kuhdasht plain)

Abstract This study has been investigated using the new criterion measuring complexity Hilbert-Huang entropy (HHE), changes, and fluctuations in groundwater levels under the influence of climate change factors and humans. In this regard, the 120-month time series of groundwater level, precipitation, temperature, and runoff parameters (period 2009–2018) are first divided into five 24-month periods, then by applying the Hilbert-Huang transform. Eachperiod is decomposed down into smaller periods, and finally, the HHE criterion measures the complexity of the time series. The examination of changes in the HHE complexity criterion shows a decrease of 2.02, 2.93, and 29.58% in the groundwater level in the second, third, and fourth periods. Also, according to the results of changes in the complexity of precipitation and temperature time series as climate change factors, these two parameters have decreased in the third and fourth periods, with the rate of decrease for the temperature time series being 18.8 and 8.3%, respectively, and for the precipitation time series is obtained by 30.6 and 4.4%, respectively. Hence, according to the results, the discharge parameter, as a human factor in the second and fourth periods, shows a decrease of 8.6 and 39.1%, respectively.

Synthetic Time Series Data in Groundwater Analytics: Challenges, Insights, and Applications

This study presents ‘Synthetic Wells’, a method for generating synthetic groundwater level time series data using machine learning (ML) aimed at improving groundwater management in contexts where real data are scarce. Utilizing data from the National Water Information System of the US Geological Survey, this research employs the Synthetic Data Vault (SDV) framework’s Probabilistic AutoRegressive (PAR) synthesizer model to simulate real-world groundwater fluctuations. The synthetic data generated for approximately 100 wells align closely with the real data, achieving a quality score of 70.94%, indicating a reasonable replication of groundwater dynamics. A Streamlit-based web application was also developed, enabling users to generate custom synthetic datasets. A case study in Mississippi, USA, demonstrated the utility of synthetic data in enhancing the accuracy of time series forecasting models. This unique approach represents an innovative first-of-its-kind tool in the realm of groundwater research, providing new avenues for data-driven decision-making and management in hydrological studies.

Groundwater level response to precipitation at the hydrological observatory of Pinios (central Greece)

To study groundwater dynamics and their response to precipitation in the Pinios river basin (central Greece), the variation trends and seasonal patterns of fourteen monitored groundwater and two precipitation time series were examined using time series decomposition, autocorrelation, and wavelet analysis. The objective is to perform indirect methodologies based on groundwater level and precipitation data to attain a detailed understanding of the hydrogeological mechanisms of the aquifer system and develop an approach that can be easily employed worldwide. The three approaches confirm local groundwater circulation understanding by dividing the aquifer system into three parts characterised by varying transmissivity and storativity properties. The extended memory effect, ranging between 70 and 230 days suggests great inertia and storage capacity of the examined groundwater system. The wavelet analysis proves that the system acts as a filter which absorbs the precipitation frequencies converting them into GWL signals, as demonstrated by the 256 days periodicity shared by groundwater level and precipitation time series. Generally, GWLs present a hysteresis to precipitation due to the geological setup and local flow conditions and are particularly affected by a seasonal component, as also highlighted by time series decomposition. An evident annual periodicity is also observed in hydrological and hydrogeological time series. Results suggest the effectiveness of the proposed combination of methods in developing hydrological understanding and their general applicability to other regions where at least groundwater level and precipitation time series are available. By utilising such data-driven approaches, hydrogeologists and water resource managers can gain valuable insights into the behavior of aquifers, groundwater flow patterns, and the impact of human activities on groundwater quality and quantity.

Groundwater level reconstruction using long-term climate reanalysis data and deep neural networks

Study regionNorthern Metropolitan France. Study focusAssessing long-term changes in groundwater is crucial for understanding the impacts of climate change on aquifers and for managing water resources.However, long-term groundwater level (GWL) records are often scarce, limiting the understanding of historical trends and variability. In this paper, we present a deep learning approach to reconstruct GWLs up to several decades back in time using recurrent-based neural networks with wavelet pre-processing and climate reanalysis data as inputs. GWLs are reconstructed using two different reanalysis datasets with distinct spatial resolutions (ERA5: 0.25° x 0.25° & ERA20C: 1° x 1°) and monthly time resolution, and the performance of the simulations were evaluated. New insightsLong term GWL timeseries are now available for northern France, corresponding to extended versions of observational timeseries back to early 20th century. All three types of piezometric behaviours could be reconstructed reliably and consistently capture the multi-decadal variability even at coarser resolutions, which is crucial for understanding long-term hydroclimatic trends and cycles. GWLs'multidecadal variability was consistent with the Atlantic multidecadal oscillation. From a synthetic experiment involving a modified long-term observational time series, we highlighted the need for longer training datasets for some low-frequency signals. Nevertheless, our study demonstrated the potential of using DL models together with reanalysis data to extend GWL observations and improve our understanding of groundwater variability and climate interactions.

Coupled processes of groundwater dynamics and land subsidence in the context of active human intervention, a case in Tianjin, China

To address the crisis of water shortage in the North China Plain, the Chinese government implemented the South-to-North Water Transfer Project (SNWTP). In this context, Tianjin, one of the main beneficiaries of this project, has been relieved from water shortages and begun to implement Groundwater Management Plans (GMP) since 2018, which undoubtedly have a significant effect on the groundwater recovery. Meanwhile, this provides a good case for studying the coupled process of ground settlement and groundwater dynamics, especially the soil deformation pattern driven by groundwater level (GWL) rebound. To analyze these issues in detail, field well data was collected to depict groundwater flow field. Moreover, geodetic data was also collated, including leveling, GPS, and InSAR, so that a vertical deformation field with high spatiotemporal resolution could be generated. The results reveal that the GWL of the third confined aquifer which is the main exploitation layer in Tianjin recovered significantly since 2018 with a rate of 2.1 m/yr. The dynamic deformation patterns indicate that the area of land subsidence cones in Tianjin has reduced significantly, accompanied by a sharply declining subsidence rate (decreased from −32.2 mm/yr to −4.5 mm/yr.). Particularly, a significant poroelastic rebound has occurred in the Wuqing and Beichen districts since 2020. Furthermore, due to the delayed pore pressure dissipation in the aquitard, we find a time delay of 0.3–5.5 years between land subsidence and GWL time series, which is far less than that estimated by hydrogeological parameters, as the latter ignored the recharge and recovery capacity of the aquifer system. Finally, an evolution models in Tianjin was presented to illustrate interactive process among the deformation, pore pressure, and hydraulic head. In general, the SNWDP and the GMP has restored the pore pressure of aquifer, reduced the land subsidence, and alleviated the groundwater storage depletion of Tianjin.

Metamodelling of Naturalised Groundwater Levels at a Regional Level in New Zealand

Groundwater is under pressure from increasing demands for agriculture, industry, domestic uses and support of ecosystems. Understanding the natural state of a groundwater system helps policy makers manage groundwater sustainably. Here we developed a metamodelling approach based on stepwise linear regression that emulates the functionality of physically-based models in the three primary aquifers of the Greater Wellington region of New Zealand. The inputs for the metamodels included local weather data, and nearby river flow data. The metamodels were calibrated and validated against the available simulations of naturalised groundwater level time series from physically-based models for 47 selected wells. For 36 of these wells, the metamodels had Nash-Sutcliffe Efficiency and coefficient of determination over 0.5, showing that they could adequately mimic naturalised groundwater level dynamics as simulated by the physically-based groundwater models. The remaining 11 wells had unsatisfactory performance and were typically located far away from rivers or along the coast. The results also showed that modelled groundwater levels in the aquifer’s recharge zone were more sensitive to short-term (less than 2 weeks lag) than long-term river flow (above 4 weeks to 1 year lag), whereas the converse pattern was observed for the aquifer’s discharge zone. Although some special considerations are needed, this metamodelling framework can be generally applied to other aquifers to support groundwater resource management at a lower cost than updating physically-based models.

Excessive groundwater withdrawal and resultant land subsidence in the Küçük Menderes River Basin, Turkey as estimated from InSAR-SBAS and GNSS measurements

Küçük Menderes River Basin (KMRB) is located in the western part of Turkey in the city of İzmir. The basin hosts about half a million people and 90% of the population makes their living from farming and animal husbandry. Most of the water needs for agricultural irrigation, animal husbandry, and drinking are met from groundwater resources in the basin. Since the water pumped from the aquifers cannot be naturally recharged by surface waters and precipitation, the groundwater level is rapidly receding and a decrease of up to 150 meters has been observed in the historical groundwater level time series since the 1990s. For this reason, land subsidence occurred in the basin and as a result of this, crustal surface cracks have formed and infrastructure and superstructure facilities have been endangered. To prevent loss of life and property that may occur in the future, it has become necessary to estimate the magnitude and the expansion zone of the land subsidence by carrying out geodetic monitoring studies in the region.In this study, the analyzes performed by Interferometric Synthetic Aperture Radar-Small Baseline Subset (InSAR-SBAS) and Global Navigation Satellite System (GNSS) methods to monitor land subsidence in the KMRB are presented. For this purpose, monthly 74 SAR images from 2017 to the end of 2022 were evaluated according to the SBAS procedure in the GMTSAR academic analysis software. In order to, both validate the InSAR-SBAS results and convert the deformation rate in the Line of Sight (LOS) to the ellipsoid normal direction, 3 periods of GNSS campaign were performed between 2021 and 2023 in 5 different GNSS stations and evaluated with the GAMIT/GLOBK software. According to the analyses, the land subsidence rate of up to −165 mm/year has been estimated in the KMRB, especially in the district centers of Bayındır and Ödemiş. In addition, unlike many studies, different rates and propagation of subsidence were determined in different parts of the basin depending on the aquifer type and precipitation. Monthly time series of the deformations were created and the behavior of the deformations and its relationship with the groundwater withdrawal and the lithological structure in the region was scrutinized. Thus, the spatio-temporal behaviour of the land subsidence and the associated geological factors were revealed.

Data‐Driven Estimation of Groundwater Level Time‐Series at Unmonitored Sites Using Comparative Regional Analysis

AbstractA new method is presented to efficiently estimate daily groundwater level time series at unmonitored sites by linking groundwater dynamics to local hydrogeological system controls. The proposed approach is based on the concept of comparative regional analysis, an approach widely used in surface water hydrology, but uncommon in hydrogeology. Using physiographic and climatic site descriptors, the method utilizes regression analysis to estimate cumulative frequency distributions of groundwater levels (groundwater head duration curves, HDC) at unmonitored locations. The HDC is then used to construct a groundwater hydrograph using time series from distance‐weighted neighboring monitored (donor) locations. For estimating times series at unmonitored sites, in essence, spatio‐temporal interpolation, stepwise multiple linear regression (MLR), extreme gradient boosting (XGB), and nearest neighbors are compared. The methods were applied to 10‐year daily groundwater level time series at 157 sites in unconfined alluvial aquifers in Southern Germany. Models of HDCs were physically plausible and showed that physiographic and climatic controls on groundwater level fluctuations are nonlinear and dynamic, varying in significance from “wet” to “dry” aquifer conditions. XGB yielded a significantly higher predictive skill than nearest neighbor and MLR. However, donor site selection is of key importance. The study presents a novel approach for regionalization and infilling of groundwater level time series that also aids conceptual understanding of controls on groundwater dynamics, both central tasks for water resources managers.

Value and limitations of machine learning in high-frequency nutrient data for gap-filling, forecasting, and transport process interpretation

High-frequency monitoring of water quality in catchments brings along the challenge of post-processing large amounts of data. Moreover, monitoring stations are often remote and technical issues resulting in data gaps are common. Machine learning algorithms can be applied to fill these gaps, and to a certain extent, for predictions and interpretation. The objectives of this study were (1) to evaluate six different machine learning models for gap-filling in a high-frequency nitrate and total phosphorus concentration time series, (2) to showcase the potential added value (and limitations) of machine learning to interpret underlying processes, and (3) to study the limits of machine learning algorithms for predictions outside the training period. We used a 4-year high-frequency dataset from a ditch draining one intensive dairy farm in the east of The Netherlands. Continuous time series of precipitation, evapotranspiration, groundwater levels, discharge, turbidity, and nitrate or total phosphorus were used as predictors for total phosphorus and nitrate concentrations respectively. Our results showed that the random forest algorithm had the best performance to fill in data-gaps, with R2 higher than 0.92 and short computation times. The feature importance helped understanding the changes in transport processes linked to water conservation measures and rain variability. Applying the machine learning model outside the training period resulted in a low performance, largely due to system changes (manure surplus and water conservation) which were not included as predictors. This study offers a valuable and novel example of how to use and interpret machine learning models for post-processing high-frequency water quality data.

Exploring the potential for groundwater-related ground deformation in Southern New South Wales, Australia

Unsustainable groundwater extraction can lead to aquifer compaction, damages to infrastructure, changes of water accumulation in rivers and lakes and to a decrease of the aquifer's ability to store water for future generations. While this phenomenon is well identified across the globe, the potential for groundwater-related ground deformation is still largely unknown for most of the heavily exploited aquifers of Australia. This study fills that science gap by exploring signs of this phenomenon across a large region comprising seven of Australia's most intensively exploited aquifers, in the New South Wales Riverina region. To detect ground deformation, we processed 396 Sentinel-1 swaths acquired during 2015–2020 using a multitemporal spaceborne radar interferometry (InSAR), leading to the production of a near-continuous ground deformation maps covering ~280,000 km2. To explore potential groundwater-induced deformation hotspots, four criteria are used in a multiple-line of evidence approach: (1) the amplitude, shape, and extent of the InSAR ground displacement anomaly, (2) the spatial correspondence with groundwater extraction hotspots. (3) The correlations between InSAR deformation time series and change in head levels in 975 wells. Four areas are identified as potentially prone to inelastic, groundwater-related deformations, with average deformation rates ranging from −10 to −30 mm/yr, high rates of groundwater extraction, and ample critical head drops. Comparison of ground deformation and groundwater level time series also suggests potential for elastic deformation in some of these aquifers. This study will help water managers mitigating the groundwater-related ground deformation risk.

Understanding the Spatiotemporal Characteristics of Land Subsidence and Rebound in the Lianjiang Plain Using Time-Series InSAR with Dual-Track Sentinel-1 Data

The Lianjiang Plain, renowned for its position as ‘China’s textile hub’ and characterized by its high population density, has experienced considerable subsidence due to excessive groundwater extraction in recent years. Although some studies have investigated short-term subsidence in this plain, research on long-term subsidence and rebound remain understudied. In this paper, the characteristics of surface deformation in the Lijiang Plain during two periods (2015–2017 and 2018–2021) have been investigated using the time-series interferometric synthetic aperture radar (TS-InSAR) technique, and the correlation with the changes in groundwater level, geological factors, and urban construction are discussed. The InSAR-derived results are cross-validated with the adjacent orbit datasets. Large-scale and uneven subsidence ranging from −124 mm/year to +40 mm/year is observed from 2015 to 2017. However, a significant decrease in the subsidence rate during 2018–2021, with local rebound deformation up to +48 mm/year in three regions, is also observed. Groundwater level changes are found to be the major cause of the ground deformation, and the intercomparison between groundwater level and ground displacement time series from TS-InSAR measurements also indicates a clear relationship between them during 2018–2021. Geological factors control the range of deformation area over the study period. The impact of urban construction on surface subsidence is evident, contributing to high deformation. Our findings could improve the understanding of how deformation is affected by groundwater rebound and offer valuable insights into groundwater management, urban planning, and land subsidence mitigation.

Groundwater flooding hazard assessment in a semi-urban aquifer through probability modelling of surrogate data

Many areas of the world are experiencing groundwater flooding (GF) events with serious impacts on the environment and human health. In Europe, GF has received specific attention with the Directive 2007/60/EC, which requires Member States to map GF hazard and risk and propose mitigation measures. In Italy, despite the commitment assigned by the Legislative Decree 49/2010 to River Basin District Authorities, no models have been developed for groundwater flooding hazard (GFH) evaluation also for the lack of standard methods based on commonly available data. In this paper, a new method has been developed through a specific procedure that accounts for i) probability of GF events estimation based on available time series of groundwater levels and ii) probability mapping through a newly developed MATLAB™ code, high-resolution topographic data and groundwater table model. The methodology has been tested and validated in the Metropolitan City of Naples (Italy), affected by GF since 2010, for which i) sixty-two-year long time series of groundwater levels, ii) high-resolution Digital Elevation and Digital Terrain Models and iii) basement/building geometry, are available. The results show that about 59% of the buildings present in the study area have potentially floodable underground structures. In the 2010–2021 period, only 2% of these buildings were flooded and most of them were built in the period 1980–2000, when the piezometric levels were deeper due to over-pumping of the groundwater. The obtained GFH map shows that only 2% of the built area is involved by 1- to 2- year events, while 6% is affected by 1- to 10- year events. A good spatial correspondence between the return period and locations of GF events occurred during 2010–2021 period was observed, thus indicating the significant performance of the model. The obtained GFH map output can support decisions of authorities and policy-makers to reduce the GF risk.