Abstract
Freshwater lenses and their freshwater–saltwater transition zones are affected by climate change. Both sea-level rise and groundwater recharge influence freshwater volume and transition zone thickness. This study used a semi-generic approach to investigate climate change effects on freshwater lenses: a hypothetical island cross-section was combined with real-world boundary conditions. Sea-level projections including tides and storm surges, annual mean sea-level rise data, and monthly recharge projections of several climate models of the German barrier island Norderney in the North Sea were used to evaluate changes in freshwater lens and transition zone size between 1971–2000 and 2071–2100. Firstly, impacts of sea-level and recharge boundary conditions were investigated on islands of different widths. Secondly, a multi-parameter study was conducted focussing on variations of several relevant hydrogeological parameters. Results showed that it is very likely but not certain that freshwater lens volume and depth will decrease and transition zone thickness increase as a consequence of climate change. Model predictions revealed a strong dependency on the employed climate models and to a lesser extent on the hydrogeological parameters, at least for the parameter ranges used in this study. Of all hydrogeological parameters tested, the largest effects were caused by the hydraulic conductivity and its anisotropy. Furthermore, the study showed that boundary conditions have larger impacts on smaller islands. These results illustrate the importance of using projections from climate models in a sufficiently high resolution. Furthermore, their uncertainties and changes in variability of boundary conditions should be considered in studies about climate change impacts on freshwater lenses.
Résumé
Les lentilles d'eau douce et leurs zones de transition entre l'eau douce et l'eau salée sont affectées par le changement climatique. L'élévation du niveau de la mer et la recharge des eaux souterraines influencent le volume d'eau douce et l'épaisseur de la zone de transition. Cette étude a utilisé une approche semi-générique pour étudier les effets du changement climatique sur les lentilles d'eau douce : Une section transversale d'île hypothétique a été combinée avec des conditions limites réelles. Les projections du niveau de la mer, y compris les marées et les surcôtes de tempête, les données sur l'élévation annuelle moyenne du niveau de la mer et les projections de la recharge mensuelle de plusieurs modèles climatiques de l'île-barrière allemande de Norderney dans la mer du Nord ont été utilisées pour évaluer les changements de la taille de la lentille d'eau douce et de la zone de transition entre 1971-2000 et 2071-2100. Tout d'abord, les impacts du niveau de la mer et des conditions limites de recharge ont été étudiés sur des îles de différentes largeurs. Deuxièmement, une étude multiparamétrique a été menée en se concentrant sur les variations de plusieurs paramètres hydrogéologiques pertinents. Les résultats ont montré qu'il est très probable, mais pas certain, que le volume et la profondeur des lentilles d'eau douce diminuent et que l'épaisseur de la zone de transition augmente en raison du changement climatique. Les prévisions des modèles ont révélé une forte dépendance à l'égard des modèles climatiques utilisés et, dans une moindre mesure, à l'égard des paramètres hydrogéologiques, du moins pour les gammes de paramètres utilisées dans cette étude. De tous les paramètres hydrogéologiques testés, les effets les plus importants ont été causés par la conductivité hydraulique et son anisotropie. En outre, l'étude a montré que les conditions limites ont un impact plus important sur les petites îles. Ces résultats illustrent l'importance d'utiliser les projections des modèles climatiques à une résolution suffisamment élevée. En outre, leurs incertitudes et les changements dans la variabilité des conditions limites devraient être pris en compte dans les études sur les impacts du changement climatique sur les lentilles d'eau douce.
Resumen
Las lentes de agua dulce y sus zonas de transición agua dulce-agua salada se ven afectadas por el cambio climático. Tanto el aumento del nivel del mar como la recarga de las aguas subterráneas influyen en el volumen de agua dulce y en el espesor de la zona de transición. En este estudio se utilizó un enfoque semigenérico para investigar los efectos del cambio climático en los lentes de agua dulce: Se utilizó una sección transversal de una isla hipotética y se combinó con las condiciones de contorno del entorno natural. Se utilizaron proyecciones del nivel del mar que incluían mareas y mareas de tormentas, datos de la elevación media anual del nivel del mar y proyecciones mensuales de recarga de varios modelos climáticos de la isla barrera alemana de Norderney, en el Mar del Norte, para evaluar los cambios en el tamaño de las lentes de agua dulce y de la zona de transición entre 1971-2000 y 2071-2100. En primer lugar, se investigaron los impactos del nivel del mar y las condiciones límite de recarga en islas de diferentes anchos. En segundo lugar, se realizó un estudio multiparamétrico centrado en las variaciones de varios parámetros hidrogeológicos relevantes. Los resultados mostraron que es muy probable, aunque no seguro, que el volumen y la profundidad de las lentes de agua dulce disminuyan y que el espesor de la zona de transición aumente como consecuencia del cambio climático. Las predicciones de los modelos revelaron una fuerte dependencia de los modelos climáticos empleados y, en menor medida, de los parámetros hidrogeológicos, al menos para los rangos de parámetros utilizados en este estudio. De todos los parámetros hidrogeológicos probados, los mayores efectos fueron causados por la conductividad hidráulica y su anisotropía. Además, el estudio demostró que las condiciones límite tienen mayores repercusiones en las islas más pequeñas. Estos resultados ilustran la importancia de utilizar proyecciones de modelos climáticos con una resolución suficientemente alta. Además, sus incertidumbres y los cambios en la variabilidad de las condiciones límite deberían tenerse en cuenta en los estudios sobre los impactos del cambio climático en las lentes de agua dulce.
摘要
淡水透镜体及其淡水-咸水过渡带受气候变化的影响。海平面上升和地下水补给共同影响淡水体积和过渡带厚度。本研究采用一种半通用的方法来调查气候变化对淡水透镜的影响:将一个假想的岛屿剖面与现实世界的边界条件相结合。为了评估1971-2000年与2071-2100年之间淡水透镜体和过渡带大小的变化,使用了包括潮汐和风暴潮的海平面预测数据、年平均海平面上升数据,以及北海德国诺德奈沙洲几个气候模型的月补给预测数据。首先,研究了海平面和补给边界条件对不同宽度岛屿的影响。其次,进行了一项多参数研究,重点研究了几种相关水文地质参数的变化。结果表明,淡水透镜体体积和深度很可能会减少,而过渡带厚度会因气候变化而增加,但这并不确定。模型预测显示,结果在很大程度上依赖于所使用的气候模型,在较小程度上依赖于水文地质参数,至少在本研究使用的参数范围内。在所有测试的水文地质参数中,渗透系数及其各向异性对结果的影响最大。此外,研究表明边界条件对较小岛屿的影响更大。这些结果表明,在研究气候变化对淡水透镜体的影响时,使用足够高分辨率的气候模型预测是重要的。此外,应考虑这些预测的不确定性和边界条件变化的可变性。
Resumo
As lentes de água doce e suas zonas de transição de água doce e água salgada são afetadas pelas mudanças climáticas. Tanto o aumento do nível do mar quanto a recarga de água subterrânea influenciam o volume de água doce e a espessura da zona de transição. Este estudo usou uma abordagem semigenérica para investigar os efeitos das mudanças climáticas nas lentes de água doce: Uma seção transversal de ilha hipotética foi combinada com condições de limite do mundo real. As projeções do nível do mar, incluindo marés e tempestades, dados de elevação média anual do nível do mar e projeções de recarga mensal de vários modelos climáticos da ilha-barreira alemã de Noderney, no Mar do Norte, foram usadas para avaliar as mudanças nas lentes de água doce e no tamanho da zona de transição entre 1971-2000 e 2071-2100. Em primeiro lugar, os impactos do nível do mar e das condições de limite de recarga foram investigados em ilhas de diferentes larguras. Em segundo lugar, foi realizado um estudo multiparâmetro com foco nas variações de vários parâmetros hidrogeológicos relevantes. Os resultados mostraram que é muito provável, mas não certo, que o volume e a profundidade das lentes de água doce diminuam e a espessura da zona de transição aumente como consequência das mudanças climáticas. As previsões do modelo revelaram uma forte dependência dos modelos climáticos empregados e, em menor grau, dos parâmetros usadas neste estudo. De todos os parâmetros hidrogeológicos testados, os maiores efeitos foram causados pela condutividade hidráulica e sua anisotropia. Além disso, o estudo mostrou que as condições de contorno têm impactos maiores em ilhas menores. Esses resultados ilustram a importância de usar projeções de modelos climáticos em uma resolução suficientemente alta. Além disso, suas incertezas e mudanças na variabilidade das condições de contorno devem ser consideradas em estudos sobre os impactos das mudanças climáticas nas lentes de água doce.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
Introduction
Climate change impacts have been observed over the last few decades. The CO2 concentration in the atmosphere, the global surface temperature as well as global mean sea level have increased (IPCC 2014, 2021). Representative concentration pathways (RCP) describe several climate change scenarios depending on greenhouse gas emissions, atmospheric concentrations, air pollutant emissions, and land use until the year 2100 (IPCC 2014). The stringent mitigation scenario RCP2.6 represents a scenario that aims to restrict global warming compared to preindustrial temperatures likely below \(2^\circ{\rm C}\) , whereas the very high greenhouse gas emission scenario RCP8.5 assumes no additional efforts to reduce emissions (IPCC 2014). Global mean sea-level rise (SLR) is projected to reach 0.26–0.55 m for scenario RCP2.6 and 0.45–0.82 m for RCP8.5 until 2100 compared to the 1986–2005 period, whereby local sea-level projections will deviate (IPCC 2014).
In addition to sea-level rise, changes in storm-surge activity could contribute to future changes in extreme water levels (Muis et al. 2023). Increased storm-surge activity and an increase in storm-surge levels are projected for the North Sea (Vousdoukas et al. 2016; Muis et al. 2023). Furthermore, regions not only on the north coast of Australia, Alaska, but also parts of the coasts of China, northern Caribbean, or eastern Africa, may experience higher storm-surge levels in the future as well (Colberg et al. 2019; Muis et al. 2023).
Climate change impacts on precipitation patterns will depend on climate and region. The annual mean precipitation in high latitudes and the equatorial Pacific will likely increase until the end of this century under scenario RCP8.5. Whereas many mid-latitude and subtropical dry regions will likely experience a mean precipitation decrease, mean precipitation in many mid-latitude wet regions will likely increase in the RCP8.5 scenario (IPCC 2014). Evapotranspiration is likely to increase as a result of temperature increase, and thus impact groundwater recharge (IPCC 2014).
Climate change effects on freshwater lenses (FWLs) and coastal aquifers are well studied. Several studies examined the effect of SLR (e.g. Masterson and Garabedian 2007; Masterson et al. 2014; Gulley et al. 2016), the effect of changes in groundwater recharge (e.g. Mollema and Antonellini 2013), or both (e.g. Werner and Simmons 2009; Rozell and Wong 2010; Sulzbacher et al. 2012; Unsal et al. 2014; Ketabchi et al. 2014, 2016; Mahmoodzadeh et al. 2014; Holding and Allen 2015; Babu et al. 2020). However, some studies modelled SLR as one or few instantaneous shifts (Rozell and Wong 2010; Ketabchi et al. 2014, 2016; Mahmoodzadeh et al. 2014; Holding and Allen 2015; Babu et al. 2020) instead of a continuous process.
Under unconfined aquifer conditions, the impact of SLR on freshwater resources varies according to the hydrogeological setting, specifically, topography-limited and recharge-limited aquifers (Werner and Simmons 2009; Michael et al. 2013). In topography-limited cases, the vadose zone above the water table is thin; thus, SLR, which causes the FWL to ascend within the aquifer, accompanied by a rise in the water table, can result in surface inundation by fresh groundwater. Consequently, the freshwater volume decreases (Rotzoll and Fletcher 2013; Masterson et al. 2014; Manda et al. 2015; Gulley et al. 2016). In recharge-limited cases, the FWL rises within the aquifer without inundating the surface as the vadose zone is sufficiently thick (Michael et al. 2013; Morgan and Werner 2014; Holding and Allen 2015). Some studies modelled the change of annual recharge rate as a certain percentage change (Rozell and Wong 2010; Sulzbacher et al. 2012; Unsal et al. 2014; Mahmoodzadeh et al. 2014) instead of using temporally variable recharge projections.
Many of the aforementioned studies are case studies of islands with FWLs (Rozell and Wong 2010; Sulzbacher et al. 2012; Holding and Allen 2015; Babu et al. 2020), others are generic analyses. Ketabchi et al. (2014) simulated a generic circular island with a two-layer aquifer based on characteristics of atoll islands. They investigated the effect of SLR with and without land surface inundation on the FWL volume for several ratios of hydraulic conductivity of the upper and lower aquifer layers. Furthermore, a sensitivity analysis examined the impact of additional parameters such as recharge in combination with SLR on the FWL volume. Unsal et al. (2014) investigated the effect of different pumping rates in combination with SLR and/or decreased recharge on the FWL volume and the freshwater–seawater interface position of a circular island. They also examined the impact of different hydraulic conductivity and porosity values. Gulley et al. (2016) modelled a cross-section of a generic topography-limited strip island on which SLR led to the creation of a lake in the island centre. They analysed the effect of this lake on the FWL volume of the island. Strip islands feature a length significantly greater than their width, and they have parallel boundaries (Vacher 1988; Chesnaux et al. 2021). Therefore, strip islands can be represented by a cross-sectional model as the three-dimensional (3D) affects can be disregarded.
In general, the size of a FWL depends on several island characteristics including island size (Fetter 1972; Vacher 1988) and shape (Thissen et al. 2024). Additionally, FWLs grow with increasing groundwater recharge and shrink with greater hydraulic conductivity (Fetter 1972; Vacher 1988; Bailey et al. 2009).
The interface between the freshwater and the underlying saltwater is characterised by groundwater mixing and the formation of a transition zone of brackish water. This zone is geochemically active and may feature enhanced dissolution of carbonates (Hanshaw and Back 1980; Back et al. 1986; Mylroie and Carew 1990). The transition zone can be very thick, extending almost up to the surface of the water table (Fratesi 2013).
The effect of variable recharge on FWLs and their transition zones has been investigated in several studies using conceptual hydrogeological models. Post et al. (2019) found that applying transient meteoric groundwater recharge caused a wider transition zone to form relative to the use of a constant mean recharge value when comparing several model approaches for the German North Sea island of Langeoog. Mollema and Antonellini (2013) also noted that invoking larger recharge variability results in smaller lenses and thicker transition zones relative to scenarios with lower recharge variability when they compared the influence of several recharge patterns from different climate regions on FWLs. These results agree with Eeman et al. (2012), who demonstrated that larger amplitudes of seasonal recharge variation resulted in a greater transition zone thickness.
Tidal signals are important to consider due to their impact on the transition zone, and the phenomenon of tidal overheight. Tidal overheight is the average tide-induced increase of the near-shore groundwater level above mean sea level (Nielsen 1990). Its magnitude depends on beach slope, tidal amplitude and hydraulic diffusivity (Nielsen 1990; Chang et al. 2010; Haehnel et al. 2023) and effects the thickness of the FWL (Pauw et al. 2014). However, many studies that investigated the effect of climate change on island or coastal freshwater resources neglected the effects of tides (e.g. Werner and Simmons 2009; Rozell and Wong 2010; Sulzbacher et al. 2012; Mollema and Antonellini 2013; Green and MacQuarrie 2014; Mahmoodzadeh et al. 2014; Holding and Allen 2015; Ketabchi et al. 2016; Babu et al. 2020).
To summarise, the majority of climate change studies on FWLs rely on simplified recharge and/or sea-level boundary conditions. Specifically, these studies did not take into account possible changes of variability in boundary conditions by using detailed projections based on climate models.
Therefore, this study aims to analyse in which way (1) future sea-level dynamics without SLR, i.e., only caused by tides and storm surges, (2) future sea-level dynamics with SLR, and (3) future sea-level dynamics with SLR and future recharge will change the state variables (a) FWL volume, (b) FWL depth, and (c) the thickness of the transition zone as compared to the past under several hydrogeological conditions. Approaches (1), (2), and (3) were chosen to distinguish between the effects of projected storm surges, SLR, and future groundwater recharge on FWLs, respectively. Therefore, a semi-generic approach was used that has, at least to the best of the authors’ knowledge, never been used before: a generic strip-island was modelled, employing climate projections of monthly recharge and high temporal resolution sea-level projections which are projected for an existing barrier island. This study aims to examine climate change impacts to FWLs of the East Frisian islands in north-west Germany, and to provide qualitatively transferable information to other regions and islands. Therefore, data on the island Norderney and parameter ranges typical of the East Frisian islands were used in an otherwise generic approach. To assess the results, the relative change of the considered state variables between a reference (1971–2000) and a future period (2071–2100) is shown for several combinations of projections from sea-level and recharge models available to further assess model uncertainties in the results.
Methods
Model
Model setup
The cross-sectional model represented a hypothetic, symmetrical sandy barrier strip island of 500, 1,000, and 2,000 m width with homogeneous geology. These widths are in the range of widths of the East Frisian barrier islands. The model was built using FloPy version 3.3.5 (Bakker et al. 2022) and the software SEAWAT (Langevin et al. 2008). For simplicity, only half of the respective symmetrical island was modelled (Fig. 1) and a flat island topography of 10 m above sea level was assumed; thus, the FWL was recharge-limited. Dunes were represented by a strong gradient of the topography towards the sea with a base at 3.5 m, which is approximate to the dune base on the East Frisian island Spiekeroog as described by Beck et al. (2017) and Seibert et al. (2019a). The beach was represented by a constant sloping morphology and the topography was considered to be temporarily invariant. The model had a horizontal and vertical discretisation of 10 and 1 m, respectively. A preliminary modelling experiment using spatial discretisations of 5 and 0.5 m in the horizontal and vertical directions, respectively, did not lead to significantly different results. Note that the hybrid method of characteristics for solving the advection–dispersion equation was used, which is robust against artificial dispersion and oscillation. It is also essentially independent regarding the grid-peclet number and, therefore, functions well even with coarser grids (Zheng and Bennett 2002). An exception to the general spatial discretisation was the top layer, which had a larger vertical discretisation that included the ground level (10 m) to –1 m. The base of the top model layer was set to –1 m to avoid the SEAWAT-specific numerical difficulties that often arise during drying and re-wetting of model layers in transient, unconfined aquifer simulations. Towards the sea, the upper 10 layers featured a decreasing vertical extent to maintain a constant number of layers in each column with the top layer representing the topography. The base case, shown in Fig. 1, featured a depth of 170 m, thus, 170 layers, one row, and 150 columns in the case of a 1,000-m-wide island. The beach slope was 0.4°. The length of the beach was chosen in order to keep the tidal boundary condition of the tidal phase with the lowest low tide within the model domain. The selected depth of the aquifer allowed the FWL to grow freely. All chosen parameter ranges and fixed parameters (Table 1) were based on those typical for the East Frisian islands—Fig. S1 in the electronic supplementary material (ESM). The East Frisian islands form a barrier island chain in the North Sea in north-west Germany that is exposed to semidiurnal tides with tidal ranges from 2.2 to 2.8 m (Wang et al. 2014). Boundary conditions were based on those of the island Norderney. All software settings that differ from default values and are not mentioned otherwise are provided in Table S1 in the ESM.
Boundary conditions
For the tidal boundary, water-level projections from 1951 to 2100 with a temporal resolution of 20 min at a position north of Norderney (latitude: 53.731197, longitude: 7.179276) were used. These data were derived from the hydrodynamic TRIM-NP model based on the regional climate projections obtained with the regional climate model REMO from two global climate models: MPI-ESM and HadGEM2 (Gaslikova 2023). For both global climate models, scenarios RCP2.6 and RCP8.5 (IPCC 2014) were available. Due to the large memory and time requirements of tide-resolved simulations, a phase-averaged approach based on Pauw et al. (2014) was applied (Appendix 1). In this tidal boundary condition, a stress period, i.e., a period of time-invariant boundary conditions, has the length of one tidal cycle; thus, models with HadGEM2 and MPI-ESM as sea-level models had 103,646 and 105,862 stress periods, respectively. Land surface inundation was considered implicitly. For simulations including mean SLR, annual SLR values were added (Fig. 2)—for the historical period 1961–1990, the difference between the 30-year moving average from observed annual mean sea level at Norderney (WSV 2021) and the observed mean sea level averaged for the 1961–1990 period was used; for the projection period 2020–2100, the median SLR estimates for RCP2.6 and RCP8.5 scenarios at Delfzijl, The Netherlands, from the FACTS model (Fox-Kemper et al. 2021; Garner et al. 2021) were applied; the data were linearly interpolated between those periods for the intermediate period 1991–2019.
In the flow model the groundwater recharge was modelled by assigning a second type boundary condition to the cells of the top layer above 0 m. The corresponding salinity was set to zero. Groundwater recharge projections with monthly data from 1970 to 2100 were provided by the regional hydrological model mGROWA22 for the German state of Lower Saxony (Hajati et al. 2022). For both climate change scenarios RCP2.6 and RCP8.5, three models out of eight or eleven, respectively, of the mGROWA ensemble were used (Fig. 3 and Fig. S2 of the ESM ). These models covered the whole range of the climatic water balance of the model ensemble: the model with the largest change in climatic water balance between the reference period 1971–2000 and the future period 2071–2100 on Norderney (“max”), the one with the smallest change (“min”), and the one nearest to the mean change (“mean”). These changes in mean monthly recharge are shown in Fig. 3c. Mean monthly recharge decreases between the reference period 1971–2000 and the future period 2071–2100 for recharge scenarios “min”. There is almost no change between analysed periods for recharge scenarios “mean”, and an increase for recharge scenarios “max”. The names of corresponding climate models are provided in Table S2 of the ESM. The spatial mean over all cells covering Norderney was used per month (Fig. S3 of the ESM ). If a stress period started in 1 month and extended into the following, it was attributed to the month in which it began. To adjust the length of recharge projections to that of sea-level projections, the monthly means of the first 10 years (1970–1979) were assigned to the corresponding months of the years 1951–1969. No-flow boundary conditions were assigned to the bottom and vertical boundaries in both flow and transport models.
Initial conditions
To investigate the effect of changes to boundary conditions over the time span of 150 years without including the effects of initial conditions, the model started in a dynamic equilibrium. This was reached by repeatedly running the model over the years 1951–1960. The first run started with initial conditions described in the following, and consecutive runs always started with the final situation of the previous run. These consecutive runs were repeated until the mean of all three investigated quantities over these 10 years visually reached a dynamic equilibrium (Fig. S4 of the ESM). Hydraulic heads and salinity distribution of the final time step of the last of these previous runs were then used as “true” initial conditions for the run over all 150 years.
Initial conditions of the first run over the years 1951–1960 depended on the sea-level model. Initial heads \({h}_{\text{ini},\text{FWL}}\) between island centre and high tide mark \(\text{HTM}\) of the first high water \(\text{HW}\) were calculated corresponding to Pauw et al. (2014):
where \(x\) is the location of the hydraulic head and \(\delta\) is the relative density difference between saltwater and freshwater \(\left(\frac{{\rho }_{\text{s}}-{\rho }_{\text{f}}}{{\rho }_{\text{f}}}\right)\). \({\rho }_{\text{f}}\) and \({\rho }_{\text{s}}\) are the densities of freshwater and saltwater, respectively. For \(K\) the horizontal hydraulic conductivity was used. \(N\) was originally the precipitation rate; however, the mean recharge rate over the years 1970–1979 was used, and \(\widetilde{L}\) was the theoretical distance from one side of the FWL to the \(\text{HTM}\) on the other side. The corresponding head at \(\text{HTM}\) \({h}_{\text{HTM}}\) was the mean of all \({h}_{\text{HTM}}\) of the years 1951–1960. Seaward of the \(\text{HTM}\), the heads \({h}_{\text{ini},\text{sea}}\) were linearly interpolated between \({h}_{\text{HTM}}\) and 0. The initial FWL depth \(D\) was calculated using the Ghyben-Herzberg relation (Herzberg 1901) and the initial heads \({h}_{\text{ini},\text{FWL}}\) from the island centre to \(\text{HTM}\): \(D= \frac{{\rho }_{\text{f}}}{{\rho }_{\text{s}}-{\rho }_{\text{f}}}\bullet {h}_{\text{ini},\text{FWL}}\). After testing with the base case model, the thickness of the initial transition zone was chosen to depend on the island width as follows: 30 m in the case of \(L=500 \text{m}\), 25 m in the case of \(L=1000 \text{m}\), and 20 m in the case of \(L=2000 \text{m}\). In all cases, the transition zone started 5 m above the calculated initial FWL depth \(D\) and extended down towards the seawater. The salinity was linearly interpolated from 0 to 35 kg/m3. Between \(\text{HTM}\) and the seaward boundary, and below the transition zone, all cells were filled with seawater.
The aim of this study was to investigate the effect of the change of boundary conditions due to climate change on FWLs. Running the model with boundary conditions of the first 10 years until a dynamic equilibrium is reached and using this situation as the initial condition for the run over all 150 years allows for extracting the effect of the change in boundary condition from other influences. It should, however, be mentioned that natural systems often are not in a dynamic equilibrium with their boundary conditions—for example, in the Netherlands, it was found that the density distribution of a groundwater system may need several millennia to reach a dynamic equilibrium (Oude Essink 1996). Similarly, a model of the East Frisian mainland in Germany reached steady-state conditions of salinities after ~3,000 simulated years (Seibert et al. 2023).
Approach
The study was divided into two parts:
-
Part 1.
Examination of the influence of several combinations of boundary conditions on different island widths. For the three island widths 500, 1,000, and 2,000 m, the effect of boundary conditions was evaluated in three steps: effects of (1) meteorically driven sea-level dynamics only, disregarding actual SLR, (2) sea-level dynamics with SLR, and (3) sea-level dynamics with SLR and recharge projections. In cases (1) and (2), monthly means of recharge from the reference period 1971–2000 of the models “mean” of the RCP scenario corresponding to the respective sea-level model were applied to account for seasonality but not for changes between past and future.
-
Part 2.
Investigation of the effects of sea-level dynamics with SLR and recharge projections (3) for \(L=1000\, \text{m}\) in combination with variations of hydrogeological parameters (Table 1). These parameters were changed separately to smaller and larger values, while unchanged parameters maintained the base case value.
In both parts, the results were examined by calculating the mean of (a) FWL volume, (b) FWL depth, and (c) transition zone thickness over all stress periods in the reference period 1971–2000 as well as in the projection period 2071–2100. (The projection period is 2070–2099 for the HadGEM2 sea-level model, since it ends in 2099. However, for better readability, in the following, the future projection period is referred to as 2071–2100.) The relative difference between these periods was then compared to model runs with similar values of hydrogeological parameters but boundary conditions from different climate models. Thus, per RCP scenario and island width, two (sea level) models were run for approaches (1) and (2) and all possible combinations of two sea-level models and three recharge models; hence, six models for (3). In total, 204 models were run. An overview of both study parts, the approaches with used parameters, investigated boundary conditions, and number of models are given in Table 2.
In the following, FWL depth refers to the depth of freshwater where 2.5% of the salinity of seawater is reached. The value is taken at the island centre, which corresponds to the leftmost column of the model (Fig. 1). The transition zone is defined as the brackish water zone from 2.5 to 95% salinity of seawater, likewise taken at the island centre. The FWL volume is the volume of freshwater (i.e. <2.5% of the salinity of seawater) below the island in all model columns that never get inundated by seawater. Only these columns were considered, since the freshwater volume below the beach is highly dependent on the saltwater fingering which occurred in the intertidal zone. Developing saltwater fingers are highly dynamic on a relatively short time scale and are of variable size (e.g. Greskowiak 2014; Fang et al. 2021) and, hence, strongly affect the freshwater volume within the intertidal zone. Note that salt fingering flow below beaches has so far not been observed in the field, but its occurrence has been described in experimental and modelling studies (e.g. Greskowiak 2014; Röper et al. 2015; Fang et al. 2021). This paper, however, aims to investigate the long-term effects of climate change; furthermore, the freshwater below a beach is usually not considered for use as drinking water because pumping wells are typically installed in more protected areas. Therefore, the beach was disregarded when calculating the freshwater volume (Fig. 1).
Calculation, representation, and interpretation of model results
Results are presented as relative differences between the averages over all time steps in the 30-year future projection period 2071–2100 and in the reference period 1971–2000. Thus, decreases in values of the investigated quantities from the reference to the future period are described with negative values, while positive values indicate an increase from past to future. Using averages over 30-year periods to calculate climate change impacts is typically done and recommended by the World Meteorological Organisation (2017). To the authors’ knowledge, however, this approach has so far never been used in hydrogeology to investigate climate change impacts on coastal aquifers and freshwater lenses.
Furthermore, it is recommended to use a model ensemble instead of one single climate model to account for uncertainties (Benestad et al. 2021). For the sea-level dynamics, the two existing models (Gaslikova 2023) and, thus, the largest possible model ensemble was used. For groundwater recharge those three models per RCP scenario were used as a subsample of the mGROWA22 model ensemble, representing the full range of projected climate change signals (Benestad et al. 2021; Hajati et al. 2022).
To investigate the uncertainties in the model projections the bandwidth of the results of the model ensemble have to be defined by either calculating mean and standard deviation or median and lower and upper percentiles (Benestad et al. 2021). The second possibility is less influenced by outliers. Therefore, results in “Results” section are represented as boxplots.
Sea-level projections and groundwater recharge models were combined in all possible combinations to include uncertainty resulting from using model projections. Due to this approach, however, boundary condition models originating from different climate models driven by different climate forcings were also combined. Therefore, model runs in this study cannot be associated with a single climate model.
Results
Impact of boundary conditions and island width
The positions of the transition zone in the first stress period of the reference period and the last time step of the future period are compared in Fig. 4 for models with sea-level model HadGEM2 RCP8.5 and recharge model “mean” RCP8.5 as boundary conditions and an island width of 1,000 m. With sea-level dynamics only (approach (1)), the transition zone maintains its position and its thickness increases only very little (Fig. 4a). However, when SLR is considered in addition to sea-level dynamics (approach (2)), the transition zone shifts upwards and expands slightly (Fig. 4b). When adding recharge projections to sea-level dynamics and SLR in approach (3), the transition zone shifts upwards and its thickness increases strongly (Fig. 4c). In all cases, an unstable, salt finger forming upper saline plume exists below the sloping beach.
The time series of the model in Fig. 4c is shown in Fig. 5, together with the time series of models with the same island width and sea-level model but different recharge projections. Results of models with recharge models “mean” and “min” are similar to each other and project a decrease in FWL volume and depth. If the average monthly recharge is projected to increase in the future (recharge model “max”, Fig. 3), FWL volume and depth are projected to increase first and then decrease slightly. The transition zone thickness is projected to increase for all three models. Note that initial conditions are not identical because the models started each with their dynamic equilibrium over the years 1951–1960 (see section “Initial conditions”).
For all models and all island widths for approaches (1), i.e., sea-level dynamics only, and (2), i.e., sea-level dynamics + SLR, relative differences in per cent between 30-year averages of reference and future period are displayed in Fig. 6. Negative values describe a decrease in the variables from past to future, while positive values describe an increase. FWL volume and depth are maintained for all cases in approach (1), while they are projected to decrease slightly for all cases in approach (2). The relative loss of freshwater volume is stronger for smaller islands. The effect on the transition zone thickness likewise depends on the island width. Smaller islands experience an increase in transition zone thickness. The increase is larger but still rather small in cases with SLR, approach (2). For larger islands, the transition zone thickness hardly changes at all—approaches (1) and (2)—or may even slightly decrease, approach (1). There are almost no differences between climate change scenarios RCP2.6 and RCP8.5.
Relative differences between averages of reference and future period of all model combinations of sea-level and recharge boundary conditions of approach (3), i.e., sea-level dynamics + SLR + recharge projections, are plotted in Fig. 7. FWL volume and depth are projected to decrease in most cases. Their mean and median decrease is projected to be only slightly larger in the case of the high emission scenario (RCP8.5) compared to the stringent mitigation scenario (RCP2.6). For islands with \(L=2000\, \text{m},\) mean and median of FWL volumes and depths show a decrease for the RCP8.5 scenario, while they predict no or only minor changes for scenario RCP2.6. In almost all model runs for approach (3), the transition zone increases. However, boxplot sizes are large; thus, model predictions differ strongly. Means and medians of changes in transition zone thickness are greater for scenario RCP8.5 than for scenario RCP2.6.
Impact of hydrogeological parameters and model setup
Results of model runs with varied hydrogeological parameters for an island with a width of 1,000 m are shown in Fig. 8. Corresponding values of mean, median, minimum, and maximum are listed in Table S3 of the ESM for each parameter setting. For comparison, boxplots in the centre of each pair of three, thus in dark blue and light red, show results with base case parameters equivalent to those in Fig. 7. For all model runs, mean and median show a decrease in FWL volume and depth from reference to future period, while mean and median of the transition zone thickness show an increase in almost all cases. Predicted decreases and increases are more pronounced for the high-emission climate change scenario RCP8.5 when considering means and medians. Also, the sizes of boxplots are usually larger for climate change scenario RCP8.5 than for scenario RCP2.6.
The strongest observed decreases from the reference period to the future period in FWL volume and depth, when looking at means and medians, are projected for larger hydraulic conductivity values (Fig. 8, top rows). If either \({K}_{\text{x}}\) is set to a greater value (\({K}_{\text{x}}=5\bullet {10}^{-4} \text{m}/\text{s}\)) or if \({K}_{\text{z}}\) is changed to result in isotropy (\({K}_{\text{x}}/{K}_{\text{z}} =1\)), mean and median of the relative change of FWL volume and depth become smallest. However, boxplot sizes in both cases, especially for a greater \({K}_{\text{x}}\), are large. Note that if \({K}_{\text{x}}\) is enlarged, also \({K}_{\text{z}}\) has a greater value, since the anisotropy remains constant. For the larger \({K}_{\text{x}}\) the projected mean relative increase in transition zone thickness has a smaller absolute value than the projected mean relative decreases of FWL volume and depth, but there is again a wide boxplot for the climate change scenario RCP8.5. A greater \({K}_{\text{x}}\) combined with scenario RCP2.6 is the only parameter combination for which mean and median are negative and, thus, project a maintenance or even a slight decrease of transition zone thickness. Both model cases with greater hydraulic conductivities than the base case, thus, \({K}_{\text{x}}=5\bullet {10}^{-4} \text{m}/\text{s}\) and \({K}_{\text{x}}/{K}_{\text{z}} =1\), have huge transition zones (Fig. S5 of the ESM). This explains the large absolute change in transition zone thickness and very large boxplots in Fig. S6 of the ESM. If \({K}_{\text{x}}\), however, is smaller than in the base case (\({K}_{\text{x}}=5\bullet {10}^{-5} \text{m}/\text{s}\)) or the anisotropy is greater (\({K}_{\text{x}}/{K}_{\text{z}} =20\)), relative and absolute changes are similar to those of the base case for all three investigated state variables. However, boxplots of FWL volume and depth of absolute values are a bit larger if \({K}_{\text{x}}\) is set to a smaller value.
Greater decreases in FWL volume and depth as well as greater increases in the transition zone thickness than in the base case are also predicted for larger dispersivities (\({\alpha }_{L}=5 \text{m}\)). Smaller dispersivities (\({\alpha }_{L}=0.5 \text{m}\)) result in smaller changes. If the beach slope is smaller (\(\beta =0.1^\circ\)), projected relative mean and median changes and their boxplots are small for FWL volumes and depths. For the transition zones, however, relative changes are projected to be smaller as well, but the boxplots are as large as in the base case scenario. In the case of a steeper beach \((\beta =0.8^\circ\)), relative changes are similar to those of the base case. The same accounts for different aquifer depths (\(B\)) or values of specific yield (\({s}_{\text{y}}\)). For these, all results are similar to those of the base case, only the boxplot sizes of relative differences of the transition zone thickness differ slightly.
Discussion
Climate change impact of boundary conditions and island width
The results describing the impact of boundary conditions and island width illustrate the importance of including both SLR and projections of groundwater recharge simultaneously in calculations of future FWL size. Almost no changes in average FWL volume, depth, and transition zone thickness are projected from reference to future period for approach (1), i.e., sea-level dynamics only. In contrast, changes are projected for approaches (2), i.e., sea-level dynamics + SLR, and (3), i.e., sea-level dynamics + SLR + recharge projections, whereas they are greater for the latter. Uncertainties of changes, represented by the difference between the largest and the smallest projected relative difference and, thus, the size of boxplots, are greatest for approach (3). Changes in the transition zone thickness in approach (1) for islands of width 500 or 1,000 m are mostly less than 1 m, which is the vertical resolution of the model. Thus, these changes are not significant and the effect of the changing dynamics is thus only minor as compared to the effect of SLR. Note that approach (A) was realised to differentiate the effect of mean SLR from the effect of potentially changing water-level variability. Clearly, approach (1) is not a realistic scenario. In approach (2), however, the largest absolute decrease of FWL depth is more than 3 m in scenarios RCP2.6 and RCP8.5 for \(L=500\text{ m}\). The projected increase of the transition zone is close to 2 m in three of four models with \(L=500\text{ m}\). These changes are not negligible as they clearly exceed the vertical model resolution. A slight increase in transition zone thickness by SLR was also observed by Mahmoodzadeh et al. (2014). They modelled the long-term response of an island cross-section based on Kish Island in the Persian Gulf, following an instantaneous SLR until the FWL reached steady-state conditions again. Figures 6 and 7 show that the impact of the boundary conditions is more pronounced for smaller islands than for larger ones. Decreases in volume and depth of the FWL and increases in the transition zone thickness are in most cases predicted to be largest for islands with \(L = 500\) m. For transition zones in approach (3), the impact of the island width is unclear and results show a large uncertainty. Transferred to the East Frisian islands, these results indicate that narrower islands like Juist, with a width of maximum 900 m, are more vulnerable to climate change impacts than wider islands like Norderney or Borkum.
Comparing aproaches (2) and (3) clearly shows the importance of including groundwater recharge projections into models. Groundwater recharge seasonality was already included in approach (2), but it is projected to become more pronounced in the future in all recharge models (Fig. S7 of the ESM). Such changes enhance the increase of transition zone thickness in approach (3) compared to approach (2) (Fig. 4). These results also underline the importance of modelling groundwater recharge as a transient process as stated by Post et al. (2019). This is in agreement with Mollema and Antonellini (2013) who compared FWL size and transition zone thickness of modelled FWLs using potential recharge values of several climatic regions from all around the world. Mahmoodzadeh et al. (2014) stated that under recharge-limited conditions, a change in recharge rate greatly affects FWLs. Depending on whether groundwater recharge increases or decreases, it may both mitigate or increase saltwater intrusion impacts due to SLR. SLR itself, however, has a smaller impact in recharge-limited aquifers since it simply results in a rise of the FWL within the aquifer, whereas the hydraulic gradient between sea and land does not change (Michael et al. 2013). This study shows that changes in groundwater recharge have a greater influence on FWLs than SLR under the given conditions, which was also reported by Holding and Allen (2015) and Ketabchi et al. (2014). Moreover, adding recharge projections in approach (3) results in differences between climate change scenarios RCP2.6 and RCP8.5 that were not observed in approaches (1) and (2). Recharge projections also increase uncertainties of changes between past and future periods (Figs. 6 and 7). Furthermore, a comparison of approaches (1), (2), and (3) shows that projected groundwater recharge has a larger impact on the transition zone thickness than sea-level dynamics. Projected increase of transition zone thickness underlines the importance of including mixing processes into models instead of using sharp-interface models as, for example, done by Babu et al. (2018; 2020) for the island of Tongatapu in the Kingdom of Tonga. Results in Figs. 6 and 7 show that neglecting the transition zone may lead to an underestimation of climate change impacts on FWL volumes and depths.
Climate change impact of hydrogeological parameters
The second part of this study investigated the impact of several hydrogeological parameters on changes in FWL size and transition zone thickness. Performing a sensitivity analysis for a two-layered circular island model, Ketabchi et al. (2014) discovered that FWL volume is more sensitive to groundwater recharge than to beach slope or hydraulic conductivity across the tested parameter range. Results given in Fig. 8 likewise show that future relative changes of FWL volumes and depths are more sensitive to external factors such as, e.g., groundwater recharge than to hydrogeological conditions or beach slope. The study by Ketabchi et al. (2014) is based on atoll islands but results are nevertheless similar to those of this study. Thus, it is likely that the findings of the present study are transferable from the North Sea region to other parts of the world.
The hydrogeological factor with the largest impact but also highest uncertainty is the hydraulic conductivity. A larger \(K\) increases the transmissivity of the aquifer and, thus, decreases the damping of the tidal amplitude during its motion within the aquifer (Ferris 1952). A larger vertical hydraulic conductivity affects the vertical flow and allows for stronger changes of the vertical extend of FWLs. The recharge model “max”, therefore, leads to a stronger relative increase, while “min” results in a stronger relative decrease of FWL volume and depth, if \({K}_{\text{z}}\) is greater than in the base case. Therefore, uncertainties of the changes of FWL volume and depth are largest if both, \({K}_{\text{x}}\) and \({K}_{\text{z}}\), have greater values or if the aquifer is isotropic with the same \({K}_{\text{x}}\) as in the base case. Another relevant parameter is the dispersivity. When FWLs expand or contract, longitudinal dispersivity dominates the mixing process (Eeman et al. 2011). Therefore, because the vertical transverse dispersivity is increased correspondingly due to a constant \({\alpha }_{\text{VT}}/{\alpha }_{\text{L}}\)-ratio, a larger \({\alpha }_{\text{L}}\) results in a stronger increase in transition zone thickness and a decrease of FWL volume and depth than smaller dispersivities.
It should be noted that changes in Figs. 6, 7 and 8 are those which occur during the simulation time. However, since climate change impacts FWLs slowly, the system is likely not in equilibrium with the new boundary conditions after 150 model years. This could be the reason why there are almost no differences between the RCP2.6 and RCP8.5 scenarios in approach (B), although SLR between both scenarios differs strongly. It should further be considered that the tidal boundary condition in this study includes land surface inundation. However, the effect of land surface inundation is not reflected in the results in Figs. 6, 7 and 8 and Table S3 of the ESM , because only the volume below the island and not below the beach was considered. Neglect of land surface inundation due to SLR would lead to an underestimation of climate change impacts on FWL volumes and depths (Ataie-Ashtiani et al. 2013; Ketabchi et al. 2014; Mahmoodzadeh et al. 2014).
Conclusions
A cross-sectional model of a hypothetical island was modelled with boundary conditions representative of the German North Sea barrier island Norderney. High-resolution sea-level projections, annual SLR data, and monthly groundwater recharge values of several climate models were used to evaluate climate change impacts on freshwater lens volume, depth, and transition zone thickness. Differences between the reference period 1971–2000 and the future period 2071–2100 were quantified. In the first part, the impact of boundary conditions on islands of 500, 1,000, and 2,000 m width was analysed. The effects of the three external factors—sea-level dynamics, SLR, and groundwater recharge variations—were assessed. A multi-parameter study additionally evaluated the impacts of the hydrogeological factors of aquifer depth, hydraulic conductivity, anisotropy of hydraulic conductivity, specific yield, beach slope, and dispersivity on 1,000-m-wide islands.
Key conclusions of this study are:
-
It is very likely that FWL volume and depth on islands will decrease, while transition zone thickness will increase in the future as a consequence of climate change.
-
The smaller the island, the greater the effect of the boundary conditions.
-
The external factors groundwater recharge and SLR have a larger influence on FWLs and their transition zones than most hydrogeological parameters within parameter ranges realistic for the East Frisian islands.
-
From the hydrogeological parameters investigated, hydraulic conductivity has the greatest effect but also the largest uncertainty associated with it. Its anisotropy and dispersivity have a moderate and beach slope a minor influence. Specific yield and aquifer depth have a negligible effect.
-
Recharge projections have a larger effect on the transition zone thickness relative to sea-level dynamics.
-
With respect to the employed recharge and sea-level projection models, recharge projections lead to greater uncertainties regarding the projected changes in the considered state variables.
-
Observed differences between climate change scenarios RCP2.6 and RCP8.5 within the 150 model years largely result from groundwater recharge projections.
In summary, this paper shows that under all (geological) conditions tested, climate change will very likely lead to smaller freshwater lenses in the future, consequently affecting the amount of freshwater available for drinking-water production on sandy barrier islands. This study confirms results of previous studies that predict a decrease in the amount of freshwater due to climate change, but in addition it also illustrates the importance of including variability and uncertainties of projections resulting from climate models.
References
Ataie-Ashtiani B, Werner AD, Simmons CT, Morgan LK, Lu C (2013) How important is the impact of land-surface inundation on seawater intrusion caused by sea-level rise? Hydrogeol J 21(7):1673–1677. https://doi.org/10.1007/s10040-013-1021-0
Babu R, Park N, Nam B (2020) Regional and well-scale indicators for assessing the sustainability of small island fresh groundwater lenses under future climate conditions. Environ Earth Sci 79(1):47. https://doi.org/10.1007/s12665-019-8773-3
Babu R, Park N, Yoon S, Kula T (2018) Sharp interface approach for regional and well scale modeling of small island freshwater lens: Tongatapu Island. Water 10(11):1636. https://doi.org/10.3390/w10111636
Back W, Hanshaw BB, Herman JS, Van Driel JN (1986) Differential dissolution of a Pleistocene reef in the ground-water mixing zone of coastal Yucatan. Mexico Geol 14(2):137–140. https://doi.org/10.1130/0091-7613(1986)14%3c137:DDOAPR%3e2.0.CO;2
Bailey RT, Jenson JW, Olsen AE (2009) Numerical modeling of atoll island hydrogeology. Groundwater 47(2):184–196. https://doi.org/10.1111/j.1745-6584.2008.00520.x
Bakker M, Post VEA, Hughes JD, Langevin CD, White JT, Leaf AT, Paulinski SR, Bellino JC, Morway ED, Toews MW, Larsen JD, Fienen MN, Starn JJ, Brakenhoff D (2022) FloPy v3.3.5: U.S. Geological Survey Software Release. US Geological Survey, Reston, VA
Beck M, Reckhardt A, Amelsberg J, Bartholomä A, Brumsack H-J, Cypionka H, Dittmar T, Engelen B, Greskowiak J, Hillebrand H, Holtappels M, Neuholz R, Köster J, Kuypers MMM, Massmann G, Meier D, Niggemann J, Paffrath R, Pahnke K, Rovo S, Striebel M, Vandieken V, Wehrmann A, Zielinski O (2017) The drivers of biogeochemistry in beach ecosystems: a cross-shore transect from the dunes to the low-water line. Mar Chem 190:35–50. https://doi.org/10.1016/j.marchem.2017.01.001
Benestad R, Buonomo E, Gutiérrez JM, Haensler A, Hennemuth B, Illy T, Jacob D, Keup-Thiel E, Katragkou E, Kotlarski S, Nikulin G, Otto J, Rechid D, Remke T, Sieck K, Sobolowski S, Szabó P, Szépszó G, Teichmann C, Vautard R, Weber T, Zsebeházi G, Horányi A (2021) Guidance for EURO-CORDEX climate projections data use: Version 1.1. https://www.climate-service-center.de/imperia/md/content/csc/cordex/guidance_for_euro-cordex_climate_projections_data_use__2021-02.pdf
Chang Y-C, Jeng D-S, Yeh H-D (2010) Tidal propagation in an oceanic island with sloping beaches. Hydrol Earth Syst Sci 14(7):1341–1351. https://doi.org/10.5194/hess-14-1341-2010
Chesnaux R, Marion D, Boumaiza L, Richard S, Walter J (2021) An analytical methodology to estimate the changes in fresh groundwater resources with sea-level rise and coastal erosion in strip-island unconfined aquifers: illustration with Savary Island, Canada. Hydrogeol J 29(3):1355–1364. https://doi.org/10.1007/s10040-020-02300-0
Colberg F, McInnes KL, O’Grady J, Hoeke R (2019) Atmospheric circulation changes and their impact on extreme sea levels around Australia. Nat Hazards Earth Syst Sci 19(5):1067–1086. https://doi.org/10.5194/nhess-19-1067-2019
Eeman S, Leijnse A, Raats PAC, van der Zee SEATM (2011) Analysis of the thickness of a fresh water lens and of the transition zone between this lens and upwelling saline water. Adv Water Resour 34(2):291–302. https://doi.org/10.1016/j.advwatres.2010.12.001
Eeman S, van der Zee SEATM, Leijnse A, de Louw PGB, Maas C (2012) Response to recharge variation of thin rainwater lenses and their mixing zone with underlying saline groundwater. Hydrol Earth Syst Sci 16(10):3535–3549. https://doi.org/10.5194/hess-16-3535-2012
Fang Y, Zheng T, Zheng X, Yang H, Wang H, Walther M (2021) Influence of tide‐induced unstable flow on seawater intrusion and submarine groundwater discharge. Water Resour Res 57(4):e2020WR029038. https://doi.org/10.1029/2020WR029038
Ferris JG (1952) Cyclic fluctuations of water level as a basis for determining aquifer transmissibility. US Geological Survey, Reston. http://pubs.er.usgs.gov/publication/70133368. Accessed July 2023
Fetter CW (1972) Position of the saline water interface beneath oceanic islands. Water Resour Res 8(5):1307–1315. https://doi.org/10.1029/WR008i005p01307
Fox-Kemper B, Hewitt HT, Xiao C, Aðalgeirsdóttir G, Drijfhout SS, Edwards TL, Golledge NR, Hemer M, Kopp RE, Krinner G, Mix A, Notz D, Nowicki S, Nurhati IS, Ruiz L, Sallée J-B, Slangen ABA, Yu Y (2021) Ocean, Cryosphere and Sea Level Change. In: Masson-Delmotte V, Zhai P, Pirani A, Connors SL, Péan C, Berger S, Caud N, Chen Y, Goldfarb L, Gomis MI, Huang M, Leitzell K, Lonnoy E, Matthews JBR, Maycock TK, Waterfield T, Yelekçi O, Yu R, Zhou B (eds) Climate Change 2021: the physical science basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, NY, pp 1211–1362
Fratesi B (2013) Hydrology and geochemistry of the freshwater lens in coastal karst. In: Lace MJ, Mylroie JE (eds) Coastal karst landforms. Springer, Dordrecht, pp 59–75
Garner GG, Hermans T, Kopp RE, Slangen ABA, Edwards TL, Levermann A, Nowicki S, Palmer MD, Smith C, Fox-Kemper B, Hewitt HT, Xiao C, Aðalgeirsdóttir G, Drijfhout SS, Golledge NR, Hemer M, Krinner G, Mix A, Notz D, Nurhati IS, Ruiz L, Sallée J-B, Yu Y, Hua L, Palmer T, Pearson B (2021) IPCC AR6 sea-level rise projections. Version 20210809. https://podaac.jpl.nasa.gov/announcements/2021-08-09-Sea-level-projections-from-the-IPCC-6th-Assessment-Report. Accessed 23 Feb 2022
Gaslikova L (2023) coastDat TRIM-NP-2d CMIP5 hydrodynamic transient scenarios for the North Sea and the Northeast Atlantic for the period 1950–2100 with REMO MPI-ESM and HadGEM2 forcing. World Data Center for Climate (WDCC) at DKRZ. https://doi.org/10.26050/WDCC/cD_C5_sc
Gerhardy H, Hoffmann B, Meyer HH, Wichmann K (1976) Wasserhaushalt der Ostfriesischen Inseln: Geologische und Hydrogeologische Datenbestandsaufnahme für die Insel Norderney [Water balance of the East Frisian islands: geological and hydrogeological data inventory for the island Norderney]. Technische Universität Hannover, Hannover, Germany
Green NR, MacQuarrie KTB (2014) An evaluation of the relative importance of the effects of climate change and groundwater extraction on seawater intrusion in coastal aquifers in Atlantic Canada. Hydrogeol J 22(3):609–623. https://doi.org/10.1007/s10040-013-1092-y
Greskowiak J (2014) Tide-induced salt-fingering flow during submarine groundwater discharge. Geophys Res Lett 41(18):6413–6419. https://doi.org/10.1002/2014GL061184
Gulley JD, Mayer AS, Martin JB, Bedekar V (2016) Sea level rise and inundation of island interiors: assessing impacts of lake formation and evaporation on water resources in arid climates. Geophys Res Lett 43(18):9712–9719. https://doi.org/10.1002/2016GL070667
Günther T, Müller-Petke M (2012) Hydraulic properties at the North Sea island of Borkum derived from joint inversion of magnetic resonance and electrical resistivity soundings. Hydrol Earth Syst Sci 16(9):3279–3291. https://doi.org/10.5194/hess-16-3279-2012
Haehnel P, Greskowiak J, Robinson CE, Schuett M, Massmann G (2023) Efficient representation of transient tidal overheight in a coastal groundwater flow model using a phase-averaged tidal boundary condition. Adv Water Resour 104538. https://doi.org/10.1016/j.advwatres.2023.104538
Hajati M-C, Harders D, Petry U, Elbracht J, Engel N (2022) Dokumentation der niedersächsischen Klimaprojektionsdaten AR5-NI v2.1 [Documentation of climate projection data AR5-NI v2.1 of Lower Saxony] - Geofakten 39. Landesamt für Bergbau, Energie und Geologie (LBEG), Hannover, Germany. https://doi.org/10.48476/geofakt_39_1_2022
Hanshaw BB, Back W (1980) Chemical mass-wasting of the northern Yucatan Peninsula by groundwater dissolution. Geol 8(5):222–224. https://doi.org/10.1130/0091-7613(1980)8%3c222:CMOTNY%3e2.0.CO;2
Herzberg A (1901) Die Wasserversorgung einiger Nordseebäder [The water supply of some North Sea spas]. J Gasbeleucht Wasserversorg 44:815–819 and 842–844
Holding S, Allen DM (2015) From days to decades: numerical modelling of freshwater lens response to climate change stressors on small low-lying islands. Hydrol Earth Syst Sci 19(2):933–949. https://doi.org/10.5194/hess-19-933-2015
Holt T, Greskowiak J, Seibert SL, Massmann G (2019) Modeling the evolution of a freshwater lens under highly dynamic conditions on a currently developing barrier island. Geofluids 2019(Article ID 9484657):1–15. https://doi.org/10.1155/2019/9484657
IPCC (2014) Core Writing Team, Pachauri RK, Meyer LA (eds) Climate Change 2014: Synthesis report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Intergovernmental Panel on Climate Change, Geneva, Switzerland
IPCC (2021) Masson-Delmotte V, Zhai P, Pirani A, Connors SL, Péan C, Berger S, Caud N, Chen Y, Goldfarb L, Gomis MI, Huang M, Leitzell K, Lonnoy E, Matthews JBR, Maycock TK, Waterfield T, Yelekçi O, Yu R, Zhou B (eds) Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge. https://www.cambridge.org/core/product/identifier/9781009157896/type/book. Accessed July 2023
Ketabchi H, Mahmoodzadeh D, Ataie-Ashtiani B, Werner AD, Simmons CT (2014) Sea-level rise impact on fresh groundwater lenses in two-layer small islands. Hydrol Process 28(24):5938–5953. https://doi.org/10.1002/hyp.10059
Ketabchi H, Mahmoodzadeh D, Ataie-Ashtiani B, Simmons CT (2016) Sea-level rise impacts on seawater intrusion in coastal aquifers: review and integration. J Hydrol 535:235–255. https://doi.org/10.1016/j.jhydrol.2016.01.083
Langevin CD, Thorne Jr. DT, Dausman AM, Sukop MC, Guo W (2008) SEAWAT version 4: a computer program for simulation of multi-species solute and heat transport (No. 6-A22). US Geological Survey, Reston. https://pubs.er.usgs.gov/publication/tm6A22
Mahmoodzadeh D, Ketabchi H, Ataie-Ashtiani B, Simmons CT (2014) Conceptualization of a fresh groundwater lens influenced by climate change: a modeling study of an arid-region island in the Persian Gulf, Iran. J Hydrol 519(Part A):399–413. https://doi.org/10.1016/j.jhydrol.2014.07.010
Manda AK, Sisco MS, Mallinson DJ, Griffin MT (2015) Relative role and extent of marine and groundwater inundation on a dune-dominated barrier island under sea-level rise scenarios. Hydrol Process 29(8):1894–1904. https://doi.org/10.1002/hyp.10303
Masterson JP, Garabedian SP (2007) Effects of sea-level rise on ground water flow in a coastal aquifer system. Groundwater 45(2):209–217. https://doi.org/10.1111/j.1745-6584.2006.00279.x
Masterson JP, Fienen MN, Thieler ER, Gesch DB, Gutierrez BT, Plant NG (2014) Effects of sea-level rise on barrier island groundwater system dynamics: ecohydrological implications. Ecohydrol 7(3):1064–1071. https://doi.org/10.1002/eco.1442
Michael HA, Russoniello CJ, Byron LA (2013) Global assessment of vulnerability to sea-level rise in topography-limited and recharge-limited coastal groundwater systems. Water Resour Res 49(4):2228–2240. https://doi.org/10.1002/wrcr.20213
Mollema PN, Antonellini M (2013) Seasonal variation in natural recharge of coastal aquifers. Hydrogeol J 21(4):787–797. https://doi.org/10.1007/s10040-013-0960-9
Morgan LK, Werner AD (2014) Seawater intrusion vulnerability indicators for freshwater lenses in strip islands. J Hydrol 508:322–327. https://doi.org/10.1016/j.jhydrol.2013.11.002
Muis S, Aerts JCJH, Á. Antolínez JA, Dullaart JC, Duong TM, Erikson L, Haarsma RJ, Apecechea MI, Mengel M, Le Bars D, O’Neill A, Ranasinghe R, Roberts MJ, Verlaan M, Ward PJ, Yan K (2023) Global projections of storm surges using high‐resolution CMIP6 climate models. Earth’s Future 11(9):e2023EF003479. https://doi.org/10.1029/2023EF003479
Mylroie JE, Carew JL (1990) The flank margin model for dissolution cave development in carbonate platforms. Earth Surf Process Landforms 15(5):413–424. https://doi.org/10.1002/esp.3290150505
Naumann K (2005) Eine hydrogeologische Systemanalyse von Süßwasserlinsen als Grundlage einer umweltschonenden Grundwasserbewirtschaftung [A hydrogeological system analysis of freshwater lenses as a basis for environmentally friendly groundwater management]. PhD Thesis, Technische Universität Carolo-Wilhelmina zu Braunschweig, Germany
Nielsen P (1990) Tidal dynamics of the water table in beaches. Water Resour Res 26(9):2127–2134. https://doi.org/10.1029/WR026i009p02127
Oude Essink GHP (1996) Impact of sea level rise on groundwater flow regimes: a sensitivity analysis for the Netherlands. Delft University of Technology, The Netherlands
Pauw PS, Oude Essink GHP, Leijnse A, Vandenbohede A, Groen J, van der Zee SEATM (2014) Regional scale impact of tidal forcing on groundwater flow in unconfined coastal aquifers. J Hydrol 517:269–283. https://doi.org/10.1016/j.jhydrol.2014.05.042
Post VEA, Houben GJ (2017) Density-driven vertical transport of saltwater through the freshwater lens on the island of Baltrum (Germany) following the 1962 storm flood. J Hydrol 551:689–702. https://doi.org/10.1016/j.jhydrol.2017.02.007
Post VEA, Houben GJ, Stoeckl L, Sültenfuß J (2019) Behaviour of tritium and tritiogenic helium in freshwater lens groundwater systems: insights from Langeoog Island, Germany. Geofluids 2019: Art. no. 1494326. https://doi.org/10.1155/2019/1494326
Post VEA, Zhou T, Neukum C, Koeniger P, Houben GJ, Lamparter A, Šimůnek J (2022) Estimation of groundwater recharge rates using soil-water isotope profiles: a case study of two contrasting dune types on Langeoog Island, Germany Hydrogeol J 30(3):797–812. https://doi.org/10.1007/s10040-022-02471-y
Röper T, Greskowiak J, Massmann G (2015) Instabilities of submarine groundwater discharge under tidal forcing. Limnol Oceanogr 60(1):22–28. https://doi.org/10.1002/lno.10005
Rozell DJ, Wong T (2010) Effects of climate change on groundwater resources at Shelter Island, New York State, USA. Hydrogeol J 18(7):1657–1665. https://doi.org/10.1007/s10040-010-0615-z
Rotzoll K, Fletcher CH (2013) Assessment of groundwater inundation as a consequence of sea-level rise. Nat Clim Change 3(May):477–481. https://doi.org/10.1038/nclimate1725
Seibert SL, Holt T, Reckhardt A, Ahrens J, Beck M, Pollmann T, Giani L, Waska H, Böttcher ME, Greskowiak J, Massmann G (2018) Hydrochemical evolution of a freshwater lens below a barrier island (Spiekeroog, Germany): the role of carbonate mineral reactions, cation exchange and redox processes. Appl Geochem 92:196–208. https://doi.org/10.1016/j.apgeochem.2018.03.001
Seibert SL, Böttcher ME, Schubert F, Pollmann T, Giani L, Tsukamoto S, Frechen M, Freund H, Waska H, Simon H, Holt T, Greskowiak J, Massmann G (2019a) Iron sulfide formation in young and rapidly-deposited permeable sands at the land-sea transition zone. Sci Total Environ 649:264–283. https://doi.org/10.1016/j.scitotenv.2018.08.278
Seibert SL, Greskowiak J, Prommer H, Böttcher ME, Massmann G (2019b) Modeling of biogeochemical processes in a barrier island freshwater lens (Spiekeroog, Germany). J Hydrol 575:1133–1144. https://doi.org/10.1016/j.jhydrol.2019.05.094
Seibert SL, Greskowiak J, Bungenstock F, Freund H, Karle M, Meyer R, Oude Essink GHP, Van Engelen J, Massmann G (2023) Paleo‐hydrogeological modeling to understand present‐day groundwater salinities in a low‐lying coastal groundwater system (northwestern Germany). Water Resour Res 59(4):e2022WR033151. https://doi.org/10.1029/2022WR033151
Sulzbacher H, Wiederhold H, Siemon B, Grinat M, Igel J, Burschil T, Günther T, Hinsby K (2012) Numerical modelling of climate change impacts on freshwater lenses on the North Sea Island of Borkum using hydrological and geophysical methods. Hydrol Earth Syst Sci 16(10):3621–3643. https://doi.org/10.5194/hess-16-3621-2012
Thissen L, Greskowiak J, Massmann G (2024) Estimating freshwater lens volume based on island circularity. Groundwater 62(2):250–259. https://doi.org/10.1111/gwat.13323
Unsal B, Yagbasan O, Yazicigil H (2014) Assessing the impacts of climate change on sustainable management of coastal aquifers. Environ Earth Sci 72(6):2183–2193. https://doi.org/10.1007/s12665-014-3130-z
Vacher HL (1988) Dupuit-Ghyben-Herzberg analysis of strip-island lenses. GSA Bull 100(4):580–591. https://doi.org/10.1130/0016-7606(1988)100%3c0580:DGHAOS%3e2.3.CO;2
Vandenbohede A, Lebbe L (2006) Occurrence of salt water above fresh water in dynamic equilibrium in a coastal groundwater flow system near De Panne, Belgium. Hydrogeol J 14(4):462–472. https://doi.org/10.1007/s10040-005-0446-5
Vandenbohede A, Lebbe L (2007) Effects of tides on a sloping shore: groundwater dynamics and propagation of the tidal wave. Hydrogeol J 15(4):645–658. https://doi.org/10.1007/s10040-006-0128-y
Vousdoukas MI, Voukouvalas E, Annunziato A, Giardino A, Feyen L (2016) Projections of extreme storm surge levels along Europe. Clim Dyn 47(9–10):3171–3190. https://doi.org/10.1007/s00382-016-3019-5
Wang Y, Yu Q, Gao S (2014) Modeling interrelationships between morphological evolution and grain-size trends in back-barrier tidal basins of the East Frisian Wadden Sea. Geo-Mar Lett 34(1):37–49. https://doi.org/10.1007/s00367-013-0349-8
Wasserstraßen- und Schifffahrtsverwaltung des Bundes (WSV) [Federal Waterways and Shipping Administration] (2021) Sea level rise at Norderney from 1961–1990, provided by Bundesanstalt für Gewässerkunde (BfG), Koblenz, Germany
Werner AD, Simmons CT (2009) Impact of sea-level rise on sea water intrusion in coastal aquifers. Groundwater 47(2):197–204. https://doi.org/10.1111/j.1745-6584.2008.00535.x
World Meteorological Organisation (WMO) (2017) WMO guidelines on the calculation of climate normals. WMO, Geneva, Switzerland
Zheng C, Bennett GD (2002) Applied contaminant transport modeling. Wiley, New York
Acknowledgements
The authors would like to thank the LBEG (State Authority for Mining, Energy, and Geology, Lower Saxony, Germany) for provision of groundwater recharge projections of mGROWA22. Furthermore, thanks to Patrick Haehnel from the Universität Oldenburg for his help regarding the tidal boundary condition. The authors would further like to thank him and all colleagues from the WAKOS project for fruitful discussions. Lastly, the authors also would like to thank the associate editor Nils Moosdorf and three anonymous reviewers for their comments, which helped to improve this manuscript.
Funding
Open Access funding enabled and organized by Projekt DEAL. The authors would like to thank the German Bundesministerium für Bildung und Forschung (BMBF) for funding this work (No. 01LR2003E) within the project ‘WAKOS—Wasser an den Küsten Ostfrieslands: Basis für maßgeschneiderte Klimaservices für die Anpassung [WAKOS—Water on the coasts of East Frisia: Basis for customised climate services for adaptation]’. The simulations were performed at the HPC Cluster CARL, located at the Universität Oldenburg (Germany) and funded by the German Research Foundation (DFG) through its Major Research Instrumentation Programme (INST 184/157–1 FUGG) and the Ministry of Science and Culture (MWK) of the Lower Saxony State.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflicts of interest
On behalf of all authors, the corresponding author states that there is no conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Electronic supplementary material
Below is the link to the electronic supplementary material.
Appendices
Appendices
Appendix 1
Each stress period was defined between two consecutive low waters, i.e. its duration is equivalent to the length of the respective tidal cycle (Fig. 9a). For each stress period the following steps were conducted: Mean tide level \(\overline{{h }_{\text{s}}}=\frac{\text{HW}+\overline{\text{LW}}}{2 }\) was calculated, where \(\text{HW}\) is the high water and \(\overline{\text{LW} }\) is the mean of both low waters. The tidal amplitude resulted in \(A=\text{HW}-\overline{{h }_{\text{s}}}\). Furthermore, those model cells of the top model layer were evaluated where \(\text{HW}\) and \(\overline{\text{LW} }\) meet the beach, corresponding to the high tide mark \(\text{HTM}\) and the low tide mark \(\text{LTM}\) (Fig. 9b). The hydraulic head at \(\text{HTM}\) \({h}_{\text{HTM}}\) was calculated using an empirically derived formula by Pauw et al. (2014). Since this formula was developed for a mean sea level at 0 m, the mean tide level \(\overline{{h }_{\text{s}}}\) of each stress period was added to account for changes in \(\text{HW}\) and \(\text{LW}\). Because Pauw et al. (2014) used an isotropic hydraulic conductivity \(K\), it was tested whether the use of the vertical or horizontal hydraulic conductivity predicts hydraulic heads at \(\text{HTM}\) better by comparing calculated heads of this approach with heads calculated using a tide-resolved approach. Within the parameter ranges investigated in this study, good results were obtained using the horizontal hydraulic conductivity \({K}_{\text{x}}\) (Fig. S8 of the ESM), thus, this was used. For parameter set-ups with \(\beta =0.1^\circ\), \({K}_{\text{x}}=5\bullet {10}^{-5}\, \text{m}/\text{s}\) and \({K}_{\text{x}}=5\bullet {10}^{-4} \,\text{m}/\text{s}\) the mean offsets over 2 months of –0.15, –0.23, and 0.36 m were added to \({h}_{\text{HTM}}\) (Fig. S8 of the ESM ). At and below \(\text{LTM}\) the phase-averaged hydraulic head corresponds to \(\overline{{h }_{\text{s}}}\) (Vandenbohede and Lebbe 2006, 2007). Within the intertidal zone, thus between \(\text{HTM}\) and \(\text{LTM}\), heads were linearly interpolated between \({h}_{\text{HTM}} (+\text{offset})\) and \(\overline{{h }_{\text{s}}}\) (Fig. 9b). These heads were specified in the flow model for each stress period using a third-type boundary condition. Above \(\text{HW}\) only outflow of water was allowed. Therefore, all cells above \(\text{HTM}\) but below or equal to the highest high tide mark \(\text{HHTM}\) were set to drainage cells (DRN-package), while for cells below \(\text{HTM}\) the general head boundary package (GHB) was used. Using this tidal boundary approach, the model accounted for the overheight of the water table due to tides (Nielsen 1990; Pauw et al. 2014) and for lateral movement of the sea, i.e., land surface inundation due to storm floods and SLR. In the transport model, inflowing water was set to have the salinity of seawater, while outflowing water has the salinity of the water within the respective model cell.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Thissen, L., Greskowiak, J., Gaslikova, L. et al. Climate change impact on barrier island freshwater lenses and their transition zones: a multi-parameter study. Hydrogeol J (2024). https://doi.org/10.1007/s10040-024-02800-3
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s10040-024-02800-3