Skip to main content

Advertisement

Log in

Coastal groundwater model calibration using filtered and amplified hydraulic information retained in the freshwater–saltwater interface

Calage d’un modèle des eaux souterraines côtières utilisant l’information hydraulique filtrée et amplifiée contenue dans l’interface eau douce–eau salée

Calibración de modelos de aguas subterráneas en zonas costeras utilizando información hidráulica filtrada y amplificada en la interfaz agua dulce–agua salada

利用过滤和放大后的咸-淡水界面保留的水力信息校正海岸带地下水模型

Calibração de modelo de águas subterrâneas costeiras utilizando informação hidráulica filtrada e ampliada retida na interface água doce–salgada

  • Paper
  • Published:
Hydrogeology Journal Aims and scope Submit manuscript

Abstract

Coastal groundwater flow is driven by an interplay between terrestrial and marine forcings. One of the distinguishing features in these settings is the formation of a freshwater lens due to the density difference between fresh and saline groundwater. The present study uses data collected on Sable Island, Canada—a remote sand island in the northwest Atlantic Ocean—to highlight the potential of exploiting freshwater lens geometry for calibration of numerical groundwater flow models in coastal settings. Three numerical three-dimensional variable-density groundwater flow models were constructed for different segments of the island, with only one model calibrated using the freshwater–saltwater interface derived from an electromagnetic geophysical survey. The other two (uncalibrated) models with the same parameterisation as the calibrated model successfully reproduced the interpreted interface depth and location of freshwater ponds at different parts of the island. The successful numerical model calibration, based solely on the geophysically derived interface depth, is enabled by the interface acting as an amplified version of the water table, which reduces the relative impact of the interpreted depth uncertainty. Furthermore, the freshwater–saltwater interface is far more inertial than the water table, making it less sensitive to short-term forcings. Such “noise-filtering” behaviour enables the use of the freshwater–saltwater interface for calibration even in dynamic settings where selection of representative groundwater heads is challenging. The completed models provide insights into island freshwater lens behaviour and highlight the role of periodic beach inundation and wave overheight in driving short-term water-table variability, despite their limited impact on the interface depth.

Résumé

L’écoulement des eaux souterraines côtières est déterminé par une interaction entre les forçages terrestres et maritimes. L’une des caractéristiques distinctives de ces configurations est la formation d’une lentille d’eau douce provoquée par la différence de densité entre eau douce et eau salée souterraine. La présente étude utilise les données recueillies sur l’Ile de Sable au Canada—une île sableuse isolée du Nord-Ouest de l’Océan Atlantique—dans le but de démontrer la possibilité d’utiliser la géométrie de la lentille d’eau douce pour le calage des modèles numériques d’écoulement souterrain dans les milieux côtiers. Trois modèles numériques d’écoulement souterrain tridimensionnels, avec densité différentielle des eaux souterraines, ont été construits pour différents secteurs de l’île, un seul modèle calé utilisant l’interface eau douce–eau salée déduite d’un levé de géophysique électromagnétique. Les deux autres modèles (non calés) soumis à la même paramétrisation que le modèle calé ont reproduit avec succès dans différentes parties de l’île la profondeur de l’interface et la localisation des lentilles d’eau douce interprétées. La réussite du calage du modèle numérique, basé uniquement sur la profondeur de l’interface déduite de la géophysique, est rendue possible par l’interface, qui se comporte comme une version amplifiée de la surface piézométrique, ce qui réduit l’impact relatif de l’incertitude sur la profondeur interprétée. De plus, l’interface eau douce–eau salée est beaucoup plus inertielle que la surface piézométrique, ce qui la rend moins sensible aux forçages de court-terme. Un tel comportement de “filtrage du bruit” rend possible l’utilisation de l’interface eau douce–eau salée pour le calage, même dans les situations dynamiques où le choix de la charge représentative des eaux souterraines est difficile. Les modèles réalisés donnent un éclairage du comportement de la lentille d’eau douce de l’île et mettent en évidence le rôle de l’inondation périodique de la plage et de la très forte hauteur de la vague dans la détermination de la variation de la surface piézométrique sur le court terme, malgré leur impact limité sur la profondeur de l’interface.

Resumen

El flujo de las aguas subterráneas en las zonas costeras es resultado de la interacción de factores terrestres y marinos. Uno de los rasgos distintivos en estos ambientes es la formación de una lente de agua dulce debido a la diferencia de densidad entre las aguas subterráneas dulces y salinas. El presente estudio utiliza datos obtenidos en la isla de Sable (Canadá), una remota isla de arena situada en el noroeste del Océano Atlántico, para resaltar el potencial de la explotación de la geometría de la lente de agua dulce para la calibración de modelos numéricos de flujo de agua subterránea en ambientes costeros. Se construyeron tres modelos numéricos tridimensionales de flujo de agua subterránea de densidad variable para diferentes segmentos de la isla, con sólo un modelo calibrado utilizando la interfaz agua dulce–agua salada derivada de un estudio geofísico electromagnético. Los otros dos modelos (no calibrados), con la misma parametrización que el modelo calibrado, reprodujeron con éxito la profundidad de la interfaz interpretada y la ubicación de los depósitos de agua dulce en diferentes partes de la isla. El modelo numérico calibrado con acierto, basado únicamente en la profundidad de la interfaz derivada geofísicamente, es posible gracias a que la interfaz actúa como una versión amplificada del nivel freático, lo que reduce el impacto relativo de la incertidumbre de la profundidad interpretada. Además, la interfaz agua dulce–agua salada es mucho más inercial que el nivel freático, lo que la hace menos sensible a los factores de influencia a corto plazo. Este comportamiento de “filtrado de ruido” permite el uso de la interfaz agua dulce–agua salada para la calibración incluso en entornos dinámicos en los que la selección de cargas de agua subterránea representativas es un desafío. Los modelos completados proporcionan información sobre el comportamiento de la lente de agua dulce de la isla y ponen de relieve el papel de las inundaciones periódicas de las playas y la sobrecarga de las olas en el impulso de la variabilidad de la capa freática a corto plazo, a pesar de su limitado impacto en la profundidad de la interfaz.

摘要

海岸带地下水流受陆地和海洋作用力的相互作用所驱动。在此背景下, 其最显著的特征之一就是由于咸淡水间的密度差异形成的淡水透镜体。本研究使用了在加拿大塞布尔岛(大西洋西北部一个偏远的沙岛)采集的数据, 以突出淡水透镜体的几何形态在海岸带环境下地下水流数值模型修正中的利用潜力。针对该岛的不同部分建立了三个三维可变密度的地下水流数值模型, 其中只有一个模型采用电磁地球物理法确定的咸淡水界面来修正。采用该修正模型对其他两个(未经修正的)模型进行相同的参数化, 这两个未经修正的模型成功地重现了该岛不同部位解译的“淡水体”界面深度和位置。仅基于地球物理方法获得的界面深度, 可以通过将延伸的潜水面作为界面成功修正数值模型。该方法可以降低解译深度不确定性导致的相对影响。此外, 咸淡水界面比潜水面受惯性影响更大, 使其对短期营力作用的响应更为灵敏。这样的“干扰过滤”使得在变化环境下利用咸淡水界面修正代表性地下水水头具有一定挑战性。完整的模型有利于深入了解岛屿淡水透镜体特性, 并突出了周期性的海滩淹没和海浪过高在短期水位变化中的作用(尽管其对界面深度的影响有限)。

Resumo

O fluxo da água subterrânea costeira é orientado pela interação de forças terrestres e marinhas. Um dos fatores que distingue essas características é a formação de lentes de água doce devido a diferença de densidade entre águas subterrâneas doces e salinas. O presente estudo utiliza dados coletados na Ilha Sable, Canada—uma remota ilha de areia a noroeste do Oceano Atlântico—para destacar o potencial em explorar a geometria das lentes de água doce na calibração de modelos numéricos de fluxo das águas subterrâneas em cenários costeiros. Três modelos de fluxo das águas subterrâneas numéricos, tridimensional e de densidade variável, foram construídos para diferentes segmentos da ilha, com apenas um modelo calibrado utilizando a interface água doce–salgada derivada de um levantamento geofísico eletromagnético. Os outros dois modelos (descalibrados) com a mesma parametrização que o modelo calibrado reproduziram com sucesso a interpretação da profundidade da interface e localização de lagos de água doce em diferentes porções da ilha. A calibração do modelo numérico bem-sucedida, baseada unicamente na profundidade da interface derivada geofisicamente, é possível pela atuação da interface como uma versão amplificado do nível d’água, que reduz o impacto relativo da incerteza da profundidade interpretada. Além disso, a interface água doce–salgada varia mais por inércia do que o nível d’água, tornando-a menos sensível a forças de curto prazo. Este comportamento como “filtro de ruídos” permite o uso da interface água doce–salgada para calibração inclusive de configurações dinâmicas onde a seleção de cargas hidráulica representativas das águas subterrâneas é desafiadora. O modelo final fornece perspectivas no comportamento das lentes de água doce na ilha e destaca o papel periódico da inundação da praia e o das ondas excessivamente altas em influenciar a variabilidade do nível d’água a curto prazo, apesar do impacto limitado na profundidade da interface.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11

Similar content being viewed by others

References

Download references

Acknowledgements

Dr. Dan Kehler (Parks Canada Agency) and Mr. Terry Hennigar are particularly thanked for their invaluable technical support and local knowledge and experience relating to Sable Island. We also thank two anonymous reviewers for their comments, which helped to improve the readability of the manuscript.

Funding

This research was undertaken thanks in large part to funding from the Canada First Research Excellence Fund through the Ocean Frontier Institute (postdoctoral fellowship to I. Pavlovskii) and financial contributions and logistical support from Parks Canada Agency. Field campaigns were supported by a MEOPAR early-career award/grant to B. Kurylyk and scholarship/grant support to J. Cantelon from NSERC, Killam, the Geological Society of America, and the Canadian Water Resources Association Dillon Scholarship. B. Kurylyk is supported through the Canada Research Chairs Program.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Igor Pavlovskii.

Ethics declarations

Conflicts of interest

Authors declare no conflicts of interest.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

ESM 1

(DOCX 21950 kb)

Appendix

Appendix

A separate set of zone 1 Monte-Carlo model runs was performed to evaluate the impact of anisotropy on calibration. The zone 1 model was run 100 times with randomly generated horizontal and vertical hydraulic conductivity values from 1 × 10–4 to 1 × 10–3 m/s. The vertical and horizontal hydraulic conductivity values were generated independently of each other resulting in an anisotropic model. The narrow hydraulic conductivity range was selected based on the results of the isotropic Monte-Carlo runs showing the objective function minimum at a hydraulic conductivity value of ~3.1 × 10–4 m/s (Fig. 6). These runs used the same boundary conditions, objective function, and initial conditions as zone 1 Monte-Carlo runs described in section ‘Model initial conditions, calibration, and evaluation’.

The objective function minimum in the anisotropic runs was achieved for the 2.5 × 10–4 to 4 × 10–4 m/s range of horizontal hydraulic conductivity values, which is in the vicinity of the 3.1 × 10–4 m/s value obtained through isotropic calibration (Fig. 12). The corresponding range of vertical hydraulic conductivity values is wider but still reasonably well constrained between 3 × 10–4 and 9 × 10–4 m/s. Similarly, low objective function values can be obtained for different combinations of vertical and horizontal conductivity values. These combinations follow an inverse relationship: lower values of horizontal hydraulic conductivity require higher values of vertical hydraulic conductivity to maintain the same objective function value.

Fig. 12
figure 12

Results of the Monte-Carlo calibration (colours indicate objective function values) based on the freshwater–saltwater interface depth along the Main Station transect (zone 1) with enabled anisotropy in hydraulic conductivity. The objective function is a sum of squared errors. Results for the isotropic runs are from Fig. 6

The Monte Carlo runs highlight the nonuniqueness of the anisotropic calibration which likely arises from the nonnegligible vertical flow component for a freshwater lens in a thick permeable aquifer. Such vertical flows can invalidate commonly used analytical solutions (e.g., Fetter 1972) for freshwater lens geometry. At the same time, the flow remains predominantly horizontal as indicated by the steep gradient between optimal vertical and horizontal conductivity values (Fig. 12). That is, any change in horizontal hydraulic conductivity requires a larger change in the vertical hydraulic conductivity in the opposite direction to maintain a similar lens thickness.

A trend towards lower objective function values for cases where vertical hydraulic conductivity is larger than horizontal (Fig. 12) minimises the effect of nonuniqueness on the calibration results used in the main body of the present paper. Such a combination of parameters is unlikely given the sediment lithology and depositional setting at the study site, making the isotropic case a reasonable approximation.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Pavlovskii, I., Cantelon, J.A. & Kurylyk, B.L. Coastal groundwater model calibration using filtered and amplified hydraulic information retained in the freshwater–saltwater interface. Hydrogeol J 30, 1551–1567 (2022). https://doi.org/10.1007/s10040-022-02510-8

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10040-022-02510-8

Keywords

Navigation