Mapping Coastal Groundwater Quality with Ordinary Kriging: A Spatial Modelling Approach for Sustainable Water Resource Management

Protecting groundwater quality in coastal regions has become an increasingly important challenge as urban expansion, groundwater abstraction, and seawater intrusion continue to threaten freshwater resources. Reliable spatial information is essential for identifying vulnerable areas and supporting evidence-based environmental management. In this study, the authors investigate the spatial distribution of shallow groundwater quality along the coastal corridor of Padang, Indonesia, using Ordinary Kriging interpolation. Four operational water-quality parameters—potential of hydrogen (pH), electrical conductivity (EC), total dissolved solids (TDS), and salinity—were analysed through variogram modelling, anisotropy testing, and leave-one-out cross-validation. The resulting groundwater quality maps provide practical decision-support tools for monitoring priorities, groundwater protection, and sustainable coastal management while contributing to the achievement of Sustainable Development Goals (SDGs) 6 and 11.

Bibliographic Information

Item	Information
Article Title	Spatial modelling of shallow groundwater quality in coastal areas with Kriging interpolation
Authors	Yaumal Arbi; Nurhasan Syah; Iswandi Umar; Indang Dewata; Mulya Gusman; Nevy Sandra
Journal	Teknomekanik
Volume	9
Issue	1
Publication Date	February 2026
Pages	60–76
DOI	https://doi.org/10.24036/teknomekanik.v9i1.47272
Publisher	Universitas Negeri Padang
License	Creative Commons Attribution 4.0 International (CC BY 4.0)
e-ISSN	2621-8720
p-ISSN	2621-9980
Keywords	coastal groundwater; seawater intrusion; ordinary kriging; clean water and sanitation; sustainable cities and communities

Research Background

Shallow groundwater serves as one of the primary freshwater resources for coastal communities. However, increasing groundwater extraction, rapid urbanization, and climate-related changes have intensified the risk of seawater intrusion, gradually degrading groundwater quality in many coastal environments. These pressures make continuous monitoring and accurate spatial assessment essential for protecting freshwater resources and supporting sustainable urban development.

Previous studies in Padang and other Indonesian coastal regions have commonly evaluated seawater intrusion using indicators such as electrical conductivity, salinity, and total dissolved solids. While these investigations have successfully identified areas affected by salinization, many relied primarily on descriptive analyses or deterministic interpolation methods that do not explicitly represent the spatial dependence inherent in groundwater quality observations.

Recent advances in geostatistics have demonstrated that kriging-based interpolation generally provides more reliable spatial predictions than deterministic techniques when spatial autocorrelation is appropriately modelled. Nevertheless, many applied groundwater studies provide limited discussion of variogram development, anisotropy assessment, or model validation, reducing confidence in the resulting spatial prediction maps.

Recognizing these limitations, the present study develops a comprehensive spatial modelling framework for shallow coastal groundwater quality in Padang City, Indonesia. The research integrates empirical variogram analysis, directional anisotropy testing, Ordinary Kriging interpolation, and leave-one-out cross-validation to generate reliable groundwater quality maps. The study further translates these prediction surfaces into three groundwater quality zones that can assist environmental monitoring, groundwater protection, and coastal resource management.

Research Objective

To map the spatial distribution of shallow coastal groundwater quality in Padang using four operational parameters: pH, electrical conductivity (EC), total dissolved solids (TDS), and salinity.
To develop empirical variogram models for each groundwater quality parameter and evaluate their spatial dependence.
To examine directional anisotropy so that alongshore and cross-shore spatial continuity can be accurately represented.
To generate groundwater quality prediction surfaces using the Ordinary Kriging interpolation method.
To evaluate prediction reliability through leave-one-out cross-validation (LOOCV).
To establish a three-level groundwater quality zoning system that distinguishes lower-risk, transitional, and higher-risk coastal areas associated with seawater intrusion.
To provide spatial information that supports groundwater monitoring, well protection, and evidence-based coastal environmental management.

Why This Research Matters

Supports sustainable groundwater management. Accurate groundwater quality maps help authorities identify vulnerable coastal areas before water quality deteriorates further.
Improves environmental monitoring. Spatial interpolation enables monitoring efforts to focus on locations with the greatest potential risk of seawater intrusion.
Strengthens scientific reliability. By combining variogram modelling, anisotropy analysis, and cross-validation, the study provides a robust geostatistical framework for groundwater mapping.
Provides practical decision-support tools. The three-level groundwater quality zoning offers a clear basis for prioritizing monitoring locations and protecting vulnerable wells.
Advances coastal hydrogeological studies. The research demonstrates how Ordinary Kriging can effectively represent spatial variability in multiple groundwater quality indicators within complex coastal aquifer systems.
Contributes to sustainable development. The findings directly support SDG 6 (Clean Water and Sanitation) through improved groundwater monitoring and SDG 11 (Sustainable Cities and Communities) by strengthening evidence-based coastal environmental planning.
Offers a transferable methodology. Although conducted in Padang, the proposed spatial modelling approach can be adapted to other coastal regions facing similar groundwater quality challenges.

Research Methodology

This study employed a quantitative geostatistical approach to model the spatial distribution of shallow groundwater quality along the coastal corridor of Padang, West Sumatra, Indonesia. The methodology integrates field measurements, variogram modelling, anisotropy analysis, Ordinary Kriging interpolation, and statistical validation to generate reliable groundwater quality maps suitable for environmental monitoring and decision-making.

Study Area

The research was conducted along the coastal zone of Padang City, where shallow unconfined aquifers extend parallel to the Indian Ocean. The study area consists primarily of coastal alluvial deposits and includes residential settlements, commercial districts, estuarine environments, and transportation corridors. Observation points were distributed from the northern to the southern coastal segments and extended inland to capture the transition between coastal and inland groundwater conditions.

Groundwater Sampling

Groundwater samples were collected between 2 September and 20 October 2025 from 207 shallow wells distributed throughout the study area. Each well was sampled once, and four operational groundwater quality parameters were measured directly in the field:

Potential of Hydrogen (pH)
Electrical Conductivity (EC)
Total Dissolved Solids (TDS)
Salinity

Site observations regarding proximity to coastlines and river mouths were also recorded to support interpretation of seawater intrusion patterns.

Field Instrumentation and Quality Control

Groundwater pH was measured using a Mettler Toledo SevenCompact S220 pH meter, while EC, TDS, salinity, and water temperature were measured using an AZ Instrument 86031 Water Quality Meter equipped with automatic temperature compensation. Instrument calibration followed standard operating procedures using certified calibration solutions before field measurements. Repeated measurements were performed whenever readings had not stabilized, ensuring data consistency and reliability.

Geostatistical Analysis

Spatial dependence among groundwater observations was analysed using empirical semivariograms computed with the Matheron estimator. Coordinates were projected into UTM Zone 47S (WGS 84), allowing separation distances to be calculated in metres. The analysis employed twelve lag classes with a cutoff distance corresponding to the 90th percentile of all inter-point distances.

Three theoretical semivariogram models were evaluated for each groundwater quality parameter:

Spherical
Exponential
Gaussian

The optimal model was selected based on the lowest Root Mean Square Error (RMSE) between the empirical and theoretical semivariograms.

Anisotropy Analysis

Directional semivariograms were generated at azimuths of 0°, 45°, 90°, and 135° with an angular tolerance of ±22.5°. This analysis examined whether groundwater quality exhibited different spatial continuity parallel and perpendicular to the coastline, allowing anisotropic spatial structures to be incorporated into the interpolation model.

Ordinary Kriging Interpolation

Ordinary Kriging was used as the primary interpolation technique to estimate groundwater quality at unsampled locations. Prediction surfaces were generated using a fixed spatial grid with sector-based neighbourhood searching to improve interpolation stability under anisotropic conditions. Search radii were adjusted according to the practical range obtained from each fitted semivariogram model.

The resulting interpolation maps describe the spatial distribution of pH, EC, TDS, and salinity across the entire coastal area, enabling continuous groundwater quality assessment rather than isolated point observations.

Model Validation

Prediction performance was evaluated using Leave-One-Out Cross-Validation (LOOCV). Each observation was removed sequentially and predicted using the remaining observations. Model performance was assessed using Mean Error (ME), Root Mean Square Error (RMSE), Root Mean Square Standardized Error (RMSSE), and the coefficient of determination (R²). This validation procedure ensured that the selected variogram models produced reliable spatial predictions suitable for groundwater quality mapping.

Key Findings

Groundwater Quality Exhibited Clear Spatial Variability

The analysis demonstrated substantial spatial variation in groundwater quality across the Padang coastal corridor. While groundwater pH remained relatively stable within neutral to slightly alkaline conditions, electrical conductivity, total dissolved solids, and salinity displayed much greater variability, reflecting differences in seawater influence and hydrogeological conditions.

Among the four parameters, salinity showed the highest relative variability, indicating that seawater intrusion does not affect the coastal aquifer uniformly but instead forms localized zones of elevated salinity.

Ordinary Kriging Produced Reliable Spatial Predictions

Leave-One-Out Cross-Validation confirmed the robustness of the Ordinary Kriging models. Mean prediction errors were close to zero for all groundwater quality parameters, while coefficients of determination approached one, demonstrating strong agreement between observed and predicted values. These results indicate that the selected variogram models accurately represented the spatial dependence within the groundwater dataset.

Distinct Spatial Structures Were Identified

Variogram analysis revealed that each groundwater quality parameter possesses a unique spatial structure. The optimal semivariogram model differed among parameters, with Spherical, Gaussian, and Exponential models providing the best fit depending on the variable analysed. The estimated ranges indicate that groundwater quality exhibits spatial continuity over distances of several kilometres, although salinity showed a considerably shorter spatial range than the other parameters.

Strong Directional Anisotropy Was Observed

Directional variogram analysis identified pronounced anisotropy, particularly for electrical conductivity, total dissolved solids, and salinity. Spatial continuity was generally stronger along the coastline than across it, suggesting that coastal hydrogeological processes exert a dominant influence on groundwater quality distribution. Incorporating anisotropy into the interpolation process therefore improved the realism of the resulting prediction maps.

Three Groundwater Quality Zones Were Successfully Delineated

The kriging prediction maps enabled groundwater quality to be classified into three spatial zones representing lower-risk, transitional, and higher-risk conditions. Northern coastal areas generally exhibited better groundwater quality, whereas southern coastal segments showed stronger indications of seawater intrusion. Transitional conditions occurred between these two zones, reflecting gradual changes in groundwater chemistry rather than abrupt boundaries.

The Results Support Evidence-Based Groundwater Monitoring

The spatial prediction maps provide a practical framework for identifying monitoring priorities and protecting vulnerable groundwater resources. Instead of relying solely on discrete sampling locations, the interpolated maps allow environmental managers to visualize groundwater quality continuously across the coastal landscape, facilitating more effective planning and resource allocation.

Scientific Contribution

Introduces a comprehensive geostatistical framework that integrates empirical variogram modelling, anisotropy assessment, Ordinary Kriging interpolation, and statistical validation into a unified groundwater mapping workflow.
Demonstrates the importance of anisotropy analysis for accurately representing groundwater quality variation in coastal aquifer systems influenced by shoreline orientation.
Provides validated spatial prediction models supported by rigorous Leave-One-Out Cross-Validation, increasing confidence in the resulting groundwater quality maps.
Produces practical groundwater quality zoning that translates complex geostatistical outputs into information readily applicable for environmental management.
Strengthens coastal hydrogeological research by showing that multiple groundwater quality indicators can be modelled simultaneously using a consistent geostatistical methodology.
Offers a transferable methodological framework that can be adapted for groundwater quality assessment in other coastal regions experiencing similar environmental pressures.

Industrial Implications

Supports groundwater resource management by providing spatial information that enables more efficient monitoring programmes.
Assists local governments and environmental agencies in identifying groundwater sources requiring priority protection against seawater intrusion.
Improves urban water supply planning through accurate identification of freshwater and vulnerable groundwater zones.
Enhances environmental risk assessment by revealing spatial patterns that cannot be detected using isolated sampling points alone.
Provides a scientific basis for sustainable coastal development by integrating groundwater quality into environmental planning.
Supports GIS-based environmental decision-support systems through the production of continuous groundwater quality prediction maps.
Contributes to long-term groundwater conservation by enabling proactive monitoring before groundwater quality deterioration becomes widespread.

Research Limitations

Single geographical case study.
The research was conducted exclusively along the coastal corridor of Padang City, West Sumatra. Although the study area represents an important coastal aquifer system, the resulting spatial models should be interpreted within the environmental and hydrogeological conditions of the study location. Additional studies in other coastal environments are necessary to evaluate the broader applicability of the proposed methodology.
Limited groundwater quality parameters.
The spatial modelling focused on four operational groundwater quality indicators: pH, electrical conductivity (EC), total dissolved solids (TDS), and salinity. Other hydrochemical parameters, including major ions, trace metals, nutrients, or microbiological indicators, were outside the scope of the present investigation.
Single sampling campaign.
Groundwater samples were collected during a single observation period between September and October 2025. Consequently, the resulting maps represent groundwater conditions during that sampling period and do not capture possible seasonal variations associated with rainfall, groundwater abstraction, or tidal fluctuations.
Dependence on geostatistical assumptions.
The Ordinary Kriging approach assumes local stationarity and relies on the selected semivariogram model to represent spatial dependence. Although the selected models demonstrated excellent validation performance, prediction accuracy remains dependent on the validity of these statistical assumptions.
Interpolation uncertainty.
Prediction accuracy generally decreases in locations with fewer neighbouring observations. Although cross-validation indicated strong overall model performance, uncertainty may still increase in sparsely sampled areas or locations exhibiting highly localized groundwater variability.

Future Research Opportunities

Conduct multi-season groundwater monitoring to evaluate temporal changes associated with rainfall variability, tidal influence, and groundwater abstraction.
Expand groundwater quality assessment by incorporating additional hydrochemical parameters, including major ions, heavy metals, nutrients, and microbiological indicators.
Compare Ordinary Kriging with alternative spatial interpolation techniques such as Universal Kriging, Co-Kriging, Regression Kriging, Empirical Bayesian Kriging, Inverse Distance Weighting (IDW), and machine learning-based spatial prediction methods.
Develop three-dimensional groundwater quality models that integrate aquifer depth and hydrogeological characteristics.
Combine geostatistical modelling with remote sensing, Geographic Information Systems (GIS), and Internet of Things (IoT) monitoring technologies for near real-time groundwater assessment.
Investigate long-term groundwater quality changes under climate change scenarios, urban expansion, and increasing groundwater extraction.
Integrate groundwater quality maps into coastal environmental risk assessment and sustainable urban planning frameworks.
Evaluate the economic implications of groundwater degradation to support more effective environmental management and infrastructure investment.
Apply the proposed geostatistical workflow to other Indonesian coastal cities and international coastal aquifer systems to assess methodological transferability.

Potential for Public Policy Citation

This research demonstrates considerable potential for citation in public policy documents because it directly addresses one of the most important environmental challenges faced by rapidly developing coastal cities: protecting groundwater resources from seawater intrusion. The study provides scientifically validated spatial information that can support groundwater monitoring programmes, environmental protection strategies, and sustainable urban water resource management.

The groundwater quality zoning generated through Ordinary Kriging offers practical evidence that may assist local governments, environmental agencies, water utilities, and regional planners in prioritizing monitoring activities, protecting vulnerable groundwater sources, and developing adaptive coastal management policies. Because the study explicitly aligns its outcomes with Sustainable Development Goal (SDG) 6 (Clean Water and Sanitation) and SDG 11 (Sustainable Cities and Communities), the findings also have relevance for sustainability reporting and environmental governance.

Who Should Read This Paper?

Hydrogeologists investigating coastal aquifer systems and groundwater quality.
Environmental engineers working on groundwater protection and water resource management.
Researchers specializing in geostatistics, spatial modelling, and environmental GIS.
Government agencies responsible for groundwater monitoring and environmental conservation.
Urban planners involved in sustainable coastal development.
Water utilities responsible for protecting freshwater supplies in coastal regions.
Graduate students and academics studying hydrogeology, environmental science, geography, civil engineering, and spatial analysis.
Decision-makers interested in evidence-based groundwater management and climate adaptation.

Final Thoughts

This study presents a rigorous application of geostatistical techniques for mapping shallow groundwater quality within a coastal environment. By integrating empirical variogram analysis, directional anisotropy assessment, Ordinary Kriging interpolation, and comprehensive cross-validation, the research establishes a reliable framework for representing the spatial variability of groundwater quality indicators.

Beyond its methodological contribution, the study demonstrates how scientifically validated spatial models can support practical environmental decision-making. The resulting groundwater quality maps provide an effective basis for identifying vulnerable coastal zones, prioritizing monitoring efforts, and strengthening groundwater protection strategies against seawater intrusion.

The research also illustrates the value of combining field observations with modern geostatistical modelling to transform discrete groundwater measurements into continuous spatial information that is readily interpretable by environmental managers and policy makers. As coastal cities continue to face increasing pressure from urbanization, groundwater abstraction, and climate change, methodologies such as those presented in this paper will become increasingly important for supporting sustainable groundwater management and protecting freshwater resources for future generations.

Suggested Citations

Teknomekanik (UNP) Style

Arbi Y, Syah N, Umar I, Dewata I, Gusman M, Sandra N. Spatial modelling of shallow groundwater quality in coastal areas with Kriging interpolation. Teknomekanik. 2026;9(1):60–76. https://doi.org/10.24036/teknomekanik.v9i1.47272

APA (7th Edition)

Arbi, Y., Syah, N., Umar, I., Dewata, I., Gusman, M., & Sandra, N. (2026). Spatial modelling of shallow groundwater quality in coastal areas with Kriging interpolation. Teknomekanik, 9(1), 60–76. https://doi.org/10.24036/teknomekanik.v9i1.47272

IEEE Style

Y. Arbi, N. Syah, I. Umar, I. Dewata, M. Gusman, and N. Sandra, "Spatial modelling of shallow groundwater quality in coastal areas with Kriging interpolation," Teknomekanik, vol. 9, no. 1, pp. 60–76, Feb. 2026, doi:10.24036/teknomekanik.v9i1.47272.

Harvard Style

Arbi, Y., Syah, N., Umar, I., Dewata, I., Gusman, M. and Sandra, N. (2026) 'Spatial modelling of shallow groundwater quality in coastal areas with Kriging interpolation', Teknomekanik, 9(1), pp. 60–76. doi:10.24036/teknomekanik.v9i1.47272.

Vancouver Style

Arbi Y, Syah N, Umar I, Dewata I, Gusman M, Sandra N. Spatial modelling of shallow groundwater quality in coastal areas with Kriging interpolation. Teknomekanik. 2026;9(1):60-76. doi:10.24036/teknomekanik.v9i1.47272.

Chicago (Author–Date)

Arbi, Yaumal, Nurhasan Syah, Iswandi Umar, Indang Dewata, Mulya Gusman, and Nevy Sandra. 2026. "Spatial modelling of shallow groundwater quality in coastal areas with Kriging interpolation." Teknomekanik 9 (1): 60–76. https://doi.org/10.24036/teknomekanik.v9i1.47272.

MLA (9th Edition)

Arbi, Yaumal, et al. "Spatial modelling of shallow groundwater quality in coastal areas with Kriging interpolation." Teknomekanik, vol. 9, no. 1, 2026, pp. 60–76. Crossref, https://doi.org/10.24036/teknomekanik.v9i1.47272.

Editorial Note

This article review is part of the Engineering Research Insights series, which highlights recent engineering research with potential scientific, industrial, and societal impact. The review is an independent scholarly interpretation prepared for educational and knowledge dissemination purposes. Bibliographic metadata have been verified using the official journal webpage, while the scientific discussion is based exclusively on the published research article. Readers are encouraged to consult the original publication for complete methodological details, experimental procedures, datasets, and supplementary information.

SEO Meta Description

Discover how Ordinary Kriging spatial modelling maps shallow groundwater quality in Padang's coastal aquifer. Learn how variogram analysis, anisotropy testing, and geostatistics support groundwater monitoring, seawater intrusion assessment, and sustainable coastal water management.

SEO Keywords

spatial groundwater modelling, groundwater quality mapping, Ordinary Kriging, Kriging interpolation, coastal groundwater, seawater intrusion, variogram analysis, anisotropy analysis, groundwater quality assessment, electrical conductivity, total dissolved solids, salinity mapping, pH mapping, geostatistics, environmental GIS, hydrogeology, coastal aquifer, groundwater monitoring, sustainable water management, Padang Indonesia, SDG 6, SDG 11, environmental engineering, water resources engineering

Article Summary

The reviewed study demonstrates how geostatistical modelling can transform field-based groundwater observations into reliable spatial information for environmental management. Using empirical variogram modelling, anisotropy assessment, Ordinary Kriging interpolation, and rigorous cross-validation, the researchers successfully identified groundwater quality patterns associated with seawater intrusion along the Padang coastline. The resulting groundwater quality zoning provides practical guidance for monitoring, groundwater protection, and sustainable coastal planning. As freshwater resources in coastal cities face increasing pressure from urbanization and climate change, this study illustrates the important role of spatial modelling in supporting evidence-based groundwater management.

Engineering Research Insights
Making engineering research accessible through evidence-based scholarly reviews.