Introduction
Degradation of water quality has emerged as one of the most critical challenges facing national water resources. Surging water demand has led to the over-exploitation of groundwater and a continuous decline in aquifer levels, subsequently triggering environmental degradation and a deteriorating trend in groundwater quality. Consequently, analyzing temporal trends and spatial variations of groundwater quality is vital for sustainable water resource management. This study aims to investigate the temporal trends of groundwater quality parameters using the Mann-Kendall test and Sen’s Slope Estimator, while analyzing spatial variations through the Groundwater Quality Index (GQI) in the Bahabad aquifer.

Materials and Methods
In this study, a total of 16 observation wells were initially considered for groundwater quality assessment. Among them, 9 wells with complete and reliable datasets were selected for detailed analysis. The required hydrochemical data were obtained from the Yazd Regional Water Authority and included major groundwater quality parameters: calcium (Ca²⁺), magnesium (Mg²⁺), sodium (Na⁺), chloride (Cl⁻), sulfate (SO₄²⁻), and total dissolved solids (TDS). A uniform temporal period from 2002 to 2023 was considered for all selected wells to ensure consistency in trend analysis.
To evaluate temporal variations in groundwater quality, non-parametric statistical methods, including the Mann–Kendall test and Sen’s slope estimator, were employed. These methods are widely used in hydrochemical studies due to their robustness against non-normal data distributions and their ability to detect both the direction and magnitude of trends over time.
Spatial variations in groundwater quality parameters were analyzed using a range of interpolation techniques, including both deterministic and geostatistical approaches. The applied methods consisted of Inverse Distance Weighting (IDW), Radial Basis Functions (RBF), Local Polynomial Interpolation (LPI), Global Polynomial Interpolation (GPI), and Kriging. All spatial analyses were performed in a GIS environment. The optimal interpolation method for each parameter was selected based on statistical performance indicators, primarily the lowest Root Mean Square Error (RMSE) and the highest coefficient of determination (R²).
Prior to spatial modeling, the normality of the dataset was evaluated using the Kolmogorov–Smirnov and Shapiro–Wilk tests, which are commonly applied to assess data distribution characteristics. The results of these tests were used to determine whether data transformation was required before applying interpolation and statistical analyses.
Finally, the Groundwater Quality Index (GQI) was calculated to provide an integrated assessment of groundwater quality conditions. By aggregating multiple hydrochemical parameters into a single index, GQI enables a comprehensive evaluation of groundwater quality and facilitates spatial comparison across the aquifer. This index was used as a key tool for interpreting overall water quality status and supporting groundwater management decisions.

Results and Discussion
The results indicate that while the groundwater quality of the Bahabad aquifer remained within the “Acceptable” category during the study period, a slight but statistically significant downward tendency in quality was observed. Spatial distribution analysis revealed that salinity patterns and ionic concentrations are influenced more by anthropogenic pressures such as over-pumping, water table drawdown, and inter-basin water transfer than by natural geological factors.
Temporal analysis highlighted significant spatial heterogeneity; for instance, Mg²⁺ exhibited a significant increasing trend across all stations (Sen’s slope ranging from 0.05 to 0.23), whereas Na⁺ and SO₄²⁻ showed decreasing or insignificant trends in certain wells. TDS displayed variable behavior, showing generally stable or slightly decreasing trends, while a sharp increase was observed at one specific station.
Furthermore, a comparison of interpolation techniques showed that under conditions of limited sampling density and weak spatial correlation, deterministic methods often yielded higher predictive accuracy than geostatistical models.

Conclusion
Despite only minor overall fluctuations in the GQI (ranging approximately between −0.5% and +0.1%), the index maps did not fully capture the intensity of localized quality degradation in critical parameters such as TDS, Na⁺, and SO₄²⁻. This indicates that composite indices like GQI may partially neutralize opposing trends among individual parameters, thereby providing a more generalized representation of groundwater quality conditions.
Trend analysis of individual parameters revealed considerable spatial heterogeneity across the aquifer. Magnesium and chloride exhibited consistent increasing trends in most monitoring wells, indicating progressive salinization, whereas sodium and sulfate showed mixed or insignificant trends in several locations. Total dissolved solids demonstrated spatially variable behavior, with localized sharp increases despite generally stable or declining trends elsewhere. This variability reflects the complex interaction between local hydrogeological conditions and human-induced pressures and confirms that groundwater quality evolution cannot be adequately described using a single parameter or uniform trend assumption.
The comparison of interpolation techniques further highlighted an important methodological implication: under conditions of limited sampling density and weak spatial autocorrelation, deterministic methods can provide more reliable estimates than geostatistical approaches. This emphasizes the need to adapt modeling strategies to data availability and hydrogeological settings, particularly in arid and semi-arid regions where monitoring networks are often sparse.
Although the Groundwater Quality Index proved effective in providing an overall assessment of groundwater status, it was less sensitive to localized deterioration in key parameters such as TDS, sodium, and sulfate. This confirms that composite indices, while useful for communication and management purposes, may mask emerging local-scale risks if used in isolation. Therefore, they should be applied alongside parameter-specific analyses rather than as standalone decision-making tools.
Overall, the findings highlight the urgent need for improved groundwater management in the Bahabad aquifer. Controlling excessive abstraction, regulating industrial withdrawals, and strengthening long-term monitoring programs are essential to prevent further degradation. Integrating spatial analysis with trend-based assessments can provide a more reliable framework for sustainable groundwater management, particularly in arid regions experiencing increasing water stress.