Document Type : Research Article
Authors
Department of Geography, Faculty of Social Sciences, University of Mohaghegh Ardabili, Ardabil, Iran
Abstract
Understanding the spatial variation in river water quality and the influencing factors is essential, given the growing demand for safe water in drinking, agricultural, and industrial sectors. This study employed a systematic approach to examine the interactions between watershed characteristics and the spatial distribution of water quality parameters in the Talesh region. The research utilized geological maps, a digital elevation model (DEM), and water quality data from 12 hydrometric stations. To assess the interactions, correlation and regression analyses were conducted between geomorphological variables and water quality indicators, specifically total dissolved solids (TDS) and electrical conductivity (EC). The correlation analysis revealed significant associations between eight geomorphological variables including watershed area, mean elevation, average slope, main river length, total stream length, bifurcation ratio, drainage density, and the Gravelius coefficient and the two water quality variables. Correlation coefficients ranged from 0.60 to 0.88, indicating strong relationships. Most associations were direct. Notably, the length of the main river exhibited the strongest correlation with both TDS and EC, suggesting that it is the most influential hydro-geomorphic factor in explaining spatial water quality variability in the Talesh catchments. Regression analysis further demonstrated that reliable predictive models for spatial variation in water quality can be developed using geomorphological variables. These models accounted for up to 98% of the variance in water quality parameters (TDS and EC), underscoring the critical role of watershed characteristics in controlling river water quality.
Extended Abstract
Introduction
Knowledge of the spatial variation in river water quality and the influencing factors is crucial due to the urgent demand for safe water for various uses, including drinking, agriculture, and industry. Recently, river water quality has deteriorated as a result of intensified human activity in densely populated areas. The initial step in the sustainable management of river water quality is to identify the fundamental factors that influence its spatial variation. Among the environmental variables, geomorphological factors play a significant role. The geomorphic characteristics of a watershed influence the transformation of precipitation into runoff and the transport and accumulation of solids and dissolved substances in rivers. The topic of river water quality as an outcome of the watershed system is central to geomorphology, particularly in the context of preserving and restoring river ecosystems. Analyzing the relationships between geology, hydrology, and geomorphology and their connection to river water quality through terrain analysis offers an opportunity to elucidate the specific capacity of geomorphology in modeling hazardous environmental changes. This study aimed to investigate the relationship between catchment characteristics and water quality components to develop predictive models of spatial variations in water quality within the Talesh catchments, which are experiencing increasing water pollution due to the expansion of agriculture and livestock activities, coupled with unstable hydro-geomorphic conditions.
Material and Methods
This study was based on correlation and regression analyses between geomorphometry variables and water quality parameters. The study area consists of 12 mountainous catchments with a total area of 1993 km² in northwestern Iran. The data comprised monthly water quality records from hydrometric stations and an Aster Digital Elevation Model (DEM). Analytical tools included ArcGIS Desktop 10.8, SPSS 19.0, SAGA 7.0, and Excel 2016. The research involved several steps:
First, average values for total dissolved solids (TDS) and electrical conductivity (EC) were calculated for 12 stations over a 10-year period (2011–2021). Next, 15 geomorphometric variables were derived for each catchment using the 30-meter DEM within GIS and SAGA environments. These variables included: area, mean altitude, mean slope, mean main stream slope, main stream length, total stream length, bifurcation ratio, drainage density, Gravelius coefficient, elongation ratio, concentration time, mean LS factor, mean ruggedness index, mean terrain convexity index, and the percentage of erodible rocks. After compiling the values for the two dependent variables and 15 independent variables, a correlation matrix was used to determine statistically significant relationships (p ≤ 0.05). To prepare predictive models for TDS and EC, both the Enter and Stepwise regression methods were employed. The Enter method was used to construct initial models, while the Stepwise method was applied to resolve multicollinearity issues and determine the most influential variables.
Results and Discussion
The correlation analysis identified eight geomorphometric variables catchment area, mean elevation, mean slope, main river length, total stream length, drainage density, concentration time, and ruggedness index that were significantly correlated with TDS and EC. Correlation coefficients ranged from 0.60 to 0.88. Moreover, most relationships were direct, indicating that increases in geomorphometric values were generally associated with increased water quality variables. The strongest correlation was observed between main river length and both TDS and EC, indicating that this variable was the most important hydro-geomorphic factor influencing the spatial variation in water quality across the Talesh catchments. Longer river lengths appeared to degrade water quality, possibly due to greater exposure to diverse land use types, increased pollutant input, and enhanced erosion processes facilitated by the steep gradients typical of the region.
The regression analyses confirmed that reliable predictive models could be developed using geomorphometry variables. The Enter method explained up to 98% of the variance in TDS and EC; however, these models suffered from multicollinearity. This issue was resolved in the Stepwise models, which yielded two effective predictive models. The first model, using only main river length (MRL), explained 78% of the variance in TDS and EC. The second model, incorporating both MRL and slope (S), explained 88% of the variance. A comparison of the regression model statistics revealed that the Stepwise models had higher F-values (31–35) and lower significance levels (p = 0.000) compared to the Enter models (F = 10, p = 0.02), indicating their greater validity and predictive accuracy.
Conclusion
Analyzing the spatial variation in river water quality without a systematic watershed-based approach results in incomplete and potentially misleading interpretations. Because watershed characteristics interact with one another and with land use in complex ways, understanding their effects on water quality is essential for effective forecasting and management. This study demonstrated that terrain and regression analyses can be used to construct reliable predictive models of water quality, which can guide planners and resource managers in addressing urban, rural, and industrial water demands. Correlation results also help identify vulnerable regions by evaluating the strength and direction of relationships between geomorphometry variables and water quality indicators (TDS and EC), thereby aiding in the prioritization of watershed management, conservation, and restoration efforts.
The results suggest that large, high-elevation catchments with long main rivers, low slopes, and high drainage density and concentration time are more susceptible to degraded water quality, primarily due to increased TDS and EC. Such catchments require focused conservation measures.
While both regression methods provided useful models, the Stepwise method was more effective in identifying the key predictive variables, particularly main river length and slope, which were consistent with previous research. Conversely, the lithology variable was not a significant predictor in this study unlike in earlier studies due to the low proportion (13%) of erosion-prone formations in the catchments and limitations in the definition of this variable. For future work, we recommend incorporating the percentage area of each geological formation as a more refined approach to modeling the influence of lithology on water quality.
Keywords
Main Subjects
)
Send comment about this article