Geographically, the Akmola Region lies on the western edge of the Kazakh Uplands (Kazakh Hummocks), between the Ulytau Mountains in the southwest and the Kokshetau Hills in the north. Its terrain generally slopes from east to west, following the direction of the Ishim River valley (Akmola region, 2004; Dictionary of modern geographical names/Rus, 2006). The Ishim River traverses the central part of the region before turning sharply northward near its western border (Kotlyakov, 2006). The Akmola Region borders the Kostanay Region in the west, the North Kazakhstan Region in the north, the Pavlodar Region in the east, and the Karaganda Region in the south (State Archives of Akmola Region, 2022).

The relief of the region can be divided into three main zones:

The northwestern part: it is a flat plateau adjacent to the Ishim Valley, dissected by dry ravines and gullies, and ends in a ledge toward the valley.

Closed basins between hills, ranging from several dozen meters to several dozen kilometers in diameter, are often occupied by lakes. The extreme northeastern part of the region lies within the West Siberian Lowland. The highest point in the Akmola Region is Mount Kokshe (947 m above the sea level), whereas the lowest is Lake Sholaksor (67 m above the sea level) (Kazakhstan, 2005; Land Resources Management Committee, 2022).

The rural area of the Akmola Region was selected for this research due to its diverse landscape, which ranges from steppe to forest ecosystems. Most of the region is characterized by steppe areas, whereas certain parts lie within forested and wetland areas. This geographical diversity formed the basis for selecting three districts within the study area: Korgalzhyn, Zerendy, and Tselinograd.

Three villages were selected from each district: Ortaagash, Karabulak, and Malik Gabdullin from the Zerendy District, which are located in the north-central part of the Akmola Region; Uiyaly, Zhanteke, and Karaegin from the Korgalzhyn District, located in the south-central part of the region; and Ilyinka, Karazhar, and Taitobe from the Tselinograd District, located in the southeastern part of the region (Figure 1).

Figure 2 illustrates the nine villages selected from the three districts, along with the locations where household water samples were taken. Three villages were randomly chosen from each district. The villages are as follows: Ortaagash (V1), Karabulak (V2), Malik Gabdullin (V3), Uiyaly (V4), Zhanteke (V5), Karaegin (V6), Ilyinka (V7), Karazhar (V8), and Taitobe (V9). In each village, five households were selected for water sampling using an “envelope method.” These households were chosen to reflect potential variation in drinking water sources, as residents may rely on different supply types depending on their location within the village, including three households in Ortaagash and Uiyaly villages that have two drinking water sources. Despite the limited size, the selected households are considered representative of the local community level as they cover the typical socioeconomic and everyday characteristics of the rural population.

General information about the selected villages and surveyed households in the rural Akmola Region is presented in Table 1. The table shows data on the geographical location of each household, the center of each village, the landscape type (steppe, forest, and wetland), the primary sources of drinking water, population data by gender, and whether water purification is carried out. Purification practices are considered one of the key parameters influencing respondents’ perceptions of water quality.

Drinking water sources in households were identified as the following categories: BWS, bottled water from a store; PWIPS, public water intake pump on the street; PWWOWSH, private water well owned without supply inside the house; PWWOSH, private water well owned with supply inside the house; PB, public borehole; SW, spring water; WPS, water pumping station; CWSSIT, centralized water supply system with inside tap; PW, public well; PBWWSH, private borehole water supply to the house; PBWOWSH, private borehole without water supply to the house; PWS, purified water from the store; FTW, free tracked-in water, and other water sources that were encountered sporadically were combined into the category “other” (возможно надо убрать).

These categories encompass all types of drinking water sources used by surveyed households. The table shows the drinking water sources of households from where the water samples were taken. Among these sources, we decided to analyze the wells, boreholes, and centralized water-supply systems within the study areas.

Water samples were collected from different households with 1–11 residents, and their water consumption is 3.6 thousand liters per person per day (National Bureau of Statistics, 2024).

Nine villages distributed across different parts of the region representing different landscapes were randomly selected to sample drinking water for quality assessment during September 2024. September 2024 was characterized by high average temperatures, average air humidity, average wind speed, and the lowest pressure of the year. These weather conditions showed that this month was convenient for the research period of our object as it was neither a dry nor a rainy season compared to other years. In addition, September in the Akmola Region, as in many regions at this latitude, is the middle of the vegetation period, which is also the subject of the study with regard to the types of landscape. Therefore, this period was chosen for the research.

The nine villages, where samples were collected and the population was interviewed, are inhabited by 5,127 residents. Five households located along the perimeter of the rural area and at its center were selected for drinking water sampling. For the population survey, a cluster sampling method was used near the households where drinking water samples were collected. In total, five clusters were selected at the sampling sites, and 453 households were interviewed, representing 2,081 residents.

A questionnaire survey was used as the main method of collecting primary data, allowing the acquisition of quantitative and qualitative data on the perception, behavior, and level of awareness of the respondents on the issue under study. The structure of the questionnaire included both closed and open questions, which ensured a balance between data standardization and the ability to better understand the opinions of the participants. The survey was conducted among the target group with predefined sociodemographic characteristics, and the results of the questionnaire formed the basis for the analysis of the relationships between individual factors and the stated attitudes of the respondents. The questionnaire items are given in Table 2.

Data preparation for analysis was carried out according to the flowchart given in Figure 3. The research data flowchart illustrates the workflow of the study, from data collection (survey and drinking water sampling) to chemical analysis, perception analysis, evaluation of correlation, and the final data interpretation.

The overall research process is presented in Figure 3, highlighting the sequential steps: namely, (1) survey data collection, (2) water sampling, (3) chemical analysis of water samples, (4) analysis of perceived water quality, and (5) correlation and final evaluation.

Water samples were analyzed in the chemical laboratory of L.N. Gumilyov Eurasian National University with analytical methods of determination of the chemical parameters. The main chemical substances examined in the drinking water samples were cations (sodium and potassium, magnesium, and calcium) and anions (hydrocarbonates, chlorides, and sulfates).

The water parameters were chosen because cations [calcium (Ca²⁺), magnesium (Mg²⁺), sodium (Na⁺), and potassium (K⁺)] and anions [hydrocarbonate (HCO₃ ⁻), sulfate (SO₄ ^2-), and chloride (Cl⁻)] are the main components of the mineral composition of drinking water, causing its hardness, mineralization, and taste. Sanitary standards (WHO, SanR&R of RK, and EU Drinking Water Directive) are based on the concentrations of these ions, as they are directly related to the safety and organoleptic properties of water. For example, high content of sulfates or chlorides is associated with diarrhea and a salty taste, and a lack of calcium and magnesium is associated with the risks of cardiovascular diseases. Residents evaluate water primarily by taste, smell, and sensations during use. These parameters are directly caused by the anion–cation composition, such as the bitter taste from high sulfates, salinity due to excess sodium/chlorine, and hardness because of calcium and magnesium. Therefore, the choice of these indicators allows us to explain the subjective complaints of the population through objective chemical data. Unlike microelements or organic pollutants, macro-ions (anions and cations) are present stably and in significant concentrations, making them suitable for statistical clustering and perception analysis. These parameters reflect the geochemical specificity of the sources and allow comparisons between different villages/regions.

Ca²⁺ and Mg²⁺ cations were detected using the complexometric method, and K⁺ and Na⁺ were detected using the ion-selective electrode technique. HCO₃ ⁻ anions were determined using acid–base and potentiometric titration methods, and SO₄ ^2- anions were analyzed using turbidimetry with barium chloride. Cl ⁻ was determined by the argentometric method. A spectrophotometer and an ionometer were used for this analysis.

Residents of the research area are not informed about the quality of drinking water. Monitoring and analysis of central water supply is carried out regularly, but these figures are not available to everyone; they are available in the sanitary and epidemiological stations and are controlled according to SanR&R. The decentralized systems are not controlled.

The length of the central water-supply pipe runs along the village perimeter, measuring approximately from 0.5 to 1.5 km. The pipe material is plastic, and the time for which water stays in the pipe from the source to the consumer depends on the intensity of water consumption from minutes to several hours. Water purification occurs only in the central water-supply system and mainly through chemical methods. The quality of drinking water in terms of disinfection is controlled by the central water-supply system.

In 1994, Piper (1994) proposed creating a graphical representation of the partitioning of relevant analytical data to assess the chemical components of water. This process is based on the assumption that most water samples contain cations and anions in chemical equilibrium. The most common cations are assumed to be two “alkaline earth” cations: Ca²⁺ and Mg²⁺, and one “alkaline” cation: Na⁺. The most common anions are one “weakly acidic” HCO₃ ⁻ and two “strongly acidic” anions, namely, SO₄ ^2- and Cl⁻. The Piper diagram is widely used in ecological and ecotoxicological studies (Jat et al., 2025; Jat Baloch et al., 2022a; Jat Baloch et al., 2022b; Jat Baloch et al., 2021).

Statistics for each water quality parameter are presented as ± SE (standard error) of the mean. One-way analysis of variance (ANOVA) was used to test the statistical differences among the villages through equal and unequal variance t-tests. Group means were compared for each water quality variable using ANOVA. Prior to ANOVA, we checked normality of the residuals (Shapiro–Wilk) and homogeneity of variances (Levene’s test). When homogeneity held, we performed a classic one-way ANOVA followed by Tukey’s HSD for pairwise comparisons. When variances were unequal, Welch’s ANOVA was used, followed by the Games–Howell post hoc test. For clearly non-normal data (especially with small n), we additionally report the Kruskal–Wallis test with Dunn’s pairwise comparisons (Holm adjustment). The results are visualized as bar charts of group means with standard error bars; distinct letters above bars denote significant pairwise differences (α = 0.05) (Beisenova et al., 2025).

The bar plots display the mean concentrations of the major ions, namely, K⁺, Na⁺, HCO₃ ⁻, Cl⁻, and SO₄ ^2- across multiple water sources, including BWIP, CWSSIT, PB, PWIPS, PWWOSH, PWWOWSH, SW, and WPS. Error annotations are indicated using lowercase letters (e.g., a, b, and c), which represent statistically significant groupings. Identical letters denote no significant difference between sources, whereas different letters indicate statistically significant differences (p < 0.05). For statistical analysis, R studio version was used.

We combined two datasets by village: one containing chemical water quality measurements and with the other containing the respective household survey responses. For the chemical dataset, we grouped the data by village and calculated the average concentration of each measured parameter: Ca²⁺, Mg²⁺, hydrocarbonate (HCO₃), Cl⁻, SO₄ ^2-, K⁺, and Na⁺. This produced a table of nine villages with water chemistry values.

For the questionnaire dataset, we similarly grouped responses by village. Key survey questions were converted to numeric form (yes = 1, no = 0), so we could compute village-level averages (proportions of respondents answering “yes”). Finally, two aggregated tables were merged, showing the combined dataset with both average water chemistry and average survey outcomes side-by-side.

With the datasets combined, we computed the pairwise Pearson and Spearman correlations between all numeric variables (the chemical concentrations and the survey response averages). Several matrices were constructed to determine the relationship between different chemical composition indicators and participants’ responses about water quality and purification practices. Additionally, the correlation between chemical parameters and landscape indicators of the areas was calculated using Spearman correlation matrix. Python version 3.12 was used for correlation analysis.

The following indicators were used in the study:

Chemical composition indicators (concentration of major cations and anions (mg/L)): Ca²⁺, Mg²⁺, Na⁺, K⁺, Cl⁻, SO₄ ^2-, HCO₃ ⁻, total dissolved solids (TDS), pH level, electrical conductivity (EC), and water hardness (calculated from Ca²⁺ and Mg²⁺).