Introduction

Groundwater is the most valuable resource in the world and it is used for household, horticultural, agricultural, hydropower generating, and other purposes by around 33% of the world's population (Macdonald et al., 2016; Nawab et al., 2016; Zhang et al., 2017; Su et al., 2020; Wang et al., 2020; Karunanidhi et al., 2021). Groundwater pollution has grown into an important element influencing the quality of all freshwater aquifer systems (Chandrasekar et al., 2021). The monitoring of a wide range of physicochemical properties, including cations and anions, is required for groundwater quality (Tiwari et al., 2018). Groundwater is an important natural resource for meeting our country’s water needs (Gajbhiye et al., 2015). Groundwater wells are an important source of water for many people in Kerala’s southwest region, and they serve as a source of water (Sapna et al., 2020a; Sefie et al., 2018). The quality of a region’s groundwater varies from area to area. Many industrial effluents and contaminants are discharged into the environment untreated, allowing groundwater supplies to deteriorate (Thangavelu et al., 2019). Human health, animal welfare, and agriculture are all affected by this unethical groundwater (Arumugam & Elangovan, 2009). Groundwater can be contaminated by precipitation and surface discharge from polluted soil (Khodapanah et al., 2009).

Groundwater quality assessment involves more than evaluating its acceptability; it also includes demanding groundwater management in a more acceptable framework to meet current and future drinking and water system needs (Islam et al., 2015). The evaluation of groundwater quality includes far more than finding its acceptability; it is also about require groundwater management in a more acceptable framework to satisfy current and future needs for drinking and water system usage (Islam et al., 2015; Shanmugamoorthy et al., 2022). Water quality evaluations are carried out on a regional and international level by a many experts from various locations (Abbulu & Rao, 2013; Magesh et al., 2021a, b; Sajil Kumar et al., 2013; Sapna et al., 2018; Thangavelu, 2013). The analysts attempted to increase the number of water quality index (WQI) that monitor groundwater quality (Chaurasia et al., 2018; Nagaraju et al., 2016; Pandian & Jayachandran, 2014; Sapna et al., 2020b; Singh & Khan, 2011).

A Geographical information system (GIS) is an emerging tool to demonstrate the accuracy of water quality mapping. In the past, a few researchers have analysed the issue of ground water quality by using GIS in different conditions (Emenike et al., 2018; Thangavelu et al., 2021). It was suggested by Balakrishnan et al. (2011) that the utilisation of GIS strategies is important in the assessment of groundwater pollution risk assessment methods and also groundwater quality planning. Reza and Singh (2010) used the water quality file approach to assess the state of groundwater quality. Emenike et al. (2018) investigated the geographical and hydrochemical links between groundwater quality.

A comparable technique was used for the evaluation of groundwater quality for drinking and water system usage in shallow hard rock springs in Kerala (Satish Kumar et al., 2016). It is difficult due to the spatial variation of the several contaminants that can be analysed. Arulbalaji and Gurugnanam (2017) have drawn more rapidly on groundwater quality evaluation in Tamilnadu, India, using geospatial and measurable devices. It was clearly focused on a physico-chemical evaluation of groundwater quality and its appropriateness for human and agricultural use. Several experts have investigated to develop the different groundwater quality index (GWQI) for groundwater analysis. It may be assessed using a number of factual approaches and models (Molla et al., 2015).

This analysis uses multivariate statistical methodologies to find new components that impact water systems. The combined measuring methodology and spatial interpolation techniques can help water systems. The water management sources elucidate better solutions for pollution issues within different parts of India and throughout the world (Halim et al., 2010; Rahman et al., 2012; Abinandan et al., 2014; Jeihouni et al., 2014). Statistical processes can be utilised in a simple and easy format to establish a numerical connection for physicochemical studies (Iyer et al., 2003). It was decided to approach the statistical processes such as principal component analysis (PCA) and hierarchical cluster analysis (HCA) understand the meaning of this complex data set. The analysis of a coefficient correlation matrix and the Pearson relationship bivariate one tailed approach is a significant tool that will assist in the study of water pollution issues. Groundwater quality limits that result in large datasets needed for analysing difficult informational matrixes in order to have a better understanding of the water quality (Sapna et al., 2020b).

The investigation is focused on physical and chemical components that are inextricably related depending on the physicochemical measurement and assessment locations. Many studies have been conducted on the physicochemical evaluation of groundwater quality in different locations. A few attempts have been made to assess groundwater quality using the correlation coefficient assessment of various water quality limits. Furthermore, a few methodologies were used to assess groundwater quality, which means it cannot directly indicate contaminated features (Bhasin et al., 2008).

The spatial distribution of groundwater quality method is used to identify areas where pollution poses a risk (Masoud, 2014). In addition, a few strategies were employed for the appraisal of groundwater quality, which cannot directly reflect contamination attributes (Adam et al., 2019. The spatial distribution of the groundwater quality index was created using spatial interpolation techniques such as the inverse distance weighted (IDW) approach (Islam et al., 2013; Thangavelu et al., 2019). IDW is approach that estimates groundwater sample values by averaging the values of sample data locations in the area of each processing and give an idea of the prediction accuracy and selection will be easy and appropriate. The objective is to understand and categorise groundwater quality in rural and industrial regions for spatial analysis using various physicochemical parameters. The study was compared physicochemical characteristics, multivariate statistics, WQI, and GIS to evaluate groundwater quality in the Kannur district.

Materials and methods

The study area’s description

The Kannur region is surrounded by the Malabar territory of Kerala. The northern part of the zone is covered by Kasaragod district, the southern part of the territory is limited by the Kozhikode region, and the south-east part of the region is covered by Wayanad district. The Western Ghats, which form the boundary between the Karnataka state region of Kodagu and the Lakshadweep Sea toward the western part of the Arabian Sea. The total area covered is 2966 sq. km. It lies between 11° 40′ 00'' and 12° 20′ 27" northern latitude and 75° 10′ 00'' and 75° 56′ 30" eastern longitude. High lands, midlands, and lowlands are the three geological districts that make up the area. The district’s elevation ranges from 50 to 60 m above sea level (amsl). Paithalmala is the most elevated point in the Kannur District (1372 m), with over half of its inhabitants living in metropolitan zones. Kannur has an approximate population of 26 lakhs. The highest temperature is over 35 °C, while the lowest temperature is about 20 °C. The annual average rainfall is 34–38 mm. There are different soil types, i.e. lateritic soil, coastal alluvium soil, riverine alluvium, brown hydromorphic soil, and forest loam. The geography, geology, soils, climate, and natural vegetation all contribute to the physical characteristics. The area is defined geologically by Archean crystalline rocks, the majority of which are Chernockites. Coastal and alluvium, sand stone, and clay with intercalation make up the coastal area. The district has three reservoirs which serve as surface water irrigation sources (CGWB, 2007). The study area map is as shown in Fig. 1.

Fig. 1
figure 1

Study area map of Kannur district

Full size image

Groundwater data used

Groundwater samples were collected from the Kozhikode-based Centre for Water Resources Department and Management (CWRDM). In 2011, 25 area samples were obtained from the investigation zone. For 2019, 25 groundwater field samples were collected and done with physicochemical analysis in the Kannur University laboratory. The physico-chemical estimation of the 25 locations in 2011 and 2019 was not statistically significant between the two seasons.

Groundwater sampling collection and analysis

The groundwater samples from various bore wells were collected from the lower parts of wells in five taluks (Thalassery, Kannur, Iritty, Taliparamba, and Payyanur) in rural regions and close to industrial zones. The groundwater samples were collected in a clean 1-L plastic container and examined at the laboratory. The samples were examined by the American Public Health Association (APHA) 2005 using established procedures (2005). The investigation was carried out for physico-chemical parameters. The testing locations were used to measure the geographical coordinate system with the help of global positioning system (GPS). The physico-chemical parameters such as hydrogen ions (pH), electrical conductivity (EC), total dissolved solids (TDS), and significant cations and total hardness (TH), sodium carbonate (CO3-), bicarbonate (HCO3-), chloride (Cl−), sulphate (SO2−), nitrate (NO3-), calcium (Ca2+), magnesium (Mg2+), sodium (Na+), potassium (K+), and iron (Fe2+) these anions were used to analyse the laboratory analysis. It was examined by the standard methodology followed by APHA (2005).

The pH metre was used to test the water for hydrogen ions. A water analyser was used to determine the EC and TDS levels (Microprocessor pH-EC-TDS Meter – 1615). Turbidity was measured using a turbidity metre (Spectralab Model NT 4000 Turbidity Meter). EDTA was used to monitor the concentration rather than control the water quality parameters in TH, Ca2 + , and Mg2 + . Chloride was measured by silver nitrate titration. Carbonate and bicarbonate were analysed by standard trimetric analysis;sulphate was dissected utilising a Nephelometer (Systronics 132). Nitrate was investigated using the Ion Selective Electrode Method (V-780 UV–Visible/NIR Spectrophotometer). Iron was determined using the Colorimetric Thiocyanate Method, whereas sodium and potassium were determined using a flame photometer (Microprocessor Flame Photometer S-935) compound test. For this study, statistical analysis was used to investigate the relationships between various chemical characteristics. The level of the water table affects the activity of groundwater everywhere around it. The results also include the extent of each limit, as well as insights and comments on groundwater quality. The study area map is presented in Fig. 1.

Multivariate statistical analysis

Multivariate statistical approaches such as descriptive statistics, principle components analysis, cluster analysis, correlation matrix analysis, and Pearson Correlation bivariate one tailed analysis were used in the SPSS 20 version environment to identify in groundwater sources. Figure 2 shows the entire study approach in a flow chart.

Fig. 2
figure 2

The flow chart depicts the study’s overall methodology

Full size image

Descriptive statistics

A statistical analysis of data collected for groundwater quality parameters from 25 sample locations between 2011 and 2019 was performed. Descriptive statistical analysis, such as mean, median, average, standard error, dispersion, deviation, skewness, and kurtosis, is used to acquire a basic understanding of the sample distribution.

Principal component analysis

Principal component analysis (PCA) is one of the most widely used statistical analysis techniques for analysing data patterns (Salem et al., 2022). This method for reducing groundwater data and separating a few unknown elements in order to examine links between the known factors (Farnham et al., 2003; Gou et al., 2007). PCA was used to remove the principal component (PC) from the data set and determine the geographical variations of physicochemical features in groundwater that are uncorrelated, or orthogonal in component. The first principal component (F1) absorbs and accounts for the largest possible percentage of the overall variation in the set of data, while the second component (F2) receives and provides for the variance. The percentage contribution of Fi to the overall variance in the normalised informative index, for example, is given by the formula 1.

$$\mathrm{Percentage}\;\mathrm{variance}\;\mathrm{accounted}\;\mathrm{by}\;\mathrm{Fi}=\mathrm\lambda\mathrm i/\sum\mathrm\lambda\mathrm i$$
(1)

The quantity of factors in the informative collection is equal to the sum of the latent roots of all the main components, i.e. Σλi = number of variables. Because the main component approach extracts the highest feasible variance for each preceding F, the values of latent roots get lower for future F's.

Fi is a linear combination of the initial variables, as follows:

$$\mathrm{F_1\:=\:b_{11} \ Y_1\:+\:b_{12} \ Y_2\:+\:b_{1K} \ Y_K}$$
$${\mathrm F}_2\:=\:{\mathrm b}_{21}\ {\mathrm Y}_1\:+\:{\mathrm b}_{22}\ {\mathrm Y}_2\:+\:{\mathrm b}_{2\mathrm K}\ {\mathrm Y}_{\mathrm K}$$
$${\mathrm\lambda}_{\mathrm i}\:=\:{\mathrm b}_{11} \ {\mathrm Y}_1\:+\:{\mathrm b}_{12} \ {\mathrm Y}_2\:+\:{\mathrm b}_{1\mathrm K} \ {\mathrm Y}_{\mathrm K}$$

In the above expressions, b's are factor loadings, Y's are the original variable’s standardised values, and ‘k’ is the number of variables in the above expressions. A significant association between the variable and the factor is indicated by high factor loadings (values near to 1). Kaiser–Meyer–Olkin (KMO) and Bartlett’s tests are used to assess the appropriateness of the PCA data set. KMO is a statistic for sample adequacy. If the KMO value is more than 0.5, a PCA can be performed. Bartlett’s test of sphericity is used to test the null hypothesis that the variables in the population correlation matrix are related and that the correlation matrix is an identity matrix. If the observed significance level is less than 0.05, the null hypothesis is rejected. It illustrates the existence of significant correlations between variables. Sundaray et al. (2006) used PCA on the normalised data sets and sorted those with Eigen values more than 1.0, which are regarded as having significant influence on hydro-geochemical processes in this study.

Hierarchical cluster analysis

Hierarchical cluster analysis (HCA) was used to classify the based on uniformity within a class and differences between classes (Mendiguchía et al., 2004; Han et al., 2006). It is essential for analysing data and identifying patterns. Because each ingredient of groundwater resources uses a unique approach, sample data must be normalised. Data standardization ensures that all variables in the study have almost the same impact (Rahman et al., 2014). It is most commonly used by a dendrogram to establish natural proximity links between any one sample and the entire informative collection (McKenna, 2003). The dendrogram provides a strategic distance from misclassification owing to huge variances in information dimensionality. It was employed to modify trial data using a z-scale change (Liu & Yang, 2018). The following Eq. (2) was used to normalise the data to the Z score (m = 0 and S = 1):

$$\mathrm{Z}=(\mathrm{X}-\mathrm{m})/\mathrm{S}$$
(2)

where Z represents the standardised value, X represents the sample data value, m represents the mean, and S represents the standard deviation (Rahman & Gagnon, 2014).

Correlation matrix analysis

The statistical analysis was performed using standard techniques, which included calculating correlation coefficients between different parameter variables or between the objects on which these variables were assessed (Davis, 2002). The correlation coefficients were used to identify the closely related components and associated water quality metrics.

Pearson correlation bivariate one tailed analysis

The study was investigated the connections between various groundwater physicochemical assessments. The mean and standard deviation of each measurable parameter were computed twenty-five sample locations. Pearson correlation is a way of analysing the relationship between two variables that indicates that changes in one are connected to changes in the other (Pallant, 2011). Once a correlation is created, the rest of the parameters may be determined simply by measuring a few significant parameters, and the rest of the parameters can be assessed quickly and easily (Patil & Patil, 2010). Once a correlation is established, the remaining parameters may be found simply by monitoring a few relevant values, and the remaining parameters can be assessed fast and easily (Patil & Patil, 2010). SPSS software was used for the statistical analysis, with confidence intervals of 0.05 and 0.01. The Pearson product moment correlation coefficient was calculated using Eq. (3) (Kumar & Riyazuddin, 2008).

$${r}^{2}=\frac{n\sum xy-\sum x\sum y}{\left[n\sum {x}^{2}-\left(\sum {x}^{2}\right)\right][n\sum {y}^{2}-{(\sum y)}^{2}]}$$
(3)

The coefficient of determination (r2) was calculated to explain both changes in one variable and changes in the other. HCA has been effectively used in many studies to examine and analyse groundwater quality data (Kumar & Riyazuddin, 2008; Rani & Babu, 2008). HCA was performed on a standardised data set to reduce the impact of data measurement scale.

Spatial interpolation techniques of groundwater using GIS

The groundwater information about different areas was utilised for the spatial interpolation technique. For example, kriging, inverse distance weighted (IDW), and spline techniques were used to predict and measure the geographic variation of the groundwater dataset. Apart from this study, many researchers implement the same methodologies across a wide range of topics. The method was used in groundwater-related studies and has a wide range of regional and international applications. The IDW technique was chosen for this study’s spatial analysis because of its efficacy and estimate accuracy when compared to other relation approaches such as kriging and spline (Courault & Monestiez, 1999; Gorai & Kumar, 2013).The spatial analyst-interpolation module system was attempted using the ArcGIS environment tool (Thangavelu, 2013). The IDW method is used as it handles spatially interpolated data precisely. In Eq. (4), Mitas and Mitasova (1999) used the IDW technique in a linear direction.

$${Z}_{j}=\frac{\sum_{i=1}^{n}\frac{{Z}_{i}}{{\left({h}_{ij}+\delta \right)}^{\beta }}}{\sum_{i=1}^{n}\frac{1}{{\left({h}_{ij}+\delta \right)}^{\beta }}}$$
(4)

In the equation, Zj is the result at an ungauged location, Zi is the available data, is the weight, and is a flattening factor, and hij is the spacing interval here between predictable and unpredictable point (5).

$${h}_{ij}=\sqrt{{\left(\Delta x\right)}^{2 }{+(\Delta y)}^{2}}$$
(5)

where x and y are the reference axis distances between the unknown point j and the sampled point i.

Assessment of water quality index

The water quality index (WQI) is a mathematical approach used to integrate a large quantity of data on water quality into a measure that indicates the extent of contamination (Singh et al., 2013). It is a simple method for determining the quality of water. The suitability of drinking water is determined by comparing the water quality data from the samples tested to the BIS (2012) recommended drinking water standard, which is used to calculate the WQI (Verma et al., 2020). It assists with expressing the present situation of water to public entities and decision-makers involved in water management (Alam et al., 2020; Balamurugan et al., 2020). A single number reflects an index that takes into account all of the elements that impact water quality (Vasanthavigar et al., 2010). Equation (6) gives the formula for the water quality index.

$$\mathrm{WQI}=\frac{{\sum }_{i=1}^{n}{ q}_{i}{w}_{i}}{{\sum }_{i=1}^{n}{ w}_{i}}$$
(6)
$${q}_{i}=100 \mathrm{x }\frac{{V}_{i}}{{\mathrm{S}}_{i}}$$

where qi is the Sub Index, and Vi is the observed value in the lab, n is the number of parameters taken, Si is the typical value of ith particle, and wi is the weight of ith parameter.

Results and discussion

Multivariate statistics analysis of groundwater parameters

Descriptive statistical analysis

The statistical variables range, minimum, maximum, median, standard deviation, variation, skewness, and kurtosis were examined in descriptive statistics in Tables 1, 2, and 3. The pH values of ground water between 2011 and 2019 range around 5.30 and 4.2. 2011 has a minimum pH value of 3, and 2019 has a minimum pH value of 3.7. The maximum value is represented in 2019 (8.3), and in 2011 it became 7.9 in 2019. The mean value in 2011 was 6.9, and in 2019 it is 6.52, despite the fact that the mean value in 2011 was extremely high. The standard deviation of 2011 (1.52) is highly presented when comparing the year of 2011 (0.96). The variance in 2011 was 1.48, while the variance in 2019 was 0.92, indicating that the variance in 2019 is well represented. The skewness of 2011 (−1.58) and 2019 is −1.09, when discussing highly observed in 2011. The kurtosis value of 2011 is presented at 3.13 and that of 2019 is observed at 1.76, whereas comparing the kurtosis highly observed in the year of 2011 is shown in Fig. 3. The standard deviation of 2011 (1.52) is highly presented in the year of 2011 (0.96).

Table 1 Descriptive statistics of the research area’s physico-chemical characteristics with WHO and BIS limits (2011 and 2019)
Full size table
Table 2 (a, b) Varimax with Kaiser normalization rotated method factor loadings from PCA of standardised water quality data (2011 and 2019)
Full size table
Table 3 Correlation coefficient matrix of various groundwater quality parameters (2011)
Full size table
Fig. 3
figure 3

The study area’s total physico-chemical analysis (2011)

Full size image

The variance of 2011 is observed at 1.48 and the variance of 2019 is presented at 0.92. The skewness of 2011 (−1.58) and 2019 is −1.09, whereas the comparison was highly observed in 2011. The kurtosis value of 2011 is presented at 3.13, and in 2019 it is observed at 1.76. It was highly observed in the year 2011 as shown in Fig. 3. As shown in Fig. 4, electrical conductivity, total hardness, total dissolved solids, carbonate, bicarbonate, chloride, sulphate, nitrate, calcium, magnesium, sodium, potassium, and iron were all dissolved in descriptive statistics in the same way as the previously mentioned (for example, pH) and separately. The World Health Organization (WHO, 2006), the Bureau of Indian Standards (BIS, 2003), and the United States Environmental Protection Agency (USEPA, 1975) standards were compared to physicochemical parameters for drinking purposes of groundwater. Modifying traditional cropping types and introducing drips, sprinklers, and micro-irrigation technologies to decrease groundwater depletion may be advantageous in some areas to avoid groundwater depletion, according to the study. The findings show that the recommended limits in Table 1 are the most important factors to consider when comparing groundwater samples.

Fig. 4
figure 4

The study area’s total physico-chemical analysis (2019)

Full size image

Principal component analysis

A groundwater quality study was performed in 2011 and 2019 using principal component analysis (PCA) with Varimax rotation and Kaiser Normalization component values. According to Kaiser (Liu & Yang, 2018), only factors with eigenvalues bigger than one should be considered. In the 2011 and 2019 data sets, the KMO measure of sample adequacy is 0.846, and the importance level of Bartlett’s trial of sphericity is 0.001. To improve groundwater management, we need to focus on more than the quantity of water aquifers can provide. Many of the location’s major aquifers have been losing groundwater, particularly in the nearby Kannur region. PCA was performed on normalised data sets, including fourteen water quality standards. Figure 5a and b shows that in 2011, four variables with eigenvalues larger than accounted for 84.7% of the absolute change in the data set. Four variables in the 2019 informative ranking have eigenvalues larger than one, accounting for 73.4% of the data set’s overall variability. Figure 6a and b shows the factor loadings created by varimax symmetrical rotation from the 2011 and 2019 data sets, respectively.

Fig. 5
figure 5

Component plot in rotated space (a) and scree plot of the characteristic eigenvalues (b) are used in principal component analysis (2011)

Full size image
Fig. 6
figure 6

Component plot in rotated space (a) and scree plot of the characteristic eigenvalues (b) are used in principal component analysis (2019)

Full size image

The PC loadings indicate both positive and negative relationships. The loadings approaching 1 indicate a significant relationship between the variable and the PC. A strong link has a PC value that is greater than 0.75. The values between 0.5 and 0.74 were determined to be closely connected. The processes that have been interpreted based on the factor loadings are shown in Table 2a and b. The eigenvalues were highly presented in 2011. Table 3 used scree plots to determine the number of PCs in order to understand the groundwater parameter structure. The overall variation for PC1, PC2, PC3, and PC4 for groundwater quality data in 2011 was 34.84%, 20.76%, 19.25%, and 9.81%, respectively. The factor (PC1) in the data sets showed that it was highly positively loaded with all of the physicochemical characteristics in 2011.

The factor (PC1) pH shows a very high positive correlation and a moderate correlation with TH, EC, TDS, SO2−, Ca2+, Mg+, and K+ parameters in 34.84% of the total variance. EC stands for salt concentration and is a good indicator of accessible nutrients. Calcium includes largely dissolved mineral elements, and hardness considerably affects the value of water for public and industrial uses. The study found that factor (PC1) causes groundwater dilution, with TDS, chloride, and sulphate concentrations in groundwater slightly lower than the other parameters in the 2011 samples. PC2 is moderately linked with CO3−, HCO3−, and Mg2+ and discloses 20.7% of the variability. The clusters of site-specific samples in space and their spatial distribution are confirmed by the PCs scores plot created using PC1 and PC2 components.

PC2 demonstrates the use of the physicochemical limit. CO3-, HCO3-, and Mg2+ are delivered to groundwater because of the disintegration of minerals bearing these particles during the stimulation of spring by precipitation. PC3 accounts for 19.2% of the data set’s variance and includes moderate positive correlations with pH, CO3-, HCO3-, and Fe2+, all of which are explained by combining multiple sources. Nitrate is negatively highly correlated, and Cl−is moderately correlated, with variability reduced in groundwater samples. Two separate processes have been proposed for this, with bicarbonate indicating intense weathering of carbonate settings and groundwater's alkaline character. PC4 has iron loadings and accounts for 9.8% of the variability. The loading of Fe has a positive relationship and is linked to the decomposition of organic materials by iron reduction. The screet plot (Fig. 6b) is excellent for detecting a number of positive components with a significant scale difference. For PC1, PC2, PC3, and PC4, the overall variance of groundwater quality data in 2019 was 31.52%, 20.11%, 12%, and 9.77%, respectively. The majority of the samples have visible grouping in the upper left and lower left quadrants, but sample loading in the upper right and lower right quadrants is significantly scattered.

PC1 reports 31.5% of the variance, elucidating that TH, TDS, and Ca were strongly positive correlated. The loadings EC, K+, Na+, and Cl− are in moderate correlation. The loading of TH, TDS, Ca2+, EC, K+, Na+, and Cl− comes from various sources, including groundwater salinity and the application of chemical fertilisers in agricultural fields. As a result, the study location was not far from salinity-affected areas along the neighbouring Valapattanam river system. PC1 illustrates the distribution of the physico-chemical parameters TH, TDS, Ca2+, EC, K+, Na+, and Cl−, indicating that residential waste water seeping from on-site sanitation systems has contaminated groundwater. PC2 is responsible for 20.1% of the variability and includes pH-related factors. The moderate positive association of Mg+ with CO3− means that Mg increases as the magnesium concentration of groundwater in dissolved condition with deposits of salts in the study area. The sample characteristics are represented in the PCs score plots, which help in understanding their spatial distribution.

PC3 in the data sets have 12% of variability, which are CO3-, SO2- and NO3-, which comprise three parameters moderately positive. Non-carbonate hardness scale is hard and difficult to remove, but carbonate hardness scale is porous and easy to remove. The increased amount of soap or detergent needed to generate a lasting lather is usually a sign of hardness in the water. As the hardness of the water increases, so does the amount of soap used, resulting in a disagreeable curd. However, the factor loading of NO3- is non-significant; PC3 shows that evaporation is not a main factor determining Cl− concentrations in groundwater. Anthropogenic pollution from onsite sanitation, livestock waste, and municipal landfill sites may be identified as another source of nitrate process, and it is desirable to dilute NO3- concentrations to reduce them. PC3 and PC4 show that the strong and positive loadings by one parameter with an insignificant proportion of variance (i.e. 12 and 9.7%, respectively) were given the least advantage in regulating groundwater chemistry. PC4 has 9.7% variability in the data sets and is made up of two factors, HCO3- and Na2+, which has a bit positive relationship. The bicarbonate content of groundwater causes a rise in Na+ value, which might cause organic matter in the soil to dissolve.

Hierarchical cluster analysis

Groundwater samples from various classes were used in a hierarchical cluster analysis. Squared Euclidean distances were employed to quantify item similarity, and a dendrogram employing the Centroid linkage method of hierarchical clustering was used for cluster analysis (CA). The dendrograms are made up of multiple clusters, each of which contains one or more variables. Clusters are chosen for easier understanding based on a visual analysis of the dendrogram Fig. 7a and b. CA created five types of cluster groupings by combining groundwater samples from the study region in 2011 and 2019. In 2011, EC and Ca2+ were in Group I, TDS and Na+ were in Group II, TH and HCO3- were in Group III, Mg2+ and Cl− were in Group IV. The weathering of parent rocks, ion exchange processes, and salt leaching from the rocks are the main sources of increasing Ca, Mg, and Cl concentrations in groundwater. Permanent hardness has been recognised as an issue in the study area, necessitating advanced treatment methods such as reverse osmosis, base ion exchange, and desalination procedures to treat groundwater before use. K+, Fe2+ and SO4- were in Group V, NO3- was in Group VI, and CO and pH were in Group VII. In 2019, Group I included NO3-, K+, Na+, Fe2+, CO3-, pH, and Ca2 + ; Group II included Cl− and SO4-; Group III included Mg; Group IV included HCO3- and TDS; and Group V included TH and EC. Extreme values in the data sets are sensitive to the clusters. The similarity between PCA factors and HCA clusters supports the PCA-suggested dominating processes.

Fig. 7
figure 7

a Hierarchical cluster analysis (2011). b Hierarchical cluster analysis (2019)

Full size image

Correlation coefficient matrix analysis

In most cases, the correlation coefficient is used to quantify the relationship between the two variables. The values were evaluated for the ground water quality zones shown in Table 3. There is a direct relationship when the parameters of one of the region’s variables increase or decrease. In the resultant, the aftereffects of the relationship evaluation are taken into account. A high correlation coefficient (1 or − 1) indicates a strong relationship between two parameters, whereas a connection coefficient of zero indicates no relationship. Positive values suggest a positive relationship, while negative “r” estimations indicate the opposite.

In 2011, the results show that a highly positive correlation is observed between calcium and total hardness (0.967), TDS and TH (0.926), TH and EC (0.926), Mg2+ and TH (0.888), Na+ and Cl− (0.882), K+ and EC (0.861), HCO3- and TH (0.781), CO3- and TH (0.482), SO4- and EC (0.558). Ca2+ has a highly significant correlation with TDS, Mg2+, HCO3- and CO3-, it conveyed that increases in rainfall infiltrate rate, sewage water intrusion, and uncovered septic tank increases nitrate level in groundwater respectively. TH, K+ and SO4- have low positive correlations between TH and EC, indicating that changes in the direct correlation of conductivity increase as the concentration of ions increases and electrical conductivity values can be measured in low quality water.

All of the sample locations have Na+ values below the allowed limit of 200 mg/L. In all of the sample locations, the sulphate concentration in groundwater in the study region was below the acceptable limit of 400 mg/L. For this, a matrix analysis of Cl− is performed, and a highly negative relationship between pH and (−0.025) is observed, demonstrating that changes in hydrogen particle fixation resolve the direct correlation of Cl− in groundwater.NO3- was acquired with an exceptionally helpless negative connection between pH (-0.339), demonstrating that the immediate relationship of Cl−in water is unique. About a quarter of the sample locations surpass the 300 mg/L permissible limit. Around 50% of the sample locations in the study region have nitrate concentrations above the maximum allowable in groundwater. Despite a low negative relationship with Cl−(-0.145), it is clear that higher fixation is observed in groundwater. It might be linked to groundwater contamination sources from storm basement rocks or anthropogenic sources of salinity level in the study area, as shown in Table 3.

In 2019, the results show that Fe2+ has a highly significant correlation with K+ (0.963), indicating that it has lower geochemical mobility in groundwater samples, rich fertilisers in agricultural field for increasing the crop yield. Ca2+ has a highly positive correlation with total hardness (0.896); K+ has a positive significant correlation with Na+ (0.807), indicating that it is affected by drainage water seepage into the groundwater aquifer and salt dissolution in soil; TDS has a strong positive correlation with EC (0.896). TH has a minimum positive connection with EC (0.481), despite the fact that it is dependent on ion mobility in water. Cl− and SO2- have low positive correlations with total hardness (0.431 and 0.349, respectively), indicating the hardness and scale-forming properties of water; NO3− have low positive correlation with CO3- (0.262) which indicates that rock water interaction and weathering of parental rock contribute to elevate the concentration of nitrate in groundwater; and Mg2+ have a low positive correlation with pH (0.498).The conductivity values tested in low quality water had a strong negative connection with CO3- and HCO3-. Table 4 shows the Na+ concentration for the low negative correlation with pH (-0.489).

Table 4 Correlation coefficient matrix of various groundwater quality parameters (2019)
Full size table

Pearson correlation bivariate one tailed analysis

The groundwater testing showed a strong (p 0.01) and large (p 0.05) relationship. The Pearson correlation result for pH (r = 0.325), TDS (r = 0.25), CO3-(r = 106), HCO3- (r = 0.297), SO42−(r = 0.246), Mg2+ (r = 0.133), Na+(r = 0.226), and K+ (r = 0.246) in 2011 is provided in Table 5. Groundwater testing found a strong (p 0.01) and a positive correlation (p 0.05). As shown in Table 5, the Pearson correlation output has a high level of confidence in pH (r = 0.325), TDS (r = 0.25), CO3- (r = 106), HCO3- (r = 0.297), SO42− (r = 0.246), Mg2+ (r = 0.133), Na+ (r = 0.226), and K+ (r = 0.246). Separately, with a 95% confidence level, the modest positive association of Ca2+ is noted with NO3- (r = 0.223). It suggests that variations in NO3- are linked to calcium; there is a weak positive correlation between Fe2+ and SO42− (r = 0.184). It denotes the presence of sulphate in iron. When a drop in NO3- (−0.15) is linked to a rise in carbonate levels in groundwater, the result is a negative correlation (Table 6).

Table 5 Bivariate one-tailed Pearson correlation study of various groundwater quality metrics (2011)
Full size table
Table 6 Bivariate one-tailed Pearson correlation study of various groundwater quality metrics (2019)
Full size table

The coefficient of determination for EC—pH (r2 = 0.105) and TDS—pH (r2 = 0.325) as shown in Tables 7 and 8, which indicates that changes in the level of TDS and EC (10.5%) are explained by changes in pH, respectively. CO3−- pH (r2 = 0.011), HCO3−—pH (r2 = 0.088), SO42−-pH (r2 = 0.061), Mg2+- pH (r2 = 0.017), Na+-pH (r2 = 0.051), K+—pH (r2 = 0.061), while correlation with conductivity was strong but high respectively. The coefficients of correlation of Cl−-HCO3-(r2 = 0.004), Ca2+- NO3−(r2 = 0.049), and Fe2+-SO42−(r2 = 0.033) are less positive, whereas NO3−-CO3−(r2 = 0.022) is negative. It means the presence of iron and manganese at an acceptable temperature is linked to groundwater turbidity. TDS, CO3−, HCO3−, NO3-, Mg2+, Na+, and K+ levels all have a significant positive relationship with pH, NO3-, and SO42−. Because nitrate interacts with ionic content in water, the minor negative correlation with carbonate is predictable. As shown in Table 7, nitrate and carbonate have modest negative correlations.

Table 7 Groundwater quality parameter coefficients of determination (2011)
Full size table
Table 8 Groundwater quality parameter coefficients of determination (2019)
Full size table

In 2019, the correlation analysis from the Pearson correlation output (Table 6) showed that pH has a strong correlation with SO42− (r = 0.3267) and Mg2+ (r = 0.151); pH shows the low positive correlation with HCO3- (r = 0.233), nitrate with CO3- (r = 0.25) and calcium with CO3- (r = 0.135), sodium with SO42− (r = 0.79), and iron with NO3- (0.242) respectively. A negative correlation EC involves the decrease in pH (-0.052) and TDS with pH (-0.121) and is related to an increase in TDS levels in groundwater. This indicates that the existence of EC and TDS at an acceptable pH will be related to the groundwater as indicated in Table 6. The pH has a strong positive correlation with SO42−-pH (r2 = 0.071) and Mg2+ (r2 = 0.022); pH has a weak positive correlation with CO3–TH (0.096), chloride with HCO3-(r2 = 0.054), nitrate with CO3-(r2 = 0.062), calcium with CO3-(r2 = 0.018), sodium with SO42− (r2 = 0.143), iron with NO3-. A negative correlation EC involves that a decrease in pH (0.002) and TDS with PH (0.014) is related to an increase in TDS levels in groundwater as presented in Table 8. The pH has a critical negative relationship with the greater part of the physico-chemical characteristics. As a result, the chemical processes that take place in a groundwater system are highly involved. The correlation analysis provides accurate information about these complicated systems of water–rock interactions, as well as a broad knowledge of water–rock interactions.

Groundwater quality in Kannur district: an integrated geographical analysis

Spatial analysis is used to determine the quality of groundwater variation in the study area. The spatial variation of many groundwater physicochemical parameters, such as pH, EC, TH, TDS, HCO3-,CO3−,Cl−, NO3-,SO42−, Ca2+, Mg2+, Na+,K+and Fe2+, is based on BIS (2012) and WHO recommendations (2006). Using GIS, generate spatial distribution maps based on water quality parameters.

Hydrogen ion (pH) and electrical conductivity

The hydrogen ion is used to evaluate the acidity or alkalinity of a solution (pH). It is a measure of hydrogen ion concentration, or more precisely, hydrogen ion activity. The significance of the 2011 pH ranged from 3.0 to 8.3 and 3.7 to 7.3 in 2019. Table 1 demonstrates that the BIS and WHO suggest a pH range of 7.0 to 8.5 for drinking water. At low pH levels, water becomes corrosive, causing organoleptic issues. The acidic water promotes corrosion in well service construction components such as casings and screens, as well as domestic items. The spatial distribution of pH in the investigation was calculated between 2011 and 2019 using the Interpolated Distance Weighted (IDW) maps illustrated in Fig. 8a and b. All of the water samples were confirmed to be within permissible limits. In 2011, the majority of the samples were alkaline, although within permissible limits. In addition, an acidic pH was found in a slight of the study area. In 2019, the majority of water samples will be performed as much as possible. The pH values increase as far in the northern region of the area of study. The results suggest that the pH of the study area is not affected by the various industries in the Kannur area.

Fig. 8
figure 8

a, b Spatial distribution comparison map of pH in 2011 and 2019

Full size image

In electrical conductivity (EC), most of the water samples (25) of 2011 ranging from 31 to 505 μScm, and in 2019 the range were observed to be 35 to 236 μScm respectively. The capacity of a solution to lead an electrical flow is administered by the movement of the arrangements and is subject to the nature and quantities of the ionic species in that arrangement. The significance of EC stems from its salinity measurement, which impacts the taste of water and, as a result, the user’s acceptance of it. When comparing 2011 and 2019, the value of EC fluctuates greatly in 2019, as shown in Fig. 9a and b. A few areas in the northeast have been discovered that are not permitted by WHO standards. The highest conductivity value may be attributed to the highest concentration of ionic components.

Fig. 9
figure 9

a, b Spatial distribution comparison map of EC in 2011 and 2019

Full size image

Total hardness and total dissolved solids

Total hardness is crucial, when it comes to water for domestic use. In 2011, total hardness levels ranged from 15.15 to 18.18 mg/l, whereas in 2019, total hardness levels varied from 14 to 188 mg/l. Hardness is a quality of water that prevents foam formation with cleaners and increases the water’s limits. The amount of dissolved calcium and magnesium in the water affects its hardness. It is commonly represented by a set of CaCO3 equivalents (WHO, 2006). By comparing 2011 and 2019, in certain places, water turns too hard due to several industries located in the study area. Groundwater samples clearly show the significant influence of wastes and effluents from various industries. It has been suggested that it contributes to the creation of kidney stones and causes heart issues. As illustrated in Fig. 10a and b, industries have a negative impact on ground water supplies. In general, hardness is a mix of calcium and magnesium ions. As a result, rainfall may eventually result in increased in water hardness as calcium and magnesium-containing minerals dissolve.

Fig. 10
figure 10

a, b Spatial distribution comparison map of TH in 2011 and 2019

Full size image

The TDS levels in groundwater samples were observed in the range of 21.6 to 177.6 mg/l in 2011 and 22 to 268 mg/l in 2019. The high TDS values are caused by salts eroding from the surrounding soil, affecting water quality and maybe causing gastrointestinal disturbances in humans (WHO, 2006). During the rainy season in Kannur, the maximum guideline requirement of 500 mg/L was exceeded. The main contributors to high TDS values include groundwater activity in rocks, untreated sewage, waste deposits, and agrochemicals. The presence of high TDS levels in water causes a change in the flavour of non-potable water. As illustrated in Fig. 11a and b, the regional variance in TDS values illustrates the effect of industrialization.

Fig. 11
figure 11

a, b Spatial distribution comparison map of TDS in 2011 and 2019

Full size image

Carbonate and bicarbonate

In 2011, carbonate ranged from 0 to 11.61 mg/l and in 2019 from 0.l to 5.32 mg/l. Figure 12a and b shows that the northeast and southwest regions were observed to have high values in 2011 and 2019. A few of the areas exceed the limit in both the BIS and WHO standards. As a result, groundwater is used for water system purposes, but dissipation causes an increase in ion concentration in the underground. In 2011, bicarbonate concentrations ranged from 0 to 175.74 mg/l, and in 2019, it was from 10.69 to 213.2 mg/l while comparing the entire areas that were observed within the permissible limit. The outcome map (Fig. 13a, b) shows that the northeast, southwest, and little space in the focal piece of the investigation territory have moderate levels. When the groundwater used for water system purposes evaporates, it naturally influences the underground with an increment in the concentration of ions. Further dispersion from the irrigated zone causes a high concentration of salts, which is especially visible in carbonates in soils.

Fig. 12
figure 12

a, b Spatial distribution comparison map of CO3- (2011 and 2019)

Full size image
Fig. 13
figure 13

a, b Spatial distribution comparison of HCO3- (2011 and 2019)

Full size image

Chloride and sulphate

Cl− concentrations ranged from 5.75 to 59.46 mg/l in 2011 and 0 to 25.52 mg/l in 2019. The southeast part of the region has high fixations in 2011, and the southeast and southwest areas were observed in high concentrations, as shown in Fig. 14a and b. Cl−bearing rocks such as sodalite chloroapatite, which are tiny portions of igneous and metaphoric formations and generate little of the total volume of aquifers, may be to blame for the elevated Cl. Cl−bearing rocks like sodalite and chloroapatite, which are minor components of igneous and metamorphic rocks that generate even less than the total underground water quantity, could be to blame for the high Cl. Chloride levels that are undeniably high can be dangerous for your health, affecting the heart and kidneys, as well as taste, acid reflux, erosion, and satisfaction. Sulphate fixations range from 0 to 30.6 mg/l in 2011 and in 2019, 0 to 100 mg/l as presented in Fig. 15a and b. It shows that the majority of the samples (southeast) fall within the permissible limit and a few zones fall under permissible levels.

Fig. 14
figure 14

a, b Spatial distribution comparison map of Cl in 2011 and 2019

Full size image
Fig. 15
figure 15

a, b Spatial distribution comparison map of SO4 in 2011 and 2019

Full size image

Calcium and magnesium

Calcium concentrations range from 1.62 to 47.62 mg/l in 2011, and 4 to 160 mg/l in 2019.The difference in calcium ion concentration between 2011 and 2019 shows that the southeast portion of the study region has high concentrations by WHO criteria. The human body develops kidney or bladder stones as a result of elevated calcium concentrations. In common, about 12% of the sampling locations indicated calcium content values higher than the standard. In any of the surface water samples, however, there is no additional calcium. Figure 16a and b shows the IDW calcium maps from 2011 to 2019.

Fig. 16
figure 16

a, b Spatial distribution comparison map of Ca in 2011 and 2019

Full size image

Magnesium concentrations range from (2011) 0 to 15.22 mg/l and (2019) 4 to 62 mg/l respectively. The maximum of the samples in 2011 and 2019 was under the permissible limit of the WHO and BIS standards. A few areas were observed to be exceeding the limit. As shown in Fig. 17a and b, the southwest and northwest regions of the region have high concentrations. A high Mg2+ content in animals promotes scouring disorders.

Fig. 17
figure 17

a, b Spatial distribution comparison map of Mg in 2011 and 2019

Full size image

Sodium and potassium

In 2011, Na+ values varied from 2.8 to 28.5 mg/l, while in 2019, they ranged from 0.6 to 5.4 mg/l. Figure 18a and b illustrates that levels are high in the northeast and southeast. According to WHO guidelines, the majority of the samples are within the permissible limits, but a small fraction of those slightly exceed them. High Na+ levels are harmful because they can cause difficulties with the heart, kidneys, and circulation.

Fig. 18
figure 18

a, b Spatial distribution comparison map of Na in 2011 and 2019

Full size image

Potassium concentrations ranged from 0.1 to 9.6 mg/l in 2011 and 0 to 1.2 mg/l in 2019 of the samples were found to be within BIS and WHO acceptable levels. As shown in Fig. 19a and b, very few samples are obtained that are not permissible, and the southwest and northeast parts of the study region have moderate levels.

Fig. 19
figure 19

a, b Spatial distribution comparison map of K in 2011 and 2019

Full size image

Nitrate and iron

Nitrate concentrations range from 0 to 3.47 mg/l in 2011 and 0.001 to 2 mg/l in 2019. Nitrate concentration varies in 2011 and 2019. Figure 20a and b shows that the northwest and southwest parts of the study area have concentrations according to BIS and WHO standards. Figure 21a and b indicates that iron concentrations range from 0 to 3.79 mg/l in 2011 and from 0 to 4.65 mg/l in 2019. According to WHO guidelines, the majority of samples facing northwest and southwest parts are allowed to reach the permissible limit, while less than half of samples facing northeast parts are covered in a few areas that do not exceed the limits.

Fig. 20
figure 20

a, b Spatial distribution comparison map of NO3 in 2011 and 2019

Full size image
Fig. 21
figure 21

a, b Spatial distribution comparison map of Fe in 2011 and 2019

Full size image

Water quality index

The total status of groundwater quality for drinking purposes is shown by the water quality index. The study areas can be categorised as excellent, good, moderate, or poor based on water quality index (WQI) levels. The good areas of Vayakkara, Alakkode, Vayathur, Aralam, Vekkalam, Peringathur, Koothuparamba, Eruvatty, Ancharakkandy, Koodali, and Kalliassery were recognised by WQI in 2011. The water conditions in the Cheruthazam, Irikkur, Chavassery, Kolavellur, Edakkad, and Kottiyoor areas were excellent. Water quality was moderate in Panoor, Kannur, Chirakkal, Thaliparamba, Payyannur, and Karivellur locations. Figure 22 shows the relatively poor water quality at Mattannur and Thalassery.

Fig. 22
figure 22

The study area’s water quality index map (2011)

Full size image

According to WQI, Alakkode, Vayathur, Vekkalam, Edakkad, Panoor, Irikkur, Thaliparamba, and Karivellur were among the top-rated locations in 2019. Water sources of high quality include Vayakkara, Cheruthazham, Payyannur, Koodali, Ancharakkandy, Peringathoor, Kolavellur, Vekkalam, Chavassery, and Kottiyoor. Figure 23 illustrates that while the ground water quality in the Chirakkal, Kalliassery, and Thalassery areas was discovered to be bad, it was present in the Kannur, Koothuparamba, Eruvatti, and Aralam areas in a moderate range. When comparing the WQI of the ground water samples from 2011 and 2019, more areas are found to be of poor quality, while other areas go from exceptional to good to moderate range. The Thalassery region was noted to be more WQI in both 2011 and 2019.

Fig. 23
figure 23

The study area’s water quality index map (2019)

Full size image

The lower WQI in the study’s overall findings denotes the high quality of the groundwater. Based on the results from 2011, it can be assumed that most of the sampling sites in the study area had acceptable water quality, with the exception of Mattannur and Thalassery, with a WQI value of 0 to 25. The majority of the sampling locations in the study area fall within the 0 to 25 range, which indicates excellent water quality, according to the results obtained. However, numerous groundwater quality indices that had previously shown high water quality have changed when compared to 2011. In the Chirakkal region, one indicator of low water quality is displayed. According to the WQI value obtained for the various samples in 2019, most of the area around industrial sites in the Kannur district’s water is safe for human consumption.

The WQI’s overall results were evaluated well. For this study, groundwater quality in the northern regions is usually excellent. Overall, the 2011 and 2019 analyses reveal that the western part of the district has poor potable water quality. The higher values of iron, nitrate, total dissolved solids, hardness, sodium, potassium, calcium, magnesium, bicarbonate, chloride, and pH at this location have been discovered to be the main causes of the high WQI value. The overall results of the WQI were observed to be in good condition. Most of the regions in the northern part of the district have good water quality. The study illustrates (in 2011 and 2019) that the western part of the district shows poor potable water quality. The overall quality of groundwater is excellent in the entire study area.

Conclusion

The study of 25 groundwater samples taken from various portions of rural and industrial areas of Kannur district. It was studied that various physico-chemical analyses completed in the study area. The spatial distribution of several groundwater quality indicators was communicated using statistical and GIS innovations in an effort to evaluate the nature of groundwater. Additionally, the groundwater quality of the years 2011 and 2019 was compared. Most of the groundwater samples in the study region are below the allowed level. In some instances, there are exceptions. According to the overlay of thematic maps for essential aspects, a wider part of the district in the study region has portable groundwater.

In PCA, four factors from the 2011 dataset were recovered with eigen values greater than unity, while four factors from the 2019 dataset were retained with eigen values greater than unity, accounting for 73.4 percent of the total variance of the data set. The data sets’ extreme values are sensitive to HCA. The loading of the factor on the factor reveals natural pollution and soil erosion phenomena carried on by drifting fluctuations. The major instructive latent components for each season were extracted using PCA. The results of groundwater valuable studies can be improved by using the Pearson connection coefficients as a multivariate statistical tool to help identify measurably relevant features of information variance. All of the data were analysed using a Pearson correlation matrix. The purpose of the groundwater samples is to identify any statistical relationships between various pairs of ground water quality indicators. As observed in the comparison between 2011 and 2019, several locations have declined in value from 2011 and it cause of the poor industrialization.

This study reveals that almost all the samples collected and tested from Kannur district are portable, but there are some samples that show some variation in some of the properties from the permissible limit in BIS and WHO standards. The water quality index, as calculated, provides useful information by classifying specific places as excellent, good, poor, or unsuitable for drinking. WQI and GIS assist in the provision of more valuable data for water quality evaluation and problem solving. The study area is under threat due to some critical issues of environmental pollution. It is concluded that all the samples will be portable if proper remedial measures are put into practice. It is to assess the groundwater quality in Kannur district from a GIS perspective, and to create a geo-referenced groundwater database and maps that may be used to establish sustainable groundwater use strategies. Both governmental and non-governmental organisations have attempted to use GIS to locate the problem areas. The study was providing more useful information to the organisers and chiefs in order to develop strategy rules for effective management of groundwater assets.