Assessment of Soil and Water Quality Indices in Agricultural Soils of Manouba Governorate, North-East Tunisia

Abstract

Assessing soil and water quality in irrigated farming is vital for sustainable agriculture management. Low-quality irrigation water, particularly in semi-arid regions, poses environmental challenges and leads to soil salinization. This study was conducted in the Jedaida district, Manouba province, NE Tunisia. Forty-three soil and water samples were collected to develop indices for assessing soil quality. Sixteen indicators were selected using principal component analysis (PCA) for the minimum soil data set (MSD), including electrical conductivity, sand, organic soil carbon, and pH. The linear method shows a correlation with physical and chemical properties, classifying Jedaida soils into three quality metrics: good, moderate, and poor. The non-linear method displays the lowest indicator contribution in Zahira soils, followed by Mansoura soils (high and moderate). MSD combined with linear scoring is the most acceptable method of assessing the soil quality index (SQI). Water quality indices (WQIs) identify the suitability of irrigation. The results show a Kelly’s ratio > 1, a sodium adsorption ratio (SAR) > 10, and a sodium soluble percentage (SSP) varying from 40 to 60%. This highlights the negative effects of long-term irrigation with poor-quality water on soil health. Accordingly, groundwater was found to be unsuitable for irrigating surface soils. This finding emphasizes the importance of selecting suitable irrigation water to ensure soil quality.

1. Introduction

Soil is the pillar of fertility and agricultural productivity. It plays a vital role in water regulation, food security, and carbon sequestration [1]. It is formed of interacting organic, solid, liquid, and gaseous phases that provide services for ecosystems and life [2].

Salinization and sodicity, two major threats, cause irreversible soil damage and plant water uptake inhibition, negatively affecting crop yields [3,4,5,6,7]. Studies show that salinization causes structural change and disruption to soil aggregates that are extremely difficult to reverse. This leads to aggravated soil degradation over years, especially in semi-arid and arid regions where water scarcity is a problem [8,9,10,11].

Assessing soil quality using indicators that include biological, chemical, and physical properties is a crucial approach for environmental development, especially in semi-arid areas, where intensified agriculture has accelerated [12,13,14]. Due to limited water resources and demographic development, as well as an increase in food production in these regions, irrigated areas will be extended using poor-water quality, including wastewater and saline groundwater [15]. These areas, where rainfall naturally leaches salt after irrigation, are currently experiencing a water crisis due to climate change [16,17]. Studies have shown that providing water quality indicators is useful for classifying water for irrigation, reflecting mineralization processes, and assessing the potential recharge rate and flow direction of groundwater. This helps to ensure the efficient utilization of groundwater supply [11,18]. Various water hazard potential indices are used to categorize irrigation suitability, primarily the sodium adsorption ratio (SAR). This is the most commonly used index because it is a ratio of sodium amounts to combined magnesium and calcium. Other indices used include the soluble sodium percentage (SSP), permeability index (PI), and Kelly’s ratio (KR) [19,20]. They are often used to manage the effects of salinity and permeability problems. Generally, water contains nutrients such as ammonia and orthophosphate in mineral form. However, the suitability of water for irrigation depends on the nature and concentration of dissolved mineral elements [21]. Although nitrite is found alongside nutrients, it is a hazardous form of nitrogen that can pose serious risks to plant health and soil microbial activity. The Manouba province is located in northeastern Tunisia, where crops and arboriculture depend heavily on irrigation. Since 1960, this province has been known as an agriculture granary due to the expansion of water management infrastructure (conventional irrigation) and drainage systems in irrigated land [22]. These areas were irrigated by the Medjerda river, a primary source of irrigation in the region, which is subdivided into two parts: the Beja sub-basin and the Zarga river sub-basin, from Medjez El Beb to the sea, and the Lahmar river and Chafrou watershed sub-basin [23]. A study conducted along the Medjerda river found that the water quality varies significantly between the humid and dry season, with high concentrations of magnesium and carbonates threatening soil health and crop development [24]. Further research showed that seasonal changes impact the water quality of the Medjerda River [25]. However, the river faces challenges such as pollution (industrial, urban, agricultural) and evaporation (climate change risks). Similarly, the Chafrou river, another vital water source for the Manouba province, requires a hydrochemical assessment of its water quality for irrigation purposes. Poor water quality may disrupt ecological functions and increase salinity levels [26].

Soil quality is defined as “the capacity of a soil to function within ecosystem and land use boundaries, sustaining biological productivity and maintaining environmental quality, while promoting plant and animal health” [27]. The crucial factor of environmental quality combines physical, chemical, and biological qualities [28,29,30,31]. Worldwide, several studies have quantified soil quality using test kits [32], geostatistical methods [33], and soil quality index methods [27,28]. The most widely used method, the soil quality index (SQI), was applied in this study due to its flexibility. The notion of the evaluation of soil quality by implementing a rating criteria score was first suggested in earlier studies [34]. Many studies have proposed statistical methods such as principal component analysis (PCA), a minimum soil data set (MSD), and factor analysis (FA) to cluster indicators from a total data set (TDS), and the soil quality index (SQI) is then calculated [27,35,36]. Overall, the SQI calculation involves three steps, starting by choosing the appropriate indicators to establish the TDS; second, the indicators are scored and their weights are determined; and finally, all the indicator scores are combined into an SQI. However, due to the complexity of soil functions, it is essential to select soil quality indicators based on the specific soil functions of our region of interest, land uses, and the diversity of crop types, such as arboriculture, wheat, and artichokes [37,38]. In a study [39], physical and chemical indicators were used to quickly calculate the SQI for agricultural soils due to the difficulty in obtaining soil biological indicators. The two methods used to select soil indicators from the TDS were expert opinion (EO) [28] and statistical methods [40,41]. PCA is an effective method for the evaluation of a data set and is commonly used to identify the relationships between original soil indicators and transform them into independent principal components [28,40]. Other studies have achieved good results when calculating the weights of soil indicators using PCA [28,42], while others have used PCA to establish an MSD to quantify the SQI [43,44,45,46,47,48]. Linear and non-linear scoring models are commonly used in many studies [49]. Hence, studies have shown interesting results using the linear scoring model [14,50], while others have demonstrated that the non-linear scoring model is a very effective way to create a soil quality index [49]. Among the existing data standardization approaches, the scoring method is less complex because it first rates soil indicators based on actual measured values and then allocates scores to each one [51].

The innovative aspect of the study involves combining the SQI and WQI across diverse crop cultivation in the north-east of Tunisia. The use of both indices has never been investigated in any studies carried out in Tunisia.

The aims of this study were (i) to identify a minimum data set (MSD) of key soil indicators in the province of Manouba; (ii) to develop a general soil quality index (SQI) using two scoring methods (linear and non-linear); (iii) to categorize the suitability of irrigation water using multiple quality indices; and (iv) to generate a spatial distribution map of calculated SQI values using IDW interpolation based on field data.

2. Materials and Methods

2.1. Study Area

The study was conducted in the Zahira (36°49′24″ N, 9°49′24″ E) and Mansoura (36°47′49″ N, 9°54′20″ E) regions, located within the Jedaida district of Manouba province, in the northeastern region of Tunisia. It covered 19,074 hectares and accounted for about 60% of national wheat production [52]. The two regions under investigation are both agriculturally active and have similar environmental conditions, although they differ in terms of their sources of irrigation (the Medjerda River and the Chafrou groundwater system) [53,54]. The vegetation in the area includes a variety of crops and arboriculture, such as Cynara scolymus L., Triticum aestivum L., Malus sylvestris Mill., and Vitis vinifera L., and these species rely heavily on irrigation from the Medjerda river and Chafrou groundwater. The study area (Figure 1) is situated within a Mediterranean bio-climate, characterized by warm and temperate conditions, with an annual average temperature of approximately 20 °C. The Jedeida district receives a mean annual precipitation of 504 mm [54]. Previous studies of the Manouba province indicate a predominance of Fluvisol formations, characterized by limestone series dating from the continental Mio-Pliocene to the middle-upper Pleistocene [55,56]. The region is mostly occupied by quaternary alluvium consisting of an alternation of clay deposits and silt, which varies based on topography. There are three major categories of soils in this region: Lithosols, Cambisols, and Calcisols. Lithosols are widespread in the shallow foothills, along the Medjerda river, well-drained and not suitable for irrigation. Cambisols and Calcisols, which occupy the lower part of the Chafrou watershed, are developed on silt–argillaceous materials with a clayey texture and hydromorphic properties.

Salinization processes have been observed in these soils due to irrigation with saline water, which contains elevated concentrations of dissolved salts [53,54]. The conversion of these flattening lands from wheat cultivation to rainfed fields, mainly transformed to gardens, orchards, and vineyards, has led to a decline in soil quality. This land use change reduced organic matter and disrupted the soil structure because the lower biomass returns and more intensive management in perennial systems caused disruption.

Random soil sampling was used in agricultural parcels irrigated by water from the Medjerda river and the Chafrou watershed to ensure representative coverage of soil conditions across the region and to capture the natural variability in soil properties influenced by human activities. A total of 20 surface soil samples (0–20 cm depth) were collected in December 2022 and June 2023 from the Zahira and Mansoura regions, respectively, to assess their suitability for annual crop production. The sampling was carried out during the humid seasons when soil is influenced by rainfall and irrigation is reduced. June corresponds to the dry season, which is characterized by intensive evaporation and irrigation. This allows the seasonal impacts on soil quality to be assessed using a shovel in different land use types and in areas with different sources of irrigation water, reflecting the diversity of practices in Zahira and Mansoura. Farmers rely on a mix of surface water (rivers, drains), groundwater, and river water treated with chemical products to reduce salinity, all of which have distinct chemical characteristics. Following a multi-step preparation process, samples were air-dried and sieved to 2 mm for various analyses, as shown in

Table 1. Soil and water analysis methods used in the laboratory.

Soil Properties	Methods of Analysis
Texture (Sand, Silt, and Clay Content)	Twenty grams of air-dried surface soil were mixed with 20 mL of 30% hydrogen peroxide (H2O2) in a 250 mL Erlenmeyer flask to oxidize the organic matter. The mixture was left to react at room temperature for 48 h and then heated on a hot plate at approximately 90–100 °C for 3 h to complete the oxidation process. After cooling, 50 mL of distilled water was added, and the suspension was filtered using Whatman No. 42 filter paper to remove residual debris. The remaining soil was transferred to a clean beaker, and 10 mL of 1 M potassium chloride (KCl) was added, followed by 40 mL of a sodium hexametaphosphate solution (25 g.L−1). To adjust the pH and enhance dispersion, a few drops of 0.1 N ammonia (NH3) were added. The mixture was left to stand for 24 h after being stirred thoroughly and having distilled water added to bring the total volume to 100 mL. After dispersion, 20 mL of the suspension was transferred to a pre-weighed container, dried at room temperature until it reached a constant mass, and weighed. The sample was then oven-dried at 150 °C for 2 h and weighed again. The mass difference between the room-dried and oven-dried samples was used to estimate the number of dispersed particles [57].
Soil and Water pH (Suspension (1:2.5 Soil/Water))	To determine the pH of the water directly in situ, a CyberScan pH 110 Portable pH/mV Meter (Eutech Instruments, Singapore, Singapore) calibrated pH meter was used to measure the pH of the water sample at the site without any prior treatment. For soil pH determination, 20 g of air-dried soil was weighed and placed into a beaker and then mixed with 50 mL of distilled water. To ensure a uniform suspension, the mixture was thoroughly stirred and then left to settle undisturbed for 2 h at room temperature. After settling, the suspension was filtered using Whatman No. 42 filter paper and the clear filtrate was collected. The pH of the filtration was then measured with the CyberScan pH 110 Portable pH/mV Meter. The pH meter was allowed to stabilize for approximately one minute, and then, the reading was taken.
ECe (Saturated Paste (SP) Method)	A 200 g mass of air-dried soil was gradually moistened with distilled water in a container, and thoroughly mixed until a saturated paste consistency was achieved. The saturated paste was then covered to prevent evaporation and left to equilibrate for 24 h at room temperature. This allowed the soluble salts to fully dissolve into the pore water. After equilibration, the pore water was extracted by suction filtration using a Buchner funnel lined with Whatman No. 42 filter paper. This solution, representing the soil saturation extract, was collected, and its electrical conductivity was measured immediately using a calibrated bench-type conductivity meter Model: BCT-4308 (Lutron Electronic Enterprise, Taipei, Taiwan) [58].
EC (Water)	A direct in situ measurement via a bench-type conductivity meter (Model: BCT-4308).
Soil Organic Matter (SOC)	The soil organic carbon content in the soil was determined by placing 1 g of air-dried soil into a conical flask and mixing it with 10 mL of 1 N potassium dichromate (K2Cr2O7) solution. Then, 20 mL of concentrated sulfuric acid (H2SO4) was added to initiate the oxidation reaction. The mixture was then gently swirled for 1 min and left to react for 30 min at room temperature. A standardization blank (without soil) was run at the same time. Following the reaction period, a small quantity of silver sulfate crystals (approximately 0.5 g) was added to catalyze the oxidation. After half an hour, 200 mL of distilled water was added, along with 0.5 mL of ferroin indicator (typically 0.025 M in 1,10-phenanthroline and iron (II) sulfate). The solution was titrated with 0.5 N ferrous ammonium sulfate [Fe (NH4)2(SO4)2·6(H2O)2] until the color changed from olive green to reddish-brown, indicating the endpoint. The volume of titrant consumed was used to calculate the amount of oxidizable organic carbon [59].
Total Nitrogen Content (TN)	The Kjeldahl digestion–distillation method was used to determine the total nitrogen content in the soil. A 1 g mass of finely crushed air-dried soil was placed into a digestion tube. A pinch of selenium powder (approximately 0.1 g) was added. This was used as a catalyst. Then, 5 mL of 30% sulfuric acid (H2SO4, ~5.4 M) was added under a fume hood to initiate digestion. The mixture was heated at 300 °C for 10 min to ensure the complete mineralization of the organic nitrogen into ammonium sulfate. Once digestion was complete, the tube was left to cool to room temperature, after which 20 mL of distilled water was added to dilute the digestate. The solution was then transferred to a distillation unit. To monitor the titration endpoint, we added five drops of methyl red indicator (typically 0.02% w/v in ethanol). During distillation, if boric acid was used, the ammonia released was captured in this, and the final solution was titrated with standard 0.01 N sulfuric acid until the color changed from green to pink, indicating the endpoint. The total nitrogen content in the soil sample was calculated using the amount of acid that had been consumed [60].
Iron (Fe)	A 0.5 g mass of finely ground, air-dried soil was placed in a Teflon digestion beaker. Under a fume hood, 10 mL of hydrofluoric acid (HF, 48%) and 5 mL of perchloric acid (HClO4, 70%) were added to the sample to initiate complete mineral digestion. The beaker was loosely covered and left to react for 24 h at room temperature to allow the silicate matrices to break down gradually. The mixture was then heated on a digestion plate at approximately 150–200 °C for 15 min to accelerate the reaction. After cooling slightly, 2 mL of nitric acid (HNO3, 65%) was added, and the solution was stirred and reheated for an additional 15 min to ensure the complete oxidation and dissolution of iron compounds. Once the reaction had finished and the sample had reached boiling point, the digest was cooled, transferred quantitatively into a 100 mL volumetric flask, and then diluted to volume with deionized water. The iron concentration in the final solution was measured using a Bench atomic absorption spectrophotometer (AAS-900), manufactured by Labtron Scientific Ltd., Grand Rapids, MI, USA.
Calcium Carbonate Content (CaCO3)	The Bernard calcimeter method was used to determine the calcium carbonate content. One gram of finely ground, air-dried soil was placed in the reaction flask of the calcimeter. Then, 10 mL of concentrated hydrochloric acid (HCl, 0.5 N) was added to the flask to initiate the reaction with calcium carbonate. As the acid reacted with the carbonate compounds in the soil, carbon dioxide gas (CO2) was released. The volume of CO2 produced was measured using a burette connected to the calcimeter setup. This volume was directly proportional to the amount of calcium carbonate present in the sample.
Cation Exchange Capacity (CEC)	A 2 g mass of sieved, air-dried soil was placed in a clean container and mixed with 25 mL of a 10% barium acetate solution (0.4 M). The mixture was stirred vigorously for 90 s to saturate the exchange sites with barium ions. The mixture was then centrifuged at 300 rpm for five minutes, and the supernatant was discarded. The residue was subsequently treated with 25 mL of a 0.1 M magnesium sulfate solution (MgSO4) to displace the adsorbed barium ions. After stirring for a further 90 s, the mixture was centrifuged again under the same conditions. A 10 mL volume of the resulting extract was collected and diluted with 150 mL of distilled water. To buffer the solution, 10 mL of an ammonium chloride (NH4Cl, 0.1 M) buffer solution at pH 10 was added, followed by a few drops of an Eriochrome Black T indicator solution (0.5% w/v in ethanol). Initially pink due to the presence of magnesium ions, the solution was titrated with 0.01 M EDTA (ethylenediaminetetraacetic acid) until it changed color from pink to blue, indicating the endpoint. The volume of EDTA consumed was then used to calculate the soil cation exchange capacity [61].
Water and soils Calcium (Ca2+), Magnesium (Mg2+), Sodium (Na+), Potassium (K+), Carbonates (CO32−), Bicarbonates (HCO3−), Chloride (Cl−), Sulfates (SO4−)	To determine the concentrations of calcium (Ca2+) and magnesium (Mg2+) in the water and soil extracts, complexometric titration using ethylenediaminetetraacetic acid (EDTA) was employed. A 10 mL sample was buffered to pH 10 using 10 mL of ammonium chloride buffer solution. A few drops of 0.5% w/v Eriochrome Black T indicator solution in ethanol was added, producing a pink coloration in the presence of divalent cations. The solution was then titrated with 0.01 M EDTA until the color changed from pink to blue, which indicated the endpoint. Titration was performed using a burette and magnetic stirrer, with the pH monitored using a CyberScan pH 110 Portable pH/mV Meter. The sodium (Na+) and potassium (K+) concentrations were determined using the BWB XP Plus Flame Photometer (BWB Technologies, Newbury, UK). Samples were filtered and diluted appropriately prior to analysis. Calibration curves were prepared using standard sodium and potassium solutions (1000 mg/L stock solutions, serially diluted). Carbonates (CO32−) and bicarbonates (HCO3−) were quantified by acid–base titration. A 25 mL sample was titrated with 0.1 N hydrochloric acid (HCl), using phenolphthalein indicator (0.5% w/v in ethanol) to detect the carbonate endpoint, indicated by a color change from pink to colorless. Following this, methyl orange indicator (0.1% w/v in water) was added to determine the bicarbonate endpoint. The bicarbonate endpoint was marked by a color change from orange to pink. Titrations were performed using a burette and monitored with a CyberScan pH 110 Portable pH/mV Meter to confirm the accuracy of the endpoints. Chloride (Cl−) concentration was measured using the Mohr argentometric method. A 25 mL volume of the sample, placed in a beaker, was titrated with 0.1 N silver nitrate (AgNO3) in the presence of a 5% w/v potassium chromate indicator. The endpoint was indicated by a color change from yellow to reddish-brown, signaling the formation of silver chromate (Ag2CrO4) after all the chloride ions had precipitated as silver chloride (AgCl). Sulfate (SO42−) was determined using a turbidimetric method by mixing 10 mL of sample with 20 mL of conditioning reagent containing 30 g/L sodium chloride (NaCl) and 5 mL/L glycerol, followed by the addition of 1 mL of 0.25 M barium chloride (BaCl2) under stirring. The absorbance was measured at 420 nm using a Bench atomic absorption spectrophotometer (AAS-900). The sulfate levels were calculated from a calibration curve prepared with Na2SO4 standards ranging from 0 to 40 mg/L [62].
Water Nitrate (NO3-N) Ammonium (NH4-N)	The concentration of nitrate (NO3−-N) in water samples was determined using the cadmium reduction method. A 10 mL water sample was treated with a powdered reagent containing cadmium particles and sulfanilic acid (commercially known as NitraVer® 6). In this method, nitrate (NO3−) is chemically reduced to nitrite (NO2−) by cadmium. The nitrite then reacts with sulfanilic acid under acidic conditions to form a diazonium salt, which couples with a second aromatic compound to produce pink azo dye. After 5 min of reaction time, absorbance was measured at 543 nm using a Double Beam UV-Visible Spectrophotometer (Model: UV1720) manufactured by Shanghai Yoke Instrument Co., Ltd., Shanghai, China. The nitrate concentrations were calculated from a standard calibration curve. Ammonium (NH4+-N) was measured by colorimetry using the indophenol blue reaction. A 10 mL volume of water sample was mixed with 1 mL of alkaline phenol solution (containing 0.5% w/v phenol and 0.25% w/v sodium nitroprusside) and 1 mL of sodium hypochlorite solution (0.5% available chlorine) in a beaker. The mixture was allowed to react for 1 h at room temperature to form an indophenol blue-colored complex. The absorbance was measured at 640 nm using a Double Beam UV-Visible Spectrophotometer (Model: UV1720). The ammonium concentrations were determined from a calibration curve prepared using ammonium chloride standards ranging from 0.1 to 10 mg/L NH4+-N.

. To account for seasonal variations in water quality, samples were additionally collected in December 2022 and June 2023, which matched the soil sampling dates. Twelve water samples were collected from various sources in the Zahira region, including drains, wells, groundwater, and aquifers, as well as irrigation systems fed by the Medjerda River. In Mansoura, 11 water samples were taken from similar sources, including wells and the Chafrou River (Figure 1).

Soil quality indicators encompassing 16 properties were identified and categorized into three groups: (1) physical properties, including the percentage of clay, silt, and sand; (2) chemical properties, including soil organic matter (SOC), soil pH (1: 2.5 w/v), electrical conductivity (EC, dS.m⁻¹) by saturated paste (ECe), calcium carbonate content (CaCO₃), and cation exchange capacity (CEC); and (3) soil nutrient properties, including total nitrogen (TN) and plant nutrients (Na, Ca, Mg, K). These parameters were measured in soil samples to establish a soil quality index (SQI) and to evaluate the influence of irrigation water on the SQI in these regions. The SQI is a metric that has been used to quantify soil health and entails three primary steps: selecting soil indicators, assigning scores through standard functions, and integrating them into an SQI using the weighted method [63,64] The total soil data set was statistically reduced to a minimum soil data set (MSD), using principal component analysis (PCA), due to its objectivity and flexibility as mentioned in the previous studies of [14,50,65,66,67].

Only principal components (PCs) with an eigenvalue ≥ 1 were considered meaningful for the MSD, and only soil properties with loading values ≥ 0.5 were considered [68]. For each PC, the soil indicators with a loading value within the top 10% of scores in each group were retained for the MSD. A representative indicator must be chosen to reflect soil functions [66,69].

Pearson’s correlation analysis was also used to identify redundant soil indicators [70]. Linear (L) and non-linear (NL) scoring functions are widely applied to normalize the MSD from different units in to a unitless soil indicator scale [66,70,71]. According to [65], indicators were scored based on their impact on soil quality. A good indicator with a positive effect on soil quality was labeled as “More is better”, whereas that with a negative effect was labeled “Less is better”. If it had both positive and negative effects, it was assigned an ‘Optimum’ score. This criterion defines a range of parameters that support soil fertility. Values outside this range may lead to soil deterioration [14,63,72]. The following linear scoring functions (Equations (1) and (2)) were used as “more is better” and “less is better”, respectively [73]. Equation (3) represents the non-linear (NL) scoring model [28].

$$\mathit{S}\text{\_}\mathit{L}={\displaystyle \raisebox{1ex}{$x$}\!\left/ \!\raisebox{-1ex}{$xmax$}\right.}$$

(1)

$$\mathit{ }\mathit{S}\text{\_}\mathit{L}={\displaystyle \raisebox{1ex}{$xmin$}\!\left/ \!\raisebox{-1ex}{$x$}\right.}$$

(2)

where L is the linear score varying from 0 to 1, x is the value of the soil property, and x_max and x_min are the maximum and minimum values of each observed soil property [72].

$$\mathit{ }\mathit{S}\text{\_}\mathit{N}\mathit{L}={\displaystyle \frac{a}{1+\left(\raisebox{1ex}{$x$}\!\left/ \!\raisebox{-1ex}{$x_0$}\right.\right)b}}$$

(3)

where NL is the score of the soil indicator, a is the maximum score reached by the function (a = 1), x is the measured value of the soil indicator, x₀ denotes the mean of each soil indicator, and “b” refers to the slope of the equation. The b is set to −2.5 and +2.5 for a “more is better” and “less is better” curve, respectively [64,74].

The weight (Wi) was assigned based on the variance of each PC obtained through PCA. Each PC explained a certain percentage of the variation in the TDS. The weight additive index of each soil property chosen by the PC is obtained by dividing the percentage explained by each PC by the total percentage variation explained by all the PCs with eigenvalues (EV) > 1 [27].

The weighted SQI indicator scores were calculated with Equation (4).

$$\mathit{ }\mathit{S}\mathit{Q}\mathit{I}={\sum }_{i=1}^n{W}_i\times S_i$$

(4)

S_i is a non-linear or linear scoring function of the indicators, n is the number of variables, and W_i is the weight of variables derived from the PCA.

Regulated water quality indices (WQIs) were calculated, including the sodium adsorption ratio (SAR), soluble sodium percentage (SSP), permeability index (PI), and Kelly’s ratio (KR), to evaluate the suitability of irrigation water and its potential risks to soil permeability, sodicity, and plant health [19,20,21]. The lowest WQI value indicates water quality that is suitable for irrigation.

These indices evaluate the balance between the sodium (Na⁺), calcium (Ca²⁺), potassium (K⁺), magnesium (Mg²⁺), and bicarbonates (HCO₃⁻) concentrations in (meq/L) using the following Equations (5)–(8):

$$\mathit{K}\mathit{R}={\displaystyle \frac{Na⁺}{C{a}^{2+}+M{g}^{2+}}}$$

(5)

$$\mathit{S}\mathit{A}\mathit{R}={\displaystyle \frac{Na⁺}{\sqrt{\frac{C{a}^{2+}+M{g}^{2+}}{2}}}}$$

(6)

$$\mathit{S}\mathit{S}\mathit{P}={\displaystyle \frac{Na⁺}{N{a}^{+}+C{a}^{2+}+M{g}^{2+}+K⁺}}\times 100$$

(7)

$$\mathit{P}\mathit{I}={\displaystyle \frac{N{a}^{+}+\sqrt{HC{O}_3^{-}}}{N{a}^{+}+C{a}^{2+}+M{g}^{2+}}}\times 100$$

(8)

The sodium levels are well-balanced with calcium and magnesium, which minimizes the risk of soil degradation. However, the highest values may pose challenges for irrigation, combined with excessive sodium; this can lead to the deterioration of soil structure and reduced water-holding capacity [20,75].

2.2. Statistical Analysis and SQI Mapping

Soil and water data were analyzed using various statistical methods, including descriptive statistics, principal component analysis, and the Pearson’s correlation test, applied at the p < 0.05 significance level. All were performed using Microsoft Excel (2019), R programming software (version 4.4.1), and Statistica software (version 8.0).

Spatial distribution maps of the soil quality index (SQI) values were generated using ArcGIS 10.8 software to visualize the spatial variability of soil quality across the study area. The inverse distance weighting (IDW) interpolation method was employed to estimate unknown SQI values at unsampled points. This method is based on the idea that data points closer together have a stronger impact on the estimated values than those that are farther apart.

3. Results

3.1. Soil Quality

Table 2. Descriptive statistics of soil properties in the Zahira region.

Soil Variables	Minimum	Maximum	Mean	Std. Dev	CV%	Kurtosis	Skewness
pH	7.12	8	7.38	0.07	0.95	5.54	2.03
ECe (dS.m−1)	1.17	5.89	2.69	0.53	19.70	0.43	1.26
CaCO3 (%)	11.87	13.47	12.46	0.16	1.28	−0.36	0.66
Clay (%)	41	55	45.97	1.35	2.94	1.01	0.74
Silt (%)	15.5	39.25	25.5	2.46	9.65	−0.56	0.47
Sand (%)	7.85	35.86	26.04	2.62	10.06	1.56	−1.09
SOC (%)	1.23	2.85	1.90	0.18	9.47	−1.35	0.38
TN (%)	0.02	0.20	0.12	0.01	8.33	−0.73	0.09
CEC (meq.100 gr−1)	11.87	31.25	23.87	1.86	7.79	0.61	−0.88
Na (mg.kg−1)	3.05	29.4	11.35	3.03	26.69	1.03	1.56
Mg (mg.kg−1)	2	11	5	0.89	17.80	1.15	1.21
Ca (mg.kg−1)	3	19	9.60	1.94	20.21	−1.36	0.68
SO4 (mg.kg−1)	0.17	15.1	5.85	1.57	26.84	−0.31	0.71
Cl (mg.kg−1)	5.46	40.89	16.63	4.26	25.61	0.50	1.37
Fe2O3 (mg.kg−1)	0.02	0.23	0.17	0.02	11.76	0.96	−1.26

and

Table 3. Descriptive statistics of soil properties in the Mansoura region.

Soil Variables	Minimum	Maximum	Mean	Std. Dev	CV%	Kurtosis	Skewness
pH	7.16	7.83	7.53	0.25	3.32	−1.60	−0.03
ECe (dS.m−1)	0.694	8.15	0.694	3.846	554.17	2.04	0.82
CaCO3 (%)	31.38	41.54	36.88	3.62	9.82	−1.13	−0.33
Clay (%)	27.00	53.00	43.33	8.79	20.29	−0.14	−0.88
Silt (%)	23.00	45.00	32.89	7.47	22.72	−0.45	0.41
Sand (%)	7	50.00	23.78	14.94	62.83	−0.63	0.82
SOC (%)	0.83	3.47	1.52	0.81	53.29	4.93	2.12
TN (%)	0.06	0.28	0.12	0.07	58.33	1.80	1.46
CEC (meq.100 gr−1)	16.25	31.25	25.83	5.01	19.55	0.02	−0.89
Na (mg.kg−1)	2.09	36.75	16.81	12.49	74.30	−1.19	0.54
Mg (mg.kg−1)	5.67	60.35	28.33	17.65	62.31	−0.21	0.42
Ca (mg.kg−1)	73.89	514.27	280.07	150.65	53.79	−0.92	−0.24
K (mg.kg−1)	4.60	22.65	13.80	6.51	47.17	−0.86	−0.15

summarize the descriptive statistics for the soil parameters of the studied regions. In general, the soils in the Zahira and Mansoura regions have neutral to slightly basic pH levels varying between 7.4 and 8. The soil is mostly saline, with an ECe exceeding 0.5 dS.m⁻¹. However, the ECe increases slightly near irrigation water sources and in areas with poor drainage due to the flat topography. The soil organic carbon (SOC) was slightly higher at >2.8% in Mansoura lands but still insufficient at <2% in the majority of Zahira soils.

The soil texture varies from clayey to clay loamy. The average calcium carbonate (CaCO₃) content ranges from 12.46 to 36.88%. For CEC, Mansoura soils have an average of 25.83 (meq.100 gr⁻¹), while Zahira soils have an average of 23.87 (meq.100 gr⁻¹). Regarding nutrient elements, the total nitrogen (TN) content shows a similar trend in both regions, with an average value of 0.12%. The average iron content for Zahira is 0.17 mg.kg⁻¹, which is caused by excessive fertigation. The same is observed for sodium, magnesium, calcium, sulfates, and chloride, which present the lowest average values of 11.35 (mg.kg⁻¹), 5.0 (mg.kg⁻¹), 9.60 (mg.kg⁻¹), 5.85 (mg.kg⁻¹), and 16.63 (mg.kg⁻¹), respectively, in Zahira soils. Mansoura soils present the highest average sodium, magnesium, calcium, and potassium, at 16.81 (mg.kg⁻¹), 28.33 (mg.kg⁻¹), 280.07 (mg.kg⁻¹), and 13.80 (mg.kg⁻¹), respectively.

The PCA results for both regions are shown separately in

Table 4. Principal component analysis (PCA) of soil quality indicators, eigenvalues, and component matrix variables for soil profiles of the Ap horizon of the Zahira region.

Soil Properties	Factor 1	Factor 2	Factor 3	Factor 4	Factor 5
pH	0.039	0.097	0.485	0.837	−0.130
ECe (dS.m−1)	−0.982	−0.144	−0.039	0.100	−0.023
CaCO3 (%)	−0.225	0.153	0.319	−0.513	−0.671
Clay (%)	−0.284	0.180	−0.735	−0.172	0.494
Silt (%)	0.023	0.893	0.142	0.299	0.092
Sand (%)	0.014	−0.897	0.160	−0.033	−0.352
SOC (%)	0.460	−0.610	−0.544	0.205	−0.019
TN (%)	−0.086	0.426	−0.627	−0.441	−0.248
CEC (meq.100 gr−1)	−0.135	0.126	−0.744	0.322	−0.447
Na (mg.kg−1)	−0.935	−0.184	−0.183	0.182	−0.072
Mg (mg.kg−1)	−0.938	−0.110	−0.092	0.165	0.121
Ca (mg.kg−1)	−0.922	−0.146	0.217	0.130	0.026
SO4 (mg.kg−1)	−0.976	0.042	−0.072	−0.066	0.084
Cl (mg.kg−1)	−0.965	−0.218	0.006	0.133	−0.042
Fe2O3	−0.053	0.612	−0.428	0.338	−0.401
Eigenvalue	6.047	3.127	2.745	1.694	1.298
Variance (%)	37.797	19.548	17.159	10.591	8.118
Cumulative variance (%)	37.797	57.346	74.505	85.096	93.215

The bold underlined values were the factors included in the MSD. The bold loading values (>0.50) indicate the highly weighted factors.

and

Table 5. Principal component analysis (PCA) of soil quality indicators, eigenvalues, and component matrix variables for soil profiles of the Ap horizon of the Mansoura region.

Soil Properties	Factor 1	Factor 2	Factor 3	Factor 4	Factor 5
pH	−0.456	−0.378	−0.692	−0.038	0.078
ECe (dS.m−1)	0.569	0.548	0.490	0.290	−0.026
EC 1:5 (dS.m−1)	−0.116	0.678	0.471	−0.196	−0.159
CEC (meq.100 gr−1)	−0.233	0.071	−0.496	−0.243	−0.783
OM (%)	−0.006	0.788	−0.563	−0.125	0.198
CaCO3 (%)	−0.796	0.106	0.486	0.241	−0.012
Na (mg.kg−1)	−0.775	0.257	0.005	0.553	0.079
Mg (mg.kg−1)	−0.877	0.026	−0.042	0.450	−0.015
Ca (mg.kg−1)	−0.902	−0.109	−0.029	0.340	−0.175
K (mg.kg−1)	−0.871	0.161	−0.052	0.039	−0.106
SOC (%)	−0.006	0.789	−0.563	−0.124	0.198
TN (%)	0.775	0.107	0.189	−0.044	−0.408
Clay (%)	−0.848	0.220	0.217	−0.375	−0.148
Silt (%)	−0.533	−0.188	0.268	−0.699	0.259
Sand (%)	0.765	−0.034	−0.262	0.571	−0.042
Eigenvalue	6.37	2.36	2.29	1.86	1.03
Variance (%)	42.47	15.78	15.27	12.40	6.88
Cumulative variance (%)	42.47	58.25	73.52	85.92	92.8

The bold-underlined values were the factors included in the MSD. The bold loading values (>0.50) indicate the highly weighted factors.

. The bold loading soil properties were chosen within 10% of the factor loading; the overall cumulative percentage varied between 92 and 93%.

For the Mansoura region, parameters with high factor loading in same principal component (PC) retained through PCA were correlated using the Pearson correlation matrix (

Table 6. Correlation coefficients (Pearson’s) for highly loaded parameters in PC1 for Mansoura soils.

	ECe (dS.m−1)	CaCO3 (%)	Na (mg.kg−1)	Mg (mg.kg−1)	Ca (mg.kg−1)	K (mg.kg−1)	TN (%)	Clay (%)	Sand (%)
ECe dS.m−1	1
CaCO3 (%)	−0.12	1
Na (mg.kg−1)	−0.14	0.78 **	1
Mg (mg.kg−1)	−0.34	0.76	0.94 **	1
Ca (mg.kg−1)	−0.46	0.73	0.84	0.96 **	1
K (mg.kg−1)	−0.36	0.61	0.68	0.80	0.84 **	1
TN (%)	0.68 **	−0.58	−0.62	−0.64	−0.61	−0.53	1
Clay (%)	−0.35	0.72	0.49	0.57	0.62	0.77 **	−0.49	1
Sand (%)	0.43	−0.58	−0.29	−0.41	−0.49	−0.65	0.48	−0.93 **	1

** Correlation coefficients are significant at p < 0.02.

). Soil indicators retained from the MSD were normalized using linear (L) and non-linear (NL) scoring functions. Normalized SQIs indices for both regions using weighted additive method were presented in

Table 7. Summary of average and range values and normalized linear and non-linear equations of scoring functions for all MSDs in the Zahira region.

	ECe	pH	CaCO3	Sand	CEC
Min	1.17	7.12	11.78	7.85	11.87
Max (Xmax)	5.89	8	13.47	35.86	31.25
Average (X0)	2.69	7.38	12.46	26.04	23.87
Curve Type	Less is better	Optimum	Optimum	Optimum	More is better
Weighting Factor a	0.40	0.11	0.08	0.20	0.18
Slope (b Value)	+2.5	−2.5	−2.5	−2.5	−2.5
Linear Equation	$W\times \left({\displaystyle \raisebox{1ex}{$x$}\!\left/ \!\raisebox{-1ex}{$5.89$}\right.}\right)$	$W\times \left({\displaystyle \raisebox{1ex}{$x$}\!\left/ \!\raisebox{-1ex}{$7.12$}\right.}\right)$	$W\times \left({\displaystyle \raisebox{1ex}{$x$}\!\left/ \!\raisebox{-1ex}{$13.47$}\right.}\right)$	$W\times \left({\displaystyle \raisebox{1ex}{$x$}\!\left/ \!\raisebox{-1ex}{$35.86$}\right.}\right)$	$W\times \left({\displaystyle \raisebox{1ex}{$x$}\!\left/ \!\raisebox{-1ex}{$31.25$}\right.}\right)$
Non-Linear Equation	${\displaystyle \raisebox{1ex}{$W\times 1$}\!\left/ \!\raisebox{-1ex}{$\left(1+{\left(\frac{x}{2.69}\right)}^{+2.5}\right)$}\right.}$	${\displaystyle \raisebox{1ex}{$W\times 1$}\!\left/ \!\raisebox{-1ex}{$\left(1+{\left(\frac{x}{7.38}\right)}^{-2.5}\right)$}\right.}$	${\displaystyle \raisebox{1ex}{$W\times 1$}\!\left/ \!\raisebox{-1ex}{$(1+{\left(\frac{x}{12.46}\right)}^{-2.5})$}\right.}$	${\displaystyle \raisebox{1ex}{$W\times 1$}\!\left/ \!\raisebox{-1ex}{$(1+{\left(\frac{x}{26.04}\right)}^{-2.5})$}\right.}$	${\displaystyle \raisebox{1ex}{$W\times 1$}\!\left/ \!\raisebox{-1ex}{$(1+{\left(\frac{x}{23.87}\right)}^{-2.5})$}\right.}$

^a Derived from principal component analysis (PCA) (please see Table 4).

and

Table 8. Summary of average and range values and normalized linear and non-linear equations of scoring functions for all MSDs in the Mansoura region.

	Ca	SOC	pH	Silt	CEC
Min	73.89	0.83	7.16	23	16.25
Max (Xmax)	514.27	3.47	7.83	45	31.25
Average (X0)	280.07	1.52	7.53	32.89	25.83
Curve Type	Optimum	More is better	Optimum	Optimum	More is better
Weighting Factor a	0.46	0.17	0.16	0.13	0.07
Slope (b Value)	−2.5	−2.5	−2.5	−2.5	−2.5
Linear Equation	$W\times \left({\displaystyle \raisebox{1ex}{$x$}\!\left/ \!\raisebox{-1ex}{$514.27$}\right.}\right)$	$W\times \left({\displaystyle \raisebox{1ex}{$x$}\!\left/ \!\raisebox{-1ex}{$3.47$}\right.}\right)$	$W\times \left({\displaystyle \raisebox{1ex}{$x$}\!\left/ \!\raisebox{-1ex}{$7.83$}\right.}\right)$	$W\times \left({\displaystyle \raisebox{1ex}{$x$}\!\left/ \!\raisebox{-1ex}{$45$}\right.}\right)$	$W\times \left({\displaystyle \raisebox{1ex}{$x$}\!\left/ \!\raisebox{-1ex}{$31.25$}\right.}\right)$
Non-Linear Equation	${\displaystyle \raisebox{1ex}{$W\times 1$}\!\left/ \!\raisebox{-1ex}{$(1+{\left({\displaystyle \frac{x}{280.07}}\right)}^{-2.5})$}\right.}$	${\displaystyle \raisebox{1ex}{$W\times 1$}\!\left/ \!\raisebox{-1ex}{$\left(1+{\left(\frac{x}{1.52}\right)}^{-2.5}\right)$}\right.}$	${\displaystyle \raisebox{1ex}{$W\times 1$}\!\left/ \!\raisebox{-1ex}{$\left(1+{\left(\frac{x}{7.53}\right)}^{-2.5}\right)$}\right.}$	${\displaystyle \raisebox{1ex}{$W\times 1$}\!\left/ \!\raisebox{-1ex}{$\left(1+{\left(\frac{x}{32.89}\right)}^{-2.5}\right)$}\right.}$	${\displaystyle \raisebox{1ex}{$W\times 1$}\!\left/ \!\raisebox{-1ex}{$(1+{\left({\displaystyle \frac{x}{25.83}}\right)}^{-2.5})$}\right.}$

^a Derived from principal component analysis (PCA) (please see Table 5).

.

Pearson’s correlation (Table 6) was used to assess the linear relationships among soil parameters with high loadings within the same principal component (PC) that were identified through PCA.

The SQI histogram of the Zahira region (Figure 2) includes both linear (SQI-L) and non-linear (SQI-NL) values, and shows three types of soil quality classes for the SQI-L. The D3 (0.90) and D6 (0.89) samples have the highest SQI-L values. These soils are categorized as “good”, while others were classified as moderate soil quality.

The D2 (0.54) and D8 (0.54) are in the lower class. The SQI-NL values were uniformly low (0.28–0.35). The SQI-NL values were uniformly low (0.28–0.35). This indicates that, while some individual properties are highly favorable, overall soil quality may need improvement, particularly when using alternative measurement parameters, to ensure long-term sustainability.

The SQI histogram of the Mansoura region (Figure 3) includes both linear (SQI-L) and non-linear (SQI-NL) values. We identified three types of soil quality classes for the SQI-L: High SQI-L values, presented by P8 (0.80) and P10 (0.81), indicate strong soil quality. Moderate to high SQI-L values were observed for samples P9 (0.67), S04 (0.68), S07 (0.70), and S08 (0.65).Low SQI-L values were presented by the P4 (0.45), P5 (0.46), P6 (0.41), and P7 (0.50) soil samples. The SQI-NL displays two soil quality classes: high SQI (P4, P5, P6, and P7), indicating soil stability with values varying from 0.93 to 0.97, and moderate soil quality (P8, P9, P10, S04, S07, and S08), with values ranging from 0.89 to 0.91.

Soil is distributed as follows across the Mansoura and Zahira irrigated regions: 20% have a very low quality (SQI < 0.4), while a further 60% show a moderate quality. Only 20% of the soil has a high SQI, ranging from 0.7 to 0.8. Some samples have SQIs above 0.8, indicating exceptionally high quality. By grouping the SQI values calculated for both regions, digital maps of soil quality across the irrigated zones were generated.

3.2. Water Quality

Zahira water samples show an increase in EC values ranging from 0.74 to 13.51 dS·m⁻¹. A slightly alkaline pH level, averaging around 7.75, is present. Similarly, Mansoura samples revealed high EC values, varying between 3.85 and 10.62 dS·m⁻¹, with an average pH level of 7.8. Figure 4a illustrates the geochemical facies of water collected from the Zahira region. A predominance of Ca²⁺ is shown in the cation diagram, while the anion diagram clearly demonstrates the prevalence of HCO₃⁻. It shows a calcium bicarbonate (Ca-HCO₃) facies for the Chafrou watershed, its drains, and the Medjerda River. According to Figure 4b, the Mansoura water samples denote a distinct hydrochemical facies: sodium chloride (Na-Cl).

The water quality indices (WQIs) are represented by the sodium adsorption ratio (SAR), soluble sodium percentage (SSP), permeability hazard indices (PI), and Kelly’s ratio (KR). Figure 5a shows that the water samples in the Jedaida district were classified as excellent, as the PI values were greater than 75% [21]. The results showed that the water samples had no permeability hazard potentials but had sodicity hazard potential. However, Figure 5b shows SAR values ranging from 2.27 to 26.02, with an average of over 15, indicating the unsuitability of irrigation water in the Medjerda river and Chafrou watershed. On the other hand, only seven out of the thirty-three water sources were suitable with SAR values < 10 [76].

The SSP values of the water samples from both regions ranged from 41.42 to 76.73% (Figure 6). Only three of the water samples (D5 and D10 from the Zahira region, and S3 from the Mansoura region) were in the permissible category (40–60%), while the majority of water sources were classified as poor (SSP > 64%). When combined with SSP and soil analysis, Kelly’s ratio is an effective index for assessing water quality, with values ranging from 3.94 to 33.9 for Zahira samples and from 1.26 to 3.34 for Mansoura samples (

Table 9. Kelly index for Zahira and Mansoura water samples.

Zahira Water Samples	KR	Mansoura Water Samples	KR
D1	13.67	S1	2.01
D2	18.61	S2	2.59
D3	19.76	S3	1.26
D4	29.75	S4	1.69
D5	3.94	S5	1.80
D6	33.90	S6	1.92
D7	9.58	S7	1.82
D8	16.63	S8	1.93
D9	11.50	S9	1.55
D10	13.20	S10	2.60
D11	10.19	S11	3.34
D12	8.90

). All the samples showed highest KR values > 1. This indicates unsuitability for irrigation [21,75].

The water contaminants measured in both regions (Figure 7) show a significant level of NO₃⁻ and NH₄⁺. We found a dominance of NO₃⁻ in water collected from drains, with values of 34.76%, 34.12%, and 18.03%, respectively, for D1-D2-D5. Water samples collected directly from the watershed show a dominance of NH₄⁺, with values of 15.24% and 33.34% for D3 and D8, respectively, in the Zahira region (Figure 7a). In contrast, Figure 7 b illustrates the contaminants values of water samples from the Mansoura region, where groundwater collected from S2 shows an important percentage of 11.87% of NO₃⁻. On other hand, treated water containing chemicals (Floband, Fertilizers) used by farmers in the large basin before irrigation shows an abundance of ammonia, with 23.27%, 19.27%, and 31.44%, respectively, for S5, S6, and S8 (Figure 7b).

The permissible concentration of nitrate in groundwater for drinking purposes is 10 mg/L; nitrate may pose a threat to human health at higher concentrations in water sources. Agricultural activities were responsible for the enrichment in NO₃⁻ in the groundwater via the excessive use of fertilizers on lands [77,78].

4. Discussion

This research evaluates the impact of the quality of irrigation water on soil under different crop types and arboricultural diversity in the Manouba province. A total of 16 soil parameters were laboratory-analyzed (silt, sand, clay, pH, EC _1:5, SOC, CEC, CaCO₃, Mg, Ca, SO₄, Cl, Na, TN, ECe, and Fe₂O₃) to assess soil quality in the Zahira and Mansoura regions, in the Jedaida districts. Even though this study focused on physico-chemical indicators, biological indicators such as microbial biomass carbon and microbial respiration could enhance the future assessment of soil and water quality in the region. While previous studies [40] have widely used pH as a soil indicator with an “Optimum” range, we applied this criterion to sand, silt, calcium, CaCO₃ content, and pH, as the MSD variables were positively correlated with the SQIs, based on the literature and local soil surveys. In this study, the CEC and SOC demonstrated a validation of “More is better” criteria. ECe was the only parameter negatively correlated with SQIs, with “Less is better” criteria. Regarding the PCA results, a difference in MSD contribution in the computing of the SQI is observed from one region to another depending on crop variations and soil types. The soil texture, especially silt and sand, was retained as an MSD for both regional soil samples. Jedaida soil is rich in loam and clay, and the sand content is increased due to the intensive irrigation and the movement of fine particles through the water and soil profile. These particles, settled in depth or carried away by drainage, lead to water stress for plants [64].

Based on previous studies [79], the excess of sodium and chloride in the water can be attributed to the dissolution of anhydrite, gypsum, and halite. Several researchers have discussed the high dissolution of carbonate minerals (calcite, dolomite) along the flow [80,81]. This is mainly due to the geology of the study area and the nature of soils crossed by carbonates, as well as the variation in the mineralogical composition of the water bedrock [82]. Indeed, the attribution of the “Optimum” criteria to normalize those indicators, particularly in soils with high sand content, requires more management strategies such as organic matter incorporation, crop selection, soil amendments, and mulching to improve water retention for both plant and soil health.

For the pH, Zahira and Mansoura soils show neutral to slightly alkaline results, labeled “Optimum” as scoring validation for this indicator. Consequently, alkaline conditions can interfere with nutrient availability, particularly affecting the mobility of essential nutrients like phosphorus [83], leading to nutrient deficiencies in plants, further impacting their growth [84]. Another soil parameter, calcium carbonates (CaCO₃), chosen as an MSD with “Optimum” criteria, influences soil structure and nutrient absorption. Studies such as [64,85] show that calcium carbonate serves as a source of nutrients for plant development and soil fertility, like calcium. Calcium (Ca²⁺), as one of the major nutrients, improves soil structure. Its importance to plant growth has led to it given an “Optimum” criterion. Our soils are classified as Calcisols [53], which are naturally rich in calcium carbonate. This is positively correlated with the characteristics of semi-arid and arid regions, where low rainfall allows calcium to accumulate and influences soil structure. Finally, CEC and SOC were retained in the MSD with the attribution of “More is better” criteria. A significant increase in SOC values was detected in irrigated perimeters due to the nature of crop cultivation. Additionally, variations in soil types of influence CEC values due to differences in clay content. Soils with higher clay content indicated higher CEC because clay minerals possess negative charges that improve cation retention [86]. ECe was the only soil parameter that correlated negatively with “Less is more” score attribution. The highest electrical conductivity values in surface soil layers were associated with the highest salt accumulation, caused by poor irrigation water quality. This indicator is considered the most important parameter for modeling the SQI in semi-arid regions, as it limits agricultural sustainability, inhibits soil microbial growth, and deteriorates soil health [40,87]. Furthermore, the slightly alkaline pH levels (7 to 8) recorded in both the water and soil samples suggest the prevalence of bicarbonates and carbonates, which lead to soil dispersion and reduced permeability. Salinity directly impacts soil health by causing soil particle dispersion and reducing water infiltration. This, in turn, creates osmotic stress for plants, disturbing their absorption of water and nutrients, thereby reducing crop yields [88]. In their studies, [89] mentioned that low organic matter content and nutrient deficiencies were observed in soil samples, emphasizing the need for farmers to improve soil management practices. The Mansoura water samples showed similar salinity levels, with EC values between 3.8 and 5.2 dS.m⁻¹.

While still high, the salinity levels were slightly lower compared to those of the Zahira water samples. Outstandingly high sodium concentrations were identified in samples like D6 (103 meq/L) and D4 (95 meq/L), posing a challenge for irrigation water quality in Zahira lands. The pH remained consistent at 7.7, reflecting a slightly alkaline behavior that may reduce soil permeability and aeration. The nitrate presence in water samples from the Mansoura region is indicative of human-induced pollution, a common problem in areas with intensive farming practices and industrial installation. These contaminants add an additional layer of complexity, posing a long-term risk to soil health as opposed to the primary agricultural runoff concerns, altering the soil microbial biomass. Consequently, soil fertility has decreased over time [90]. Based on the water quality indices and Kelly ratio results, D5 and D10 from Zahira and S3 from Mansoura water sources are suitable for irrigation, whereas other water sources are unsuitable for irrigation; high EC and KR values cause these problems unless managed extensively. According to the SQI results (Figure 2 and Figure 3), the linear method had better performance, with higher values for the SQIs compared to the non-linear method. This may be caused by the influence of the MSD-selected parameters on SQI computation, which aligns with some previous studies [14,88]. Studies conducted on the lower Yellow River in China by [91] and in Egypt by [92] found that using the PCA and the linear scoring method was the most accurate method for correlating crop types with the SQI. While the SQI in the Zahira region for D3 is 0.9, suggesting good soil quality, the high salinity (ECe = 5.52 dS·m⁻¹) and sodium levels could pose challenges for irrigation and crop productivity. The soil quality classification as “good” might depend on the specific crops cultivated in Jedaida district, such as artichokes and wheat, which are salt-tolerant.

Despite the high water EC in the Zahira perimeters, these herbaceous crops are quite resilient and still thrive with proper management due to the plant’s ability to tolerate salinity up to 6.1 dS.m⁻¹ values [93,94]. Anomalies like the soil sample D3 suggest that organic matter and CEC play a critical role in buffering the negative impact of saline water. Other studies have found that soils with high organic carbon can enhance microbial activity and nutrient cycling, reducing the adverse effects of salinity [95]. In addition, organic matter can counteract these effects by enhancing cation exchange capacity and buffering salt levels. The agricultural lands in Mansoura have the highest ECe values (7.64–8.15 dS.m⁻¹) and exceed tolerance levels, particularly for woody crops. This aligns with findings that high salinity can reduce fruit quality and threaten soil fertility. The soil samples (P9, P10) with low organic matter (1.42–1.68%) align with studies suggesting that compost and organic amendments improve nutrient availability [96]. Previous research shows that grapes are moderately salt-tolerant, with salinity thresholds up to 1.5 dS.m⁻¹, while apples are more sensitive [97,98].

The soil quality spatial distributions (Figure 8) show that soil quality is best estimated by the linear model. Understanding the spatial variability of soil parameters is crucial, as it offers useful information for farmers in connection with fertilizer use and agricultural practices [99]. The IDW method was used to interpolate the spatial distribution of the SQI for both regions of the Jedaida district, Manouba province, northeastern Tunisia [100,101].

Furthermore, this research provides the advantages of (i) a comprehensive understanding of soil quality by integrating physical and chemical indicators and water parameters, and (ii) surveying soil quality.

5. Conclusions

In this research, the soil quality index (SQI) was the most-effective index for evaluating the impact of irrigation water quality on soil fertility, thereby reducing the need for time-consuming process of sampling and analysis. This was achieved by integrating several techniques, including statistical methods, the minimum soil data set (MSD) method, and digital mapping. The results represent four soil quality indices (SQIs) developed using both linear (L) and non-linear (NL) scoring methods. The total data set (TDS) of 16 soil parameters was analyzed and reduced to key parameters via the MSD approach. Each soil indicator influences the SQI scoring for Zahira soils (ECe, sand, CaCO₃, pH, and CEC) and for Mansoura soils (Ca²⁺, silt, SOC, pH, and CEC). Although these indicators differ across regions, the calculated SQI demonstrates a strong correlation with the scores produced by the linear model. Spatial mapping clearly shows a general homogeneous distribution of soil quality, with 20% classified as low, 60% as moderate, and 20% as high. Furthermore, water quality indices (WQIs) were calculated to assess the effect of different water sources on soil. Water samples collected from the lands of Zahira and Mansoura revealed high levels of salinity and poor overall quality, with a sodium adsorption ratio (SAR) > 10 and a Kelly’s ratio (KR) > 1.

Most samples with a soluble sodium percentage (SSP) of less than 75% were unsuitable. They are not recommended for irrigating cereals, artichokes, or arboriculture in the Jedaida region. Combined with the soil results, these water sources could potentially degrade soil structure and obstruct plant growth over time. Finally, SQI and WQI are effective indices for evaluating the effects of irrigation water quality on soil health. For further studies, we recommend incorporating biological indicators, exploring SQI-WQI and crop field relationships, and employing innovative remote sensing techniques to enhance soil quality monitoring.