1 Introduction

All nations worldwide have prime goals for sustainable agricultural development by balancing inherent land resources with crop requirements, emphasizing resource optimization to achieve long-term sustainable production. Finding the ideal sites for sustainable agriculture is crucial because farmlands are constantly decreasing [1]. It will be necessary to use land evaluation techniques to create models that anticipate the appropriateness of a given piece of land for various forms of agriculture [2]. The process of assessing a specific land area's degree of fit for a certain use (forest, agriculture, recreation, etc.) is known as land-use suitability analysis. The appropriateness analysis is a decision-making process that considers environmental and socio-economic factors in addition to the land unit's inherent potential to support a particular use.

The traditional land suitability approaches, like the parametric method, Storie’s index, etc., are costly and time-consuming. These are operated based on biophysical parameters, and sometimes results are incomplete due to the non-accountability of short-term changes in input parameters [3]. The parametric model is entirely based on empirical expert judgements and standard mathematical models, and the classified land suitability classes are completely separate and discrete groups that are separated from each other by a distinguished and consistent range. In addition, the decision-making steps are very complex and complicated. These are forced to choose the MCDA method for land suitability analysis [4].

Spatial Multi-Criteria Decision Analysis (MCDA) is one of the most widely used numerical modelling tools for land-use decision-making. Combining all the significant spatial features, this technique creates a map with the ideal places for a specific kind of land use. The most common objective of MCDA is to determine which location is best for a particular land use. MCDM techniques are capable of problem analysis, alternative solution generation, and alternative evaluation. The main goal of these techniques is to make it easier for decision-makers to select the best application from the available options [5].

With both spatial as well as attribute data, GIS can do a wide range of activities and overlay many maps or layers [6]. Furthermore, models that may be created from a collection of theme maps can be created through the usage of geographical information science (GIS) [7]. For land suitability assessments and developmental operations, GIS is a tool for making decisions [8]. Moreover, decision-making processes are improved by using efficient mapping and visualization capabilities when the Analytic Hierarchy Process (AHP) is integrated into a GIS. This facilitates the creation of maps of land use suitability and maximizes land use planning. The combined application of Analytic Hierarchy Process (AHP) and Geographic Information System (GIS) is one of the efficient approaches for determining land-use suitability [9]. Saaty [10] created the AHP, which is a hierarchical paradigm based on criteria and options for describing complicated situations [11].

In a traditional soil survey, physiographic and geopedological methods define the mapping unit, and the soil attributes are noted at typical sites. Even though soil surveyors are highly aware of the spatial soil variability characteristics, it is important to consider this continuous variability to correctly assess soil characteristics because, in nature, soil values are extremely changeable spatially [12]. The conventional approach to soil examination and interpretation is time-consuming and labour-intensive, which makes it costly. Kriging is a geostatistical approach that is commonly acknowledged as a crucial method for spatial interpolation in land resource inventories [13, 14]. The spatial scale of the study region, the distance between sample points, and the spatial pattern of modelling semi-variograms are all significant factors considered when geostatistical methods are used to evaluate spatial distribution and variability. An evaluation procedure that takes into account factors including terrain, climate, soil, water resources, and environmental components is used to estimate the prospective appropriateness of a piece of land for agricultural use [15].

For the most part, GIS is a useful tool for processing as well as storing the enormous amounts of data required to map the suitability of the area for particular crop farming. Additionally, using spatial analysis techniques in GIS facilitates the assessment of the relative weights of multi-criteria elements, which are thought to play significant roles in evaluating land suitability [16]. Among the most beneficial GIS applications are those that involve mapping and analysing land-use suitability [17]. The spatial structure of geo-referenced variables can be studied and predicted using geostatistics, which can also be used to map soil qualities and comprehend their distribution in this area [18]. Enhancing sustainable land use and enhancing agricultural management strategies require a deeper comprehension of the spatial variability of soil parameters. This is because it offers a working system for comparing current and future observations. The AHP method integrated with GIS application settings was applied for agricultural land suitability evaluations on a number of case study locations across the world [19,20,21,22]. The Analytic Hierarchy Process (AHP) is a mathematical technique employed to address intricate problems that include the consideration of multiple criteria [23]. According to Forman and Gass [24], the decision maker's influence in choosing the criterion weights is not as significant in other approaches as it is in AHP methods. Utilizing multi-criteria decision analysis to ascertain the significance of the parameters yields more dependable outcomes, and the Analytic Hierarchy Process (AHP) is a more widely acknowledged approach in the academic literature [25].

The study area Manimutha nadhi watershed belongs to Kallakurichi district which is an emerging agricultural district. Around the study area, about four agro-industries are operated at present. Agriculture is an important sector of the district economy and about 30 per cent of the population is involved in the agricultural sector for their livelihood [26]. However, the potential of the study area is more than the current cropping system and cropping pattern. At this juncture, it is necessary to ensure sustainable production with suitable crops in the research region through advanced techniques. In view of the above, the present study has been formulated to assess the potential of AHP and GIS to evaluate agricultural land suitability in the Manimutha nadhi watershed, south India.

2 Materials and methods

2.1 Description of the study area

The watershed of Manimutha Nadhi, Tamil Nadu, India, has been selected as a study area for this research (Fig. 1). It stretches between North Latitudes 11° 28′ and 11° 42′, East Longitudes 79° 14′ & 79° 27′ with an area of 263.74 km2. The river Manimutha Nadhi starts from the southern slopes of the Kalrayan hills. Rainfall occurs in the study area during the northeast, southwest, and non-monsoon seasons. The region experiences tropical weather, with temperatures between 26 and 41 °C. The majority of the research region is made up of flat lands with a mild slope that runs east to west toward the sea. Extrusive and intrusive rocks, both basic and ultra-basic, cover the majority of the study region. Dark-looking dolorite dykes from the youngest basic intrusive. During sub–recent to recent times, there has been the deposition of alluvial soil and lateritic soil. The total population of the study area is estimated to be 4,707,970 out of which 206,075 are males while 201,895 are females with an average sex ratio of 980 and a total of 91,183 families were residing in the study area. Urban areas, housing 3.8% of the population, have a literacy rate of 79.8% and a sex ratio of 1004. The remaining 96.2% reside in rural areas with a literacy rate of 63.5% and a sex ratio of 979. The overall literacy rate is 64.09%, with male literacy at 63.4% and female literacy at 48.49. Out of the total population in the study area, 210,805 were involved in work activities. 78.6% of the worker describe their work as Main Work (Employment for more than 6 Months) while 21.4% were involved in marginal activity providing a livelihood for less than 6 months. About 210,805 workers engaged in Main Work and 52,957 were cultivators (owner or co-owner) and 81,909 were Agricultural labourers [27].

Fig. 1
figure 1

Study area map

Full size image

2.2 Soil sample data and environmental variables

Soil samples were randomly gathered from the Manimutha Nadhi watershed study region using a GPS map 76CSx, a Garmin global positioning system model. Samples of soil (0–15 cm) were taken from the top layer. After surface litter was extracted, 57 samples in total were gathered. The reconnaissance soil survey was performed prior to soil sample collection. The soil sampling location and numbers were decided based on the soil colour and texture. At every sampling point, a "V"-shaped cut had been made with the spade at a depth of 15 cm. With the GPS device, every location detail was captured. The air-dried soil samples were powdered using a wooden mallet. A 2 mm sieve was utilized to filter the soil sample, which was then sealed in an airtight container. The samples were examined for SOC, pH, EC, texture, available nitrogen, phosphorus, and potassium. A glass electrode pH meter was used to measure the pH of the soil [28]. Soil EC using a conductivity meter [29]. Wet oxidation was used to determine the SOC concentration [30], available nitrogen was determined alkaline KMnO4 method [31], available phosphorus by Olsen method [32], and available potassium by Neutral normal ammonium acetate approach [33]. The hydrometer method was used to determine the texture of the soil [34]. The annual average soil erosion was calculated using the RUSLE method [35, 36]. Annual average rainfall (AAR) and annual average temperature (AAT) data were collected from https://globalweather.tamu.edu/. The twenty years (1995–2014) of AAT and AAR were collected, averaged and used as input data. The slope map was generated using 1° Arc DEM by SRTM data from the Earth Explorer website.

2.3 Descriptive statistics of soil parameters

Descriptive statistics is an expedient tool that gives insight into information about the soil variable. Descriptive statistics variables like minimum, maximum, mean, SD (Standard Deviation), CV (Coefficient of variation), skewness, and kurtosis of soil variables were calculated with the data analysis add-in of Microsoft Excel® 2007. In addition, the Pearson correlation coefficient is also worked out to evaluate the relationship among the soil parameters.

2.3.1 Landuse and land cover of Manimutha Nadhi watershed

The land use land cover map of the Manimutha Nadhi watershed generated from Landsat 08 OLI dt 12.01.23 from https://earthexplorer.usgs.gov/. The ArcGIS 10.4 software used to generate Land use and Landcover map. The supervised classification technique—maximum likelihood method is employed to generate LULC map. It has five classes: forest, water body, barren land, urban settlements, and agricultural land. The forest (natural) and farming areas occupied 63.11 and 90.96 km2, respectively. The waterbody, barren land, and urban settlements occupied about 52.88, 100.29, and 33.27 km2, respectively. The land use/land cover map is presented in the Fig. 2.

Fig. 2
figure 2

Land use land cover map of Manimutha Nadhi watershed

Full size image

2.3.2 Thematic map of soil data

The outlier-removed data were used to create point layer data in QGIS 3.16 environments. The thematic map of each soil parameter is created using the kriging technique. This is one of the best and most commonly used techniques for fielding out data in unknown points. All the thematic maps, including climatic, topography, and soil parameters, have a resolution of 30 m.

2.3.3 Kriging

The geostatistical wizard uses the model created in the variogram modeler and curve-fitting interface to interpolate a surface. With the help of the quantity and kind of data points and the direction and distance to other known points, the model could efficiently extract spatial continuity information for unknown points. Kriging was utilized in order to overestimate large values and underestimate low ones by fitting a smooth model of spatial variability to the sample data and lowering the fit error to the sample data. By minimizing the inaccuracy caused by the variations in the fit of the spatial continuity to every local neighbourhood, kriging was used to create a smooth surface. By substituting the mean of the closed surrounding pixels around the sampled points and substituting it in the sampled point, the kriged surface was generalized for all parameters. The variance error had been employed to spot issues with the model parameters, the definition of the local neighbourhood, and the sample data. Thus, the image obtained finally provides information to the user regarding identifying the problematic areas.

Geostatistics was used to identify spatial patterns and produce the soil characteristics prediction maps. The data have been first checked for patterns using geostatistical analysis. The semi-variogram studies were performed on the soil variables before applying standard kriging interpolation. This is so because the semi-variogram model determines the interpolation function. To evaluate a semi-variogram, use the following Eq. (1) [37].

$$\gamma \left( h \right) = \frac{1}{2|N(h)} + \mathop \sum \limits_{N(h)}^{} \left( {Z\left( {x_i } \right) - Z\left( {x_j } \right)} \right)$$
(1)

Z (xi) & Z (xi + h) signified the sample values with distance h, and γ (h) is signified as the semi-variance utilized for the interval class h. N (h) is denoted as the number of sample value pairs within the distance interval h. The variogram can then be fitted with many models (linear, Circular, Exponential, Spherical, and Gaussian) as per the given Eqs. (2–6) [38].

$${\text{Linear model}}:\gamma \left( h \right) = C_0 + ch$$
(2)
$${\text{Circular model}}:\gamma \left( h \right) = C_0 + gh\alpha$$
(3)
$${\text{Exponential model}}:\gamma \left( h \right) = C_0 + C({1} - e - h/a)$$
(4)
$$\begin{aligned} {\text{Spherical model}}:\gamma \left( h \right) & = C_0 + C\{ {1}.{5}h/a - 0.{5 }\left( {h/a} \right){3}\} \left( {{\text{for}}\,\,h \, < \, a} \right) \\ & = C_0 + C\left( {{\text{for}}\,\,h \ge a} \right) \\ \end{aligned}$$
(5)
$${\text{Gaussian model}}:\gamma \left( h \right) = C_0 + C({1} - e\left( {h/a} \right){ 2})$$
(6)

The nugget variance (variation resulting from measurement mistakes) in these models is represented by the parameter C0, and a is signified as the range (Above this point, there is just a chance, not a spatial, similarity between the attributes in the 2 samples divided by a lag of h). The sill parameter, which is equal to C0 + C (in the case of spherical, Gaussian, and exponential models), is the greatest variance between samples divided by a lag at least equal to the range. The sample variogram's goodness-of-fit statistics % variance was used to determine the spatial pattern using the best of these models [39].

2.4 Executing multi-criteria evaluation (MCE) using AHP techniques

One of the MCDM techniques, AHP, was created by Professor Thomas L. Saaty to address challenging spatial issues [40, 41]. A mathematical technique called the Analytic Hierarchy Process (AHP) can be used to handle extremely complicated decision-making problems with several criteria, elements, and scenarios [42]. In order to help decision-makers incorporate a multitude of possibilities that represent the perspectives of relevant parties into a prospective or retrospective framework, MCDM techniques were created in the 1980s. ‘How to combine the information from many criteria to generate a single index of evaluation’ is the primary focus of this approach [40].

With a higher degree of confidence, complicated spatial decisions can benefit from the hierarchical structure of AHP [10]. Every problem in the AHP technique is approached using a hierarchical model that includes criteria, objectives, alternatives, and sub-criteria [43]. There are two possible ways in MCDM. Distinct results and conclusions of observations of similar entities to analyze and correlate. The pair-wise comparisons method and Eigenvalue technique are combined to form the AHP. It offers a fundamental numerical scale, ranging from 1 to 9, to calibrate priority quantitative and qualitative performances [44] (Fig. 3).

Fig. 3
figure 3

Flow chart of the study

Full size image

2.4.1 Generation of a hierarchy structure

The first step involves building a hierarchy representing the pertinent problem elements (criteria, sub-criteria, characteristics, and decision alternatives). At the top of this hierarchy is the mission or goal of the decision-making problem. The remaining layers contain more pertinent elements (criteria, subcriteria, characteristics, etc.) [45].

2.4.2 Development of a comparison matrix at each level of the hierarchy

Criteria pairs, sub-criteria pairs (sub-sub-criteria pairs and so on), and alternative pairs are compared in the second stage. The AHP creates a matrix of pair-wise comparisons by using a basic 9-point scale measurement to describe individual judgments or preferences [10]. The expert opinions determine the relative importance between criteria, questionnaires from the farmers, and previously published literature. The relation importance of criteria is based on a scale ranging from 1 to 9. By enabling independent assessments of each factor's contribution, these pair-wise comparisons streamline the process of decision-making [46]. A = [aij] n × n is described in pair-wise comparisons as follows using Eq. (7):

$$A = \left[ {\begin{array}{*{20}c} {a_{11} a_{12} a_{13} } & \cdots & {a_{1n} } \\ {a_{21} a_{22} a_{23} \ldots \vdots } & \ddots & {a_{2n} } \\ {a_{n1} a_{n2} a_{n3} } & \cdots & {a_{nn} } \\ \end{array} } \right]d$$
(7)

where it is required that aii = 1 and aij = 1/aji for all I and j. Following the formation of each pair-wise comparison matrix, Saaty's eigenvector approach is used to calculate the vector of weights, w = [w1, w2, w3…wn]. Next, Eq. (4) normalizes this eigenvector, and Eq. (5) computes the weights. The normalized eigenvector's constituents are assessed in relation to the alternatives and weighted in terms of the sub-criteria or criteria [47, 48]. The aij and wij can be calculated as per the Eqs. (8) and (9).

$$aij = \frac{aij}{{\sum_{j = 1}^n aij}}$$
(8)
$$wij = \frac{{\sum_{j = 1}^n aij}}{n}$$
(9)

Here i, j = 1, 2, 3…. n. Additionally, the AHP offers quantitative metrics to assess the judgment matrix's consistency. Using the matrix's attributes, one can compute a Consistency Ratio (CR). In a matrix, the number of columns or rows (n)is always equal to or larger than the maximum eigenvalue (max). According to Saaty [10], a CI (Consistency Index) that gauges the coherence of pair-wise comparisons could be expressed in the Eq. (10)

$${\text{CI}} = \frac{{\lambda \max - {\text{n}}}}{{{\text{n}} - 1}}$$
(10)

here λmax indicates the greatest or major eigenvalue of the matrix, n presents the number of elements being compared in the matrix, and CI is the consistency index. The comparison answers need to be re-examined if this CI falls below a certain threshold. The consistency judgment needs to be verified by CR for the right value of n in order to guarantee the consistency of the pair-wise comparison matrix. The coefficients of CR are computed using Saaty's suggested methodology [10]. The CR coefficients indicate the pair-wise comparison matrix's” overall consistency, which should be smaller than 0.1 [49]. The definition of CR is presented in the Eq. (11).

$$CR = \frac{CI}{{RI}}$$
(11)

Thus, based on the matrix, RI is the average of the CI that is produced (Table 1) [50]. Table 1 displays the RI values for many n values. If the pair-wise comparison matrix has an acceptable consistency, its CR value is less than 0.10. AHP could not provide useful results until the judgments are re-examined and adjusted as needed to get the inconsistency below 0.10 if CR > 0.10 indicates that pair-wise consistency is insufficient [10, 51]. The pair-wise comparison matrix's resulting CR in this investigation was 0.028. These numbers show that there was absolute consistency in the comparisons of land features. The rating of every alternative is then multiplied by the sub-criteria’s weights and combined to obtain local ratings for every criterion. The global ratings are then calculated by multiplying the local ratings by the criteria's weights and averaging them [47].

Table 1 Random consistency indices (RI) for n = 13
Full size table

3 Results and discussion

3.1 Descriptive statistics of soil properties

Descriptive statistical parameters of all the soil (chemical and physical) variables are displayed in Table 2. The soil pH ranged from 6.0 to 8.80, with a mean value of 7.40 ± 0.58. The EC of the soil ranged between 0.20 and 1.23 dS m−1, with the mean value of EC being 0.71 ± 0.18 dS m−1. The sand content of the Manimutha Nadhi watershed soil mean value was 44.05% ± 2.81 and varied between 2.2 and 85.90. The silt contents of soil ranged between 2.4 and 74%, with a mean value of 38.2% ± 1.97. The soil's clay content (%) varied between 2.9% and 66.8%, averaging 34.85 ± 3.11. Organic carbon content ranged from 0.06 to 1.70% (with a mean value of 0.88). The soil available nitrogen ranged between 210 and 460 kg ha−1 with a mean value of 335.31 ± 58.18. The available phosphorus varied from 7.00 to 19.00 kg ha−1 with a mean value of 39.68 ± 73.87. The available potassium ranged from 82.00 to 200.00 kg ha−1 with a mean value of 141.45 ± 36.94. The pH and EC CV values were 9.49 and 33.37, respectively. The coefficient of variation of the soil organic carbon values was 47. The available N P K coefficient of variation values were 18.80, 27.14, and 31.15, respectively. The CV of sand, silt, and clay were 49.24, 34.17, and 54.73, respectively. Skewness and kurtosis coefficients are used to describe the shape of the data distribution. The skewness of the soil varied from − 0.42 to 1.62. For every soil property, there was a distinct coefficient of variation (CV), which ranged from low to high and is a typical statistical measure of the total variation or heterogeneity of a given soil sample. The pH in the research area had the lowest CV (9.49). Reference [52] as well as [53], reported similar findings. Other soil characteristics (EC, SOC, available NPK, sand, silt, and clay) showed moderate CV values. This could be related to differences in soil management, slope classes, and topography, resulting in moderate variability in soil characteristics. References [54] and [55] reported comparable findings.

Table 2 Descriptive statistic of soil parameters in Manimutha Nadhi watershed
Full size table

3.2 Correlation coefficient among the soil properties

The values of the Pearson correlation coefficient established among the soil parameters of the Manimutha Nadhi watershed are displayed in Table 3. The positive and negative correlation of the soil variables is submitted to illustrate similar and opposite spatial distribution patterns among the variables. Soil pH was positively correlated with available phosphorus (r = 0.358*) and negatively correlated with clay (r = − 0.186**). The EC of soil had been positively correlated with available phosphorus (r = 0.196**) and negatively correlated with Organic carbon (r = − 0.223*). The correlation values of soil organic carbon positively correlated with available nitrogen (r = 0.202**). Similar results were corroborated by [53]. The available phosphorous had been positively correlated with available potassium (r = 0.155**). The sand had been negatively correlated with clay (r = − 0.793*), and the silt was negatively correlated with clay (r = − 0.419*). A similar result was also reported by [56]. Some parameters have low correlation coefficient because of the high variability in soil characteristics.

Table 3 Correlation of coefficient soil parameters of Manimutha Nadhi watershed
Full size table

3.3 Semi-variogram analysis of soil properties

Maps for every soil variable can be generated using a simple kriging interpolation based on the best-fitted semi-variogram model. Geostatistics is a useful tool for land suitability because it can be used to predict and map soil attributes in areas that were not sampled. On the basis of the minimum RMSE (Root Mean Square Error), the semi-variogram parameters (sill, nugget, and range) for every soil physical feature that best fit the modal were determined. Isotropic variogram’s analysis denoted that the organic carbon, available phosphorus, and available potassium content semi-variograms were well explained by the exponential model, with the distance of spatial dependence being 67, 105, and 19 m, respectively, while the pH, EC, and available nitrogen semi-variogram were well explained by the spherical and circular model, with the distance of spatial dependence being 274, 147 and 195 m, respectively (Table 4). Values of < 0.25, 0.25–0.75, and > 0.75 indicate strong (ascribed to intrinsic causes), moderate (ascribed to intrinsic and extrinsic factors), and weak (ascribed to extrinsic factors) spatial dependence, according to [57]. The nugget (shows the micro-variability) values have been small for soil pH, EC, and SOC (v from 0.019 to 0.486) and large for available N, P, and K (varied from 8.107 to 1294). According to [58], large nugget values showed that ecological processes had a stronger impact on available N, P, and K on a small scale and that the chosen sampling distance poorly captured the spatial dependency. Conversely, given available N, P, and K, sill values—which show “the variance of the sampled population at large separation distances if the data show no trend—were likewise higher. References [53] and [59] observations provide evidence for the difference in nugget and sill values of investigated soil characteristics. Similar to the findings of [57], the study area's pH, EC, and organic carbon showed significant spatial dependence, while available nitrogen and potassium showed modest spatial dependence (Fig. 4).

Table 4 Kriging statistic parameters of soil of Manimutha Nadhi watershed
Full size table
Fig. 4
figure 4

Spatial variability map. a Slope map, b soil erosion map, c annual average rainfall map, d annual average temperature map.

Full size image

The spatial variability information and map of pH, EC, soil texture, and SOC are presented in Table 5 and Fig. 5. In the spatial variability map, the pH of the soil Acidic (< 6.5) was occupied for 65.24 km2, (6.5–7.5) (neutral) was occupied for about 88.84 km2, whereas alkaline pH (more than 7.5) was occupied about109.78 km2 and soil EC normal range (≤ 1 dS m−1) was occupied for about 217.03 km2. EC (critical to sensitive crops > 1 dS m−1) was occupied at about 45.94 km2 (Table 4; Fig. 5). The medium level of OC (5–7.5 kg−1) was occupied at about 129.08 km2. High-level soil OC (> 7.5 g kg−1) occupied about 133.89 km2 of the study area. Soil textural class clay was occupied for about 17.94 km2, sandy clay soil for 68.98 km2, and sandy clay loam soil for 176.06 km2.

Table 5 Spatial variability of the influencing factors of Manimutha Nadhi watershed
Full size table
Fig. 5
figure 5

Spatial variability map. e Soil pH map, f soil EC map, g soil texture map and h soil organic carbon content map

Full size image

The spatial variability map, the low available nitrogen (< 280) occupied about 156.26 km2, medium available nitrogen (280–440) occupied about 106.71 km2, low available phosphorus (< 11) occupied about 29.52 km2, and medium available phosphorus (11–22) was occupied about for 233.45 km2 and low available potassium (< 118) was occupied about for 175.89 km2 and medium available potassium (118–280) was occupied about for 87.09 km2 (Table 5; Fig. 6).

Fig. 6
figure 6

Spatial variability map soil properties. i available nitrogen, j available phosphorus map, k available potassium map

Full size image

3.4 Climate and topographic factor and erosion classes

The spatial variability information and map of the slope, erosion, annual average rainfall, and annual average temperature are present in Table 5 and Fig. 4. The slope map reveals a level (0–1%) observed in 52.26 km2, very gentle slope (1–3%) for 33.32 km2, gentle slope of (3–8%) occupied for36.18 km2, the marginal slope of soil (15–30%) occupied for 45.15 km2, and steep slope of soil (> 30%) occupied for 12.01 km2 and the variability of soil erosion low (< 5) occupied about for 181.64 km2, moderate (5–10%) occupied about for 51.90 km2, high (10–15) occupied about for 10.11 km2, very high (15–25) occupied about for 11.06 km2, and severe (25–50) occupied about for 2.83 km2. The erosion (tons ha−1 year−1) of the study area is classified into five class, namely low ((< 5), Moderate (5.0–10.0), High (10–15), Very high (15–25), and Severe (25–50). The low, moderate, and high erosion classes occurred at about 181.64, 51.90, and 10.11 km2, respectively. The very high and severe erosion classes were occupied about 11.06 and 2.83, respectively. The annual average rainfall > 800 mm was about 130.28 km2, whereas 800–1200 mm was about 132.17 km2. The annual average temperature class 22–26 °C and 26–32 °C were occupied about 130.73 and 132.17, respectively.

3.5 Land suitability analysis through GIS and AHP technique

The land suitability for the Manimutha Nadhi watershed was determined by Weighted Overlay Analysis utilizing the weight values of chosen factors determined by the AHP and specified scores of sub-criteria. As per the classification of FAO (Food and Agricultural Organization), land suitability for agriculture is categorized into 5 levels i.e., highly suitable, moderately suitable, marginally suitable, currently and permanently not suitable for agriculture. The main purpose of the current work was to advance the land suitability of the Manimutha Nadhi watershed using an AHP and geographical information system. In this context, three main criteria, namely, climate, soil parameter, and topography, were chosen to examine the land suitability of the study area. Among the different climate parameters, Annual average rainfall and annual average temperature were chosen as the climate criteria. The slope was chosen as a criterion for this study in the topography factor. The soil parameters like soil EC, available nitrogen, soil pH, soil organic carbon content, available potassium, available phosphorus, soil texture, and annual average erosion were chosen as criteria for AHP analysis. All criteria factors were selected per the opinions and ideas of the experts, farmers, and subject specialists. Sub-criteria were established based on literature reviews, field investigations, and other recommendations.

The most crucial stage in the multi-criteria assessment method is assigning each factor a weight according to Saaty's scale. Saaty's fundamental scale was used to rank each factor on a scale of 1–9. Table 6 displays the pair-wise comparison matrix that was utilized in this investigation. The pair-wise comparison matrix contains the AHP weight rank and its reciprocal value. The consistency of the current research explains whether it was over- or under-predicted as determined by computing the CR and CI. The CR value alone determines whether the consistency is acceptable when 0.1 or less. The consistency ratio, consistency index, and λmax values in this investigation were 1.51, 0.043, and 0.028, respectively, indicating that every matrix was consistent. The following step was calculating each criterion's weight using the pair-wise comparison matrix's rows' normalized geometric mean (Table 7). All criteria and their weights are presented in Table 8.

Table 6 Pairwise comparison matrices
Full size table
Table 7 Normalized comparison matrices
Full size table
Table 8 All criteria factor weights utilised in the study area
Full size table

3.6 Final land suitability analysis

Figure 7 and Table 9 represent the agricultural land suitability class and map of the Manimutha Nadhi watershed in Kallakurichi district. The input variable's influence and potential produce the suitability map that can be divided into different classes of agricultural land suitability with diverse capacities. The agricultural land suitability model performance specified that 150.58 km2 of the area was located in the study region in the S1 class. This area is highly suitable for agricultural production, which may be due to the lack of restriction of input parameters contributing to the potentiality of the site. In the research region, a 29.60 km2 area is defined as marginally suitable (S2) to the study results. The S3 class, i.e., moderately suitable land of the research region, is 53.3 km2. The currently unsuitable (N1) and permanently unsuitable (N2) classes are about 19.87 and 13.45 km2, respectively. After removal of forest, settlements and water body in the study area (Fig. 8), about 52.52 and 10.70 km2 area were highly suitable (S1) and moderately suitable (S2), respectively. The marginally suitable (S3) class was occupied about 18.33 km2. About 3.20 and 4.76 km2 were assigned as currently unsuitable (N1) and permanently unsuitable (N2), respectively. The land suitability class of the Manimutha Nadhi watershed may be attributed to the influence of slope, soil erosion, annual average rainfall. pH and EC accounting for more than 73%. The slope, soil erosion class and annual average rainfall are an important limiting factor affecting the agricultural suitability directly and all other parameters in the study area. The high degree of the slope in the western part of the study area that prevents soil formation and increase the erosion of the study area that affects the adequate root depth for plant production. This will influence the suitability (permanently not suitable-N2) in the study area. Depth parameter was considered the most essential criterion for wheat production due to the importance of providing sufficient moisture and plant nutrients [60]. The results indicated that the slope and erosion rate are an important parameter affecting the suitability of the research region.

Fig. 7
figure 7

Agricultural Land Suitability map of Manimutha Nadhi watershed

Full size image
Table 9 An overview of the various agricultural land suitability maps for the Manimutha Nadhi watershed's suitability index classes
Full size table
Fig. 8
figure 8

Agricultural Land Suitability map of Manimutha Nadhi watershed after masking

Full size image

4 Conclusion

Land suitability assessment for agriculture serves the purpose of selecting suitable land for suitable crops for sustainable production to attain sustainable development goals for food security and eradicate poverty. The study area, Manimutha nadhi watershed was forced to increase food production due to high population growth. So, it is necessary to improve food production with advanced land use suitability techniques with suitable crops. The tropical watershed Manimutha Nadhi is taken as a research region to evaluate the suitability of the land for agriculture. The suitability for agriculture is employed by the AHP technique and GIS (Geographical Information System) technique. The LSA results for agriculture in the Manimutha Nadhi watershed indicated that the most influential factors are slope, erosion, annual average temperature, and annual average rainfall. The study revealed that 57.04% of the area was moderately suitable class (S1), 10.08% of the area was moderately suitable (S2) class, whereas marginally suitable (S3) class occupied the area of 20.20%. The currently unsuitable (N1) and permanently unsuitable (N2) classes cover about 7.53% and 5.10%, respectively. The total area of forest water body and settlement and urban area was about 167.3 sq. km. After removing these land use cover areas, about 52.52, 10.70, and 18.33 km2 area was highly suitable (S1), moderately suitable (S2), and marginally suitable (S3), respectively. About 3.20 and 4.76 km2 were assigned as currently unsuitable (N1) and permanently unsuitable (N2), respectively. This study proves the capacity of the AHP as an excellent tool for land suitability analysis. This assessment will help farmers, land managers, institutes, and the government to prioritize the land. The proper water management scheme, judicial nutrient application, and inter-cultivation operation will support further attaining sustainable development goals of the Manimutha nadhi sub-watershed.