Introduction

Agriculture is vital to the economic growth of developing countries like India. The agriculture industry has proved that it is more effective than other sectors in decreasing hunger and poverty. Approximately half of the population is directly or indirectly dependent on agriculture (Khan et al. 2020). However, the agricultural sector’s gross domestic product (GDP) contribution has been facing down for the past 50 years. The fall in agriculture's percentage of GDP is primarily due to increased contribution from the manufacturing and services sectors. However, population growth and urbanization have increased the strain on agricultural resources. Land degradation may occur as a result of the increased pressure on available land resources (Ahmed et al. 2016). The forced land-use change from possible bare land to agricultural land could be compensatory for the degradation of cultivable area. Bare land with free access to the road network shall be given higher priority for the land-use conversion process (He et al. 2021).

Even though there are myriad reasons for the decrease in agricultural activities, reducing yield results in lower desired profit for farmer's society was identified as a significant cause. Getting a good yield always boosts the farmer's confidence, which increases the cultivation area in the upcoming season. A crop’s yield in a specific area depends on various factors. These are classified as technological factors (agricultural practices, managerial decisions), biological factors (diseases, pests, weeds, insects), and environmental factors (soil fertility, climatic condition, water quality, terrain topography) (Ngoune Liliane and Shelton Charles 2020; Susaiappan et al. 2021). The advanced science and technology application in agriculture aids in achieving higher yields with minimal resource usage. Data on a crop's performance can be collected in the field, analyzed, and used to make decisions that maximize yields (Van Alphen and Stoorvogel 2000). Geographical Information Systems (GIS) are techniques that assist farmers in increasing production while decreasing costs by enabling better management of land resources.

Geographic information systems (GIS) can assist in selecting the ideal places for sustainable agricultural production for food supply based on environmental potential (Bozdağ et al. 2016). Geospatial technology is used to find ideal sites for farming based on various factors such as soil quality, geology, drainage, terrain soil (slope, drainage), and availability of roads (Mishra et al. 2015; Chozom and Nimasow 2021). Farmers may use this technology to map and forecast the present and future swings in precipitation, temperature, crop yield, etc. Choosing the optimal location for agricultural production in the context of land suitability analysis (LSA) is a complicated process that involves technical criteria and physical, economic, social, and environmental needs, which may result in competing aims (Saha et al. 2021).

The process of establishing the appropriateness of a particular land area for a specific type of use (agriculture, forest, recreation, etc.) and the level of suitability is known as land-use suitability analysis (Kazemi and Akinci 2018). However, because not all criteria are equally important, GIS does not consider preference. It cannot overcome the problem of inconsistency in judging and assigning a degree of importance to the criteria required for land suitability assessment (Kahsay et al. 2018). However, it might be difficult to assign relative weights to the various criteria involved in determining the suitability of a land mapping unit for a specific land-use category.

Multiple-criteria decision-making (MCDM) technologies were developed due to improvements in the geospatial domain to expand the decision-support capabilities of GIS. These methods assist decision-makers in formally structuring multi-faceted decisions and evaluating alternatives for problem-solving by rating sets of other options. Analytical Hierarchy Process (AHP) is one of these techniques, which is recognized as one of the Multiple Criteria Decision Analysis (MCDA) methodologies, particularly in spatial or GIS-based problems (Bagheri et al. 2013; Chakrabortty et al. 2018). Integrating multi-criteria decision analysis methodologies in GIS creates a powerful spatial decision-support system that promptly allows for the production of land suitability maps (Elaalem et al. 2011; Mendas and Delali 2012; Ullah and Mansourian 2016; Seyedmohammadi et al. 2019).

The analytic hierarchy process is a structured technique for organizing and analyzing complex decisions, considering mathematics and psychology. It is beneficial in multi-criteria decision-making. The AHP transforms these evaluations into numerical values that can be processed and compared across the entire problem range (Ishizaka and Labib 2011). All hierarchy elements are assigned a numerical weight or priority, allowing diverse and often incommensurable elements to be compared logically and consistently. This feature distinguishes the AHP from other decision-making methods (Ramamurthy et al. 2020).

This study focused on identifying land suitability for agricultural purposes with the help of AHP and GIS at Sivanthipatti village of Tirunelveli district. The primary goal of this study was to demonstrate the potential of an integrated AHP and GIS approach in determining suitable agricultural areas from existing bare land in the study area.

Study area

The study area considered for agricultural land suitability analysis is Sivanthipatti village of Tirunelveli district. The bare land traced by processing the satellite image with supervised classification for land-use detection bound between the geographical location 77° 48′ 10.8′′–77° 56′ 42" East and 8° 39′ 14.4′′–8° 41′ 52.8′′ North. The Tirunelveli Taluk experiences annual rainfall ranging from 550 to 775 mm. Last year (2022), the maximum rainfall at Tirunelveli was in November, with an average of 195 mm (7.68′′) of precipitation. The city's monthly average relative humidity ranges from 29 to 54%. Its hydrometeorological feature separates the year into two broad classifications: the monsoon season between June and December and the non-monsoon season between January and May (Jeykumar and Chandran 2019). Tanks, reservoirs, wells, and bore wells are the primary irrigation sources in and around the study area.

The cropping pattern of the entire district varies spatially due to different climatic conditions. The study area landscape is classified as semi-arid (Kirubakaran et al. 2018). The black cotton soil is found in the blocks of Tirunelveli and Palayankottai and is suitable for growing paddy, ragi, and maize, among other crops. In the Tirunelveli block, river alluvial soils are found along the Thaamirabharani covering river channels, and they are ideal for growing millets, groundnuts, and chillies. Around the study area, a small patch of porous formations includes Tertiary to Quaternary sandstones, limestones, laterite, and clays.

The soil samples were collected from 20 sample points spatially around the study area with its geographical coordinates. The geographical position of the study area and the sample point is shown in Fig. 1. An agricultural extension center was located in Sivanthipatti village, where agrarian inputs like seeds, bio-fertilizers, and bio-pesticides can be purchased. This establishment offers additional benefits by increasing cultivable land through suitability analysis.

Fig. 1
figure 1

Study area—Bare land of Sivanthipatti village

Full size image

Methodology

The workflow for this agricultural suitability study starts with downloading the recent satellite image. Identifying barren land may not be regarded in the scope of the study, which is a part of preparatory work. However, at this stage is necessary to pick an appropriate study area from Sivanthipatti village. Landsat-8 is a popular satellite that collects and stores medium-resolution spectral bands image data, allowing for seasonal observation of the whole world. Since the Landsat has 16 days as a repeat cycle with a spatial resolution of 30 m, it is often used to perform land-use and land-use change detection (Kumar et al. 2021). In terms of acquisition geometry, calibration, coverage characteristics, spectrum features, output product quality, and data availability, these data are sufficiently compatible with data from previous Landsat missions.

Supervised image classification is a process for recognizing spectrally comparable areas on an image by finding 'training' sites of known targets and then extrapolating those spectral characteristics to additional areas of unknown targets. It involves training data that are typical and homogeneous and applying a set of techniques or rationalities (Chu et al. 2020; Zhong and Zhang 2012). The polygon remarked as bare land: Sivanthipatti was the processed output of preparatory work for the study. Figure 2 depicts the sequence of activities in mapping suitability for agricultural purposes.

Fig. 2
figure 2

Methodology for site suitability study

Full size image

Collection of sample

The characteristics of soil directly relate to the yield of any crop. Soil fertility refers to the soil's ability to support plant development and maximize crop yield. The extracted circumscribed bare land has an area of 17.54 sq. km and a perimeter of 20.64 km. Twenty soil samples were collected in a spatially distributed pattern within this geometry. The environmental and spatial aspects are used in the suitability analysis. The geographic coordinates of the sites where the samples were gathered must be recorded for spatial aspects. The location details of sample points are presented in Fig. 3.

Fig. 3
figure 3

Spatial distributions of sample points

Full size image

Criteria for suitability analysis

Terrain, climatic, and pedological parameters were generally considered for agricultural site suitability. Since the entire study area is almost flat and smooth, the terrain parameters like slope and elevation will not cause any variability (Atul et al. 2021). The climatic parameter, rainfall, is highly uncertain in semi-arid regions. Organic carbon content, dissolved solids, pH, conductivity, and soil texture are the key features of soil science, and the same were examined for this agricultural suitability analysis (Mendas et al. 2021). The collected samples were tested for organic carbon content, pH, electrical conductivity, and texture in the public health laboratory at Government College of Engineering, Tirunelveli. The results are tabulated with geographical coordinates of the sample location and used as input for further analysis in Arcmap, a GIS package.

Organic carbon content

Soil organic matter (SOM) is one of the necessary constituents of soil as it governs practically all the qualities of soil that control agricultural and environmental functions. Soil organic matter is complicated to measure; thus, laboratories commonly measure and report its most prominent component, soil organic carbon (SOC) (Paul et al. 2006; Prout et al. 2021). Organic matter helps with nutrient retention and turnover, moisture retention and availability, pollutant breakdown, carbon sequestration, and soil resilience (Lal 2016).

pH

Soil pH, often known as soil reactivity, is a measurement of the acidity or alkalinity of the soil (Penn and Camberato 2019). The negative logarithm of the hydrogen ion concentration calculates soil pH. The pH scale ranges from 0 to 14, with a pH value of 7 designated as neutral. The pH is significant because it impacts the availability of necessary nutrients.

Conductivity

Soil salinity is a major constraint to agricultural output all around the world. Electrical conductivity tests in soil are precious for assessing environmental effects (Ghorbani et al. 2019; Mazur et al. 2022). Conductivity helps identify soil physicochemical parameters (moisture contents, color, pH, organic matter content, etc.) that influence crop yield patterns.

Soil texture

The soil texture reflects the relative content of particles of various sizes, such as sand, silt, and clay. Soil texture significantly impacts agriculture because it affects aeration, water movement and storage, nutrient retention capacity, and susceptibility to erosion (Pilevar et al. 2020).

Spatial interpolation

Spatial interpolation is the technique of estimating values at other sites by using known values at points with known values. Spatial interpolation is often used as a raster in GIS applications, with estimates made for all cells. Spatial interpolation is, thus, a technique for generating surface data from sample points (Ikechukwu et al. 2017). Interpolation is based on the concept that spatially distributed objects are spatially correlated; in other words, things that are close together tend to have similar features. Inverse Distance Weightage is a popular deterministic and multivariate interpolation method adopted in this study (Venkatesh and Mannar Thippu Sulthan 2022). The criterion map was created using the values of 20 sample points for each of the four criteria.

Determination of weight by AHP

The AHP technique allows users to utilize effective mathematical procedures to estimate the weights of the parameters in addressing the multi-criteria problem. A hierarchical model with objectives, criteria, sub-criteria, and options was used to answer each problem using the AHP technique. The weights of the criteria and scores of the sub-criteria can be calculated after arranging the problem in a hierarchical framework (Saaty 1987).

The pairwise comparison matrix was built to establish the weights of parameters according to the AHP preference scale based on the relevance of one criterion over another in the pair (Benítez et al. 2012; Hamidy et al. 2016). A scale of 1 denotes equal importance, whereas a scale of 9 indicates that one component is clearly more essential than the others, as shown in Table 1. The reciprocals of 1–9 (1/1 and 1/9) demonstrate that one is less significant than the other. A team of experts' opinions was used to determine the relative value of parameters in the pairwise comparison matrix (Kılıc et al. 2022).

Table 1 Scales for pairwise comparison
Full size table

The consistency ratio index (CR) will be produced using Eq. 1 to assure the legitimacy of the relative impact ranking utilized. Table 2 contains a list of Saaty’s Random Index (RI) estimates for the positive reciprocal matrix of orders 1 to 10 obtained from the integer set 1–9 and their enforced reciprocals.

$${\text{CR}} = \frac{{{\text{CI}}}}{{{\text{RI}}}},$$
(1)

where CI is the consistency index and RI is the random matrix.

Table 2 Random Index
Full size table

All selected criteria are compared in a pairwise comparison matrix, a measurement used to express the relative preference among the elements (Akinci et al. 2013). Thus, numerical values reflect a judgment of one factor's relative favor over another.

Soil quality assessment entails thoroughly examining the best agricultural land location. The selected acceptable soils must have the significant nutrients for fundamental plant nutrition to reduce organic and mineral chemical inputs while minimizing economic expense and environmental damage. We produced the pairwise comparison matrix judgments (relative levels of importance of the characteristics) by consulting the opinions of a committee of experts, which included local agronomists and faculty members from the Departments of Soil Science and Plant Nutrition in agricultural faculties (Odu 2019). The arrived pairwise comparison matrix for the considered criteria is shown in Table 3.

Table 3 Pairwise comparison matrix for selected criteria
Full size table

The consistency index obtained through the consistency matrix, as shown in Table 4, is 0.05463. Finally, the consistency ratio was calculated as 0.0607, less than 0.1. Hence, the pairwise comparison matrix is reasonably consistent. So we can use the determined criteria weights in the weighted overlay analysis.

Table 4 Normalized pairwise comparison matrix
Full size table

Reclassification of criterion

Reclassification changes the output values of one or more values in a raster dataset. It is used to assign values to a raster based on choice, priority, sensitivity, or other similar criteria and to set particular values to NoData to eliminate them from the analysis (Parry et al. 2018). Our considered site suitability criteria like pH, organic carbon, electrical conductivity, and texture were reclassified by assigning new values for the pixels between 1 and 5 (5 classes) based on the suitability of corresponding criteria for agricultural purposes.

Weighted overlay suitability analysis

Weighted overlay analysis (WOA) is carried out on an ArcGIS environment after weights for each raster layer have been calculated using AHP (Hama et al. 2019). WOA is a helpful method for resolving spatial complexity in site selection and suitability analysis based on a broad measurement of various disparate impacts. A weighted overlay is the intersection of standardized and variously weighted layers during suitability analysis (Chakrabortty et al. 2018). The weights assigned for each criterion were multiplied by the suitability scores assigned for the sub-criteria within each criteria layer to create the final suitability map using the WOA technique as given in Eq. 2.

$$S = \mathop \sum \limits_{i = 1}^{n} \left( {W_{i} X_{i} } \right),$$
(2)

where S is the final suitability score, Wi is the weight of the layer of suitability criteria, Xi represents the assigned sub-criteria score of i suitability criteria, and n refers to the total number of agricultural suitability criteria.

Results and discussion

The analytical hierarchy process enabled the mathematical assignment of the agricultural suitability weightage. The weightage of 54.86%, 28.05%, 11.51%, and 5.58% was allotted to pH, organic carbon content, salinity, and soil texture, respectively. The pairwise comparison has determined the maximum weightage for pH, which is even more than 50%. It shows the sensitivity of land pH in agricultural suitability (Husson et al. 2018). The least weightage evaluated to the texture detected that soil's chemical characteristics are relatively more crucial than physical characteristics in choosing a better site for agricultural purposes (Santhosh and Sivakumar Babu 2018).

We created criterion raster layers from the point features of the sample location using Inverse Distance Weightage interpolation technique, as presented in Fig. 4. The distance weightage power value was assumed as 2, which is a default value for IDW interpolation (Fazeli Sangani et al. 2019; Yanto et al. 2022).

Fig. 4
figure 4

Criterion map for the study area (OC, pH, EC and Texture)

Full size image

The deciding criteria for suitability analysis always have different slabs of value ranging from favorable to unfavorable. With the reclassify option in the spatial analyst toolset from GIS software, one can rapidly reassign a value, a range of values, or a list of values in a raster to new output values (Tashayo et al. 2020). We reclassified all four criterion map layers with a new value of range between 1 and 5. The soil parameter values responsible for better crop yield and unfavorable ranges for poor or unsuitable yield were reassigned to 1 and 5, respectively. The in-between classified values are assigned proportionately as per the suitability for agriculture (Zandi et al. 2011).

The agricultural suitability map was finally derived with a weighted overlay analysis tool. In the output file, pixels with a value of five could be highly suitable for agriculture. The suitability gets decreases as the pixel values decreases. In this order, raster cells having values 1 is nothing but spaces that are highly unsuitable bare land for land-use conversion. For the entire study area, the spatial variation of agricultural suitability is presented in Fig. 5.

Fig. 5
figure 5

Agricultural suitability map of the study area

Full size image

Our study area has shown the distributed soil suitability characteristic. Either unsuitable or moderate suitability was witnessed along the boundary of bare land where other land-use classes intersect. These two suitability categories were observed as dominating the entire study area. Finally, from our suitability map, it is estimated that only 8.5% of bare land in the study area is highly suitable for the land-use change process technically. At the same time, 40.4% of unsuitable and 3.5% of highly not suitable areas consolidating concluded that almost 50% of the area is against transformation. Around 50% of the landscape is ready for the desired purpose in the study area, as presented in Fig. 6. If any additional agriculturally relevant soil characteristics are considered for the study, the proportion of suitability of the five different categories may differ from the current level (Pramanik 2016; Bhunia et al. 2018).

Fig. 6
figure 6

Distribution of suitability area in the bare land

Full size image

The southeastern part of the study area is technically opposed to land-use conversion, but there are plenty of opportunities for farming on the northwestern side. The transformation of 31.8% area of moderate suitability is practically hard to achieve. A significant effort is needed to prepare moderately suitable regions for agricultural suitability. The trade-off between cost and benefits could be a systematic way of deciding on conversion. This study has considered only a few soil characteristics as a decider for suitability and assumes the climatic conditions and water availability are favored throughout the study area.

Conclusion

Agricultural land usage must be enhanced and expanded using a scientific methodology and a thorough feasibility analysis to fulfill rising population demands for food and other agricultural commodities. It is critical to achieving a harmonious balance among the many uses of land that does not jeopardize the availability of life support systems and improves the diverse groups of people who rely on the land. As a prime agricultural input, land occupies a significant position among all the resources required for a modern economy. Generally, the agricultural suitability analysis is performed on a region. However, it is practically not possible to convert impermeable urban areas (residential, commercial, and road), water bodies (rivers, lakes, and ponds), and restricted areas (reserve forests) into farmland. Hence the only non-functioning bare land extracted using supervised classification was decided as our study area.

The criterion weightage determined from AHP for pH, organic carbon content, salinity, and soil texture is 54.86%, 28.05%, 11.51%, and 5.58%, respectively. The percentages of our study areas that were highly unsuitable, not suitable, marginally suitable, moderately suitable, and highly suitable for agricultural purposes, respectively, were 3.5%, 40.4%, 15.8%, 31.8%, and 8.5%. It has been identified that out of 17.51 sq. km, 1.49 sq. km of the land area is highly suitable for agriculture. These were followed by the land area of 5.57 sq. km classified under moderately suitable, implying a modest technical effort on this land might turn completely to highly suitable. We could restore more than 1750 acres for agricultural conversion. While having a quick glimpse at the final map, there is clear evidence that a few proportions of bare land are immediately available for agricultural use conversion.

This preliminary study on suitability with gathering and analyzing the field data will be part of the most successful farming management concept called precision agriculture. Our suitability modeling approach can be applied to a larger area. But it requires an intensive data collection process. Once we successfully collect the appropriate soil parameters of agricultural suitability from the identified bare land across the vast area, we shall do the same analysis for the land conversion process.