Land Suitability Analysis for Afforestation in Semi-arid Watershed of Western Ghat, India: A Groundwater Recharge Perspective

ABSTRACT Land suitability analysis (LSA) for afforestation is an approach that stands amongst the most able frameworks revealing the concern for cultivable land and predicting the availability for sustainable development in semi-arid areas. In view of this, the objective of the present study is to propose a conceptual procedure for LSA that would help enhancing the green cover to combat environmental threats rendering groundwater recharge. LSA involves various thematic layers such as distribution of land use/landcover, slope, soil depth, soil type, pH, soil calcium, soil magnesium, sodium, bulk density, organic matter, boron and run-off that have been derived using satellite images and collateral data. Food and Agriculture Organization framework and guidelines have been followed for LSA and it has been found that only 9.16% land is ‘highly suitable’, 14% of land is ‘moderately suitable’ whereas, marginally suitable lands for afforestation are estimated about 14% of total area and non-suitable areas for afforestation are about 61% of the total area. The multi-parametric based decision making AHP tool integrated with GIS proposes a novel methodology and outcomes of the research could be useful to identify suitable lands for agriculture in any various parts of the world.


Introduction
Land under afforestation is a limited natural resource that facilitates sub-surface moisture, surface runoff, vegetation, timber land, soil and mineral deposits which are getting degraded by heavy erosion, water logging, increasing acidity and salinity, decreasing nutrients and soil productivity. Out of total land, 1.68% productive land is reduced and 47% land has degraded in the last twenty years (FAO). About 14.9% of population of developing countries are suffering from malnutrition (Fao, 2012). Demand of agricultural products for food, fodder, industry, etc. is increasing rapidly while productive land and productivity are decreasing remarkably. Therefore, the studies undertaken in this regard (Dumanski, 1997;Nyeko, 2012;Schwilch et al., 2011) have been focussed on sustainable land use development and management.
Forestation, reforestation and protection of existing forests are suggested for conservation and enhancement of CO 2 storage in forest cover to reduce the atmospheric carbon (Patenaude et al., 2005). Forests increase subsurface recharge and help conservation of soil and groundwater resources. However, the area under forest and qualitative measures in terms of density, diversity, greenness and production are decreasing with alarming rate since centuries (Bhagat, 2009). The green-covered area mainly the forest is reduced to 21.31% due to soil deprivation, cut down of trees and minimum soil moisture during scarcity situations, etc. (Sonawane & Bhagat, 2017). Therefore, afforestation on degraded lands is primely suggested for environmental regulation of carbon, protection of natural resources such as soil and water, biodiversity protection, etc. Many governmental, non-governmental organisations, and forest conservation activists and movements are working on the improvements in quality of soil and water resources and land restoration. Land suitability analysis (LSA) detects the inherent capacities, potentials and suitability for plantation. Multi-criteria land suitability has been widely used for identification of potential parcels for agriculture (R. B. , plantation (Bhagat, 2009;R.B. Zolekar & Bhagat, 2014), watershed management (Gaikwad & Bhagat, 2018, A. K. Kadam et al., 2018A. Kadam et al., 2019;Rajasekhar et al., 2020), settlements, development projects, industries, etc.
The LSAs in existing methods in this paper is having the following parts such as the examination of soil possessions is a significant stage. In view of this soil samples were analyzed in laboratory to understand the physicochemical properties. To know the present LULCs supervised categorization algorithm was used for nine land utilization types. This map and information were further used to derived the soil depth in connection with slope. Also, the other influencing factors were also derived from base maps and fieldbased data. Finally, multi-criteria evaluation tool was used for detection of land suitable for afforestation in semi-arid watershed of western Maharashtra, India. In the study area, surface runoff utilized for artificial groundwater recharge has been considered as a probable source for "plantation". The groundwater recharge will be enhanced after the afforestation. Hence, land with groundwater recharge potential is taken to be a "preferable land for afforestation". Thus, the land covered by grass/shrubs shows moderate runoff potential and hence could be considered as suitable for green growth. With this hypothesis, LSA for afforestation with the perspective for groundwater conservation has been carried out in Shivganga river basin.

Study area
The area considered for study is drained by a tributary of Nira River (Shivganga River) that flows in Bhor, Haveli and Purandar Tahsil of the Pune district, Maharashtra, India. It is covering an area of about 17600 ha present in the eastern undulating part of the Western Ghats ( Figure 1). The study area experiences innumerable problems such as soil deprivation, water scarcity, deforestation and soil pollution (A. K. Kadam et al., 2018;A. Kadam et al., 2019), groundwater quality depletion (A. Kadam et al., 2019) and change in hydrological response (A. K. Kadam et al., 2018). Majority of the area under study possesses irrigated land covered by scanty to thick plantation. The forest cover is very poor and occupy mainly higher elevations, while local plantations are also scanty due the low soil cover in scrub land area. The government department planted the trees but the scientific approach is missing in the action, resulting into wilting of plants, poor plant growth with low recharge of water, etc., leading to failure of the project. The study area incorporates hilly undulating terrain in the peripheral proximity with high rainfall runoff ratio, depicting poor forest/ green cover therefore has been selected for the land suitability analysis for afforestation along with groundwater recharge perspective.

Data and software used
Present study aims to evaluate the LS for afforestation or plantation in Shivganga River basin, Maharashtra. Study necessitates analysis of biophysical components such as soil properties, satellite image-based indices, weather conditions, etc. The soil samples were collected from well-distributed locations within the study area and analyzed for physico-chemical properties such as soil depth (SD), soil texture (ST), soil moisture (SM), soil organic matter (SOM), soil pHand primary soil nutrients such nitrogen, phosphate and potassium to recognize the categories of LS. Further, remote sensing data i.e. LANDSAT-TM imagery with 30 m spatial resolution was used to generate the land use/land cover classes. The bands of satellite image were compiled, fused and stacked using the open-source image processing software. All the thematic parameters were assigned a common coordinate system and projection and weighted overlay analysis (WOA) was performed.

Soil analysis
The analysis of soil properties is an important step in LSA. Total 34 sites in the proximity of dug wells were selected using random sampling technique for collection of soil samples. These soil samples were analyzed in laboratory to understand the physicochemical properties (Table 1). Along with these physicochemical soil parameters, land slopes were also considered in finding and assessing suitable area for plantation and afforestation. Moisture content in soil has vital impact on vegetation growth and distribution. Bhagat (2009) considered SM as an important parameter for finding potential areas for afforestation. The methodology involved in this study is shown in Figure 2.

Remote sensing image analysis
Supervised categorization algorithm was used for LULC analysis considering nine land utilization types. The map of soil depth was derived from the slope and LULC classes. Land with and without scrub is comprised into the shallow and thin soil class respectively. Irrigated crop lands, fallow lands and reserve plantation were classified into deep, moderately deep and marginally deep soils, respectively.

Baseline data creation
The study area map was derived using the drainage divide from contour of survey of India and updated with watershed atlas map. Physical elements has direct relation with land efficiency and agriculture activities (Adgo et al., 2013;Adimassu et al., 2016;Moeletsi et al., 2011). The decisive factor such slope, LULC, SD, ST, SM, soil EC, N, P and K have been frequently used for land suitability of plantation area (R. B. Jamali, 2014;Mustafa et al., 2011;Romano et al., 2015;Yalew et al., 2016) but the surface runoff firstly used for such analysis. The effect and wealth of these parameters are diverse according to the land physiognomies. The thematic layers such as spatial variation of soil parameters (ST, SM, SOM, pH. N, P and K) were generated by the IDW interpolation algorithm in GIS. The runoff is generated using intersect of LULC and rainfall data in ArcGIS environment.

MCDM analysis
The LSA have been done using MCDM-based AHP method. The procedure for LSA in the current study ( Figure 2) is performed through six steps: (1) generation of base maps with other thematic layer, (2) assortment of criterion, (3) defining of sub-criteria, (4) formation of pairwise comparison matrix, (5) creation of normalized pairwise judgement matrix and finally, (6) weights calculation.
The researcher's expertise inputs were taken into consideration while deciding the ranks of influential factors. Analytical hierarchy process establishes judgments between criterions that help to arrange complete list of factors (Saaty, 1997). The comparison matrix supports the assessment to allocate various  stages of significance of the features that present in land suitability analysis (R. B. . The weights to the LSA studies have been given for interpretation of comparative rank and land physiognomies (Bandyopadhyay et al., 2009;Mendas & Delali, 2012;Reshmidevi et al., 2009;Romano et al., 2015;Yalew et al., 2016). FAO (1976) guidelines have been used prerequisite of land-use classification. The high land suitability shows maximum effect of sub-condition while lowest values display least LS for afforestation (Tables 3). Weights for the parameters namely % slope, LULC, SD, ST, soil erosion, OM Ca, Mg and Na were allotted, as per their significance. Finally, land suitability map was prepared using overlay technique in GIS to detect suitable sites for afforestation to improve infiltration rate for groundwater conservation in the region. The resultant map was categorized into highly suitable, moderately suitable, marginally suitable and notsuitable classes.

Land use and land cover (LULC)
LULC plays major role in identification of area potential for LSA, for these two seasons' satellite images were used for detail understanding. The LULC map was generated using level-II classification system that includes agriculture land as major class that has been further classified into horticulture land, crop land, fruit orchard, etc., thus encompassing the whole gamut of green growth. Similarly, the waste land major class has been further classified into barren land, land with scrub, land without scrub and stony waste helpful in identification of land suitable for afforestation ( Figure 3).

Slope
The elevation-based slope is responsible for variation in soil humidity. For understanding LSA suitability for plantation, it is essential to consider percent slope and its direction for amassing of water (Olshevsky, 2008). The elevation of the study area ranges from 560 to1257m above MSL. The % slopes are classified into seven categories according to IMSD specification (IMSD, 1995). The nearly level 0-1% area is measured as large storage land owing a little rolling landscape with less surface flow. The part with comparatively moderate runoff and excellent penetration of water is very gently slope 1-3% favourable for plantation. The moderate surface flow causes relatively moderate penetration of water area with slope of 3-5%. The area with 10-15% slope having moderate soil depth and high surface flow causes relatively low penetration of water in sub-surface. This area is suitable for afforestation ( Figure 4). The area with 15-35% slope is main concern for afforestation to reduce soil loss and increase groundwater recharge.

Soil depth
Soil depth is a function of slope, runoff and rainfall pattern of the area (Bandyopadhyay et al., 2009;Rejani et al., 2017;Worqlul et al., 2017). Slope, LULC map, soil sample and field-based measurement were defining the depth of soil layer. The soil depth was also crosschecked with the surrounding well section as well as the geophysical investigation done in the study area validation of  the field experimental methods. Based on field observation the agriculture lands with nearly levelled land having deep soil depth were found at lower part of watershed. While the peripheral part of basin having forest as land use with moderate to steep slope having moderate to marginal depth of soil ( Figure 5). The cliff as well as near pediment part having steep slope show shallow soil. The soil depth represents the cropping pattern in the area, deep soil having rice as well as sugarcane, while parsley, etc., are found on thin soils.

Soil type
Soil has an important role in supporting the plant life in that area. The soil type is also validating with the experimental methods in laboratory analysis. As per the NBSS and LUP categorization and the experimental methods, the area is characterizing into sandy loamy soil, clayey loamy soil and clayey soil types ( Figure 6). The study region is majorly covered with well-drained sandy loamy soils having 48% of area with low nutrient present on moderate strong to steep slope goo for afforestation. The moderately drained loamy clay soil covers 43% of area, sandy soil has high infiltration rate due to greater porosity and permeability therefore it will be assigned highest priority. Clayey soil is compact and impervious hence given low priority rank.

pH of soil
The utmost informative features of soils are its pH. It defines the comparative contents of hydrogen ions in liquid. The pH of a regular soil is measured as an indicator of its transferable cation's saturation (USDA, 1954). The pH of soil ranged from 7.39 to 8.63 reflecting alkaline nature (Table 2) and slight difference in the pH range in the study area ( Figure  7). All superficial soil samples excluding S6, S8 and S16 having pH higher than 8, showing alkaline type of common soil samples (Table 2), is the results of high CaCO 3 concentration in host rock. Cumulative quantity of CaCO 3 in soil, may affect in increasing the heavy metal absorption. This could result into elevated buffering volume of soils and non-appearance of CO 3 in the saturation extract (Bagyaraj et al., 2013). The presence of solvable and transferrable Na with bicarbonate ions results in such high values. Further, there is a precipitation of CaCO 3 and MgCO 3 carbonates during evaporation contributing to such high pH values. High pH contents result into the growth of salinity/sodicity in soil of study area.

Soil calcium
Ca 2+ is main cation over other found cations in soil. This is because of the more adsorption potency of Ca 2+ ion in soil related to other positively charged ion (Miller, 1995). The Ca 2+ ion present in abundant amount to form the major transferable complex in soil media specially in the study area to occupy a major position on the exchange complex (Sharma & Raghubanshi, 2009). A brief look at the information from Table 2 shows that the transferrable Ca 2+ ion varies from 40.08 to 180.63 ppm ( Figure  8). The high proportion of Ca 2+ in the majority of the samples is suggestive of the occurrence of Ca-clay. Such soils with high Ca 2+ possess a good physical condition, develop a good crumb structure and allow the free movement of water and air. The soil Ca 2+ ion helps in maintaining the soil pH with playing important role in cell wall development for plant growth.

Soil magnesium
Magnesium is plentiful in the earth's shell. It is present in a varied variability of minerals. Weathering or break down of these minerals leads to the availability of magnesium for plant use. Though magnesium (Mg 2+ ) is a vital element for vegetal growth, its use in a manure activity obtains only slight importance. Mg 2+ is held on the superficial part of clay soil particle and organic matter particles. A brief look at the data from Table 2 shows that the Mg 2+ ion varies from 9.75 to 73.09 ppm. The less residence time of water in upstream parts of study region shows that low leaching of Mg 2+ ion (Figure 9). The prevalence of Mg 2+ in the majority of the soil samples due to photosynthesis, as it is a building block of the chlorophyll, which makes leaves appear green. This is particularly observed in the irrigated areas (S4, S7, S8, S13, S21 and S24).

Soil sodium
Plants are provided with sodium (Na + ) from the soil naturally, this is obtained from manures, insecticides, runoff from phreatic salt-loaded waters and the weathering of minerals which dissolve salt into the soil. Na + content in soil that is not at poisonous levels can simply be seeped out by reddening the soil with fresh water. The less percolation of water in upstream parts of study area shows comparatively high concentration of Na + ion due less leaching (Figure 10). The concentration of sodium in soil is generally less in the watershed, the sodium concentration shows the standard deviation of 4.72 ppm in soil, while overall concentration ranges from 7.25 to 25.20 ppm in the watershed (Table 2).

Soil bulk density
The ideal bulk density (BD) of sandy clay loam or loam for plant growth should be less than 1.60 gm/ cm 3 and for clay loam it should be less than 1.10 gm/ cm 3 . Soil sample numbers from 1 to 21 have bulk densities in the ideal range for plant growth; whereas, soil samples 22-25 have clay content >30% with maximum value 65.52% and sand <35% indicating that the soil samples with higher percentage of clay than sand show possibilities of affecting the root growth of plants. Because of the pore spaces, clay has the characteristics to hold water molecules; hence, the density of soil mass declines with increasing clay content, while with more compact soil with less pores, the bulk density of soil is higher. The bulk density is higher in the upstream part of basin but there is not much variation in it, spatial distribution map portraits similar pictures (Figure 11). This moderate value of BD shows the modest water holding capacity.

Soil organic matter
The organic matter (OM) is a vital constituent of soil that is main component to soil productiveness. According to Piccolo and Stevenson (1982), Soil OM includes all organic compounds in soils, excluding undecayed plant and mammals' tissues, their incomplete decay products, and soil biomass. The biological and physical characteristics of the soil are influenced by the organic matter. It improves WHC, aggregation, ventilation and soil structure. It is considered as a significant source of nutrients (N, P, S, B and Mo). The detailed investigation of data showed in Table 2 illustrations that OM concentration varies from 0.48 to 2.41% in the soils from the area. It was observed that 45% of samples have low OM ( Figure  12). The moderate soil OM shows that modest will be the water-holding capacity of soil. Therefore, these soils need adequate fertilizers through organic    compost, agricultural compost, plant manure, etc. The upper part of river basin sample S2 having lowest OM (0.48%) is moderately suitable for agriculture.

Soil boron
Boron (B) is a distinctive micronutrient necessary for regular development and progress of vegetation. The total content of B varies from 20 to 200 mg/kg (Mengel & Kirkby, 1987), and its content also differ significantly from one type of soil to other ( Figure 13). Various soil parameters that affect the uptake of Boron are soil pH, organic matter content and type of soil (Welch & Eswarakrishnan, 1991). The boron concentration in soil varies from 0.0874 to 0.967 mg/ kg, while high concentration is observed in lower part of basin due accumulation in soil particle.

Runoff
Run-off is a function of soil type and LULC through undulating topography. Meadows with gradient normally have small shallow soil storing space to retain run. Normally, agriculture yield and water inferiority problems are directly related to superficial overflow. Surface overland flow can remove and transfer top soil units, manures and insecticides from the ground location. Additional possible problem related with overland flow comprises of an absence of soil moistness in local zones of the field, crop nutrient deficit and eroded out seeds or plant (Rajasekhar et al., 2019). Study area is categorised into three zones as high runoff potential, moderately run-off potential and low run-off potential ( Figure 14).

Land suitability analysis
The ground signatures were used to allocate the scores to sub-criteria units. Lands having low to moderate gradients with thick soil cover are most suitable for agricultural practices. Uncultivated lands are dispersed in pediment regions with moderately deep soils; hence 4 th rank has been assigned to it which shows potentials for extension of agricultural activity. Third rank to moderate suitability land which is observed to be occupying sparse forest characterized by modest gradients and shallow soil depth.   Undergrowth lands representing sharp gradients and thin topsoils are marginally fit whereas waste and stony outcrop lands are unsuitable for green growth. The watershed is a representative of the semi-arid region from Western part, Maharashtra (India) that includes environmentally protected biodiversity hot spots. Hence, the thick-forested parts are categorized into the "null class" with rank one. In the present attempts, FAO specification has been used for LS analysis and found that only 9.16% land is "highly suitable" for plantation. These lands have 3-5% slopes, HSG-C and B soil type, more waterholding capability, soil moisture and low EC value. Nutrients like N-P-K are abstemiously available and exterior contributions are essential for plantation.
About 14% of the lands are of "moderately suitable" class ( Figure 15). The physiognomies of these lands are high slopes with small plain land, having thick soils in pediment region (50-90 cm), moderate waterholding capacity, soil moisture and runoff ( Table 5). The area under vegetation and very scant scrub have also been recommended for afforestation. However, it needs extra efforts for planning, natural resources conservation, etc.
Marginally suitable area for afforestation is assessed as around 14% of total river basin. It particularly shows thin soil cover, high slope, little water-holding volumes, poor soil moistness and potencies and high runoff.
Stepped lands having the problem of high soil erosion, so the use of this land for afforestation can also be done after proper perquisite conservation work.  Unsuitable areas for afforestation were assessed about 61% of the studied area. These areas have very steep gradients with stony area, waste area, shallow, dry soils, etc. (Table 4 and Figure 15). Farming area and medium to thick forests are not recommended for afforestation.

Conclusions
Land suitability maps for irrigated and dry farming in study area from Western India were generated using MCDM tool supported by GIS. Various derived thematic layers such as LULC, slope, soil depth, soil type, pH, soil calcium, soil magnesium, soil sodium, BD, OM, soil boron and run-off have been evaluated through RS and field data. The expert's opinion and analyses were used to identify the impact of specific criterion and calculated based on MCDM using the AHP that was opted to assign the weights. The present research has been carried out using FAO framework for LS analysis and it is found that only 9.16% land is "highly suitable", 14% of lands are of "moderately suitable", marginally suitable lands for plantation are assessed as about 14% of total study region and unsuitable areas for plantation were estimated about 61% of the studied area. Further, these zones were compared with an updated LULC map of the study region.
The results show that the areas that are categorised as "highly suitable" for plantation are already cultivated. However, areas categorised as "moderately suitable" are currently only being marginally used. The land suitability types namely "moderate suitable" and "not suitable" in suitability map are accurately assessed than the marginally and highly suitable in producers and users' point of view. The methodology, techniques and results of the study may be useful to assess the suitability of the land for plantation in the study area. This technique provides an effective method for creating land suitability maps based on complex evaluation criteria. This can considerably facilitate negotiations between land users and other stakeholders. Using the resulting set of suitability maps, one for each possible land use at a certain location, planners can quickly compare the scenarios. Therefore, the decisions could be taken with a good knowledge of the consequences, as well as the limitations in future land use development. Finally, the LSA is conducive to negotiate the objectives and constraints of data, making it an excellent tool to promote for democratic decision-making in the areas under spatial planning.

Disclosure statement
No potential conflict of interest was reported by the authors.  Medium suitability for afforestation under cautious management. It is likely to be done on flat land but there is need to conservation practices to be done on the land to reduced intensive erosion and drainage