Introduction

Land suitability analysis not only optimizes the use of land but also preserves natural resources for future generations. In recent decades, appropriate and comprehensive environmental planning has been designed based on identifying potentials and assessing land suitability. Comprehensive consideration of ecological capability can reduce the risk of conflicting natural and socio-economic interests and lead to sustainable development (Majnoniyan, 2000). Forest ecosystems, which have been significantly impacted by human activities, play an important and effective role in balancing land use (He et al., 2021). In many developing countries, extensive use of forests due to grazing, preparing firewood, and converting forest lands to agriculture has led to vast deforestation (Doggart et al., 2020; Paul & Banerjee, 2021).

This has led to afforestation planning by government and non-governmental organizations in response to the increasing demand for wood and wood fibers as well as to prevent further deforestation (Mohammadi et al., 2015; Zhao et al., 2021). Afforestation will also improve the hydrological performance of degraded forests and their surrounding areas. On the other hand, the employment of rural people in afforestation activities can improve the living standards of local communities and lead to sustainable development. Therefore, afforestation planning requires considerable attention to all the ecological and socio-economic characteristics of the region and inappropriate decision-making and planning can cause ecosystem instability and social conflicts (Mohammadi et al., 2015; Gholizadeh et al., 2020; Chen et al., 2021). In other words, afforestation planning without an in-depth consideration of ecological capabilities not only does not improve the environmental situation in the region, but also leads to more environmental degradation. Ignorance of ecological conditions and habitat characteristics in the past has led to failure and the unsustainability of afforestation. For instance, the quantitative and qualitative examinations of the afforestation with Cypress (Cupressus sempervirens var. horizontalis) in the eastern part of Mazandaran province, Iran, concluded that this afforestation has not been successful due to the lack of attention to habitat characteristics and planting in wet slopes and severe cold weather condition (Kiasari et al., 2010). Afforestation in Chah Afzal, Ardakan County, Yazd province, Iran, is another unsuccessful example due to the high salinity of soil and cold climate of the region (Amiraslani & Dragovich, 2011). To avoid similar scenarios, planners are encouraged to use new approaches such as multi-criteria decision-making techniques that can incorporate heterogeneous data and variables to make more informed and less subjective decisions (Greene et al., 2010).

Multi-criteria decision-making (MCDM) techniques can integrate diverse opinions and handle large amounts of complex information in the decision-making process (Liu et al., 2022a, b). Therefore, the practical application of MCDM techniques has become more common in land suitability studies in Iran such as afforestation planning (e.g., Hajjarian et al., 2016; Mohammadi & Limaei, 2018; Szulecka & Zalazar, 2017). The AHP (analytic hierarchical process) is one of the most common MCDM methods. This method, in combination with geographic information systems (GIS), is widely used to determine the relative weight of decision criteria and to assess ecological capabilities in land suitability and natural resource management (Malczewski, 2004; Ownegh et al., 2006). For example, Alemi et al. (2014) used AHP to identify the suitable area for afforestation of endangered species of yew (Taxus baccata) in Pooneh Aram reserve, Golestan province, Iran. Hashemi et al. (2014) used AHP to assess afforestation in Darab Kola, Miandorud County, Mazandaran province, Iran. Gholizadeh et al. (2020) also examined the AHP method to assess two afforestation plans with Quercus robur and Pinus sylvestris in northeastern Iran.

The fuzzy analytic hierarchical process (FAHP) and technique for order of preference by similarity to ideal solution (TOPSIS) are the other two common MCDM methods. FAHP method was derived from the AHP method and uses fuzzy numbers instead of absolute values. This method aims to overcome ambiguity and reduce uncertainty in the decision-making process. The TOPSIS method is based on the distance measure and was developed by Hwang and Yoon (1981). This method has less sensitivity to weighting the criteria (Malczewski, 1999) and chooses the option with the shortest geometric distance from the positive ideal solution and the longest geometric distance from the negative ideal solution (Mafi-Gholami et al., 2019, 2020).

Few studies in Iran have implemented the use of fuzzy AHP and TOPSIS in afforestation and foresty planning (e.g., Fazlollahi Mohammadi et al., 2014; Rahdari et al., 2019; Vatani et al., 2019). However, comparative analysis of the use of different MCDM methods in the field of afforestation in Iran is rare. The present study, therefore, uses a case study to compare the outcomes of AHP, fuzzy AHP, and TOPSIS in afforestation planning for the Siahpoosh Watershed, located in Ardabil province, Iran.

Materials and methods

Study area

Siahpoosh Watershed is located in the southern part of the city of Koraim, one of the southern cities of Nir city, Ardabil province, Iran (Fig. 1). The main access route to the watershed is through the Ardabil to Koraim main road, which leads to the watershed by passing through the city of Kuraim through the Khademloo side road. This region is located between 46° 06′ 35"–48°16′ 46" E longitude and 37° 46′ 37"–37° 54′ 37" N latitude with the total area of 10,103.4 ha.

Fig. 1
figure 1

Geographical location of Siahpoosh Watershed, Ardabil province, Iran

Full size image

This region is a semi-arid and cold area with the average annual temperature of 7.05 °C and 339.1 mm average of precipitation. The slope of terrain in this area varies between 0 and above 60%. The main soil textures observed in the region are sandy-loamy, loamy, clay, clay-sandy, clay-silty, and sandy-loamy-clay textures, and soil depth varies from very shallow to semi-deep.

Methods

The land suitability criteria used in this study were selected based on a comprehensive literature review of previous studies and an analysis of the regional characteristics (Szulecka & Zalazar, 2017; Zhang et al., 2019a, b; Rahdari et al., 2019; Liu et al., 2020; Xie et al., 2021; Quan et al., 2022). The suitability criteria were identified at three levels (i.e., main criteria and two sets of sub-criteria). The first level of this structure indicates the aim, which is the ecological capability assessment for afforestation in the study area. At the second level, the effective criteria for afforestation are presented, and the 2nd and 3rd sub-criteria (Fig. 2). These include physical and environmental factors such as the slope, aspect direction, elevation height, temperature, rainfall, soil texture, drainage, depth, erosion and vegetation cover, vegetation type, and vegetation density.

Fig. 2
figure 2

The flowchart of study

Full size image

In the next step, the suitability criteria were ranked by a panel of experts and the relative importance of each criterion was calculated using the AHP, FAHP, and TOPSIS methods. Finally, the suitable areas for afforestation were identified using the WLC equation (\(LS={\sum }_{i=1}^{n}\mathrm{Wi}\)). In this formula, LS is the suitability for particular land-use, n is the number of evaluated criteria, and Wi is the weight of each criterion.

AHP method

In the AHP method, the panel of experts is asked to rank the criteria and sub-criteria by referring to the numerical scale of 1–9, with a score of 1 representing indifference between the two criteria and 9 representing absolute importance (Saaty, 1980).

In this study, the data for pairwise comparisons were then analyzed using the EXPERT CHOICE software based on AHP algorithms to obtain the final ranking for each criterion as per the following steps:

  1. 1.

    Preference judgment (pairwise comparisons)

    The respondent measures the relative importance or priority of each criterion by making two-way comparisons between the decision elements and by assigning numerical scores indicating the priority or importance between the two decision elements (Ülengin et al., 2001).

  2. 2.

    Weighting the criteria and calculating their relative weight

    Relative weights are then calculated by the arithmetic mean method, in which the scores of each column in the paired matrix comparison are summed and then each score of the column is divided by the sum of the scores of that column (Eq. (1)). The resulting matrix is the “normalized comparison matrix.”

    $${\overline a}_{jk}=\frac{a_{jk}}{\sum_{l=1}^ma_{lk}}$$
    (1)

    where \({a}_{jk}\): score of the column, \(\sum_{l=1}^{m}{a}_{lk}\): the sum of the scores of that column, and \({\overline{a} }_{jk}\): normalized comparison matrix.

  3. 3.

    Average relative weight

    The scores of each row in the “normalized comparisons matrix” are averaged (Eq. (2)), and this mean represents the relative weight of the decision elements in the rows of the matrix.

    $${w}_{j}=\frac{\sum_{l=1}^{m}{\overline{a} }_{jl}}{m}$$
    (2)

    where \(\frac{\sum_{l=1}^{m}{\overline{a} }_{jl}}{m}\): mean row, and \({w}_{j}\): relative weight.

  4. 4.

    Calculating the final weight

    The final weight is obtained by multiplying the relative weight of each element by the weight of the higher elements (Eq. (3)).

    $$v=\mathrm {S.w}$$
    (3)

    where S: the weight of the higher elements, w: the relative weight of each element, and \(v\): the final weight.

  5. 5.

    Calculation of consistency

    The consistency ratio (CR) shows the consistency of comparisons and indicates the level of correctness of priorities resulting from group members or their combination. This index is measured using Eqs. (4) and (5).

    $$\mathrm{CI}=\frac{{\lambda }_{max}-n}{n-1}$$
    (4)

    where CI: consistency index, \({\lambda }_{max}\): the principal eigenvalue of the judgment matrix, and n: matrix measure.

    $$CR=\frac{CI}{RI}$$
    (5)

    where CR: consistency ratio, CI: consistency index, and \(RI\): the random consistency index (see Table 1).

    Table 1 Random consistency index
    Full size table

    Finally, the potential areas for afforestation were identified and classified using the WLC method and obtained coefficients from the information layers.

FAHP method

The FAHP is a systematic method that uses fuzzy set theory and hierarchical structure analysis. Its graph and paired matrix comparison in the fuzzy form are similar to the non-fuzzy form. However, comparisons were carried out using the fuzzy method (Table 2) and weights were calculated by the improved fuzzy AHP (Buckley technique) (Buckley, 1985).

Table 2 Triangular fuzzy number (Lin, 2010)
Full size table

To create a fuzzy layer, the raster layers were first standardized in the IDRISI operating environment using membership functions (user-defined, decremental line, incremental line, and decremental S-shape) and were converted to values (0, 1) in the raster format, in which 0 and 1 indicate the most and least priority, respectively (Table 3). Then, the standardized layers were multiplied by each of the relative weights obtained by Buckley’s (improved fuzzy) method and turned into fuzzy weighted layers.

Table 3 The function used for criteria fuzzification
Full size table

The steps of Buckley’s fuzzy method are as follows:

  1. 1.

    Fuzzification (triangular): To evaluate the importance of criteria, real scalar values are converted into a triangular fuzzy value with 3 elements whose membership function is shown in Eq. (6). In this model, the value of the membership function is 1 for (m) (Kaufmann & Gupta, 1991).

    $$\mathrm{A\ triangular\ fuzzy\ number}\ \;\widetilde T=\left(l, m, u\right):$$
    (6)

    where: \({\mu }_{\tilde{T }} \left(X\right)=\begin{cases}\frac{x-l}{m-l}.\ l\le x\le m\\ \frac{u-x}{u-m}.\ m\le x\le u\\ 0.\ otherwise \end{cases}\)  

    Algebraic operations on fuzzy numbers are similar to those on real numbers. Equation (7) shows these calculations, including the addition and multiplication of two fuzzy numbers.

    $$\begin{array}{cc}{\widetilde T}_1\oplus{\widetilde T}_2=\left(l_1+l_2.m_1+m_2.u_1+u_2\right)&\mathrm{where}:T_1=\left(l_1.m_1.u_1\right):\mathrm a\;\mathrm t\mathrm r\mathrm i\mathrm a\mathrm n\mathrm g\mathrm u\mathrm l\mathrm a\mathrm r\;\mathrm f\mathrm u\mathrm z\mathrm z\mathrm y\;\mathrm n\mathrm u\mathrm m\mathrm b\mathrm e\mathrm r.\\{\widetilde T}_1\otimes{\widetilde T}_2\cong\left(l_1\times l_{2.}m_1\times m_{2.}u_1\times u_2\right)&\mathrm{where}:T_2=\left(l_2.m_2.u_2\right):\mathrm a\;\mathrm t\mathrm r\mathrm i\mathrm a\mathrm n\mathrm g\mathrm u\mathrm l\mathrm a\mathrm r\;\mathrm f\mathrm u\mathrm z\mathrm z\mathrm y\;\mathrm n\mathrm u\mathrm m\mathrm b\mathrm e\mathrm r.\end{array}$$
    (7)
  2. 2.

    The geometric mean of rows: it is calculated using Eq. (8). This step is the first step of the improved fuzzy AHP method, in which the geometric mean of the rows should be calculated based on the following equation. The geometric mean of the first, second, and third elements is considered because the numbers in each row are fuzzy.

    $$\widetilde{r_i}=\left.\left({\prod_{j=1}^n}{\widetilde t}_{ij}\right.\right)^{1/n}$$
    (8)

    where \({\tilde{t }}_{ij}\): fuzzy weight criterioni from expert n, and \({\tilde{r }}_{i}\): geometric mean of rows.

  3. 3.

    Multiplying the geometric mean of the rows by the inverse of the sum of the geometric mean: First, the geometric mean calculated in the previous step is summed, and then each geometric mean is multiplied by the inverse of this sum (Eq. (9)).

    $${w}_{i}={r}_{i}\otimes {\left({r}_{1}\oplus {r}_{2}\oplus \dots \oplus {r}_{m}\right)}^{-1}$$
    (9)

    where \({\tilde{r }}_{i}\): geometric mean of the rows, and \({w}_{i}\): multiplying the geometric mean of the rows by the inverse of the sum of the geometric mean.

  4. 4.

    Defuzzification of weighted fuzzy mean: Eq. (10) was used for the defuzzification of the weighted fuzzy mean obtained in the previous step.

    $${w}_{crisp}=\frac{l+2m+u}{4}$$
    (10)

    where (l, m, u): a triangular fuzzy number, and \({w}_{crisp}\): weights defuzzed.

  5. 5.

    Normalizing the weight of the criteria by the linear normalization method: Each weight defuzzed in the previous step is divided by the sum of the weights to obtain the normalized weight (Eq. (11)).

    $${\tilde{r }}_{ij={\tilde{w }}_{i}=}\frac{{z}_{i}}{{\sum }_{i=1}^{n}{\tilde{z }}_{i}}$$
    (11)

    where \({z}_{i}\): weights defuzzed, \({\sum }_{i=1}^{n}{\tilde{z }}_{i}\): sum of the weights, and \({\tilde{r }}_{ij={\tilde{w }}_{i}}\): normalizing the weight.

  6. 6.

    The final weight of each sub-criterion is determined by Eq. (12).

    $$u_{i=}{{\sum}_{j=1}^n}w_jr_{ij}$$
    (12)

    where \({w}_{j}\): weight of each sub-criterion, \({r}_{ij}\): normalizing the weight of the sub-criteria, and \({u}_{i}\): final weight of sub-criterion.

  7. 7.

    Then, the layers were overlapped using the WLC method and acquired coefficients to obtain the final map.

TOPSIS method

TOPSIS is used to prioritize options based on their similarities to the ideal solution. The prioritized option should have the shortest distance from the ideal solution and the farthest distance from the anti ideal solution. This method is a suitable compensatory multi-criteria decision-making technique for prioritizing options based on the similarity to the ideal solution and has very little sensitivity to weighting. The selected option has the shortest distance from the ideal solution and the farthest distance from the anti ideal solution. The fuzzy hierarchical analysis was used to extract pairwise comparisons between criteria, sub-criteria, and relative weights. The final ordering of the options was obtained using the TOPSIS technique in Excel. The final map was obtained after overlapping layers using the WLC method.

The steps are carried out as the following:

  1. 1.

    Creating a data matrix based on n indices and m options (Eq. (13)).

    $$\begin{array}{ccc}{X}_{1}& {X}_{2}& \begin{array}{cc}\dots & {X}_{n}\end{array}\end{array}$$
    $$\mathrm X=\begin{array}{c}A_1\\A_2\\\begin{array}{c}.\\.\\.\\A_m\end{array}\end{array}\begin{bmatrix}X_{11}&X_{12}&\dots&X_{1n}\\X_{21}&X_{22}&\dots&X_{2n}\\.&.&&.\\.&.&&.\\.&.&&.\\X_{m1}&X_{m2}&&X_{mn}\end{bmatrix}$$
    (13)

    where \({A}_{I}\): m options, and \({X}_{IJ}\): the numerical value obtained from options i relative to the indices j.

  2. 2.

    Non-scaling the decision matrix (normalizing the decision matrix) is done through Eq. (14).

    $${r}_{ij}=\frac{{x}_{ij}}{\sqrt{{\sum }_{i=1}^{m}{x}_{ij}^{2}}}$$
    (14)

    where \({x}_{ij}\): the numerical value obtained from options i relative to the indices j, \(\sqrt{{\sum }_{i=1}^{m}{x}_{ij}^{2}}\): the square root of the squares is the numerical value obtained from options i relative to the indices j, and \({r}_{ij}\): normalized matrix.

  3. 3.

    Weighting each criterion: the sum of weights (W) obtained in Eq. (15) is multiplied by the normalized matrix (rij).

    $$W=\left({w}_{1}, {w}_{2}.\dots {.w}_{j}.\dots {.w}_{n}\right)$$
    (15)

    where \(\sum_{j=1}^{n}{w}_{j}=1\)

  4. 4.

    Determining the distance of option (i) from the ideal point (highest performance of each criterion) (Eq. (16)).

    $${d}_{i}^{+}=\sqrt{{\sum}_{j=1}^{n}{\left({v}_{ij}-{v}_{j}^{+}\right)}^{2}}$$
    (16)

    where \({v}_{ij}\): numeric value of option, \({v}_{j}^{+}\): positive idea, \({d}_{i}^{+}\): distance from the positive idea.

  5. 5.

    Determining the distance of the option (i) from the anti ideal point (lowest performance of each criterion) (Eq. (17)).

    $${d}_{i}^{-}=\sqrt{{\sum}_{j=1}^{n}{\left({v}_{ij}-{v}_{j}^{-}\right)}^{2}}$$
    (17)

    where \({v}_{ij}\): numeric value of option, \({v}_{j}^{-}\): negative idea, \({d}_{i}^{-}\): distance from the negative idea.

  6. 6.

    Developing a distance measure over each criterion to both ideal point (Ai+) and nadir point (Ai−) (calculating the similarity index) and prioritizing the options: This index represents the score of each option which is equal to Ai− divided by the total distance of Ai− and Ai+ (denoted by Ci*) (Eq. (18)).

    $${c}_{i*}=\frac{{d}_{i}^{-}}{{d}_{i}^{+}+ {d}_{i}^{-}}$$
    (18)

    where \({d}_{i}^{-}\): distance from the negative idea, \({d}_{i}^{+}\): distance from the positive ideal, \({c}_{i*}\): similarity index.

Results

As discussed previously, Buckley’s AHP and FAHP methods were used to weigh the land suitability criteria defined in this research. Fifteen questionnaires were prepared and sent to experts to perform pairwise comparisons. The weight of each sub-criteria was calculated using AHP and FAHP methods for forestry suitability evaluation (Table 4). As shown in Table 4, the highest and lowest weight is assigned to the rainfall and the erosion criteria, respectively. The consistency ratio is less than 0.1, confirming the accuracy of this step. In the FAHP method, the inconsistency rate is also lower than 0.1, indicating the consistency of the fuzzy pairwise comparison matrix. After measuring the final weights of each layer, the spatial database of the study area was formed in the ArcGIS 10.3 software and the layers were overlapped using the WLC method.

Table 4 Weight of each sub-criteria using the AHP and FAHP (Buckley) method
Full size table

The final map of land suitability for afforestation in the region was prepared. The results of the AHP method showed that about 65.7% of the area (about 6639 ha) was medium suitable, and 20.6% (about 2084 ha) was low suitable and very low suitable (Fig. 3 and Table 5).

Fig. 3
figure 3

Final map of afforestation with AHP method

Full size image
Table 5 Area of different classes in the two AHP and FAHP methods
Full size table

In the results of Buckley’s FAHP method (Fig. 4), about 39.2% of the study area (3973.5 ha) was very suitable. Moreover, 0.14% of the area (15 ha) was unsuitable and very unsuitable, 38% (39.2 ha) was suitable, and 8.8% (893.5 ha) was medium suitable (Fig. 4 and Table 6).

Fig. 4
figure 4

Final map of afforestation with FAHP method

Full size image
Table 6 Rank of the polygon for afforestation
Full size table

Table 6 shows the comparison of the area and its percentage in the two methods.

The final maps prepared by these two methods were examined using ecological criteria, Google earth images, and field observations. The results showed that the map prepared by the FAHP method is more realistic and consequently was used as the basis of the TOPSIS method.

TOPSIS method

After determining areas with high suitability for afforestation, it is necessary to determine the priority of options. Although there are various methods and techniques for the MCDM, the TOPSIS method is less sensitive than the weighting method (Malczewski, 1999). Therefore, the TOPSIS method was used to rank the options (117 polygon areas) selected by the FAHP method. It is necessary to use weights obtained from FAHP to make the calculations. After creating the matrix and entering the homogenous data in Excel software, the results of the weights with homogenous units were obtained shown in Fig. 5 and Table 6.

Fig. 5
figure 5

Results of the TOPSIS method rank the options for afforestation

Full size image

According to Table 6, polygon areas with No. 267, 268, 269, 270, 249, 244, 245, 246, 247, 248, 300, 301, 296, 219, 294, 295, 297, 220, 263, 264, 265, 262, 266, and 185 (647 ha) are the best area for afforestation (Table 7).

Table 7 The suitable areas for afforestation
Full size table

Figure 6 shows the area with high suitability (in 5 priorities) in an implementation plan for afforestation.

Fig. 6
figure 6

Priority map of the most suitable polygons for afforestation

Full size image

Discussion

The land suitability criteria used in this study were defined based on a comprehensive review of previous studies reported in the literature. Ecological criteria including slope, aspect, altitude, soil (depth, texture, and drainage), climate, vegetation (type), and land use were used to assess and classify the study area for afforestation (Babaei, 2006). Hosseinzadeh (2007) used slope, aspect, altitude, soil (depth, texture), geology, climate, and vegetation (type) to assess ecological capability in the Galanderoud 48 (Kodir Sar, Nur County, Mazandaran Province, Iran). Loi and Tuan (2008) used slope, aspect, altitude, soil suitability, climate, and vegetation (type) within a GIS to perform the land suitability assessment in the forest of Tatin, Vietnam. Slope, aspect, altitude, soil (depth, texture, and erosion), and climate were used for the ecological capability assessment of afforestation (Shamseh, 2010). Slope, aspect, altitude, soil (depth, texture, drainage, PH, EC, OM, and Caco3), climate, and vegetation (type) were used in a GIS to assess the land suitability for afforestation (Dengiz et al., 2010). Rahimizadeh et al. (2012) determined suitable species for afforestation in the southern part of the Alborz mountains based on aspect, altitude, soil (texture and drainage), and climate. Zare et al. (2011) used slope, aspect, altitude, soil (depth, texture, and drainage), and climate. Moradzadeh et al. (2011) assessed the ecological capability for afforestation using slope, aspect, altitude, and soil (depth, texture, organic matter, and erosion) in the Dadabad watershed forest, Lorestan province, Iran.

Figure 2 lists criteria used in this study, which include physical, biological, and socio-economical factors such as the slope, aspect, altitude, soil (depth, texture, drainage, and erosion), climate, vegetation (type and cover), and land use. Based on the literature review, in this study, a comprehensive set of criteria was used to assess the land suitability. The land cover map, which has been used in few studies (Babaei, 2006), was incorporated into the analysis. This criterion is very important because it determines the socio-economic restriction of the region for afforestation.

Previous studies showed that the use of AHP and FAHP is the most straightforward approach for either land suitability or land vulnerability assessments. Amiri et al. (2009) assessed the ecological capability for forestry use in the northern part of Iran. After determining the ecological parameters, the fuzzification of effective criteria in forestry use was carried out using linear and nonlinear membership functions. Then, these criteria were weighted by the AHP method and the final map was prepared in GIS. They concluded that the weighting and fuzzification of criteria using the MCDM methods have an important role in the land suitability assessment. In another study, Amir Amadi and Mozaffari (2012) analyzed appropriate zones for ecotourism development using GIS-based techniques that prepared the required information layers and then overlapped them using the AHP weights. Greene et al. (2010) and Phua and Minowa (2005) have also used a combination of MCDM and GIS for forestry. The WLC method enables decision-makers to involve more important factors in the land suitability assessment, and the results are more accurate and reliable compared to its other spectra, confirmed by other studies (Malik & Bhat, 2015). The WLC method has been used to produce a land capability map for forestry in the Behbahan suburb) Rahimi et al., 2015) and a suitable place for establishing a forest park in the Badreh county of the Ilam province (Piran et al., 2013).

This study demonstrated that the results of the FAHP were closer to the reality and the ecological condition of the region, which support the findings in some of the previous studies. For example, Chan and Kumar (2007), Hamzeh et al. (2014), and Rezaei and Jamshidi Zanjani (2017) also reported that the FAHP method produced more accurate and reliable results in land suitability analysis.

This study went one step further in using a triangular improved fuzzy method (Buckley) to perform FAHP. Buckley’s method is known to overcome some of the limitations in commonly used Chang’s fuzzy model. The TOPSIS method was then used to rank the priority of the suitable areas determined by the FAHP method, as suggested in previous studies (e.g., Chu, 2002; Alavi & Alinejad-Rokny, 2011; Fazlollahi Mohammadi et al., 2014; Patawaran et al., 2019; Sabir et al., 2020). According to the results, the western portions of the study area were identified as the most suitable for afforestion, whereas the eastern, northern, and eastern north portions ranked as the least priority.

Conclusions

The natural environment has a specific potential for human use. Thus, the ecological capability assessment should be carried out with principled planning before land use planning. In this study, three MCDM methods, namely, AHP, FAHP, and TOPSIS, were used to identify the most suitable areas for afforestation. This study combined the results of FAHP and TOPSIS within a GIS environment to locate suitable locations for afforestation and demonstrated that the MCDM techniques can be a great help in ranking the best available solutions. Therefore, afforestation projects informed by MCDM will make afforestation more efficient. In future research, the utility of other MCDM methods (e.g., ANP, SAW, PROMETHEE, and ELECTRE) can be compared to provide better insights for method selection.