1. Introduction
Land suitability evaluation is a procedure for assessing the prospective use of land for alternative purposes. The use of land for agriculture is considered critical areas in the scientific literature [
1,
2,
3,
4]. In general, land evaluation is the process of assessing the performance of land when used for a given purpose [
5,
6,
7]. An evaluation of the climate, soil, and topographical components, along with an understanding of the biophysical constraints, are used to determine land suitability. The evaluation process is a crucial stage in the advancement of agricultural activities. The principal objective of the suitability evaluation of agricultural land forecasts the land’s prospects and constraints of the land for the production of crops [
8,
9,
10,
11].
The multi-criteria evaluation (MCE) is an operational instrument for addressing decision-making issues using multiple criteria [
12,
13,
14]. The objective of the MCE is to examine selected possibilities based on multiple criteria and objectives [
15]. The MCE technique, along with the geographic information system (GIS) can assist to categorize, examine, and appropriately organise the accessible information regarding choice-possibilities for spatial planning [
16,
17,
18,
19,
20,
21].
The analytic hierarchy process (AHP) is a multi-criteria decision-making (MCDM) method that helps the decision maker facing a complex problem with multiple conflicting and subjective criteria. AHP was evolved to optimize decision making when one is faced with a mix of qualitative and quantitative factors. It has been very effective in making complicated decisions, especially in project management. AHP methods roughly make up a fourth of MCDM studies in various fields [
22]. The application of the AHP technique is profoundly considered; refer to [
23].
The AHP framework is a popular MCDM tool for formulating and analyzing decisions, which can be used for ranking alternatives and estimating criteria weights through pair-wise comparisons [
24,
25]. Moreover, Ishizaka and Nemery, 2013 [
26] pointed out that in problems involving the selection and ranking of multiple criteria, one of the most widely used methods is the analytic hierarchy process, or AHP.
The analytic hierarchy process (AHP) method proposed by Saaty, 1991 [
27] is a recognized multi-criteria technique that is incorporated into the GIS-based procedures for determining suitability [
28,
29,
30]. It computes the required weighting factors through a preference matrix that compares all the identified relevant criteria against one another. This is achieved through the use of reproducible preference factors, which then aggregate the weights of criterion map layers in the ordered weighted averaging (OWA) method. The AHP is particularly advantageous where it is impractical or difficult to identify the precise interactions between several evaluation criteria. Therefore, the OWA is a structure for the multi-criteria procedure in the GIS environment for aggregation [
31,
32,
33]. The use of OWA operators is influenced by their ability to execute various selected operators by stipulating an opposite set of ordered weights, and hence its application is widely reported in the literature [
34,
35,
36,
37,
38,
39,
40].
According to Yager, 1996 [
41], the standard operators in OWA have limited uses in large-set criteria for evaluation. Based on the multifaceted spatial decision conditions, the decision maker might experience difficulties providing detailed numerical parametric information on OWA [
12]. However, various strategic decision maps can be created by varying the weights of the orders in OWA.
Currently, the most popular techniques or decision rules in GIS are Boolean logic and weighted linear combination (WLC) [
42,
43,
44], which are comprehensively described within the context of OWA [
43]. However, [
45] proposed the renowned procedure of AHP. It is an important procedure applied to problems of spatial decision characterized by large numbers of conditions [
46]. As a multi-criteria decision-making method, the AHP has been applied for solving a wide variety of problems that involve complex criteria across different levels, where the interaction among criteria is common [
47,
48]. The AHP technique as an effective method is applied in the fields of soil management approach, e.g., a new method for soil health assessment [
49] based on landslide [
50], land-use change [
51], regional resources in soils [
52], and the determination of soil remediation schemes [
53].
Furthermore, the AHP technique pools the significance of every level of a “criteria tree” comprising representative criteria levels. For this setting, an insignificant number of criteria can be calculated [
54]. The integration of the OWA with AHP presents an effectively robust multi-criteria decision analysis (MCDA) instrument for problems regarding spatial decisions.
Although the use of MCDA for indistinctly assessing the suitability of agricultural sites is recognized, its adoption for mapping the suitability of the citrus site is lacking, at least to the best of the author’s knowledge. In addition, no studies have adopted the hydroclimatological impact assessment approach for categorizing suitable sites.
The aim of this study is to assist land managers with identifying areas with biophysical limitations based on the research criteria developed by the FAO (Food and Agriculture Organization) and modified by stakeholders. This study contributes to developing a GIS-based OWA approach for the citrus suitability assessment of the Ramsar district to identify the limitations that restrict the suitability classes for citrus production and recommend optimum citrus alternatives (scenarios) for improving citrus production.
The objectives of this research are to develop a GIS-based MCDA approach using the OWA and AHP methods for a citrus land suitability assessment model based on five (5) biophysical factors, and integrate the MCDA concept, as a core of GIS multi-criteria decision analysis, for the citrus-suitable regions in Ramsar district, Iran. To date, no research has been conducted on the MCDM approach, in particular, using the OWA method to access citrus site suitability in Iran, which will be discussed in this paper.
Although the developed methods can be used for citrus production management (CPM) through a set of spatial data, the issue of combining MCDM methods in the context of a GIS package is still considered as an important issue that few studies have tackled. Hence, GIS-based MCDA is an intelligent system that utilizes and converts spatial and non-spatial data into valuable information, which in addition to the judgment of the decision maker can be used to make a critical decision [
55,
56].
Overall, the contributions of this research are as follows:
An approach that combined AHP with OWA in a GIS environment is proffered. The present study focuses on an integrated citrus management strategy. Specifically, a GIS-based overlay analysis was performed to identify the optimum site for the citrus production that fulfilled all of the desired attributes.
Accordingly, this research tries to develop a new method, which is proposed by applying a novel GIS-based MCDM methodology for the assessment of citrus production.
A spatial framework adopting the AHP and OWA into the ArcGIS is used to evaluate the potential for the future expansion of citrus in Ramsar, Iran.
GIS is used for an important improvement to the conventional map overlay approaches.
Specifically, we appraise two key questions:
How can the citrus susceptibility problems be solved using a GIS-based OWA operator with a multi-criteria approach?
How can the critical factors by their relative weights, which are imported in GIS-based OWA capabilities, help decision makers with the citrus planning procedure now and in the future?
The remainder of this paper is organized as follows.
Section 1 examines the existing research trends regarding the site suitability location, investigates five (5) crucial factors, and sets the scope of the methodology of the decision-making procedure.
Section 2 proposes an OWA model in the GIS software for citrus growth potential and site values. The details of the OWA, including their comparisons, are described.
Section 3 describes the test results of the OWA model for the potential for citrus orchards in the case study area by utilizing five (5) critical factors in the GIS and IDRISI Selva software. In
Section 4 and
Section 5, we discuss the application of this paper for strategy and future work. It is envisaged that the results of this work will enable planners and policy makers to better understand the key aspects of citrus sustainable management.
3. Results
The suitability maps for citrus trees were derived from the criteria maps, and must be correlated to the findings of the AHP, based on the comparative significance of every criterion that was evaluated.
Table 4 presents the pair-wise conditions developed in AHP for each adopted standard. Eigenvalues were used to designate the relative importance weight of each criterion.
Table 4 represents that elevation; lowest temperature and rainfall are the most important site factors for citrus production in Ramsar, Iran.
Table 5 denotes that the most important condition for policymaking is elevation (46%). This is followed by a minimum temperature (23%), rainfall (14%), maximum temperature (10%), and slope angle (4%). As evaluated from Equation (4), the consistency ratio (CR) of the pair-wise matrix is 0.04, which indicates that the accumulated judgments derived from the pair-wise matrix from
Table 5 are satisfactory. The comparative significance of individual standards is denoted by the eigenvalue [
71]. Subsequently, the importance of individual site valuation standards was evaluated based on the comparative significance of each criterion and preferences according to specified decision rules. The selected criteria were: elevation, maximum temperature, minimum temperature, slope angle, rainfall, runoff coefficient, and slope, which were calculated by AHP through its pair-wise comparison matrix procedure.
The weights of the criteria are calculated by Equations (6–8), where the inconsistency index = 0.05; it is desirable to have a value of less than 0.1.
where a calculated CR value below 0.1 indicates a ratio that designates a rational level of regularity in the pair-wise appraisal by an expert judgment in AHP. The random index is denoted by RI, which is the consistency index of a randomly generated pair-wise comparison matrix. It is evident that the RI is influenced by the number of factors under comparison (
Table 6).
The OWA approach was employed to compute and aggregate various factors, weights, and constraint maps. In contrast to WLC, the approach of OWA requires additional weights that are termed order weights, and determine the combination of factor weights [
72,
73,
74]. Therefore, the terms are defined as;
Si = Suitability Index,
Wi = criteria weights,
Vj = criteria factor,
Xi = criteria score of weight, and
k = the number of criteria. OWA was obtained using the following function, Equation (9):
Consequently, the scale of one to three was selected such that one represents the least suitable, whereas three represents the most suitable. For the suitability analysis, each conditioning factor and the data layers listed were classified using OWA, as demonstrated in
Figure 5. The results of this study provide the optimal regions for citrus productivity deduced by superimposing the five crucial layers depicted as in
Figure 5.
The most potential sites evident from the lower elevation, optimum temperature, gentle slopes, and amount of rainfall regarding citrus trees requirement are given in
Figure 5. The general suitability map generated by constraint mapping demonstrates the relative ranking of the potential sites based on the important criteria. The suitability scores specify that citrus orchards are highly suitable on the site located in the lowland regions. Based on the general suitability score (see
Figure 6), 6.7% of the entire area is suitable for citrus growing, whereas 3.2% is unsuitable, and lastly, 90.1% is marginally suitable. This implies that the most suitable areas are located in northern parts, which have low elevation and flat topography (slope angle less than 15 degrees) and more than 800 mm year
−1 of rainfall, which provides sufficient rain-fed irrigation, particularly regarding the monthly rainfall in late spring and mid-summer. After superimposing all five (5) conditions from ArcGIS (version 10.1), the three (3) zones of productivity that are appropriate for the sustainable production of citrus were mapped out as depicted in
Figure 6.
From
Figure 5, most of the region under study lacks a suitable value for citrus, as evident by the values approaching zero in the fuzzy map. The suitability analysis specifies an alteration in citrus site suitability based on the OWA method, which is ascribed to the decision maker’s changeable acceptance of risk. Therefore, the attitudes of different decision makers need to be considered in the process of citrus planning. The MCDA is a distinctive comparative tool that contributes to the better integration of various activities during horticultural management. As a result, nearly the whole feasible area (6.7% of the total area) is categorized as “suitable”. In this case, 90.1% of the region is categorized as being “moderately suitable”.
Thus, it is evident that the MCDA based on the OWA approach can reliably simulate various predilections of the decision maker based on citrus implementation. The results show that the proposed method i.e., OWA, is more applicable to assessing suitable citrus cropland. It can be applied to the other regions using these techniques and predominant factors in agronomy and horticultural sciences. Besides, the final map that was adopted indicated that the southern parts of the area under study were unsuitable for cultivating citrus. Lastly, the northern area was deemed the most appropriate with regard to citrus fertility. The percentage of optimal areas for citrus production is summarized in
Table 7.
Thus, the final OWA map was generated by the GIS function for five parameters (see
Figure 5). Accordingly, as shown in
Table 7, the area with 6.7% (2348.01 hectares) is categorized as suitable in the OWA method. Second, in OWA, about 90.1% (66,359.46 hectares) of the predicted area was categorized as marginally suitable, and finally, in OWA, an amount of 3.2% (4,956.98 hectares) was determined to be unsuitable for cultivating citrus.
4. Discussion
According to the critical effective factors in the citrus site of each zone, GIS provides a suitable condition for data analysis. The GIS framework was adapted to determine the local priority of citrus activity in this study. AHP methodology has an important role in the quality and quantity classification. Accordingly, the result of OWA makes it possible to select a suitable area for cultivating citrus from the entire area in the Ramsar with three classes of relative priority.
The application of the OWA procedure can be adapted in a single operation, or as part of a hybrid scheme as a significant advantage of using this method. It entails that if one had five factors, where it was assigned that three should trade-off, but that two should not, the trouble can be resolved in two processes. In the first process, OWA would be run using order weights with the assigned degree of trade-off, followed by a second run without a trade-off in which the result of the first run was included as a supplementary factor. Factor weights can easily be set in the two runs to keep the original weights of all the factors.
The interesting feature of OWA is that it is possible to control a continuous degree of ANDORness and trade-off. In this case, OWA was derived from factor importance and different classification priorities along with a set of factor priorities and their internal classification priority; if the subject under study belongs to the highest class, decision making needs to be accomplished.
Therefore, OWA indicates that hydroclimatological evaluation classes are dividable to a different classification, and MCDM methods have only been used for cultivating citrus trees in determined places. It highlights that the determined weight has an important role in citrus cultivation factors. The result of this research in citrus can be used as an effective map of different factors in the OWA model. However, this may require finding justification for the non-acceptance of the proposed model in research for producers. More importantly, the OWA method can compliantly adhere to the diverse attitudinal decisions of the decision maker. It is theoretically effective, practically useful, and beneficial for dealing with problems in multiple criteria analysis. Conversely, a wide variety of strategic decisions can be made by altering the factors in the OWA technique.
5. Conclusions and Perspectives
In the GIS–MCE approach, two basic methods are used for environmental assessment: the WLC operators and the Boolean overlay operations. These procedures can be extracted through the scheme of the OWA. Hence, the study adopted a GIS-based framework based on the MCDM approach to examine the suitability of citrus cropland on a regional scale in Ramsar, Iran. It combines the proficiencies of OWA within the ArcGIS environment, and improves existing AHP modules. Furthermore, it improves the functionalities of ArcGIS by incorporating the multi-criteria decision analysis model. Lastly, the approach includes the opinions of experts on the standards and their weights, thereby providing a technique for directing the planner through the procedures required for the multi-criteria combination.
This study also reinforced the use of ordered weighted averaging (OWA) operators in AHP, which introduces the dimension of decision-maker’s attitude in the aggregation based on [
75,
76].
Suitability analysis allows identifying the main limiting factors for citrus production and enables decision makers to develop citrus management methods that enable increasing the citrus productivity. This paper proposes a quantitative and qualitative approach to suitable citrus cropland using GIS and MCE. A GIS-based MCE technique, using OWA strategies, examines the possible choices selected for problematic decisions by taking into account multiple conflicting criteria and objectives.
Hence, this study has outlined a GIS implementation of the ordered weighted average (OWA) operator that allows the exploration of a full range of operators such as intersection (AND) and union (OR). The GIS–OWA results are valuable tools that assist decision makers in order to determine suitable sites for optimum citrus production. In addition, the GIS-based MCE technique using OWA examined siting problems based on the various possible choices by addressing multiple criteria and contradictory objectives. The GIS–AHP module demonstrated that the tool was simple, spatial, and flexible. The results showed that the northern regions (mainly adjacent to the Caspian Sea) are the most suitable locations for citrus production in Ramsar. It allocates much more lands to the second level of suitability evaluation.
The hierarchical framework of AHP that was considered in this work presents an improved understanding of the problem and the importance of judgment required to reduce errors. The study successfully obtained numerous decision strategies and scenarios by altering the critical factors in the integrated method of OWA. The results derived from the OWA technique demonstrated that there are just a few areas, 6.7% of the total area, that were suitable for implementing citrus production. Lastly, the results highlighted that through altering the input criteria, a significant set of data, and the rules of decision, the multi-criteria analysis instrument is effective for varied disciplines. Accordingly, these include vulnerability valuation, land-use forecasting, and site assortment for agricultural (horticultural) crops.