Life cannot exist without water. By the middle of the twenty-first century, water shortages are predicted to be widespread due to possible climate change and present usage patterns [1
]. Due to the exponential growth in the population over the past few decades, the demand for water has increased dramatically. The Food and Agriculture Organization (FAO) reports that the global water demand is growing twice as quickly as the global population [2
]. Drought conditions have impacted many countries in the Middle East recently (e.g., Iraq, Iran, Syria, and Turkey) [3
]. Insufficient water storage and the lack of irrigation water negatively impact agriculture production. In addition, increased industrial and domestic water use has exaggerated the drought conditions [4
]. According to several studies, Iraq is anticipated to face additional difficulties as the country’s water deficit crisis worsens with time [5
]. RWH is a promising water conservation technique, and it can deliver accessible, inexpensive, and safe water for drinking, residential usage, and agricultural purposes [7
]. RWH techniques have been recognized as an efficient alternative water supply for the agriculture sector to deal with water scarcity, lowering the reliance on groundwater in arid and semi-arid regions [8
RWH refers to any methods for gathering and storing runoff water for home and agricultural uses [11
]. It is an age-old strategy that is still employed to keep rainfall in reserve. Planning and creating techniques for gathering, storing, and utilizing rain for agricultural purposes has been a priority for many nations [12
An RWH site’s suitability is determined by several factors, including biophysical (rainfall, slope, soil type, drainage network, and land use) and socioeconomic conditions [13
]. The FAO established six criteria to identify RWH locations: hydrology, climate, topography, soils, socio-economics, and agronomy [15
GIS and RS data have had significant roles in managing and planning water resources. RS data and GIS are essential to maximizing water availability and land productivity [16
]. RS is an economically valuable tool for locating potential RWH locations because it frequently covers vast, difficult-to-reach areas. GIS and RS can solve the missing data problems for finding viable RWH sites [17
]. The accuracy of the results obtained from RS and GIS techniques depends on the quality (i.e., temporal, spectral, and spatial resolution) and data availability. Satellite-based RS data offer trustworthy baseline characteristics such as soil type, land use/cover, geomorphology, lineaments, etc. [18
]. These thematic layers are often combined using the GIS approach [19
Moreover, several site selection techniques or methodologies are available, and some of the literature has used GIS and RS data with different decision-making analyses [9
]. In this context, GIS and RS data are used [16
], Multi-Criteria Analysis (MCA) integrated with GIS [24
], AHP with GIS [27
], and Weighted Linear Combination (WLC) and Boolean techniques were also used [7
]. Several factors, including the available data, the opinions of experts, and the study area’s ecology, have been considered in locating RWH sites [15
]. Mbilinyi et al. [24
] utilized rainfall, soil texture, slope, soil depth, land use/land cover, and drainage as input data for finding suitable RWH sites. Khashei et al. [34
] used AHP and FAHP multicriteria decision-making to determine suitable zones for drinking water in the Birjand aquifer in Iran. The study demonstrated that the map of FAHP is much closer to reality than the AHP.
Numerous recent droughts have hit the study area. As a result, the Kurdistan Regional Government (KRG) decided to adopt RWH techniques there because many agricultural lands failed to produce crops because of insufficient rainfall throughout the growing season. Most of the villagers in the area also raise animals, which has presented difficulties for them due to a water shortage, notably in the summer of 2022. Evaluating RWH potential and selecting appropriate RWH sites is a time and labor-intensive task for water resource managers and planners when applied to large scales. The site selection process for RWH structures has been made simpler and quicker using GIS and RS technologies, which have reduced the number of recommended sites and allowed the selection of only the best sites.
To the authors’ knowledge, there is no surface RWH study in the Garmian region. The present study aims at identifying appropriate sites for RWH in the Garmian region by integrating MCA, GIS, and RS. For comparison, two MCA methods are utilized to find possible RWH sites: AHP and FAHP. The WLC and fuzzy gamma approaches are then used to create a prospective RWH suitability map. The novelty of this study is that, in contrast to most past research in the field of RWH, which has depended on the lowest and maximum local rainfall values, a new approach for rainfall classification is suggested to develop representative suitability maps [7
]. This study will be helpful for decision-makers in constructing dams and ponds in the region.
3. Results and Discussion
3.1. Selection of Multi-Criteria
WLC was used to compute the sum of the weighted criteria. To carry out the WLC, an AHP was employed, known as a pairwise comparison. AHP is an MCA method executed within Arc-GIS, defining the criteria weights. The weight of each criterion is determined using a matrix of pairwise comparisons. To make sure that the comparison of the criteria is acceptable, the consistency ratio should be less than 10% [49
]. The pairwise matrix for the five layers used in this study is illustrated in Table 4
The precision of the pairwise comparison was evaluated by calculating the Consistency Ratio (CR
). The consistency ratio was used to calculate the relative weight of each criterion. The Consistency Index (CI
) to Random Index (RI
) Ratio is known as the Consistency Ratio [30
]. A consistency ratio lower than 10% indicates that the comparison between the factors was considered acceptable [4
where the consistency index (CI
) depends on the consistency vector (ʎ
) and a number of criteria (n
The criteria weight is determined by dividing each cell in the matrix by its column summed value in Table 4
. Table 5
shows the generated weight of the criteria.
The consistency ratio is calculated using the following procedure.
The consistency indices are determined from Equation (3), which equals 0.2125.
According to Hameed [4
], the random index corresponding to the five criteria is 1.12 (see Table 6
The Consistency Ratio is determined from Equation (2), which equals 1.89%.
From the pairwise matrix result, the rainfall is the most effective layer, and the LULC has the lowest effect, as shown in Table 7
3.2. Weighted Linear Combination
WLC is used in this study for overlaying the thematic maps for determining the potential RWH map for the Garmian region. The WLC method for selecting potential sites for water harvesting has been widely used recently [7
]. Implementing the WLC technique involves standardizing each criteria map, determining relative importance weights, and integrating the standardized maps following the related weights to produce a final map, as in Equation (4).
S = the suitable site.
= the ith criteria weight.
= the membership value of ith criteria
3.3. Potential RWH Map from AHP and WLC
The selection of feasible sites for RWH is essential in maximizing water availability and recharge. The process of locating RWH structures relies on the watershed’s characteristics. Five thematic layers, such as soil texture, slope, drainage density, rainfall, and land use land cover, are considered to determine proper sites for RWH structures.
RWH suitability maps were determined using MCA. The five thematic layers were added (Rainfall, Slope, Soil texture, LULC, and drainage density) using the WLC method. Thematic maps were divided into five categories according to their appropriateness for water harvesting: very low, low, moderate, high, and very high.
The results indicated that 8%, 20%, 35.6%, 30%, and 6.4% of the sites have very high, high, moderate, low, and very low suitability for rainwater harvesting, respectively, as shown in Figure 11
. The northeast of the Garmian Region has a very high level of suitability for RWH.
The area north of Kalar city is classified as classes 3 and 2 (due to the lower rainfall class). Despite the low degree of suitability, RWH is urgently needed in that area since the effects of the drought are more severe there.
It is evident from the results that the area along the Sirwan river is identified as a very highly suitable zone for RWH, as most of the drainage networks drain into this river, and the slope and topography of the valley promote a good place for water harvesting.
The areas with low or moderate suitability could also be used for rainwater harvesting because those areas receive higher rainfall depths (i.e., 377 to 493 mm); however, the cost should be considered. Figure 12
shows the potential RWH sites.
3.4. Potential RWH Map from FAHP and Fuzzy Gamma
The weights for the criteria used to choose the harvesting locations have been assigned using the fuzzy AHP. Each number in the AHP pairwise matrix has been changed to a triangular fuzzy number to find the weight of the criteria by using the method introduced by Buckley in 1985 [50
], as shown in Table 8
. Each fuzzy criteria’s geometric mean value was obtained using the geometric mean method by Equation (1). The result of the geometric mean of the criteria is illustrated in Table 9
Fuzzy AHP aids decision-makers in focusing on more crucial factors as opposed to traditional AHP. The traditional AHP technique does not allow for situations where criteria are “zero weighted”. The numerical weight of the criteria will, however, be close to zero if one criterion is assigned less weight than all of the others. [53
]. Using Table 8
, the geometric mean is calculated for each criterion (see Table 9
For the above table, the summation of each column is calculated. On the other hand, the inverse of the summation row is generated and assorted from small to large inverse. The center of area (COA) method, introduced by Chou and Chang in 2008 [53
], is then used to apply defuzzification, which turns the fuzzy results obtained by aggregation methods into crisp values. Because fuzzy sets include incomplete ranking, a crisp value is more understandable and more straightforward for the final comparison [57
]. The weights are normalized by dividing each weight by the sum. The resulting weight of the FAHP approach is illustrated in Table 10
The results indicate that the weight of the rainfall criteria decreased, while the slope and soil texture weights increased. There is little difference between the AHP and fuzzy-AHP weights for the five criteria used in the study. The thematic maps are standardized by linear fuzzy membership to overlay by fuzzy gamma to generate the suitability map for RWH in the Garmian region. The fuzzified maps of the layers are presented in Figure 13
The final step is overlaying the layers by fuzzy gamma overlay. A spatial analyst tool is used in ArcGIS to overlay the layers. The potential RWH map is generated by overlaying the layers using the weights obtained from FAHP, as shown in Figure 14
. The map resulting from the Fuzzy overlay has a high spatial indication of the Sirwan river, which is classed as a water body (i.e., very low suitability class); however, the area around the stream is suited with very high suitability. The area with very high suitability covers approximately 10.2% of the entire region. The area with high suitability is located in the north part of the study area and covers 18.7%, while moderate, low, and very low suitability zones constitute 24.5%, 21.3%, and 25.3%, respectively.
There were considerable variations between the suitability map produced by the AHP and FAHP techniques; particularly in classifying class 1 (see Table 11
). In general; they indicated that most of the areas in the south and southwest are within class 1 (i.e., low suitability).
However, based on the authors’ experience with this area, such a result is unrealistic, as it indicated that most areas in the south and southwest of the Garmian Region are of low suitability. The rainfall criteria, which has a weight of about 40%, greatly affected the results. The lowest annual rainfall depth in the study area is 313.8 mm, and this value has been given grade 1 in the classification of the rainfall criteria as it is usually adopted in most of the literature.
The NPRC will be adopted to generate a new suitability map. The new classification for rainfall is illustrated in Table 12
. The WLC and Fuzzy Gamma overlay is conducted to construct the new suitability map for RWH in the Garmian region, as presented in Figure 15
and Figure 16
The suitability maps obtained using the new rainfall classification are significantly different and look more realistic. It is found that the classification of the criteria has a vital role in finding suitable sites for RWH. The criteria weights found by both AHP and FAHP are not much different. However, the resulting maps are very different, and the Fuzzy gamma overlay map is much more accurate than the WLC overlay. The accuracy of the two overlays may be assessed using the Awa Spi dam shown in Figure 17
. Although the dam’s surrounding area has a suitability level of 4 to 5, the FAHP approach indicates the dam’s reservoir area has low suitability (water body). This result shows the superiority of the FAHP over the AHP approach.
Additionally, the FAHP-based classified locations are isolated and easily distinguishable in the suitability map. This advantage of the FAHP approach has been indicated in some literature [34
The percentage of suitable area per class is presented in Table 13
for both MCA methods used in the study before and after changing the rainfall layer’s classification.
Recently, there have been several droughts in the Garmian region, which have affected the agriculture and wildlife industries. This study has identified potential RWH sites in the Garmian region using GIS and RS. Five factors were used: slope, rainfall, land use, land cover, soil texture, and drainage density. A suitability map of five comparable classes was created; the classes were very high, high, moderate, low, and very low. In this work, two MCA approaches were investigated for their potential to discover feasible RWH systems in the Garmian region. The following points could be drawn:
ArcGIS is a useful tool that can quickly and easily determine whether a large area is suitable for an RWH intervention.
The accuracy of the produced suitability map depends on the quality and precision of the data (i.e., the procedure adopted in the data collection, processing, and production, and the MCA approach used).
Rainfall depth and the land’s slope are crucial in finding suitable sites for RWH.
Comparing the results obtained from the AHP and the FAHP methods has shown that the FAHP method is more accurate and reasonable.
In contrast to most recent studies, which employed the minimum and maximum local values, a practical minimum rainfall depth should be used as a guide to classify rainfall.
The study area can be divided into five rainwater harvesting potential zones. When using the AHP and WLC method, 9.7%, 43.4%, 33.6%, and 13.3% of the sites have very high, high, moderate, and low appropriateness for rainwater collection, respectively. However, using FAHP and Fuzzy Gamma overlay indicates that 10.6%, 15.6%, 26.6%, 21.9%, and 25.3% of the sites have very high, high, moderate, low, and very low suitability, respectively.
The NPRC used in the study has a vital role in accurately determining the suitability percentage of the area.
The suitability map will be helpful to hydrologists, decision-makers, and planners to quickly determine areas for RWH structures. Despite the promising results, it is advised that additional work be done to improve the model and expand its utility by including other essential auxiliary variables, such as socioeconomic aspects, and integrate the model findings with field observations to precisely determine the location of the proposed structures.