Development of decision support system for the identification of suitable water harvesting sites in chak-logar river basin

This study focuses on identifying suitable areas for water harvesting and planning suitable water harvesting structures in the Chak-Logar River Basin of Afghanistan using a Geographic Information Systems (GIS)-based multi-criteria analysis approach. The study incorporates 16 geophysical, hydrological, and socio-economic factors, each weighted through the Analytic Hierarchy Process (AHP) and integrated via Weighted Overlay Analysis in the GIS environment for the water harvesting suitabilty maps. The relationships between the 16 criteria identified in the water harvesting suitability maps were analyzed. Biophysical factors accounted for 82% of the total weight, while socioeconomic factors represented 18%. According to the correlation analysis, the strongest positive relationships were found among rainfall, drainage density, soil texture, and proximity to springs. In contrast, the strongest negative relationships were observed among proximity to agriculture, topographic curvature, proximity to settlements, and soil structure. The study area was classified into suitability zones using three classification schemes: five-class, three-class, and binary suitability maps. The most favorable areas for water harvesting were primarily located within Logar province. Additionally, hydrological soil group analysis based on the United States Department of Agriculture (USDA) hydrological soils classification revealed a predominantly moderate to high runoff potential. The annual average runoff volume of the basin was estimated at 2463 million m³, with 74% categorized as direct surface runoff. Based on the water harvesting suitability map for the Chak-Logar River basin, a total of 45 meso- and micro watershed harvesting structures were proposed. Which, 33 are meso- watershed and 12 are micro watershed harvesting structures. The proposed water harvesting structures have a total water harvesting potential of 1308 million m³. The total water demand for agricultural irrigation, domestic use, and livestock farming in the basin is approximately 376 million m³. Accordingly, the potential for harvestable water lies in agricultural land and socio-economic development. The prioritization of the proposed water harvesting structures was carried out using the TOPSIS and VIKOR MCDA methods. When investment opportunities are limited, the construction of small-scale sequential water harvesting structures across the site is recommended. Consequently, the total harvested water from the proposed structures remains below the theoretical maximum potential of the area. The proposed small-scale water harvesting structures were classified into three main groups based on their intended use, locations, and environmental contexts. These groups include structures for agricultural use, domestic and settlement purposes, and flood control/drought mitigation. The intended use of each proposed water harvesting structure was determined through analyses of their proximity to agricultural areas and settlement centers. At the micro-watershed scale (12 structures), the systems were proposed at the farm level, primarily for direct agricultural production and groundwater recharge. At the meso-watershed scale (33 structures), the systems were proposed at the sub-basin level for flood control, water storage, and multipurpose use. Among the proposed structures, 7 contour bunds and 10 farm ponds were considered for agricultural purposes, while 4 recharge pits were identified for groundwater recharge and agricultural use. In addition, 10 small reservoirs, 6 check dams, and 8 percolation tanks were proposed for multipurpose use, including agricultural irrigation and groundwater recharge, domestic and livestock water supply, and flood and drought management. In the evaluation of alternative scenarios based on construction costs of water harvesting structures, cost estimations were calculated using data on construction expenses within the study area, considering project similarities and an average unit construction cost of 0.8 USD/m³ per structure volume. Low-cost structures were given high priority, whereas high-cost structures were considered low priority. The analysis revealed that low-cost microstructures (1–10 million USD) should be prioritized in the initial phase of the water harvesting program, particularly given the country's political instability and the lack of national and international support. Subsequently, medium-scale projects within the 10–20 million USD and 20–40 million USD ranges should be implemented. High-cost structures (40–75 million USD) should be planned for the final phase or as part of long-term strategic investments. This stepwise investment planning approach allows for budget flexibility and provides decision-makers with a phased implementation strategy. It enhances both financial sustainability and implementation adaptability. In particular, prioritizing low-cost structures encourages local community participation and contributes to strengthening water management capacity. The total investment cost of the 45 proposed water harvesting structures is estimated at USD 1.046 billion, with a combined water harvesting potential of 1,308 million m³. Based on TOPSIS and VIKOR rankings, the top 15 structures can harvest 825 million m³ of water with 64% of the total investment cost, while the top 10 structures can provide 650 million m³ with 50% of the cost. The top 5 priority structures require only 30% of the total investment (USD 307 million) and can harvest 383 million m³ of water. Considering the basin's total annual water demand of 376 million m³, the top 5 structures alone are sufficient to fully meet the basin's water needs. Receiver Operator Characteristic (ROC) curve analysis was used to verify the prediction accuracy of the model. 12 control dams existing in the study area were used to estimate the prediction accuracy of the model. The ROC-Area Under Curve (AUC) result for this study is 62.6% (0.626). The ROC-AUC value for this model indicates a moderate level of prediction accuracy. Accordingly, the proposed model for the Chak-Logar river basin has a moderate level of suitability. Some of the reasons for the ROC-AUC value being in the medium range may be lack of water harvesting experts in Afghanistan, very limited factors are taken into account in water harvesting site selection and most of the time, decisions are made for the construction of water harvesting structures upon the request of community leaders regardless of the suitability of the sites. In addition to the ROC-AUC analysis, supplementary validations were conducted using the confusion matrix–kappa, success rate curve, and frequency ratio methods to enhance the reliability of the model. The ROC-AUC analysis for the proposed water harvesting structures yielded an accuracy of 95.6% (0.956). The ROC-AUC value obtained for the proposed check dams indicates a very high level of predictive accuracy. This comprehensive evaluation provides a scientifically grounded and practical methodology for sustainable water resources management and contributes significantly to regional water security strategies. Additionally, a survey study is conducted to assess public perception and willingness to adopt water harvesting in the Chak-Logar River Basin. A systematic questionnaire incorporating various question types, including contingency, matrix, closed-ended, and open-ended questions, was designed for data collection. 104 highly educated people living in the study area participated in the survey. The primary objectives of the study were to evaluate the current status of water resources, public knowledge of water harvesting, willingness to adopt water harvesting practices, and the potential of water harvesting as a tool for drought and flood mitigation. More than half of the participants expressed satisfaction with their current water sources. However, there has been a significant decline in water resources availability, varying across river basins. Groundwater serves as the primary domestic water source for 44% of respondents, with a notable trend toward solar-powered water pumping. Surface water (30%) and groundwater (26%) are extensively utilized for agricultural purposes. The absence of groundwater extraction policies presents a significant risk to water balance, while traditional water sources such as Karez and springs are increasingly vulnerable to depletion due to inadequate climate change mitigation and adaptation measures. Regarding water quality, 30% of respondents reported satisfaction, while 15% expressed dissatisfaction. Based on physical water quality parameters, taste (40%) and appearance (31%) were identified as the most influential factors in water use decisions. Major challenges affecting water resources include excessive groundwater extraction, drought, lack of water harvesting infrastructure, declining precipitation, climate change, water scarcity, and poor water management. Approximately 90% of participants consider water harvesting highly efficient, and 95% agree with its implementation. Despite this strong support, knowledge about the quality of the harvested water remains limited. Around 71% of respondents are familiar with water harvesting, and 90% express a strong willingness to acquire further knowledge. More than 93% are willing to adopt water harvesting practices, with significant enthusiasm for contributing financially and through workforce participation. The primary barriers to adoption include lack of government and international support, insufficient knowledge, and inadequate policies. Solutions proposed by highly educated participants include increased governmental and international support, public awareness initiatives, watershed management, public engagement, and capacity building. Additionally, 87% and 81% of respondents reported that their regions are affected by droughts and floods, respectively. About 89% believe that water harvesting can serve as an effective tool for mitigating these challenges.