A novel approach for assessing watershed susceptibility using weighted overlay and analytical hierarchy process (AHP) methodology: a case study in Eagle Creek Watershed, USA

Abstract

Watershed vulnerability and the characterization of potential risk are important inputs for decision support tools in assessing watershed health. Most previous studies have focused on the assessment of the environmental risk using physicochemical properties of surface water and mathematical models to predict the health of a watershed. Here, we present a new methodology for evaluating watershed vulnerability using the analytic hierarchy process (AHP) and weighted overlay analysis. The new methodology provides an inexpensive approach for assessing areas that need more investigation based on known factors such hydrogeological, geological, and climate parameters without the need for site-specific physicochemical data. The proposed method was implemented using six main factors that influence water quality: land use, soil type, precipitation, slope, depth to groundwater, and bedrock type. Vulnerability was predicted for ten sub-watersheds within the Eagle Creek Watershed in Indiana using publicly available data input into geographic information system. Combination of watershed susceptibility assessment and GIS spatial analysis tools was used to produce the maps that show the susceptible zones within a watershed. A comparison of the resulting vulnerability estimates showed the expected significant positive correlations with measurements of nitrate, phosphate, temperature, and electrical conductivity. Likewise, the vulnerability estimates negatively correlated with dissolved oxygen and E. coli. Furthermore, the validation of the proposed approach revealed that the areas predicted to have high vulnerability did have lower water quality indices; the results showed a high negative correlation ( r ²  = 0.77, p < 0.05) between water quality index (WQI) and vulnerability which strongly suggests this method can be used successfully to assess a watershed’s susceptibility.