Multi-criteria Spatial Evaluation and Modelling of Farm Dam Site Suitability for Water Harvesting and Conservation

Abstract

This study presents a conceptual model designed to identify and rank potential areas for siting farm dams in a section of the Hawkesbury-Nepean catchment in western Sydney, Australia. The method takes into account environmental site assessment criteria using a decision-making method known as the Analytic Hierarchy Process (AHP). Spatial data is processed by applying GIS and potential sites are ranked by a multi-criteria evaluation based on meteorologic, hydrologic, topographic, agronomic and pedologic criteria. Particular to this study is the application of the SCS-CN method using curve numbers (CN) slightly modified for Australian conditions (CNMAC) for estimating runoff along with the original NCRS-CN values for comparative purposes. The use in this research of CNMAC has indicated modified CNs could be applied successfully in Australia especially in areas with limited physical data. Spatially-explicit sensitivity analysis was used to examine the model’s robustness to the sensitivity of criteria weights resulting from AHP pair-wise comparisons. Application of the One-At-a-Time (OAT) method (Chen et al.’s (2010, 2013)) demonstrated runoff has the highest impact on the evaluation results. Most of the study catchment showed a relatively stable suitability class; therefore, the model (SSMFD) was relatively robust and flexible in identifying suitable sites. The study then focussed on climate change impacts through annual rainfall patterns and their influence on the hydrology of the catchment and the model. Investigating 130 years of rainfall pattern indicated model low sensitivity to annual rainfall amount. In addition, changing the input data resolution used in SSMFD indicated that detailed outcomes were influenced by the resolution of input datasets. According to the SCS-CN method, results indicated the most decisive factor is CN values, consequently, soil and land cover are two crucial parameters and have the greatest impact on the model performance.