Uncertainty and sensitivity analysis of West Java Water Sustainability Index - A case study on Citarum catchment in Indonesia

Water sustainability indices have been recently used to measure the sustainability of water resources within a catchment. Developing a sustainability index involves various steps, some of which have uncertainties associated with them. For the recently developed West Java Water Sustainability Index (WJWSI), three sources of uncertainties were identified, namely uncertainties in the thresholds of non-categorical indicators and sub-indicators, in the weighting schemes, and in the aggregation methods. This paper presents the uncertainty and sensitivity analysis of WJWSI, based on the application of WJWSI to Citarum catchment in West Java, Indonesia. The results of the uncertainty analysis, measured by the coefficient of variation of the thresholds and the sub-indices, indicates that minimum thresholds of Land Use Changes, Coverage, Education, Poverty, Health Impact and Sanitation, and the maximum threshold of Water Quality have higher variation when compared to variation of the other thresholds. The results of the sensitivity analysis, measured by the correlation coefficients between the final index and the thresholds, indicate that changes in the thresholds of WJWSI indicators have not significantly affected the sub-index values of most indicators and sub-indicators. The sensitivity analysis also concluded that either the equal or non-equal weighting scheme can be used for future use of the aggregation of WJWSI indicators and sub-indicators, as changes from equal to non-equal weighting scheme did not significantly affect the final index. However, it was found that the final index values were most sensitive to the aggregation method used (i.e. arithmetic and geometric methods), shown by the significant changes in the final index value when the aggregation method was changed from arithmetic to geometric. The uncertainty and sensitivity analysis presented in this study will not just assist in the efficient use of the WJWSI, but will also help undertake similar analysis for other indices.