Sensitivity Analysis of Yield Estimate of Urban Water Supply Systems

Morris method of sensitivity analysis applied to assess the importance of input variables on urban water supply yield – A case study

“…In effect, the framework can then be used to determine if climate variability and/or simulation length affects the importance of the controllable system policy variables used in the estimation of yield. The Morris method of SA (Morris, 1991) was selected for use in this paper as it is a reliable and efficient SA technique, shown to be successful when applied to a water supply system planning model (King and Perera, 2010). This paper begins by presenting a discussion of the SA framework adopted in this study for water supply systems, followed by a discussion of the Morris method, including application, indices and extensions.…”

“…The current paper follows a study by King and Perera (2010) which used the Morris method, the Fourier Amplitude Sensitivity Test (FAST) and the extended FAST methods of SA to study the importance of input variables used in the estimation of yield of a simple case study (VU and DSE, 2008). This case study dealt with an urban water supply system with two reservoirs and a single demand centre.…”

“…The main recommendation drawn from the case study presented by King and Perera (2010) was to separate the input variables used in the estimation of yield into climate dependant variables and system policy variables, and perform SA on the system policy variables using several climate scenarios. The current study follows this recommendation, considering the Barwon urban water supply system in Australia as a case study.…”

“…The yield was estimated using REALM, considering a single climate sequence consisting of 28 years of monthly climate data; consisting of streamflow, evaporation, rainfall and climate index. The SA framework adopted in King and Perera (2010) considered 26 input variables identified in the definition of yield, assuming they were subject to only data measurement and handling errors (Type II uncertainty error, as will be discussed in Section 2). The input variables included the system management rules and policies as well as climate data and demand data.…”

“…Several important disadvantages and limitations associated with the SA techniques and framework adopted in King and Perera (2010) were identified 1. Historic data use -Yield is generally estimated using a single, historic data sequence consisting of all available data.…”