Projections of storage reliability under future climate

Simulations of storage behaviour under future climate are an important component of planning for irrigation, hydro-electricity and water supply systems. The methods used for developing future climate projections of storage commonly include modelling inflows to the storages under future climate with a rainfall-runoff model. The rainfall and evaporation inputs to these models are adjusted for future climate using a variety of methods, which are dependent upon the availability of global climate model (GCM) and regional climate model (RCM) data. In this paper we compare future projections of storage reliability for storages of differing sizes used for irrigation, water supply and hydro-electric power generation. The future projections are developed using inflows generated by the SIMHYD hydrologic model. Rainfall and evaporation inputs for future climate are generated using different methods including: direct use of bias-corrected RCM outputs, application of a simple seasonal scaling factor to historic data, and quantile scaling of daily historic input values. The results show that the method used to derive future rainfall and evaporation inputs to the hydrologic model can have an impact on projections of storage reliability under climate change. In particular, simple scaling of historic rainfall and evaporation time-series does not capture the increased variability that is often a characteristic of future climate projections derived using dynamical downscaling of GCMs, which can have a marked effect on projected reliability of smaller storages.