Abstract
The water resources of the Clutha/Mata-Au catchment have an important economic value (i.e. hydropower and irrigation) for New Zealand/Aotearoa. Climate change is expected to alter both temperature and precipitation, which will then impact seasonal streamflow and water management. However, up to this point no attempt has been made to model coherently the driving natural processes as well as the major water management components. The aim of this dissertation is to assess the potential impacts of climate change on the hydro-climate of the Clutha catchment. An integral part of this study is the implementation of the fully distributed hydrological model WaSiM, which uses process oriented routines and also contains water management modules. To ensure a robust description of the meteorological drivers, existing techniques (e.g. trivariate thin-plate spline) were combined with novel elements (e.g. multi-layer lapse rate model). As a result, realistic spatial estimates of air temperature and precipitation were generated with a higher accuracy than existing products. An extensive validation of WaSiM revealed a good to very good performance when simulating daily and monthly flow during the historic period (1992-2012) at the key sites of the catchment (outflows of natural lakes and catchment outlet). Two versions of WaSiM, only differing in their snow model, were then forced with eight Regional Climate Model (RCM) simulations, stemming from two emission scenarios and four General Circulation Models (GCMs). The RCM data sets were bias corrected using both linear transformation and quantile mapping, and downscaled to the WaSiM grid (1 km2). The hydrological projections encompass substantial increases in streamflow during winter (+22 to +76% by the 2090s) and a decline in summer for most scenarios (-43 to +2% by the 2090s), with similar changes for hydropower production. The dominant drivers are increasing winter precipitation (+25 to +76% by the 2090s) and substantial decreases in the snow storage (due to temperature increase). Additionally, some of the main sources of uncertainty in hydrological projections were investigated. A quantitative approach identified GCM structure as the dominant uncertainty source followed by emission scenario, bias correction and snow model. It can be concluded that the new baseline meteorological dataset, model framework and ensemble of hydrological projections developed here constitutes the most comprehensive assessment of 21st century climate change impacts in the Clutha catchment to date, and as such also provides a new benchmark for future studies.