Howard SW, Griffiths J, Zammit C, Rouse H. 2019. Model choice effects on ecological modelling in Mataura River: SAM Programme 2018. Lower Hutt (NZ): GNS Science. 55 p. (GNS Science report; 2019/05). doi:10.21420/J32Q-W248.
This report presents the application of the River HYdraulic HABitat SIMulation (RHYHABSIM) model coupled to the suite of hydrological models considered in the Smart models for Aquifer Management (SAM) research programme. Led by GNS Science, the SAM programme aimed to investigate the implications of model simplification on prediction uncertainty, and decision-making in water resource management. Investigation is carried out through inter-comparison of different model uncertainty, in relation to key decision variables, for pre- (a priori) and post- (a posteriori) calibrated models. Models investigated within the SAM programme spanned from simple conceptual models (e.g. V-Notch), to complex semi-distributed models (such as the TopNet model suite), complex distributed models (such as LUCI), and highly complex models (such as MODFLOW). As part of the mid Mataura case study, the impact of hydrological model complexity on ecological variables was investigated at two sites: Riversdale and Otamita. The RHYHABSIM model was applied with surface water flow predictions from three hydrological models (TopNet0, TopNet-GW and MODFLOW) to produce habitat suitability time series. Differences in predicted habitat retention metrics for each model were then compared to assess the influence of hydrological model choice (and parameterisation) on water management decisions (and hence ecological variables). For the water allocation scenarios at Riversdale, the difference in number of days that specified hydro-ecological criteria were met were generally dependent on hydrological model choice. These results were clearer for the MODLFOW and TopNet0 model than for the TopNet-GW model. Reductions in the habitat retention metric at Riversdale were also generally greater than those predicted for the Otamita site. (auth)