A hydrodynamic-ecological model for Lake Rerewhakaaitu

Abstract

Lake Rerewhakaaitu is a shallow (average depth 7m, maximum depth: 15m) lake in the Rotorua region and is unique amongst the 12 Rotorua Lakes in that its catchment is the only one dominated by dairy farming as the main type of land use. At present, Lake Rerewhakaaitu is an attractive and popular lake with fairly good water quality, although over recent years the lake has shown signs of declining, water quality and biodiversity. It is essential that water quality in the lake is maintained or improved in order to retain the intrinsic, cultural and economic benefits the lake provides to the region. A computer modelling tool is required by lake managers to simulate current lake conditions and assess the potential impact of changes such as land use and/or inflows, and farm management practices. The purpose of this study was to implement, calibrate and validate the coupled hydrodynamic-ecological model, DYRESM-CAEDYM, to Lake Rerewhakaaitu. The one-dimensional (1D) hydrodynamic model DYRESM (Dynamic Reservoir Simulation Model) was coupled with the aquatic ecological model CAEDYM (Computational Aquatic Ecosystem Dynamics Model). DYRESM is a process-based model with a Lagrangian layer scheme that is used to predict the vertical distribution of temperature, salinity and density in lakes. Inputs to DYRESM are lake morphology, daily meteorological data, daily inflow temperature and volume and daily outflow volume. The CAEDYM model incorporates the major biogeochemical processes influencing water quality: primary production, secondary production, nutrient and metal cycling, oxygen dynamics and movement of sediment. The inputs to the CAEDYM model are nutrient inputs and the outputs are the oxygen distribution in the water column, nutrient distribution in the water column (N, C, P, Si), phytoplankton/zooplankton, algae, and more. Lake bathymetric data were obtained for April 2011. Meteorological data was obtained for the Rotorua Airport climate station about 20km away from the lake. There are two stream inputs to the lake: Mangakino Stream and Awaroa Stream. Surface inflows were available for the Mangakino Stream and/or filled using rated flows in the permanently gauged Ngongotaha stream. Flows from the Awaroa stream were obtained from a regression equation relating Awaroa Stream flows to the Mangakino stream flows. Daily nutrient concentrations (NO -N, NH -N, PO -P) for the Mangakino Stream were derived by interpolation between monthly samples. Suspended solid concentrations for the Mangakino Stream were obtained by using a regression equation relating flow rate and suspended solid concentrations. Labile organic nitrogen and phosphorus concentrations were calculated using monthly stream nutrient measurements and were evenly divided into dissolved and particulate fractions. For the Awaroa stream, an average daily concentration of nutrients and suspended solids were used. A further 7% of the catchment was assumed to drain directly into the lake. Lake surface outflows were assumed to be zero. To calibrate the DYRESM-CAEDYM model, data from 2005-2010 was used and to validate the model, data from 2010-2015 were used. Parameters chosen were in the range given in the literature. Simulated concentrations of temperature, dissolved oxygen (DO), total nitrogen (TN), NO -N, NH -N, total phosphorus, PO -P and chlorophyll a, at various depths were compared to measured data from observations to evaluate model fit with model error represented by a series of model performance statistics. The calibrated coupled model may be used to predict water quality in Lake Rerewhakaaitu, and ultimately, as a tool to test various management change scenarios. 3 4 4 3 4 4