Lake Ecosystem Research New Zealand
Models of lakes are used to provide insights into water quality at some future point in time, so that management actions may be targeted and cost-effective. In the past, small-scale physical models were used to simulate lake environments, but nowadays computer models are used to test potential management options. Computer models use a series of mathematical equations to describe the complex interactions amongst physical, chemical and biological processes that affect the water quality of a lake. The equations are stitched together consecutively in a computer program, allowing millions of calculations to take place in a single simulation. Use of models has become a standard practice to support community decision making for managing lake water quality. These models will increasingly be part of the National Policy Statement for Freshwater Management which will necessitate that freshwater management units are ’maintained or improved’.
New Zealand lakes are under increasing pressure from land use intensification, urbanisation, and invasive species proliferation. Lake ecosystem models express scientific knowledge of biological and physical lake processes forced by weather and climate into computer-coded equations. Model simulations can explain present day ecosystem states and predict future trajectories, and help to set aspirational water quality goals. This makes models a powerful tool to assess, manage, and protect NZ’s essential freshwater resources. They help us make management decisions such as selecting appropriate and effective actions to achieve desired outcomes.
But ecosystem models currently require expert knowledge to set up and parameterise, making them inaccessible for most practitioners. Our Smart Idea involves the use of advanced computing technologies (including supercomputing) and software to generate models for all New Zealand lakes based on knowledge gained from automated models generated for 100 lakes with known water quality. Our research will enable better management of lakes by providing councils and iwi the tools needed to simulate pathways to aspirational water quality targets.
New Zealand is a perfect place for the development of such a platform for our country boasts a diverse range of lakes, good observational records and comprehensive open geospatial data sets. However, our approach is developed with international export in mind. To ensure cutting edge science and global relevance, we connect to awesome national and international colleagues. Susie Wood (Cawthron Institute) and Marcus Vandergoes (GNS Science) will help scaling up modelling to all NZ lakes through linkages with the national-scale MBIE project ‘Lakes380’. David Hamilton (Griffith University, Australia), Dennis Trolle (Aarhus University), Annette Janssen (Wageningen University), Liancong Luo (Chinese Academy of Sciences) and Kevin Rose (Rensselaer Polytechnic Institute) form an international advisory group of world leaders in the fields of lake ecosystem models and remote sensing, ensuring global relevance, and connectedness of the research.
This interactive graphical display represents modelled Trophic Lake Index values for baseline (pre-human disturbance) and present-day lake trophic conditions. This work was produced by Abell et al. (2018) as part of the Ministry of Business, Innovation and Employment funded Lakes Resilience Programme.
Mean annual Trophic Level Index (TLI) values predicted for lakes prior to human disturbance ('reference' conditions) were estimated and compared to predicted current TLI for lakes that could be modelled. The results, expressed as deviation from reference, are displayed below as an interactive graphical representation for 1,030 mapped lakes >1 ha.
National Water Quality Monitoring Network lake data were used to develop statistical models to predict average values of the four constituent TLI variables: total nitrogen (TN), total phosphorus (TP), chlorophyll a and Secchi depth. Separate models were developed to predict water quality corresponding to current and reference states. Multiple predictor variables were used. For estimating current water quality, these included estimated TN and TP concentrations for lake inflows that were calculated using output from the Catchment Land Use for Environmental Sustainability (CLUES) model. McDowell et al. (2013) was used to predict reference nutrient concentrations for lake inflows under reference conditions.
Model predictions of TLI variables were then used to estimate TLI for reference and current states (see interactive map below), with the difference between these values interpreted as a measure of change in trophic status from reference conditions. Current TLI was greater than reference TLI in 89% of lakes. On average, lake TLI was 0.67 units greater under current conditions than reference conditions (standard deviation = 0.67 units). Based on predictions, 32% of lakes were oligotrophic under current conditions, compared with 68% of lakes under a reference state. Under current conditions, 38% of lakes were mesotrophic, compared with 24% of lakes under a reference state. The predominant lake trophic status under a reference state was oligotrophic and the predominant trophic state under current conditions was mesotrophic. Under current conditions, 28% of lakes have a trophic status of eutrophic or greater (TLI > 4) compared with 5% of lakes under reference conditions.
Research lead: Troy Baisden
Click on the image below to view interactive map