Sediment-laden river or swirling ocean current, water can carry creatures and contaminants vast distances. NIWA's hydrodynamic models, finds Veronika Meduna, are helping planners better understand nature's conveyor belts.
When the causeway on State Highway 16, between Auckland's Te Atatu Peninsula and Point Chevalier, was built on reclaimed land in the 1950s, few predicted the Herculean demands that would eventually be placed on it. Traffic volume has burgeoned over the years, says the New Zealand Transport Agency's (NZTA) Principal Environmental Specialist, David Greig, "and the capacity of that causeway is no longer adequate."
Six decades later, it's part of NZTA's biggest, most challenging and most expensive roading project. But this time, it's being built with an eye on the future.
The Waterview Connection project will integrate motorways through – and beneath – Auckland's western suburbs to complete the Western Ring Route as an alternative to State Highway 1. To keep pace, however, the ageing causeway must not only be widened, but raised after it subsided into the soft estuary sediment. Greig says it's also suffered regular inundation during storms and high tidal surges.
Because the causeway bisects the Motu Manawa Marine Reserve, flanked by the Waterview Estuary to the west, and Waitemata Harbour to the east, the resource consent called for an assessment of any effects on the coastal environment. The NZTA hired experts to assess future environmental effects, based on historic impacts. Then, to crosscheck and augment their conclusions, it asked NIWA and Tonkin & Taylor to model the effects of sediment and stormwater discharges during construction. NIWA also used its hydrodynamic models to predict the long-term effects that additional piers and a wider channel might have on current flow patterns. "We need to keep check on the sediment load into a sensitive environment," says Greig, "and we need to know the level of contaminants going into the harbour."
Modelling behaviour
Dr Rob Bell has modelled the impacts of many coastal and estuarine projects, as well as the effects of coastal climate change. The NIWA Principal Scientist says hydrodynamic models can simulate anything from the turbulence around a sand grain to the current patterns in the global ocean. "They essentially describe the physics of the flow of water, and that's a fundamental part of all transport processes, whether that's sediment, larvae or pollutants."
Generally, says Bell, the first step when developing a hydrodynamic flow model is to get good current and water-level data, because they drive all other coastal and marine processes. These basic inputs can be measured with current meters to track speed, GPS-equipped floats to measure drift patterns, and water-level gauges. "These models can be used for applications such as understanding flow patterns around reclamations, as in the SH16 motorway project, or the effects of sea-level rise, simulating coastal inundation from storm-tides and flood levels."
Then 'downstream' models, tailored to the nature of the project, are added on. Possible applications range from assessing water quality – for example, by predicting pollutant dispersion from outfalls and stormwater systems – to simulating sediment transport, or even modelling an entire ecosystem and its response to environmental changes.
When Watercare retired its Auckland oxidation ponds and replaced them with comprehensive sewage treatment, it also shifted the discharge outfall into Manukau Harbour. NIWA models were used to predict the effects, but Bell says they had to account for more than just current flows and dispersal patterns. "The main issues with outfalls are pathogens and viruses, so you have to know how well they survive in seawater, and where the treated discharge will be carried to, in order to protect recreational uses and kaimoana. Which model is being used depends on what problem is being modelled, and its space and time scales." They can range from a simple equation, he says, "to complex code running a supercomputer, but always designed to be fit for the purpose." He says models are often the only way to predict future effects, but cautions that any numbers they produce have to be checked against reality.
In 2009, Environment Canterbury asked NIWA to model flooding along parts of the Canterbury coast, following a hypothetical tsunami from a South American earthquake. Such a scenario is considered the greatest, and most likely, threat to that stretch of shore. NIWA based its model on the 1868 Peru tsunami, which reached much of coastal Canterbury.
In 2010, a tsunami triggered by a magnitude 8.8 earthquake off the coast of Chile drove strong currents into harbours and river mouths and flooded small areas of coast.
After last year's February earthquake, Christchurch regional and local authorities decided to remodel the scenario for the coast north of the city, and to extend it south to Taylors Mistake, to see if coastal evacuation plans needed adjusting, taking land subsidence into account.
Bell says good information about terrestrial and marine topography is critical to coastal models. "A model is a schematisation of reality." Researchers begin by creating a simulation of the environment they want to represent, over which they place a grid – each cell might, for instance, represent 30 square metres of the real environment.
Calculations derive averaged values for each cell, and the model is then tuned until results give the best match with real measurements. "If you don't get a good calibration of the model," says Bell, "it's often because the schematised grid is not quite representative of reality, so we go away and modify it until it is."
The next step, says Bell, is verification, where the model is run for a different period, or under different circumstances. "This time you don't touch the tuning knobs, and check it still performs well – then it's ready for predicting various scenarios or designs."
But NIWA Biosecurity Programme Leader, Dr Graeme Inglis, has reversed this process, for good reason. The Ministry for Primary Industries contracts NIWA to run a National Surveillance Programme to ensure the best chances of detecting any alien marine invaders as early as possible. "In designing this programme," says Inglis, "if we'd just allocated all our samples around the harbours, we wouldn't stand much of a chance of detecting those species, because it's just such a large area of available habitat. So we've been using the hydrodynamic modelling and habitat suitability models to try and identify the sites where species are most likely to turn up first. Those areas provide a focus for high risk of early incursion."
For the purposes of biosecurity, hydrodynamic models are applied on different scales. New Zealand-wide, Inglis says two key areas are ballast water exchange and the placement of structures on the coastal shelf. "New Zealand has an Import Health Standard for ballast water, which is one of the most frequent ways invasive species reach our shores. This standard means that ships aren't allowed to discharge ballast water in New Zealand's territorial waters, unless they can demonstrate that they've exchanged it for mid-ocean water en route ... or that they have treated the ballast water to a degree that's acceptable."
However, the standard allows exceptions: in rough weather, for instance, when a ship has to come close to shore to exchange ballast water. Models help to find stretches of the coastline where this poses the least risk.
More than 170 non-native species have already established in our coastal waters. When a new one is detected, biosecurity officers use models to estimate the maximum distances any larvae could have travelled since the incursion. "During the Mediterranean fan worm incursion in Lyttelton in 2008," says Inglis, "we used the outputs of the model that we've developed in designing the surveillance work to delimit the population. And we managed to remove it from the harbour."
Where to from here?
Dr Niall Broekhuizen, at NIWA in Hamilton, is modelling impacts on entire ecosystems, assessing the effects of fish and mussel farms in the Hauraki Gulf, the Firth of Thames and Golden Bay. While filter-feeding mussels get all the nutrients they need from surrounding water, caged fish must be fed, adding organic matter to the environment that can enrich the water.
Before he can model any impacts, Broekhuizen says he first needs to know how much nitrogen and carbon is coming out of a farm, and how much oxygen the fish will consume. "The model then predicts the fate of that material." If too much decomposing matter settles on the seabed, he says, it can create hypoxic (low-oxygen) – or even anoxic (no-oxygen) – conditions, which can cause "dramatic changes in abundance and composition of the seabed fauna in that small region."
In the water column above, the effects are more subtle. New Zealand coastal waters are generally nitrogen-limited, and any extra nitrogen excreted by the fish may trigger phytoplankton growth. If a hydrodynamic model is to predict the effects of that, it must 'know' about the stratification of the water column – essentially, how it's layered, depending on differences in salinity and temperature, which in turn cause changes in density. Less dense water ordinarily sits atop more dense layers. "The strength of stratification and the depth of the layers are important factors," says Broekhuizen. "Any changes can have a big impact on algal growth, because phytoplankton need light to grow and will essentially starve at depth."
Mostly, he needs to know how far any nitrogen will travel, and how much it will boost phytoplankton biomass. "We need to know whether these changes are larger than [natural] seasonal and inter-annual changes. Circumstantial evidence suggests that any changes at the base of the food web trickle down and can transfer to the zooplankton and affect larval survival."
But perhaps the greatest value of models, says Broekhuizen, is that they "force you to think more rigorously, and identify knowledge gaps. They're certainly not a cheap alternative to fieldwork."
