For years, Dr John Hayes has been trying to explain to decision-makers why flow rates and limits are so crucial for the survival of species like trout, which "drift-feed" on small aquatic and terrestrial bugs drifting in and on the water. Now, Hayes, a senior scientist at New Zealand's largest independent science organisation, the Nelson-based Cawthron Institute, has developed detailed models revealing how trout, and other fish including juvenile salmon and whitebait, actually require higher flows than present models allow for. He discusses his decades of research with science reporter Jamie Morton.
What was the impetus for undertaking this research programme at the outset - and what suspicions or expectations did you have going into it all those years ago?
I was interested in better understanding the effects of flow change on trout to provide better advice to regional councils and other stakeholders on environmental flow and allocation limits in rivers.
I knew from international research that the hydraulic-habitat modelling, main modelling method used to assess effects of flow change on fish and other instream life since the late 1970s-early 1980s, was too simplistic for drift feeding fish such as trout and juvenile salmon.
The most promising way of advancing our knowledge appeared to be incorporating an understanding of the energetics of drift foraging in predictive models.
Given that I had already begun working on bioenergetics models of trout growth, it made sense to build on that foundation.
My fishing and research experience led me to suspect that trout benefit from better feeding conditions that higher flows provide when river flows are receding and clearing following floods.
I had been providing that advice to regional council commissioners and judges in water consent and water planning hearings, cautioning against overlooked adverse effects of allocating large blocks of flow above minimum flow limits.
However, that advice was falling on deaf ears.
The new models and insights they provided, complemented by research in Canada, confirmed the point I had been trying to make.
Why are flow rates so crucial to drift-feeding species, and how do these ecosystems function throughout a typical year?
Flow is the volume of water moving in the river channel past a point per unit of time.
It is also termed "discharge".
Its measurement units are cubic meters per second (m3/s) or litres per second (l/s).
This is different to the water current, or velocity - or speed, which is measured as metres per second (m/s).
Some flow is needed by drift feeding, and benthic feeding fish, to provide sufficient depth and currents in which to live.
Alternative terms for this are habitat and space.
Too little flow and there is not enough space for all the fish to live in and too much flow might create conditions that are too fast for some fish.
Flow is also needed to support the benthic invertebrates that all fish in New Zealand rivers eat.
Benthic invertebrates live mainly in the fast to moderately-fast shallow places in rivers - these we refer to as riffles, or rapids in large rivers, and runs.
These are the main food-production engines of rivers, whereas pools are not.
Generally benthic invertebrate habitat has higher flow requirements than fish.
That's because riffles and runs are wide and shallow so their margins dry more rapidly with flow reduction than deep runs and pools.
Flow is also needed to entrain - dislodge or pick up - benthic invertebrates from the river bed, keep them in suspension, and transport them downstream to the hungry mouths of drift-feeding fish.
Our research has emphasised that maintaining the capacity of rivers to transport drift is vitally important for sustaining drift-feeding trout - and by inference also drift-feeding native fish.
There are two features of drift transport that are important to understand.
Firstly, entrainment of benthic invertebrates decreases with flow reduction and this decreases the concentration of drifting invertebrates in the water column.
Secondly, the flux or rate of drift concentration - the total amount of drift transported by the river - reduces even more steeply with flow reduction because the average velocity of the river also declines with flow reduction; drift flux being the product of drift concentration multiplied by average water velocity.
It is the flux, or rate, carried by the river through a channel cross section or a cross-sectional foraging area of a drift feeding fish that determines how much drift food a fish population or individual fish will experience.
How easily disrupted are stream flows, and what effect can just a minor change to these systems mean?
A minor change in stream flow - say, anything less than 10 per cent - is unimportant to fish and other instream life.
These scales of changes, and larger, naturally occur in rivers all of the time with varying rainfall.
But humans impose much larger, state-changes to flow - by water abstraction, diversion or damming.
It is not uncommon for flows to be reduced by 30 to 50 per cent in New Zealand rivers.
Some rivers are drawn lower than the 50-year return period low flow by a combination of surface and groundwater abstraction.
Canterbury has some of the most extreme examples.
It should come as no surprise to the public of New Zealand then that trout fisheries in spring-fed streams throughout lowland parts of Canterbury have substantially declined since the 1970s.
Native fish abundance and distribution has diminished along with the trout.
Over this 15-year study period, how have yourself and Cawthron colleagues attempted to investigate and model flow requirements?
We have developed a suite of computer models that operate at the reach scale doing the following.
Firstly, they model how water depths and currents change with flow throughout a surveyed river reach, and these off-the-shelf models provide the hydraulic and fish habitat platform to run the following models that we have made.
Secondly, they model invertebrate drift, or fish food, transport as a function of flow - predicting how drift concentration and rate varies at points throughout a surveyed reach of river.
Thirdly, they model the net energy intake of drift-feeding trout - predicting energy available from feeding on invertebrate drift minus energy costs of swimming and foraging, at any point in the surveyed reach.
These predictions are made for a series of flows over a flow range and summarised as predicted fish numbers versus flow graphs - which provide a new currency for assessing flow effects on fish, complementing, or perhaps substituting for, the traditional currency from habitat modelling which is predicted suitable habitat versus flow.
What trends or patterns have become apparent, and why?
We have found from the new modelling that drift feeding trout potentially benefit from higher flows than predicted by traditional habitat modelling.
For instance, whereas in fairly large rivers traditional habitat modelling might predict trout habitat could be optimised at a flow lower than the average annual low flow - apparently providing scope to abstract water and actually benefit fish - the new models suggest that optimal flows are higher than the average annual flow.
In fact, fish potentially benefit from higher flows though the annual average low flow and beyond.
The inference from this is that there is little scope for substantial water allocation without having adverse effect on drift feeding fish, if the fish are naturally at carrying capacity.
If fish are below carrying capacity naturally owing to factors other than natural low flows, for instance because of frequent floods suppressing recruitment of juveniles, then there would be scope for water abstraction.
But typically, we rarely, if ever, have sufficient knowledge on a given fish population to know that.
In the absence of this information, we advise more caution in setting minimum flows and allocation rates above the minimum flow.
NIWA research has shown that cumulative water allocation in dryland regions in New Zealand - especially Canterbury - have increased in recent decades, associated with the demand for irrigation water from intensified agriculture.
There is substantial over-allocation of flows in some places - especially Canterbury.
Does your research point to any potential solutions? Are there ways to restore flow rates to healthy levels while maintaining use of rivers by farmers, companies and other users?
An outcome of our research will be that the requirement under the Resource Management Act 1991 for the safeguarding of life supporting capacity, habitat of trout and salmon, and angling amenity, will be that trout will further constrain allocation of water for irrigation and hydropower.
Options for halting further decline and partially restoring flows, or to allow further development with current allocation, include more water storage schemes.
But large-scale water storage has its own suite of adverse effects on river natural character and ecosystems, which need careful assessment and management.
What further impact can we expect the effects of climate change to have on the issue?
Climate change will exacerbate the low-flow effects on steam ecosystems in foothill-fed and spring-fed streams and rivers in dryland regions - and further constrain the availability of allocated water to out-of-stream users.
How could this research also have global implications?
The same pressures that are bearing on New Zealand streams and rivers - from demand from agriculture, hydropower, town supply and climate change - are occurring in countries throughout the world.
Drift feeding fish, including trout and salmon, are common globally.
So our research is widely applicable.
Our models have been used to assess the effects of a large hydropower dam on trout in the Colorado River in Arizona and are currently being used to assess habitat requirements of sea-run rainbow trout and salmon in tributaries of the Columbia River, in Oregon and Washington, and options for recovery of these endangered populations.