The secret lives of orange roughy nearly 900 metres below the ocean surface have been captured on moored underwater cameras.

Niwa used the underwater time-lapse photography to film the fish on the north Chatham Rise, east of New Zealand, as part of efforts to estimate fish numbers.

Traditionally, deepwater fish numbers in the exclusive economic zone had been estimated using trawl surveys or acoustics, Niwa said.

But it wanted to enhance that work, to ensure the right fish were being counted in the right way.

So moored cameras had been used with echosounders during a 20-day fisheries experimental voyage on the research vessel Tangaroa completed in June. The aim was to provide new insights into fish populations on underwater hills on the north Chatham Rise.

Following on from a successful pilot study in 2010, video cameras were moored on two underwater hills, one known as Morgue that was closed to fishing in 2001, and another known as Graveyard which was still commercially fished, Niwa said.

Videos and lights were positioned at eight, 30, 50, and 70 metre intervals above the seabed, and timed to come on for two minutes every two hours. An echosounder was placed at the top of the mooring to provide acoustic data that could be matched with the video footage.

Niwa principal scientist and voyage leader Richard O'Driscoll said the technique had revealed a range of fish behaviours.

“In the first video frame when the lights come on, there are fish that are behaving naturally, so their distribution off the bottom, and their species composition, is as close as we have been able to get to seeing the fish undisturbed.

“Once the lights have been on for a few seconds, the fish start to dive. Some were even seen to lie on their sides on the seabed. Others froze in front of the cameras almost like possums in the headlights.”

Orange roughy were the main fish species on both underwater hills, but densities on the closed underwater hill were higher than those on the fished one, and orange roughy were seen further away from the bottom on the closed one, O'Driscoll said.

The technique avoided problems with other methods, such as trawls only effectively catching fish close to the seabed, and towed or lowered cameras performing poorly because fish moved to avoid them.

A key uncertainty with acoustic surveys - in which a pulse of sound was sent out and then pinged back from the fish - was knowing what type of fish the sound was pinging off.

It was particularly critical with orange roughy which did not have an air-filled swim bladder and so produced a lower amount of sound than many other fish species which did have such a bladder.

That raised the possibility that estimates of orange roughy numbers could be too high, if species with swim bladders were swimming among them, O'Driscoll said.

In the past the acoustic signal had been divided into different species, depending on what was caught when a trawl was put down.

But trawling was an inexact sampling method as different species reacted to it in different ways, with some easy to catch and some hard.

At Morgue hill, researchers had been concerned about whether a large aggregation of fish that showed up in acoustic surveys well off the bottom really were orange roughy.

While researchers had never caught orange roughy that far off the bottom, filming confirmed orange roughy were present at quite a high density 100 metres deep. In fact the density of the species had been higher 30 to 50 metres off the bottom than on the bottom.

The tendency of orange roughy to dive for the bottom where some lay on their side, which they did when the lights went on, was probably a good strategy when trying to avoid one of the species' main predators, sperm whales, which would be coming from above. It was not so good for avoiding a trawl being dragged along the bottom, O'Driscoll said.

At Graveyard, few orange roughy had been filmed off the bottom, showing it was not possible to make a general conclusion about fish composition across all underwater hills. The cameras showed a mix of probably nine or 10 species. As well as orange roughy, there were hoki, oreo dory, several shark species, squid, eels and cardinal fish.

The cameras appeared to show "a lot" of orange roughy, O'Driscoll said.

The moored camera technique had worked, so in future when acoustic surveys were carried out at underwater hills, a camera would also be used to provide information about the composition of species.

Acoustic surveys had been done at both hills, but work needed to be done on the data to determine the amount of fish there.

Ministry for Primary Industries (MPI) principal fisheries science adviser Dr Pamela Mace said the work had been a rare opportunity to trial some new technology, rather than standard techniques which had been used for a long time.

“We’re very excited by what we’re seeing.”

The voyage was funded by Land Information New Zealand Ocean Survey 20/20 and MPI, which commissions research to help it administer the Quota Management System.

- © Fairfax NZ News