THE “UNDERWATER CAMERA TRAP” PROJECT

Back in 2010, I joined the Marine Megafauna Foundation in Mozambique to study reef manta rays. One of the methods to study these incredible fish is to simply take photographs of them. By one photograph you can often see if it was a male or female, mature or juvenile and even exactly what individual that particular ray is. You see, manta rays have unique spot patterns on the their belly. So by going on dives and taking photos of the underside of any manta ray you encounter you can, over time, start to get an idea of population size, population structure, and possibly even movement and residency patterns.

MantaThose spots can tell us which individual it is

 

But…. In the vast ocean we are somewhat limited as primarily land-based creatures. Using regular dive equipment you may be able to spend 1 to 2 hours effectively underwater per day. And of course, you can’t survey multiple dive sites at the same time. After spending 6 months in Mozambique, I returned to Holland and started thinking if there isn’t a more effective way of collecting the same, or more, data.

DiverLobo No matter how much we are enjoying it down there, we can’t stay forever

 

For terrestrial species, photo-identification as it is known, is also often used. And by using so-called camera traps, you can survey multiple areas for extend periods of time. These specialized cameras turn on when something passes in front of it, only taking photos or a short video when it is needed, and conserving precious battery life otherwise. They can be deployed for weeks, months even.

To sense anything passing by, a sensor is needed. In conventional camera traps this is most often either a passive or active infrared sensor. So I thought, why not use this underwater?! Put one of these camera traps in a housing, put it on a reef, leave it there for a couple of weeks and pick it back up and see what animals, particularly manta rays in my case, have swum by.

If only things were this simple…. These infrared sensors, however, do not work underwater. Infrared light in an aqueous environment simply gets absorbed too quickly by the denser water. So what would kind of sensor can work in the salty sea? There are may number of possibilities, some more practical than others.

A friend of mine suggested using the fact that a large animal swimming over will cast a large shadow. Sensors that sense sudden light changes are readily available. So I started working on this. After a lot of google searching and some tinkering, this was the result. Two days before I was to leave again for Mozambique, and with the help of a good friends dad, we even made an underwater housing that would fit all of it in.

CT1After a lot of tinkering, soldering and perhaps also swearing, underwater camera trap 1.0 emerged

 

Did it work in real life scenarios? Not really. Light diffuses quickly underwater, and unless the water is clear like gin and the sun is out, the shadows created by large objects aren’t harsh enough to trigger the sensor. And it obviously could never work at night, or when it is not pointing upward towards a light source.

CTDeployedThe first ever deployment, unfortunately we came up empty handed

 

You could also set up a laser (not infrared) possibly on one side and align it with a light sensor on the other, and if anything breaks the laser it sets off the camera. But good luck putting all those different components in underwater housings and keeping the laser and sensor aligned in the often quite volatile ocean environment.

Sonar is a distinct possibility. Sound waves get sent out in pulses, and anything that it bounces off of will get picked up by a little microphone and voila, the camera takes a photo. But sonar also eats away at battery life and in reality is a lot more complicated than it sounds. And most readily available sonar components still emit at a frequency that some animals, such as cetaceans can still hear. To me, it didn’t seem like the solution.

So what next? Could you not combine lasers with sonar (evil scientist anyone?)? What if, instead of sound, you emit laserlight pulses? And if these light pulses bounce back off of anything, it gets picked up by a light sensor, in the same housing as where the camera and laser is? No tricky sonar or alignments of lasers needed! Could it possibly work?

I started scouring the internet and soon found components that might be suitable. With a lot of mucking around trying different things, I got something that worked. On land at least. With help from a friend, Ando Shah, also working at the Marine Megafauna Foundation, we improved the software and overall size of the device.

CT2.0Underwater camera trap 2.0: lasers, laser sensor and camera module all in one small package

 

We did pool tests, and it looked promising. Objects from two meters away were being sensed by our sensor. But would it work  on a reef system 25 meters deep?

AndoPool_I8A5214Preliminary tests in the pool proved promising. Here is Ando taking notes

 

In short, we don’t know yet. Turns out, electronics are not a big fan of a salty air environment. Nor do they like to be accidentally short-circuited, which does tend to happen when tinkering with a bunch of wires, lasers, cameras and batteries. Every time we had a working prototype, something would go array shortly after. And unfortunately, tiny specialized electronics are not easy to come by in Mozambique…. I once had to wait more than 6 months for a new camera module.

So it is still very much a work in progress. But I am confident that it will work one day. Could it  be used only for the benefit of manta rays? No. It could be used for anything really.  Document interesting behaviour in the deepest of seas? Assess biodiversity in remote areas? Help conserve aquatic species threatened with extinction? Perhaps this little device could do it all. Only the future will tell.

SATELLITE SUCCESS!

Ecstatic is the right word to describe how we felt after deploying five pop-up satellite archival tags (PSAT’s) on reef manta rays early July. The first ever PSAT’s to be set on reef manta rays on the continent of Africa, these sophisticated devices were programmed to stay on the animals for approximately 100 days, collecting a wealth of information on the reef manta rays’ movements and diving behavior.

Tagged mantaA reef manta with a satellite tag

 

I have been diligently checking my ARGOS satellite account for any hits every day since the tags were first deployed. These tags are only able to transmit when at the surface, so when the animal is in or in the unlucky event that a tag has shed prematurely. Not expecting any hits until the tags were programmed to pop off on the 10th of October, it was a bit of a disappointment to find a steady stream of hits of one tag after only 19 days! And another tag started transmitting only after 36 days on the animal. My ecstatic state was quickly turned into a somewhat disappointed and panicked demeanor.

ARGOS_IMG_4184A steady stream of hits from the ARGOS satellite system most likely means a tag has popped-off, as was the case with tag 128924

 

These tags are highly advanced, and this means a hefty price tag. With each tag having a total cost of around US$3500, you want them to perform well and get the most out of them, especially as the chances of seeing them ever again is pretty slim. The positions we get from these tags when floating at the surface are not that accurate, with errors of at least several 100 meters, making it quite literally looking for a needle in a haystack when it is bobbing around in an immense ocean.

However, if washed ashore, the chances of retrieval are increased tenfold. Disappointment soon turned into excitement, when it seemed that we could actually retrieve the two tags. If found back, I could access the raw data on the tag, revealing all the measurements of water depth, water temperature, light levels, magnetic levels and acceleration for every 30 seconds! Far more than what is normally transmitted through the satellite system, which is only a condensed version of the data, this will be highly beneficial to our study on the movement ecology of reef manta rays along the Mozambican coastline. And not only will we be getting this high-resolution data, we will also be able to redeploy these two tags, saving us thousands of dollars.

Path Sat Tag 128924The track of the second tag after it had popped-off. It started about 30km north of Bazaruto Island, and got blown towards Bazaruto Islands by some fierce northerly winds over the course of four days.

The tags were found to be transmitting and floating offshore off Bazaruto Island, and within a few days they seemed to have been washed ashore on the island. Lodges on the island were contacted and flyers, with information on the tags and the reward for returning them, were passed out at the local villages. We quickly got word that the tags were found, and after some phone calls, I set out to meet with the lucky beachcombers in Vilankulos.

Tags Downloading_IMG_4178After picking up the tags, the download of the data could commence

 

Looking at the data, already some interesting and very different behavior between the two animals can be observed. Tag 128922 showed data of the animal diving down to a maximum of 116 meters deep, but spending his time at an average depth of 17 meters. Tag 128924 however not only dived deeper, up to 184 meters, but also displayed a much deeper average depth of 43 meters. With these deeper depths came also colder temperatures, as low as 15 degrees Celsius. Each animal however also spent time near the surface on a daily basis, showing that they are quite the vertical swimmers.

Another incredibly interesting aspect these highly sophisticated tags record is accelerometer data. Every 1.28 seconds it record the accelerations in G per second across the x, y and z axes.  This will give us insight into how fast these animals can move, and possibly also when they display different types of swimming behaviors. It may additionally show us when they are feeding, cleaning, resting or simply on the move, information that we could previously not obtain.

 The task that lies before us now is to analyze all this data. Obtaining the most probable tracks of the animals, performing dive behavior analysis and acceleration data analysis involved a lot statistical modeling. Not an easy task, but Fabrice Jaine, our data analysis guru, and I will be cracking away at it for the next few weeks, fueling our desire to know more about these enigmatic animals.

Data_IMG_4182The large amounts of data coming from these tags will help us elucidate the movements of these large rays

 

But why have these two tags shed prematurely? Why did they not stay on the animals until the 10th of October? These questions we can also answer, or at least speculate. The first tag to pop-off, was actually initiated by the tag itself. The animal behaved so consistently for about a week that it set off the Constant-Depth-Release. A built-in feature, this enables the tag to release earlier than programmed if for example the animals has died and is lying on the seabed. Judging by the dive profile of the animal this definitely did not happen, but it was diving so consistent that the average depth was similar every day, tricking the tag into popping off.

 The second tag seemed to have been bitten off by something. This is not something uncommon with tags, as they do resemble small fish swimming with the manta ray. Predators such as game fish or sharks can mistake the tag for a fish, have a go at it, and break off the tether.

 In short, the data we get from these tags is invaluable in the conservation of reef manta rays in southeastern Africa. This is an incredibly lucky start, and let’s hope the other three tags will be retrieved as well! Many thanks to all the people involved in getting the tags back to us; Janneman Conradie of MMF and Big Blue Vilankulos, Luis, Tony, Antonio and Mapanzulo of Pestana Bazaruto Lodge, Anel of Indigo Bay Resort, and Thomas and Arnaldo for finding the tags, Muito Obrigado!

* Written by Daniel van Duinkerken. This post was originally posted on www.marinemegafauna.org.