Mapping the distribution of subtidal kelp speciesLatest update September 11, 2019 Started on July 1, 2019
Kelp, very much the forests of the ocean, are ecologically important habitats. Over the coming months we aim to deploy vessel-mounted acoustic remote sensors to map the distribution of kelp off Ireland's western coast.
'I guess I can add ROV pilot to my CV...'
For anyone who I haven't already insisted to come and have a look, the Trident ROV arrived last Thursday. I'm very grateful to those of you who following this blog and to Sofar and OpenExplorer!
Here is a brief update on my week as ROV pilot 'in-training'.
The weather sucked so I only got out twice...
...into the shallow bit of the river (I really spent most of the week staring at it and carrying it everywhere with me, I'll be a great Dad someday)
I'm talking less than 0.5m which obviously makes it difficult to fully appreciate the capabilities of the Trident.
I currently have the basic Trident package which comes with the ROV, a 25m tether, charger and support strap. I wanted to get the recommended controller but I'll need to save a little first (along with the 100m tether). Currently controlling using my phone and a cheap android gamepad I found online.
The cheap controller works ok, a little laggy and provides plenty of heart stopping moments like when the Trident completely stopped moving and it took me a while to realise the controller had disconnected, but I could still control it from my phone touch inputs!
The video quality is excellent, the attached footage won't do it justice, it's easy to identify species which is promising for kelp research. It's tough too, I spent most of my time flying headlong into rocks, barely a scratch. I did dredge up most of the algae present in the river so that's the last time i'll use it there but it hasn't seemed to have any impact after a thorough clean.
Next week I hope to take it for sea trials.
Oooo, and I'm printing a 3D GoPro mount to attached to the bottom, and I like to give it baths in the sink.
Time for a cuppa.
'Progress at last...'
'...some exciting news also....'
I've now managed to create fully accurate bathymetric layers for the site and for each of the three frequencies. I wanted these for two reasons, one to have a pretty backround for my maps (see attached image) and second, to learn all the skills needed to work with this kind of data in the future (adjusting for tides, SVPs, data cleaning etc).
The attached image (which is not to be used for navigation) is the first glimpse of my study site. To the north you can see the rocky knoll (in brown) where the kelp is, this ranges from about 6-15 m in depth. The darker turquoise areas are anywhere from 15-25 m in depth. It was important to include areas where kelp wouldn't be found so that we could properly validate the multibeam results.
The points represent the drop-down GoPro work and each one represent either and 'in' or an 'out', meaning that I recorded the GPS point when the camera went down and when it came out to try and improve the spatial accuracy of the data. The means that we I look at the video I know that it came from somewhere between each of the two points. We tried to cover all interesting features, including depth and substrate transition areas.
Green points are where dense kelp beds were observed and you can see that these match perfectly with the shallow rocky knoll in the north. Black and yellow points are either no kelp or no bottom observed (water too deep) and these correlate nicely with deeper areas. Orange points are where the bottom could just about be seen but no clear evidence of kelp or no kelp was observed.
All of this should now help me when I come to analysing the water column data for acoustic signals of kelp.
The next step will figuring out how to extract water column data and see if any quantitative measures such as area covered and canopy height can be extracted.
I have also been kindly awarded a Trident Underwater Drone and am immensely grateful to everyone who has followed thus far along with Sofar and Nat Geo!
The immediate plan for it is to verify whether it can be used to ground-truth acoustic kelp surveys. I don't have access to a boat in the near future so I will be testing it at kelp sites accessible from the shore.
More updates to follow when it arrives!
Time for a tea.
'Sometimes things work and I don't know why...'
Funny few days of processing, spent all of yesterday trying to open the raw survey line data only to find out that I was using the wrong software version (Qimera Pro and not Qimera Clean for future reference).
Today was a bit better...
Was I had the data in I could start playing with the water column and this is where things went a bit wrong. Turns out that water column data is MASSIVE and forced my computer to run at a snails pace making data processing virtually impossible. I tried a more powerful computer which was not better and I wasted an hour trying to install the software and the right drivers for the USB license key (it's always the drivers!).
After resigning myself to defeat, shame and failure I had one last look through the manual where I found it...
This lowers the resolution of the water column in the data viewer and it can now be worked with. The loss of resolution doesn't seem to be too much of a problem either (famous last words).
I had a bit of help from Fabio and Oisin which I suppose I should mention, can't take all the credit.
Hopefully tomorrow will be an even more productive day.
Time for a pint.
'Once more unto the breach...'
I have finally dipped the proverbial toe into the frozen waters of acoustic data processing.
I have been kindly provided with all the necessary software and I want to spend a bit of time fiddling around with them before invariably asking for a spot of help.
The short video is the first bit of data I successfully opened...
...onward and upwards!
'Post pilot processing...'
Ultimately the weather on Thursday and Friday meant that the boats couldn't get out and we were unable to collect more data.
That's just the way these things go sometimes.
Still, the data we collected off the RV Lir on Tuesday should allow us to fully understand the ability of this technology to map subtidal kelp. We can then use the results to inform on a future study where I would hope to map a larger area using optimal acoustic sensor settings.
I would be keen to deploy an underwater drone for these larger scale surveys as it should permit video transects to be completed which will be more efficient than drop down cameras, especially with live video feed. The Sofar Trident drone is already being used by many researchers and has been deployed in a range of environments (https://www.sofarocean.com/products/trident)..)
The internal compass will make it easy to follow a transect heading, helping to improve the accuracy of our GPS in/out markers.
I also have more underwater drone plans which I'll touch on in a later post!
So, onto data processing now. I have all the raw and processed files from the GSI and I plan to learn and use CARIS HIPS to process them and try to identify and separate the kelp canopy from the seabed to determine it's height and extent.
I'll keep you posted on my trials and tribulations.
'Weather stops play...'
'Post initial survey musings...'
Wednesday was a write off weather wise and today is looking the same. This is giving me the chance to go through the videos from yesterday and tie them in with the GPS data.
The first thing that jumps out is that, although the GoPro imagery is very clear, we're kind of flying blind when we drop the camera down. It would be more efficient to conduct video transects but this would be highly inefficient without live video feed or the ability to accurately position the camera rig. Divers are obviously able to collect the most accurate biological information but this is an expensive, time consuming and potentially dangerous method, particularly in exposed waters.
The Sofar Trident ROV is something I have had my eye on since seeing one at the University of Massachusetts last year and think it would be perfect for ground-truthing the acoustic survey data. The lights would allow us a clear image of what kelp species are present along with associated floral and faunal species. I'd say we'd have to be careful about plonking it right in the forest, less it gets caught, but it would allow for the kelp boundary to be defined and we'd have to figure out a way to locate the images without GPS.
I think if we wanted GPS data then we would have to aim for a straight transect with the ROV crisscrossing the kelp bed and then logging the GPS in and out to get a good idea of the video transect location.
In the meantime, our data looks good and, whilst we may not have a chance to collect anymore, I look forward to keeping you updated on the data processing.
The picture shows a screen grab of some of the acoustic data from Tuesday, you can see potential kelp in red!!
- Tom -
'Things went suspiciously well today...'
It's been a while trying to get out on the boats for some kelp surveying and I didn't expect to have the success we had today.
I hitched a ride on the R.V. Lir out of Baltimore this morning. After they had carried out their planned seabed mapping we turned our attention to more 'kelpy' matters. Cracking weather by the way!
We picked one of the areas that I had previously identified and dropped a GoPro down to scope out what was there.
'I'd built a simple frame out of pvc piping and fishing weights to try and create a stable and safe housing for the GoPro.'
We observed a dense kelp bed and then proceeded to map. We ran multiple passes over the bed, logging water column data at different frequencies (the file sizes are huge!) and then conducted more GoPro surveys to try and capture key features such as the kelp boundaries, sandy bottom and deeper rock. The idea being to validate the presence of kelp on shallower rock, helping us to better interpret the acoustic data.
Hopefully I might get out to do another survey but the weather isn't looking great for the week!
- Tom -
So next week I'm off down to Baltimore to join some of the team from the Geological Survey of Ireland (GSI) and will spend as much time as possible out on the survey boats developing the kelp survey methodology.
The first step, aside from coordinating will team leads and renewing my medical, is to try and identify suitable kelp locations prior to survey. This will increase the efficiency of our surveys as we, hopefully, won't waste too much time plodding around!
I didn't really have a playbook for how to go about identifying kelp without actually being there. I first got the most recent INFOMAR bathymetric and backscatter data for the region to get an idea where the rocky substrate it and also what depths it can be found at.
I found an old National Parks and Wildlife Service (NPWS) survey from the area where they had identified kelp beds using dive transects. From there I selected about 5-6 areas where all three of these datasets overlapped nicely. I didn't want kelp that was too shallow and close to rocky coasts, even though we'll be using small boats, I believe it still wouldn't be feasible.
Fingers crossed that this will help our little expedition.
The next step is going to be how to figure out the best ground-truthing method!
Kelp forests are wonderfully diverse environments support a range of ecologically and commercially important floral and faunal species. Kelp forests are found throughout temperate rocky coasts and, in Ireland, are a dominant subtidal species.
Despite their importance, little is known on kelp distribution. In order to fully protect and manage this resource it is vital that a standardised, cost-effective and efficient mapping methodology is developed.
This is where our project comes in. Over the coming months we are aiming to utilise vessel-mounted acoustic remote sensing technology to conduct small-scale mapping of kelp forests in the southwest of Ireland. We will be working with Ireland's seabed mapping programme, INFOMAR, world leading experts on acoustic remote sensing! We want to ground-truth this data using drop-down cameras and ROVs and these offer a safer, quicker and more cost-effective alternative to dive or snorkel surveys.
In the end we hope to have developed a methodology that can be expanded to much larger areas, transitioning, in effect, from scientific theory to a real-world practical application that will have direct benefits for the environment and for humans.
Contribute to this expedition
Thank You for Your Contribution!