Monitoring kelp with sonar

Between the end of my Master’s and the beginning of my PhD, I was fortunate enough to land a job working as a research assistant in a project which was joint between the Institute of Zoology (supervised by Dr. Chris Yesson) and the Natural History Museum (supervised by Dr. Juliet Brodie) in London. The project involved investigating the potential for sonar technology as a mean of continuously monitoring kelp distribution, focusing on the British Isles.

kelp-forest
image taken from: http://oceanservice.noaa.gov/facts/kelp.html

What is kelp?

Kelp are large seaweeds, usually found in shallow oceans, between the depths of 2m to about 30m. They grow together, forming kelp forests, a unique habitat which becomes home to a variety of species, ranging from invertebrates (such as crabs), to mammals (such as sea otters). Kelp forests around the British Isles are distributed along most of the coastline, covering around 19,000 km² (Yesson et al. 2015). In addition to being important habitats for some animals, kelp forests also provide important benefits, or ecosystem services, to humans. These include being harvested for commercial purposes, as well as for ecological purposes, such as reducing disturbances and erosion by waves.

Why monitor kelp?

Historically, kelp has been harvested for a range of uses, including food, fertilisers, and alginates, and demands have been growing recently. Moreover, there are now ideas of using kelp as a renewable biofuel. These have combined to lead to a growing interest in the wild harvesting of kelp. At the same time, there is also growing concern around kelp distribution and abundance, as an increase of atmospheric CO2 can impact where kelp can grow (Brodie et al. 2014).

Kelp harvesting machine
Modern kelp harvesting machine. (image from alamy.com)

As a result of this need, our project was funded by the Crown Estate to investigate potential ways of monitoring kelp distribution continuously to allow for proper management of kelp harvesting activities around Britain, ensuring the sustainability of the industry, as well as the proper conservation of the ecosystem.

With this in mind, my supervisors came up with the idea of using sonar to achieve the objective. This was a particularly attractive idea as the UK Hydrographic Office (UKHO) gathers sonar data all around the British Isles for navigation safety purposes, and have subsequently made this data freely available to the general public. If the project were to succeed, the methods developed can be used to produce data related to kelp distribution without any additional costs. Using remote sensing techniques also circumvents many problems encountered by traditional field surveys, which require huge amounts of time and resources to generate relevant data for limited areas, particularly for kelp forests which grow in difficult-to-access areas with poor visibility.

What is sonar?

Sonar stands for SOund NAvigation and Ranging, though the acronym is much more widely recognisable. It is often used in marine environments, and works by using an echosounder which emits a sound wave. This then travels towards the target surface, and is echoed back towards the echosounder, where a receiver captures and processes the echo. It reveal information related to the depth and bottom composition of the target surface.

We can broadly categorise sonar devices used in marine environments into three groups: singlebeam, side-scan, and multibeam, each with different converages.

Sonar methods
Examples of the three most commonly used sonar methods, showing their respective coverage. Taken from Penrose et al. 2005

Out of the three types of sonar methods, singlebeam data are the cheapest to produce, and are commonly used as fishfinder. On the other hand, multibeam data can provide the most information, creating detailed, three-dimensional benthic maps. This is also the data available from the UKHO.

Fieldwork

Simply relying on remotely-sensed data is not sufficient, as we also need underlying information to first train our models, telling it what signals kelp forests might produce. To do this, fieldwork was carried out at four sites: Lulworth, Devon; Wembury, Devon; Stackpole, Pembrokeshire; Falmouth, Cornwall.

Survey sites.png
Locations of the field sites in the southwest part of the UK. Produced in QGIS

At these sites, we attached a singlebeam echosounder, and a GoPro camera to our kayaks and followed a series of transects. We also carried a GPS tracker to produce a track of our exact coordinates.

We then used the footage from the GoPro camera to determine whether kelp was present at particular points of our track.

Kelp_PA_figure.png
Observable differences seen from GoPro footage

Additionally, we also looked at the images produced from the singlebeam echosounder to investigate whether it is able to detect kelp presence.

singlebeam
Image produced using singlebeam echosounder, showing the different textures as a result of kelp presence.

We can see distinct signals produced by kelp forests, and it is this signal which we hope is also captured by the multibeam echosounders. Unfortunately, as the data from the UKHO was collected for navigation purposes, the signals produced by kelp forests are actually seen as noise, and thus have been removed from the archives in some instances. This also meant that the backscatter amplitude (which provides information on bottom composition) was not important to them, and is not archived in some areas.

Additionally, since the multibeam echosounders and detectors are carried abroad much larger boats (compared to our kayaks), its coverage doesn’t always reach into shallower waters.

Pembrokeshire_figure.png
Maps showing gaps between mean low water and multibeam coverage in
southern Pembrokeshire. Green: area from mean low water; Yellow: area with multibeam coverage; Red: transect surveyed using kayaks

Results and Conclusion

Unfortunately, the data we gathered were not sufficient for us to create the detailed kelp distribution map we were hoping for at end of my time. Despite not being able to achieve the goals we originally set ourselves, we were able to develop a framework for the methods required to produce such a map if the data becomes available at a later date. Moreover, our research directly pinpoints the areas which are currently lacking, and can be used to identify areas for future development. As a result of the project, my supervisors were also able to secure funding for an extension of the project, and have hired a replacement for me as I left for Australia, to continue on towards our original goal of mapping kelp around the British Isles.

Finally, the project has given me an amazing opportunity to have a play at sonar data, and learn how to use the various open-source software. Working at London Zoo and the Natural History Museum is pretty awesome too.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s