Looking for patterns & anomalies in the data
The F-POD.exe app has built in analysis tools that make it quick and easy to see a snapshot of animal activity
We offer training in the use of the F-POD.exe freeware
​​
The files collected and donated to the CATT Project are open sourced on the Figshare+ open repository.​
If you would like to share or use this information it is published under CC 4.0.
Please credit the organisation or group that gathered the data for CATT, if in doubt, please contact us
Click on the images to magnify them
Assessing the whole file
Here we are looking at a file from St Martin's on the Isles of Scilly gathered by IoS IFCA for the university of Plymouth Fish Intel Project. PI Dr Emma Sheehan.​
When first loading the FP3 file, we see an overview of the detection trends between June and December 2022, the height of the lines shows detction rate. Dolphin detections are in orange, and Porpoise detections are in purple.
We can see that Dolphin activity increases significantly as we enter the winter months
Porpoise activity is more consistent but summer has slightly fewer detctions
These two graphs show diel activity cycles, the 0 on the left is midnight, 12 is midday and 23 is the following midnight.
The height of each yellow line reflects the number of clicks logged in that hour, on average over the 181 day file.
The left graph shows Dolphin activity, a preference for clicking at night.
The right graph shows Porpoise activity, a more even distribution
Porpoise Behaviour
Looking for social calls
Looking at the same files from St Martin's we see some click train calls that look like social interaction.
Research is underway into social communication. It is early days but we are interested in any repetitive or complex click trains.
NOTE : The F-POD.exe software generates these audio representations of click trains. The change in pitch we hear is not a change in the frequency of the clicks, rather a reflection of the changing click rate.
The three images above show Porpoise click trains, the click rate changes are represented by the red lines in the top panel, these will be Harbour Porpoises in U.K. waters indicated by the purple coloured 'species class' lines in the middle panel).
The lower panel indicates a frequency of around 120 KHz (the violet coloured blocks in the lower panel).
When we see a wavy line along the top of the violet frequency blocks, this indicates increasing and decreasing volume levels, we interpret this as the directional click beam sweeping up and down past the F-POD in time with the animal tail beats.
These repeated social calls are seen in many places, the detections below are from the Bristol Channel off North Devon
The next set of Porpoise social calls were captured off Jersey in the Channel Islands
Dolphin Behaviour
West Wales click frequencies
Here we are looking at three files from West Wales gathered by Dr Sarah Perry of the Cardigan Bay Marine Wildlife Centre in collaboration with PhD student Laura Palmer, supervisor Dr Stephanie King, University of Bristol.
CB4 is an inshore Cardigan Bay site, CB6 is an offshore Cardigan Bay site.
The third file is from an F-POD in St Brides Bay off Skomer, the sites are 20 Km to 30 Km apart.
Graphing is an easy way to spot anomalies, the images below show averages from three sites quite close together off the Welsh coast. The top 3 images show Dolphin data, the diel patterns in yellow show a preference for night time activity, the lower, colourful row of graphs show frequencies used by the Dolphins in their click trains. The left graph is Cardigan Bay offshore, a site with both Bottlenose and Common Dolphins, the middle graph is an inshore site with mostly Bottlenose Dolphins and the right graph shows a nearby site off Skomer Island.
The offshore Dolphins (left) were clicking in a narrow band around 42 KHz.
The Skomer Dolphins (right) were clicking at around 43 KHz with some activity at 120 KHz.
The Inshore Dolphins (centre) have a big frequency peak at 22 KHz... No idea why!
These are averages over many weeks and months, in the case of CB6, over 2.5 million Dolphin clicks detected.
This second set of images below show the Porpoise detections from the same files / sites over the same period.
These give us confidence that the F-PODs were working normally and logging all frequencies.
The frequency profiles are what we would expect from NBHF Porpoises
Edges of distribution
Fal River seasonal population
The following graphs are from one of the small Chelonia LF-PODs that we have been deploying in the Fal River close to the King Harry Ferry crossing at Trelissick in West Cornwall.
While testing new F-POD equipment and settings in the Fal River and not expecting to see many cetacean detections, we found evidence of Common Dolphins spending the nights through December and January 4-5 Km up the Fal River. We have had one of our small LF-PODs on Chris Rangers Oyster Farm for some years now and we would occasionally capture some cetacean activity but it does seem that some of the Common Dolphins we see in Falmouth Bay through the winter regularly head up the Fal River over night.
This graph shows Dolphin detection Positive Minutes (DPM) between 6th November 2023 and 6th April 2024. There is very little activity in this area through the rest of the year but every night through December, January and most of February, a pod of Dolphins spent all night within range of the F-POD.
The diel pattern for these files shows that the Dolphins would arrive around 17:00 and leave around 07:00. This has been confirmed by people on liveaboards in the area and commercial operators who are on the water at these times and see the Dolphins 'commuting'.
We are interested in studying the limits of distribution around the rivers and estuaries to see if these animals are returning to areas where they were common in the past but abandoned through the 1970's and 1980's.
​
We have now added the small LF-PODs to sites along the Fowey River and the Tamar. We are planning with local authorities to add further monitoring points in the Helford, the Yealm and in the 'Roads' on the Isles of Scilly.
​​
The reasons for the disappearance of the animals ​in the 70's and 80's may be due to organochlorines such as DDT or chlorinated hydrocarbons like PCB entering the waterways. Noise may be a factor, the animals are limiting their visits to winter months when few boats are in the water and overnight when most human activity decreases.
Click train phonemes and phrases
Bristol Channel Aquaculture Farm Porpoises
This set of statistics comes from the Bristol Channel four miles off Clovelly. The site is dominated by Harbour Porpoise activity.
CATT Top 5 Porpoise 'Detection Positive Minutes' (DPM) by Location DPM Porpoise
DEV Clovelly Algapelago 2024 06 06 Sea Weed Farm Bristol Channel 199
DEV Clovelly Algapelago 2024 06 06 Sea Weed Farm Bristol Channel 178.8
DEV Clovelly Alagapelago 2024 06 05 Sea Weed Farm Bristol Channel 145.6
The Algapelego Seaweed Farm four miles off the North Devon coast has three F-PODs as part of a biodiversity monitoring project run in by Ocean-OS.
This has been an interesting site, the three F-PODs have given us the highest recorded DPM (Detection Positive Minutes) for Harbour Porpoises that we have seen on the CATT project.
The first image, above, shows a fairly typical Porpoise feeding buzz. The click rate shown in the top panel increases as the animal approaches the target and the return time of the click decreases.
The Porpoises are using quite long repeated social 'calls' as seen in the earlier images, we are also seeing feeding behaviour and a blend of possible feeding, foraging and social click trains.
The next images combine long and loud clicks with upsweeps, down sweeps and some repeated calls.
These are all open to interpretation, in the future this may include A.I. machine learning or pattern recognition tools...
Interpreting these ultra-sonic click trains through visual representations is an acquired skill. We offer free training at a basic level for citizen scientists online or face to face for groups.
​
Chelonia, manufacturer of the F-PODs offers much more detailed training on scientific and commercial best practise for professional F-POD users.
Dolphin Behaviour
West Wales social calls
Here we again look at files from West Wales gathered by Dr Sarah Perry of the Cardigan Bay Marine Wildlife Centre in collaboration with PhD student Laura Palmer, supervisor Dr Stephanie King, University of Bristol.
CB4 is an inshore Cardigan Bay site at New Quay, we see high click rate dolphin calls that seem social.
Research is underway into social communication. It is early days but we are interested in any repetitive or complex click trains.
WUTS - Weak Unknown Train Sources
These mysterious clicks were first noticed in POD data from the Fal River twenty years ago.
These first WUTS are short, quiet, isolated click trains with a constant click rate or a down sweep, the image below is a Fal River WUTS. They are still there and we are still recording them.
Since those early days WUTS have been detected in POD data from mangrove swamps in Australia, in the Gulf of California, the Gulf of Maine, the Gulf of Alaska, the Baltic Sea, the Yangtze River, the Ganges river systems in Nepal, and other places.
They come in a wide variety of structures, always low amplitude, but compared to the original Fal River WUTS, some of these international WUTS are quite different, they can have click rates well into the thousands.
The KERNO-F classifier catches most of them up to around 2500 clicks per second but in the raw data, click rates up to 7000 clicks per seconds have been seen.
The ‘click trains’ can be very long and complex down sweeps, either smooth or saw tooth. WUTS ‘storms’ have been seen with large numbers of WUTS all clicking at the same time.
We don’t know what they are, best guess is that there are many different sources of weak click trains most being small animals and possibly some being physical processes.
As WUTS in Cornwall tend to occur in river estuaries and click during the day, there is interest among some students here to be the first to identify a WUTS. There are plans for F-PODs with cameras and for catch and release techniques. Watch this space...
Fish Tag Detections, a data validation example
Researchers at University of Swansea supervised by Prof. Dave Clarke and Dr. Chris Lowe found some unexpected detections in their POD data, eight distinct pulses at around 70 KHz. The Swansea researchers found that these were the data pulses emitted by fish tags that were part of an unrelated research project.
This fortuitous discovery led to the first big step forward for the CATT project, a collaboration with the Fish Intel project at the University of Plymouth. We were able to add CATT F-PODs to the acoustic telemetry on acoustic releases deployed by the regional IFCAs (Inshore Fisheries and Conservation Authorities) to capture the audio transmitted by open protocol Fish tags developed by European Tracking Network (ETN). The data shown here is from Isles of Scilly and was gathered as part of the Fish Intel Project at Plymouth, P.I. Dr Emma Sheehan.
In the example above the tag is recorded alongside Harbour Porpoise detections (purple).
In the example below the tag is recorded alongside Dolphin detections (orange).
In the next two examples we can see how fish tags have emitted their pulses at the same time as Porpoise click trains.
The KERNO-F classifier didn’t mis-classify the fish tags as Porpoise click trains.
This is not always the case, Dolphin click trains are more variable in their acoustic characteristics, so differentiating the fish tags from Dolphin click trains is harder for a classifier.
In the following two examples, at least some of the fish tag pulses seen in the raw data (FP1 in the lowest panel) have been included in FP3 Dolphin click trains by the classifier (top three panels).
This demonstrates that while the software can do much to sort the data into classes and groups, human validation is still an important part of the process and a valuable skill when assessing remotely gathered data.
The next example shows how eight pulses at around 70 KHz look like they could be a fish tag but given the activity around them are likely part of a Dolphin click train… Possibly.
As a citizen science project, CATT offers some simple, single click analysis tools to give a detailed overview of the long term, high resolution recordings of cetacean click trains.
There is scope to take this research much further, linking it to other biodiversity monitoring, and historical data.