Skip Ribbon Commands
Skip to main content

Fishing for Data

July 22
Stage 6 – Digital digest

​"If we don't manage this resource, we will be left with a diet of jellyfish and plankton stew." - Daniel Pauly

Update: We're now on a northern path, edging ever closer to the Shetland Islands and our westernmost survey rectangles. Yesterday we picked up both IBTS and saithe trawls, although we did experience a small problem with the trawl net at one station. Upon winding the net in, the crew discovered that the cod end, the slimmer enclosure at the end of the netting in which the fish bunch, had broken whilst in the water. Although several larger fish were present, the damage meant that any smaller fish that had entered the GOV would have wriggled free, rendering the trawl invalid. With the catch now unrepresentative, we had to try again.

One interesting organism – or parts of an organism – was found in the catch yesterday, puzzling the scientists for a short while. It came in the form of some glutinous, jellybean-like specimens. After some speculation and research they were found to be hagfish​ eggs. 

heggfish eggs.jpg 

A numbers game
We'll now look at the stage in the process where the handling of the physical fish makes way for the handling of the numbers that represent them. The fish have been sorted, sampled​ and (in some cases) sliced, the gloves and overalls are off and the wet lab has been cleaned of fishy residue.

The team of scientists on the shift then sit down to store and process their findings on a computer. 


First the length and weight data from the three electronic measuring boards in the lab is automatically fed into the system. This arrives with metadata from the bridge – the control deck – on the sampling station number, depth and times of the trawl, the ICES rectangle we're positioned in, trawl net mouth opening and length of wire, amongst other things. Stored together, all this information is logged for the current station (number 2421), sitting alongside the data from the 140 other stations called at since the survey started out back at the end of June.

Once the data has been uploaded, it is pooled in one table, categorised by species in the sample with various codes denoting the type of sample taken and number in the sample, fish lengths, weights of the ones that are length measured, and so on. Any age data, obtained from the otolith readings, is added later.

Other keys describe the units of weight and of length and the way the measurement has been taken. For most fish species this is from tip of snout to end of caudal fin. For octopuses and squids (cephalopods) it is the length of their mantle, or body minus the tentacles. For crustaceans the size of their shell (carapace). For the former that's because the tentacles tend to contract and for the latter because the hard shell is unchanging, both offering a better indication of size. Details of crustaceans and other seafloor non-fish can be entered into the computer manually at the end. 


This process is the same for both the surveys carried out on the Johan Hjort. However, for the IBTS and the saithe survey, two separate calculations are performed due to a key difference between them: that is, that the IBTS catch is organized into species whereas for the saithe, after the larger fish have been picked out,  a subsample of a mixture of the smaller species (such as Norway pout,  blue whiting and silver cod)   is taken. This is to give a representation of the weight of each species in the catch as well as to save time between trawls. 

For the IBTS, the catch often yields a lot of Norway pout, which when sorted can result in often numerous baskets (sometimes as many as 15!) . Although these are separated, counting them all would take forever and wouldn't make the data any more accurate.  

So, to work out the amount of, say, Norway pout in the catch, the scientists take the total weight of all baskets of the species (let's say 20 kilogrammes), multiply it by the number of pout in the sample (here, 100, as that's the maximum number that can be sampled – and there are way more than that) and then divide by the weight of the sample (let's say 5 kilos). So 20 x 100, over five – a sum that gives us 400, meaning there were 400 in the original catch.

If there was a notable size discrepancy in the pout (enough to get an adequate representation), then they might be split into two samples – big and small haddock for example.​


Finally, the data is checked against that from all other stations for incongruities, with some entries fixed now and some later. Notes can also be made on the trawl. Was it initiated because of an echogram (for the saithe)? Did it pass by without any problems? Was there any gear damage? Was it a blind trawl? Or, just like ours yesterday when the cod end came undone, did it not fish correctly?

Everything is mapped and even the smallest detail could prove important when the data is processed further and cleaned by IMR later.

In the meantime, this batch is exported to the upper deck of the Johan Hjort to be turned into graphs and maps that show how the lengths of fish are distributed across the survey area. 


There are no comments for this post.