A recent paper was done by a group of academics in the University of Santa Barbara and people I know from Global Fishing Watch publish (and promoted quite substantially) a paper named “Satellites can reveal the global extent of forced labour in the world’s fishing fleet”.
realities on bard ain’t like what you think at university sometimes
Let start with the name hat quite “eye-catching” I guess… I find it always difficult to digest generalizations based on extrapolations.
Don’t get me wrong, the paper has some very good points and I like that explore a different take on the issue of labour rights on board… which is a very complex area to deal with as I wrote many times before.
I was quite keen to see the methodology and how they got their “indicators” both from the ILO definition of forced labour and vessels behaviours, which is the only thing you can assess from satellite data in reality.
The bit that disappointed me from start is that, as usual with my former colleagues in academia (yes I’m also a proud scientist with 2 Masters in different sciences 10 years apart) is that they did not seem to speak to fisherman or to inspectors directly… which to me would be the logic first point of call: you start with the stakeholders… yet they seem to only have contacted opinion holders. The methodology says:
“We next conducted informal phone interviews with experts from several nongovernmental organizations (NGOs) working in this field, during which we asked interviewees what observable vessel behaviours they would look for if they wanted to identify suspicious activity…”
“NGO experts and investigative journalism suggest that gaps in AIS transmission, port avoidance, transshipment, and extended time at sea may indicate the presence of forced labor…”
Fisherman will tell you things from the inside, so I will take that role and provide some humble constructive criticism on features and descriptions they use for the model features used for training data (is in the supplementary information annexe)
Average daily fishing hours on days when fishing.
Has to be related to the gear…
Longline is the most time-consuming gear and is directly related to the length of the set and weather.
Squid jiggers fish only at night, so it needs to be weighed against latitude and season in the winter night in the 40-50 S (or N) last over 12 hrs, while in summer is 6 to 8.
Trawlers totally depend if they process onboard, what species which determine depth they are targeting… if you aiming deep-sea species like orange roughie setting at 500m means that you have at least 1500 of cable on the winches, so the trawl with a net during fishing of 1-2 hours setting and hauling may 3-4 hrs either side… and once the net in on sinking and fishing you don't need the whole crew on deck… so are they working? Furthermore, once the net is on deck… the vessels are not fishing…. But then is when you really work! either by processing or just freezing and even in the simplest of cases icing…
There are sooo many variables according to gear… that I don’t know how can you standardize this as to be a useful indicator.
That also applies to Vessel Engine Power, which they have a top 5 predictor of human labour abuses… to the same length of a fishing vessel, trawlers need proportionally way more engine power than a squid jigger or a longliner of the same size (or even tonnage). Just think the physics of trawling as an active gear where the boat has to drag a massive and increasing dead weight behind and below, vs LL or Jiggers where the engine is just used to bet the vessel in position and to move fishing grounds and port…. so I assume they would be some sort of correction factor by type of gear (I could not find it)
These next 3 assumptions, unfortunately I see them permeating in many papers these days:
Number of suspected transshipment events with other vessels. Suspected transshipment events are defined by events when two vessels are within 500 meters of each other, traveling less than 2 knots, and for a minimum duration of 2 hours. These suspected transshipment events may represent transshipment, refueling, or transfer of supplies or crew
Average suspected transshipment event duration (hours): Average duration of suspected transshipment events with other vessels. Suspected transshipment events are defined as events when two vessels are within 500 meters of each other, traveling less than 2 knots, and for a minimum duration of 2 hours. These suspected transshipment events may represent transshipment, refueling, or transfer of supplies or crew
Average loitering duration (hours): Average duration of loitering events. Loitering events are defined when a vessel is traveling less than 2 knots, at least 1 kilometer from shore, for at least one hour. These events may represent suspected transshipment events if the second vessel is not broadcasting AIS.
Operationally there are many valid reasons for which a carrier and a Fishing Vessel (or two fishing vessels) can get alongside that do not imply transhipment as defined in the CMM, passing food, gear, crew, parts, salt, oil, etc. name it… it happens! In one case I remember we received the ashes of the captain’s dad who wanted to be spread in the high seas.
Also honestly it takes the best of one hour to tied up securely two vessels at sea, a PS can tranship anything from 700 to 1.3 tons at the time, depending on the made and shape of the hatches (read here if interested) and they could be doing 6 to 10 per hr… so in the best case scenario we are talking 10 tons in one hour… absolutely not worth to use 30 tons as a base…
This logic can be also applied to Jiggers and Trawlers that tranship with nets from the carrier's winches... yet in their case the fish is boxed or frozen in plastic bags so it is transferred over some sort of pallet structure, yet as i said the size of the hatch is the key limiting factors (other than weather)
For longliners never been on transhipment that took less than 10 hours…. I recently peer review over 20 observer reports on longline carriers and the range was 11 to 50 hrs! so really what can you do in two hrs of which one is just getting the vessels next to each other and then away… 4 to 7 hrs (weather dependent) would be my absolute minimum…
So by having such a narrow description of what is suspected as transhipment you are expanding the risk model to genuine interactions that do not reflect at all the issues at hand. Further, I object the assumption that transferring parts or crew is to be categorised as transhipment, that has relatively clear fisheries definition (see here)
As per transferring crew at sea, while we all agree is not the best, yet depending where you want to fly in crew to transfer, you may face the issue of transit visas... for most of the North Pacific they have to go via Honolulu, for the south Australia or NZ, all of them require transit visa (yes even if you don't leave the airport) for Indonesian, Myanmar, Philippines... In the region had some hope that Air Niugini (PNG) and FIJI airways maintain their Asia flights as there more logical visa restrictions was very positive for foreign crew...at least they had a passport and rcords of travel.... Thailand has pot some better condition on passengers on carriers which is the most common way to transfer crew on purseiners... but for Chinese and Vanuatu flagged (that operate from TW or China) carriers is not really an option.
Finally and under loitering: Loitering events are defined when a vessel is travelling less than 2 knots, at least 1 kilometer from shore, for at least one hour. So this will include what happens with vessels coming in with the 2 pilots we have in Majuro o Tarawa entering the vessels to the lagoon
Flies flag of convenience: A boolean for whether or not a vessel flies a flag of convenience
Again… take the usual outs (i.e. Panama. Liberia, Vanuatu, etc) you could argue that in terms of labour CN, TW, Korea and the US, for example, act as FoCs by having a crew on board without the protections of labour conditions that would apply to their own citizens, nor the minimum immigration requirements that will apply to foreign workers coming to work in their land for example… which is the case for vessels flagged on those in principle non-FoCs nations.
Maximum distance from port (km): Maximum distance from any port.
That would be biased against any port in the pacific as it is the biggest ocean and with the lesser # of fishing ports around.
I also have a big issue with the total generalization this one:
Number of visits to ports of convenience: Number of visits to ports in countries that had not ratified the Port State Measures Agreement (PSMA) at the time of port visit.
There are many reasons why a country may choose not to sign PSMA while still provide for best standards in PSM (for example RMI, PNG, Tuvalu just to name a few) on the other side of the coin. Countries like Uruguay who has been documented receive vessels that have fished illegally in the south Atlantic was one of the 1st ones to sign PSMA, Vanuatu, a country with an abysmal record of labour infringements on board, is a signatory… furthermore, and as I wrote many times as fisheries inspector I can’t do anything in terms of labour, only since this year, and so far as the only place in the world, for the FFA member countries fishing access and labour issues are associated. In the rest of the world… I can refer the vessels to the labour agency of the port state… but unless I can prove any issues related to IUU fishing, I cannot stop the vessel from departing
So yeah, again, I’m not taking a cheap shot to the crew that wrote this paper, as they are ALL so much better qualified than me... I'm absolutely no one... But I like to think their numbers could be way more accurate if they were to adjust some of the assumptions they made. I think the better ground truthing of the methodology, the more reliable the results would be and the better the tool you trying to use.
My point is that while this paper brings very useful points, yet those could be more accurate and provide a much more realistic figures of the issues… facing that between 2,300 and 4,200 unique vessels were high- risk and between 57,000 and 100,000 crew members were working on these boats and thus potential victims of forced labour, can be quite daunting in terms of the scope of the problem to challenge, yet adjusted numbers can be more manageable with the tools and instruments we have now and this is a start on the real size problem we need to work out.
As with the book I wrote for FAO on Blockchain, it worries me when the current media discourse that seems to pin the solution to multifaceted seafood value chain problems on one tool – as AIS is. This risks hyper-inflating expectations on what this technology can offer, with potential users then walking away because it does not deliver on the hype built around it.
I hope the authors and my friends at GFW take any of these comments as nasty criticisms, is absolutely not my intention.