FAD numbers and their tracking in the WCPO / by Francisco Blaha

I never been to the meetings of the Science Committee of the WCPFC (the 14th one is happening in Busan, Korea at the present) yet I try to make time to read the papers presented (mostly by SPC) as they present the latest understanding of an incredible variety of aspects associated to the management of the world’s biggest fishery.

While by academic training I’m a fisheries biologist, after my thesis (identification of new reproductive stock of hake) I barely did anything related to stock assessments or the wider applied science field, as went back into fishing and operational work. Yet I get a kick from pushing myself to stay updated. 

Dont blame the tech... blame their missuse

Dont blame the tech... blame their missuse

Some of the papers are really dense reading and for the connoisseurs only, but most can be followed if you have a keen interest and know the basics… in any case one of my favourite topics if the one on FADs (and been writing for a while about them), and among the paper of this session there are two that brought home the staggering numbers we are dealing with, the technological advances in them, the impact they are having on the fishery and the challenges around their management (or lack of).

My brilliant colleagues at SPC author the papers “Estimates of the number of FADs active and FAD deployments per vessel in the WCPO” and “Report on analyses of the 2016/2018 PNA FAD tracking programme”, and if you are interested in FADs are compulsory reading.

I quote below the parts that impacted me the most. And then at the end, for whatever is worth, I rant a bit on my ideas for their management.

The first paper (estimates of FAD numbers) attempts to estimate the number of deployments and active FADs per vessel. Estimates were derived using two different approaches. Firstly, based on fishery data for 2011–2017, the number of deployments recorded in the observer data, the observer coverage by vessel, and a clustering of vessels based on their FAD fishing strategy were used to estimate the total number of buoy (and FAD) deployments per vessel and total in the WCPO. The number of deployments varied from 0 to 500 per vessel but few vessels (depending on the estimation method and year) deployed/redeployed more than 350 buoys per year. This corresponds to a total estimated number of deployments between 21,000 and 51,000 per year in the WCPO for the 2011–2014, but the numbers drop thereafter, likely due to delays in receiving observer data for recent years. 


The second approach combined fishery data and the PNA FAD tracking data and therefore only covered 2016 and 2017 with precise estimates only possible for some vessels. The estimated number of deployments per vessel varied between 1 and 550 (mean = 129) in 2016 and 1 and 999 (mean = 226) in 2017 and the estimated number of active FADs per vessel varied between 1 and 454 (mean = 102) in 2016 and 1 and 955 (mean = 163) in 2017. At the scale of the WCPO, this corresponds to 30,700–56,900 deployments in 2016 and 44,700–64,900 in 2017; and 26,200–37,300 active FADs in 2016 and 38,000–48,200 in 2017. The ratio between number of deployments/redeployments per year and number of active FADs per vessel and per year average at 1.48. Less than 16% of the vessels were estimated to have more than 350 active FADs per year.

Yet the sentence I’ll remember the most is “The total number of deployments varied between 30,700 and 56,900 in 2016 (based on average and quantile 95%) and 44,700 and 64,900 in 2017”. One could be inclined to think that with so many FADs there are actually no free schools… but just schools swimming in between FADs.

To improve the ability of the Scientific Committee to estimate potential FAD levels, the collection of additional information is suggested. For example, to better understand the total number of FADs in the water, this could include the submission of i) the number of new FADs deployed per year per vessel; ii) the average daily or total number of active FADs per vessel per month; and iii) the number of deactivated FADs per month. The first could be derived from fishery data, if observers can record all FAD deployments or if captains start recording these data in a FAD logsheet. In order to obtain an estimate of the average daily or total number of active FADs per vessel these data could be derived from FAD tracking data. In parallel, to better study FAD density, aggregated summaries per 1° cell and month including number of buoys activated, number of buoys deactivated and number of FAD deployments could be considered.

Smoothed kernel density of deployments of buoys per 1° grid cell during 2017. Maximum number of deployments per cell is 300 in 2017.

Smoothed kernel density of deployments of buoys per 1° grid cell during 2017. Maximum number of deployments per cell is 300 in 2017.

The 2nd paper introduced me to the issue of geo-fenced buoy transmission on FADs by which the data received by PNA are modified by fishing companies prior to submission, for example, information outside PNA Exclusive Economic Zones (EEZs) may be removed presents analyses of the PNA’s fish aggregating device (FAD) tracking programme. Approximately 35% of the fishing companies were found to have geo-fenced more than 90% of their buoys. When buoys are geo-fenced it leads to gaps in the trajectories of a few days to 1 month, limiting the analyses performed on the data.

This very interesting report includes a description of the data processing required; estimated data submission rates to the PNA; a description of the spatio-temporal distribution of buoy deployments; FAD densities; FAD connectivity; and an analysis of the fate of FADs including a focus on FAD beaching. As FADs drift in the ocean, the associated electronics can be changed making it difficult to follow individual FADs, therefore for the purposes of this analysis we followed the satellite buoys unless otherwise stated. 

To better distinguish drifting buoys from those on board vessels, data were analysed using a Random Forest model to identify, and select, the drifting at-sea section of each buoy trajectory, and at the same time identify deployment positions. In addition, using two methods matching buoy tracks and observer or logsheet data, we estimated that ~60–70% of buoy transmissions collected by fishing companies are not forwarded to the PNA. We noted that some of the data received by PNA are modified by fishing companies prior to submission, for example information outside PNA Exclusive Economic Zones (EEZs) may be removed (i.e., “geo-fenced”), which added a bias to the analyses. After undertaking the correction procedure, the cleaned dataset consisted of 14.8 million transmissions from 26,466 buoys and covered the period from 1st January 2016 to 18th March 2018. 

usual image in most Purse Seiners I board. Satlink is the clear favourite brand

usual image in most Purse Seiners I board. Satlink is the clear favourite brand

The number of deployments varied over time, with a total of 36,831 deployments in 2016–2018 (from 193 vessels including 102 buoy owner vessels and an additional 91 vessels where the fishing company was known, but the buoy ownership was not). The spatial distribution of deployments was very similar between observer data and FAD tracking data, both showed the main deployments areas to be in Kiribati South of the Gilberts Islands and East of the Phoenix Islands, Nauru, East of PNG. 

The number of transmissions from buoys almost doubled in 2017 (8.6 million compared to 4.5 in 2016) and the number of individual buoys active in the available data was 10,915 in 2016 and 18,405 in 2017. A decrease in both numbers was detected during the FAD closure in 2016 and 2017, although in 2017, both remained relatively high and constant during the first 3 months of the closure. Although influenced by the issues arising due to geo-fencing, the spatial distribution of buoy densities were investigated, with higher densities in Kiribati South of the Gilbert Islands and around the Phoenix Islands, Tuvalu, PNG and the Solomon Islands. Buoys movement between large grouped areas of the WCPO was also investigated. Patterns varied between areas; for instance, the Southwestern area mostly received buoys from the East and had a high proportion remaining or being deactivated there, compared to the Southeastern area which showed high deployment and emigration rates. 

Not much biodegradable or non entangling in this FADs

Not much biodegradable or non entangling in this FADs

Finally, at least 5% of the buoys ended up beached (probably underestimated as buoys may be deactivated before reaching coastlines), with the connected FAD potentially damaging sensitive ecosystems such as coral reefs. At least 26% of the buoys in our dataset could be considered lost, likely leading to marine pollution.

My own thoughts?

At a personal level, I think the technology associated with the FAD (i.e. the buoys) is amazing, yet as any technology need to be assessed by their impacts. 

On one side the FAD/buoys can be (and are) set to drift across Marine Protected Areas, where they drag the fish (while being monitored from the vessel) which can meet them on the other side and set around them… so it would not surprise me that a lot of effort is now being done on the eastern boundaries of the Phoenix Islands Protected Area- PIPA.  Yet on the other side, one could see it as the potential of having 65000 echo sunders providing is with an unprecedented level of understanding of the stock status in the Pacific, because the information is being collected as we speak… yet that information is not shared with SPC

Hence from the management perspective, it brings a massive challenge, and the only way I see it can be managed is through licensing conditions and costs… stick and carrots approach

Through licensing by having a cap in the number per vessel, having the vessels registering the frequencies directly to the management organizations and linking pairs in between vessels and FAD buoys, and their sharing arrangements. Complex but not impossible, FIMS is doing something along this lines with the allocation and use of Vessel Days (VDS).

Furthermore, as many newer legislations include the possibility of enforcing laws over nationals of Flag states, the buoys provider mainly based in Spain, TW and China could be required to register their units sold frequencies with the regional bodies when that FAD Buoy is sold to a vessel fishing in certain area (as is the case with VMS units). When a set is done, the logbook and observers also record the FAD buoy ID in the forms. The rest is a matter of data management and tiding loose ends.

Then is the carrot, the more transparent the vessel, the more information shared is verified… the lesser the cost of each Vessel Day or other forms of licensing agreements. 


As a mostly operational consultant I’m not in the position to influence the management thinking… but I found that when you deal with fisherman (as I was) it pays to think like them. If you what to deal with them with only the bureaucratic approach, chance are you always going to be running behind the ball