The global footprint of drifting fish aggregating devices

Abstract

Tuna are among the world’s most valuable marine life and have long been exploited by industrial fisheries. Increasingly, tuna fishing companies have shifted from targeting free-swimming fish to using drifting fish aggregating devices (dFADs): satellite-tracked rafts that move with currents while accumulating fish below. Here, we estimate the global footprint of these devices and track 30 years of progress to mitigate impacts. We estimate that 1.41 million dFAD buoys were released between 2007 and 2021, drifting across at least 134 million square kilometers, or 37% of Earth’s ocean surface. Lost dFADs have stranded in 104 maritime jurisdictions, contributing to coastal pollution and damaging sensitive habitats. Regulatory progress has been made to address data quality, entanglement, and pollution but concerns over unregulated dFAD deployments, unsustainable bycatch, and weak industry accountability persist. Our results demonstrate that the cumulative environmental footprint of dFADs reaches far beyond tuna fishing grounds and remains inadequately mitigated at the global scale.

Fish aggregating devices have transformed tuna fishing practices, with far-reaching impacts on marine species and coastal nations.

INTRODUCTION

Tuna (family Scombridae, tribe Thunnini) are the most common large predatory fishes in pelagic (open ocean) ecosystems and have supported productive coastal fisheries for centuries. Today, they remain a critical source of nutrition for millions of people and a primary source of income for many Small Island Developing States, some of which have few alternative resources. Industrial-scale exploitation of tuna by major fishing nations from Europe (EU), Asia, and North America commenced after the Second World War and led to a 12-fold increase in catches from 1950 to 2016 (1). Globally, tuna now account for ~5.2 million metric tons (mt) (2), or ~8% of wild marine fish catch destined for human consumption, representing one of few species groups with fisheries that continued to expand after the mid-1990s, when global marine seafood catches began to otherwise stagnate or decline (3).

The 1990s also saw the beginning of a progressive shift in the tuna fishing industry, when fishing practices changed from targeting free-swimming tuna schools to catching fish associated with floating objects (Fig. 1, A and B). This shift was inspired by the observation that natural flotsam, such as logs or other debris, readily attracts juvenile fish seeking protective cover in an otherwise featureless ocean environment (4). This behavior can be amplified and exploited by fisheries using artificial floating objects called fish aggregating devices (FADs). These objects, which are made of natural or synthetic materials, have traditionally been in the form of large rafts with netting extending through the water column to attract pelagic fish. Before the 1990s, most FADs were deployed close to shore and anchored to the seafloor (aFADs), so they could be easily located and fished repeatedly by local communities. Historically, aFADs have been used mostly by small-scale fisheries to target a diversity of pelagic fish, and they remain in use as a means of assuring local food security in low-income countries around the world (5, 6).

Fig. 1. Global footprint of drifting fish aggregating devices (dFADs).

(A) Industrial tuna purse seine vessels commonly use dFADs to attract and aggregate target species, which are then encircled and caught with a large net. PHOTO CREDIT: PAUL HILTON. (B) Global distribution of dFAD fishing activity, shown by reported skipjack tuna catch harvested by dFADs [Regional Fisheries Management Organizations (RFMOs) tasked with overseeing the management of tropical tuna fisheries are outlined]. (C) dFAD in the Eastern Tropical Pacific. Photo Credit: Alex Hofford/Greenpeace. (D) Global distribution of active dFADs derived from published global position system (GPS) track data. (E) Stranded dFAD in the Caribbean region. Photo Credit: Tom Pitchford/Caribbean FAD Tracking Project (F) Global distribution of reported dFAD strandings per Exclusive Economic Zone (EEZ) synthesized from published data and reports. See the Supplementary Materials for data sources and methodology.

The advent of affordable electronic tracking devices in the mid-1990s enabled the proliferation of a novel class of freely drifting FADs (dFADs; Fig. 1C), now commonly used by industrial tuna purse seine vessels (7). Purse seiners catch two-thirds of the world’s tuna by volume (2) and are typically focused on three key species (collectively the “tropical tunas”): skipjack (Katsuwonus pelamis), yellowfin (Thunnus albacares), and bigeye (Thunnus obesus). Of these, skipjack is by far the most abundant and the main source for canned tuna products globally. When outfitted with a global positioning system (GPS) transponder, a dFAD raft can be deployed from a vessel and then located weeks to months later after drifting with prevailing currents and aggregating fish. The addition of high-resolution echosounder satellite buoys in the 2000s added another technology that allows fishers to gauge the approximate biomass of fish under a given dFAD, thereby enabling them to decide when to return to it (8). Today, by using these devices, purse seine companies can monitor fish associated with hundreds of drifting devices in real time over vast ocean areas. In essence, the expansion of this technology removed a defining characteristic of all fisheries: uncertainty in the location and quantity of fish available to catch.

The transition from free-school (or “unassociated”) fishing to dFAD fishing over the past 25 years was mainly driven by a handful of large purse seine fishing companies from EU and the Americas. However, this shift has long been viewed critically by many scientists and fisheries managers due to the impacts of these devices on target stocks and other species (7, 9–13). Ultimately, dFAD drift is impossible to control once a device is in the water and where it ends up is largely contingent on regional currents and prevailing sea surface conditions (14–17), as well as deployment strategies and raft configurations. The latter often are vessel or country specific, as different crews prefer certain materials and designs (18).

In addition to accounting for these factors, effective regulation of dFADs is also challenging because tuna populations (and dependent fisheries) occupy entire ocean basins. Therefore, management decisions related to tuna fishing cannot be taken by an individual country but must be agreed upon by all countries that fish for tuna within a given ocean. The fora where management measures for the world’s tuna are determined are called Regional Fisheries Management Organizations (RFMOs), of which there are five: the International Commission for the Conservation of Atlantic Tuna (ICCAT), the Indian Ocean Tuna Commission (IOTC), the Western and Central Pacific Fisheries Commission (WCPFC), the Inter-American Tropical Tuna Commission (IATTC), and the Commission for the Conservation of Southern Bluefin Tuna (CCSBT). Each year, government representatives from RFMO member states convene to negotiate and adopt measures to ensure the long-term sustainability of the region’s tuna fisheries. Presently, 10 of 13 topical tuna populations under the purview of RFMOs are being fished sustainably (2). However, abundance of bigeye in the Eastern Pacific and the Atlantic is near biological safe limits and ~40% of the total catch in the Indian Ocean currently comes from yellowfin and bigeye stocks that are overexploited and subject to unsustainable fishing pressure (2).

In this paper, we add to the ongoing discussion on the sustainability of dFAD fisheries by taking a global perspective, and synthesizing results from numerous regional studies and datasets. While various regional issues related to dFAD fishing practices, impacts, and governance have been documented elsewhere (7, 13, 19–21), here we are primarily interested in cross-cutting issues and aggregate impacts at the global scale. Specifically, we ask: What is the approximate global footprint of dFADs based on (i) where they are fished, (ii) where they drift, and (iii) where they wash ashore? We further provide a comprehensive overview of how the knowledge, use, and governance of dFADs has evolved since their development with a specific emphasis on identifying (iv) when and to which extent major impacts associated with dFAD purse seining have been raised and addressed. From this overview and our global footprint analysis, we highlight recent progress and persistent challenges for the comprehensive mitigation of dFAD impacts at the global scale.

RESULTS

Estimating the global footprint of dFADs

The spatial footprint (extent of occurrence in square kilometers) of dFAD purse seine catches (Fig. 1A) is concentrated in major tuna fishing grounds off the Gulf of Guinea (West Africa), the western Indian Ocean, and the equatorial Pacific Ocean (Fig. 1B). When summing across all 5° × 5° cells with reported dFAD catch from 2010 to 2020, the global area affected directly by dFAD purse seine fisheries totals 117.8 million km² and intersects with the national waters [exclusive economic zones (EEZs); hereafter: “maritime jurisdictions”] of 118 coastal or island nations and territories.

However, many dFADs drift (Fig. 1C) outside of fishing grounds, where they are later lost or abandoned. dFAD track data are proprietary, and no public repository exists, but the distribution of buoy tracks from active dFADs has been summarized and published for some fleets and years and is shown as a global composite here (see Materials and Methods for data sources): Figure 1D displays the presence of tracks recorded from 88,359 dFAD buoys binned by a 1° × 1° cell, comprising a global footprint of 133.9 million km² between 2007 and 2023. This footprint intersects with 157 maritime jurisdictions.

Yet, this calculation only includes actively transmitting dFAD buoys and does not account for devices that keep drifting after their GPS has stopped transmitting. Many lost or abandoned dFADs wash ashore on the reefs or beaches of coastal nations around the world (Fig. 1E). When considering observed stranding events and those inferred from active dFAD tracks, we find that 104 maritime jurisdictions have been subject to at least one dFAD stranding event (Fig. 1F) with a total of 14,782 strandings recorded across all jurisdictions. Almost all maritime jurisdictions within the world’s main purse seine fishing grounds (Fig. 1B) have been affected by stranded dFADs, but we find the highest recorded numbers in the Seychelles, Somalia (both Indian Ocean), and French Polynesia (central Pacific), which collectively account for 43% of all known strandings (table S1). Notably, while 31 maritime jurisdictions have documented >100 dFAD strandings, only 14 of these places permit the use of dFADs in their waters. When considering the coastlines of jurisdictions with reported stranding events, we find a median density of ~1 stranded dFAD/22 km, but with large variation (table S2). Seven jurisdictions had recorded densities exceeding one stranded dFAD/km of coastline: Bassas da India (France), Tuvalu, Seychelles, Palmyra Atoll (USA), Cook Islands, Wallis and Futuna, and the Maldives.

Trends in tuna catch and dFAD deployments

Since the 1970s, the total global catch of tropical tunas has been rising steadily from 1 to ~4.5 million mt per year (Fig. 2A). Of this, dFAD purse seine catch currently contributes ~1.5 million mt annually, or 32.7% of all skipjack, yellowfin, and bigeye tuna caught from 2016 to 2020. While the total volume of tuna caught by all other fishing gears has risen modestly (22% from 1990 to 2020), catches from dFAD purse seining increased 645% at the same time (Fig. 2A), and hence contributed most of the observed increase in total tuna catch since 1990.

Fig. 2. Trends in tuna catch and drifting fish aggregating device (dFAD) deployments.

(A) Increase in global tropical tuna catch (skipjack, yellowfin, and bigeye) derived from floating objects (red) relative to all other fishing methods combined (gray). (B) Reported and (C) estimated number of dFAD deployments per year globally. (D to F) Reported proportions of mature (white) and immature (black) tuna in dFAD catch by major target species (D = skipjack, E = yellowfin, and F = bigeye) and ocean. Values above bars indicate the length at maturity (L₅₀) in each region; we considered all individuals under this length as immature. See the Supplementary Materials for data sources and methodology.

This growth was driven by a large increase in the number of dFADs deployed over time. Owing to a lack of data transparency, it has been notoriously difficult to estimate the total number of dFADs released into the ocean each year, with only one study attempting a global tally of 81,000 to 121,000 dFADs for 2013 (22). Here, we provide a global estimated timeline of dFAD buoy deployments from 2007 to 2021 using official RFMO reports (Fig. 2B and table S3) as well as scaled estimates that account for missing data (Fig. 2C and table S4). When scaling RFMO reported dFAD deployments, we estimate that a total of ~1.41 million dFAD buoys were released into the global ocean from 2007 to 2021.

One of the first concerns raised by scientists when dFAD use was rapidly increasing in the 1990s was their high selectivity for juvenile tuna, which was anticipated to contribute to growth overfishing and economic losses (23), especially for bigeye (9). When investigating the size composition of recent and historical dFAD catches, we find that reported catches of the main target species—skipjack—includes 3.7 to 59.6% immature individuals across the four fishing areas (Fig. 2D and table S5), while the proportion of immature yellowfin and bigeye exceeded 90% for seven of eight stocks (Fig. 2, E and F). We note exceptionally high dFAD capture of immature yellowfin in the Western Pacific and Indian oceans (>95.8 and >95.2%, respectively, Fig. 2E) and that >98% of bigeye caught by dFADs are below their size at maturity in three of four ocean basins (Fig. 2F).

Trends in FAD research, advocacy, and governance

A number of ecological and legal concerns around dFAD use have been documented in the scientific literature for some time, dating back to the first wave of expansion in the 1990s (Fig. 3A). Since then, there has been more attention every year and many issues have been studied and documented in detail. We conducted a comprehensive literature search extracting all peer-reviewed publications since 1970 that mentioned FADs and dFADs over time, identifying main topic areas of concerns. These can be broadly categorized according to three stages of their life at sea and how such activity is regulated (Table 1):

Fig. 3. Trends in dFAD science, advocacy, and governance.

(A) Increase in primary scientific publications related to dFADs, their use, and environmental impact. (B) Conservation and Management Measures (CMMs) adopted by RFMOs that mention dFADs. (C) Proportion of annual letters submitted by nongovernmental observer organizations in RFMO meetings that raise concerns about the use of dFADs. (D) Recent increase in tuna fisheries eco-certified under the Marine Stewardship Council (MSC) sustainability eco-label, showing dFAD purse seine fisheries (dark blue), free-school purse seine fisheries (medium blue), and other tuna fishing gears (light blue). (E) A timeline of key events during the development of dFAD fisheries in the Atlantic (AO), Indian (IO), Western and Central Pacific (WCPO), and Eastern Pacific Ocean (EO); events pertaining to a specific RFMO are colored as per the legend in (C) while events that relate to dFAD fishing advancements or improvements by industry are marked in gray. See the Supplementary Materials for data sources and methodology.

Table 1. Major conservation and management issues.

Thirteen major issues are associated with the deployment and drift of dFADs, their retrieval and harvest, their abandonment or loss at sea, and their regulation and oversight, respectively. The evidence for each issue reflects the degree to which existing literature supports the issue and highlighted effect. The primary solution reflects main policy recommendations in scientific, RFMO, and nongovernmental organization (NGO) documents. Region abbreviations and associated RFMOs: AO, Atlantic Ocean (ICCAT); IO, Indian Ocean (IOTC); WPO, Western and Central Pacific Ocean (WCPFC); EPO, Eastern Pacific Ocean (IATTC). Implementation status is current through December 2024; for measures adopted but not yet implemented, the year of implementation is also provided.

dFAD stage	Issue	Effect	Evidence	Primary solution	Implementation status				Implemented solutions
					AO	IO	WPO	EPO
Deployment and drifting	Artificial ecosystem structure	Unknown	Scarce	Science-based dFAD deployment limits	Partial	Partial	Partial	Partial	All RFMOs have limited the number of dFADs that can be active at any time (i.e., buoys transmitting) but no limits on deployments per year. Current active dFAD limits range from 340 (EPO) to 500 (AO) per vessel; however, these numbers are not science based and not reflective of actual dFAD usage (99). By 2026, IO vessels may only have 250 active dFADs each (but can acquire up of 400 dFADs per year) and AO vessels will have an active limit of 288 dFADs.
	Altered fish behavior	Changes in growth and survival	Medium	Science-based dFAD deployment limits	Partial	Partial	Partial	Partial
	Entanglement of marine life	Mortality of nontarget species	Strong	100% nonentangling dFADs*	Partial	Full	Full	Full (2025)	Fully nonentangling dFADs are mandated in the IO, WPO, and EPO, but only encouraged in AO.
Retrieval and harvest	High catch of immature tuna	Overfishing and economic losses	Strong	Seasonal fishery closures	Full	None	Full	Full	AO, WPO, and EPO have annual spatiotemporal closures for dFAD fishing ranging from 45 days to 12 weeks. In the EPO, this applies to all purse seine vessels across the RFMO area; additional prohibitions on fishing in this area are in place for vessels with high catches (>1200 mt) of bigeye. In the WPO and AO, the closure applies to dFAD fishing only and covers primary purse seining grounds.
	Bycatch of threatened species	Potential biodiversity loss	Strong	Safe release and science-based catch limits	Partial	Partial	Partial	Partial	In all regions, fleets are required to safely release incidentally caught wildlife but there are no catch limits for commonly caught bycatch species. There are retention prohibitions for certain shark species in all four regions but only the WCPO, EPO, and AO currently prohibit the retention of silky sharks, the species most frequently encountered in dFAD sets.
	Higher catch per unit effort†	Potential bias in stock assessments	Strong	Data provision and sharing	Partial	Full	Partial‡	Full	In all regions, fleets have been required to provide increasingly detailed information related to dFAD activity (e.g., deployments, visits, and losses) on a regular basis to the associated RFMO. In the IO and EPO, fleets must report the daily position of their buoys postdeployment (with a 30-day delay). In the WPO, vessels are required to report buoy locations once every 12 hours to Parties to the Nauru Agreement (PNA) countries‡ but this is not a requirement of WCPFC.
	Higher fuel use†	Increased carbon footprint	Medium	Unknown	None	None	None	None	None
Abandonment and loss	Ghost fishing	Economic and biodiversity loss	Medium	100% nonentangling dFADs	Partial	Full	Full	Full (2025)	See “Deployment” above
	Marine litter	Plastic pollution	Strong	100% biodegradable dFADs	Partial	Full (2030)	Partial	Full (2030)	In the IO and EPO, fleets are prohibited from using dFADs made entirely of nonbiodegradable materials. A process for transitioning to 100% biodegradable dFADs by 2030 has been adopted. In the AO and WPO, member states are currently encouraged (but not mandated) to use biodegradable materials with discussion of a potential phase-out in the coming years (but no timeline decided).
	Stranding of dFADs	Coastal habitat impacts	Strong	Retrieval of all dFADs at sea	None	Partial	None	Partial	No measures exist mandating that all deployed dFADs are collected at sea or that prohibit deployments in areas with high stranding likelihood. In the EPO, fleets are required to remove active dFADs from the water 2 weeks before the seasonal closure. In the IO, if a dFAD is within the national waters of a coastal nation territory when it is deactivated, a notification is sent to relevant authorities, but no protocols exist for removal.
Regulation and oversight	Unregulated drift	Aggregating fish from unauthorized areas	Scarce	Targeted spatial management	None	None	None	None	Seasonal fishery closures exist for purse seine fleets using dFADs in AO, WPO, and EPO; however, these measures were put in place to curtail total fishing effort and juvenile catch. There are no measures designed to reduce the likelihood of drift of dFADs through protected areas or regions where fishing access agreements may not exist. To determine such areas, full transparency of buoy drift data is needed.
	Unregulated deployment	Weakens oversight of fishing effort	Strong	100% observer coverage for all activation and deployment	Partial	None	Partial	Full	In the AO, EPO, and WPO, 100% monitoring of purse seine vessels at sea is required but only 5% coverage applies in the IO. In all cases, coverage does not extend to supply vessels, which can also deploy or retrieve dFADs. In the EPO, the use of supply vessels in conjunction with dFAD fishing has been prohibited since 1999.
	Opaque ownership	No accountability for unwanted dFAD impacts	Strong	Ownership rules, registry, and standardized labeling	None	Partial	None‡	Partial	Starting in 2026, all dFAD buoys and rafts deployed in the IO must have a unique identifier and be registered with the IOTC. However, these data will not be publicly available. Unique identifiers are also required for all dFADs in the EPO and, in 2022, the PNA‡ set out ownership rules and responsibilities (but no such measures exist through WCPFC). While the recovery of dFADs is encouraged, a lack of clear ownership rules and obligations for retrieval persists in all regions.

*While dFAD raft and tail components can meet this requirement, buoys attached to dFADs will likely always have electric and plastic components and therefore a source of marine pollution.

^†Fuel use per ton of catch compared to free school purse seine fishing.

^‡Measure required by the Parties to the Nauru Agreement (PNA; nine island nations covering ~30% of the WCPFC Convention area) under their Vessel Day Scheme (VDS) access arrangement. The original PNA are as follows: Federated States of Micronesia, Kiribati, Marshall Islands, Nauru, Palau, Papua New Guinea, Solomon Islands, and Tuvalu; Tokelau joined in 2012.

1) dFAD deployment and drift alters the pelagic ecosystem by introducing artificial substrates that are then colonized by a variety of species (24). The altered behavior of tuna and other species around dFADs has been repeatedly documented (25–28) and the consequences of these associations have been conceptualized as the “ecological trap” hypothesis (29), whereby juvenile fish follow a dFAD that may lead them to live in suboptimal or unproductive habitats that differ from those chosen by free-swimming schools, resulting in reduced fitness. There is evidence that tuna may be in poorer condition when associated with dFADs as opposed to those traveling in free schools (10, 30) or even those associated with aFADs (31), but causal mechanisms and population-level impacts are not yet resolved (32). Another well-documented issue related to dFAD drift is entanglement and mortality of nontarget species in dFAD netting. Traditional dFAD design relies on the use of submerged appendages (“tails”), which are made of nets, ropes, pipes, and other materials, commonly reaching depths of 40 m or more (18). These tails serve a structural role for increasing drag through the water column to slow drift speed while simultaneously serving as underwater attractants for fish (18, 33), which might become entangled, especially if netting materials are used.

2) dFAD retrieval and harvest occurs after weeks or months of drifting at sea and when echosounder data suggest that a sufficient volume of fish have accumulated under the raft. Many of these fish are immature (see Fig. 2, D to F), and juvenile bycatch (leading to growth overfishing) remains a key concern. Similarly, the aggregation of threatened nontarget species, especially pelagic sharks, has been a concern, as this leads to substantially higher bycatch rates compared to free-school tuna sets. For example, it is estimated that dFAD fleets have been catching at least 100,000 silky sharks per year (mostly juveniles) in the Indian Ocean alone (34, 35). The power of dFADs in aggregating a large variety of species (36) and continuing improvements in technology have also led to increasing catch per unit of effort relative to other methods of fishing (37). This trend, along with difficulties in quantifying nominal effort for dFAD fishing, have complicated the use of dFAD purse seine data in annual tuna stock assessments. Last, while it has been asserted that dFAD fishing has increased efficiency due to shorter search time, several studies found ~15 to 50% increased fuel use per tonnage of catch and hence a larger carbon footprint of dFAD fishing when compared with free-school purse seines (38–40).

3) As dFAD movements cannot be controlled, they often drift outside of fishing areas, where they are lost or abandoned once their transmitting function runs out. These devices then contribute to marine pollution as well as “ghost fishing” by continuing to aggregate and entangle marine life, such as sharks, turtles, and seabirds, particularly if extensive netting is attached to the raft (41). This problem is applicable to aFADs as well, in cases where they break free from their moorings (42, 43). When FADs disintegrate, they contribute to marine litter, especially where synthetic mesh and other nonbiodegradable materials are used (42). Last, the stranding of lost dFADs can cause further damage to coastal biodiversity, especially where sensitive habitats are affected (44).

4) Given the direct impacts incurred by using these devices, researchers have highlighted substantial gaps in how dFAD use is overseen and regulated, and how responsibility is assigned to companies for any damages that occur (12, 13). While the number of “active” dFADs (i.e., those transmitting) has recently been capped (see below), there are no hard limits on the number of dFADs that can be deployed by a vessel or company each year. This creates a problem in regulating total dFAD effort and assessing the dFAD fleet’s collective footprint in space and time. Concerns over the legality of dFAD abandonment (45) and unregulated drift (12) have also been raised. Tracing ownership of dFADs that wash ashore has been additionally problematic due to nonstandardized labeling schemes [and poor industry compliance with existing ones [e.g. (46, 47)], lack of dFAD ownership registries and notification protocols for losses, and inconsistencies in dFAD data reporting across regions (48).

Tracking increasing regulatory efforts over time, Fig. 3B shows the number of conservation and management measures (CMMs) adopted through the world’s tuna RFMOs that relate to dFADs specifically (table S6). We find a 17-fold increase in the number of dFAD measures adopted between 1998 and 2023. Note, however, that a single CMM may include one or more provisions related to dFAD fishing and that no measures could have been adopted at WCPFC before 2005, or at IOTC before 1996, because these RFMOs were not yet established. Across all RFMOs, there has been a notable increase in the diversity of dFAD management requirements over time (fig. S1). Notably, CMMs mentioning environmentally conscious dFAD design (i.e., biodegradable and nonentangling materials) have increased substantially since 2014, and the majority of CMM provisions relate to this topic (19%), followed by data provision requirements (12%). Further, 18% of CMM dFAD content related to ongoing or improved research on their use and impacts, much of which was to be undertaken by RFMO science providers to better inform decision-making. In recent years, RFMO FAD Working Groups, which are composed of scientists and fishery stakeholders, have also been tasked with making recommendations related to gaps in dFAD knowledge and use. Overall, their recommendations similarly focused on the need for research to better understand dFAD impacts (23%), ensuring that the necessary data for such analyses were provided (22%) and, relatedly, a need for standardizing dFAD definitions (13%; fig. S2). Table 1 provides an overview of how well each of the major issues associated with dFAD use and accountability have been addressed through existing RFMO measures, with noticeable variation by issue and region. Overall, implementation gaps related to dFAD retrieval, mitigation of unregulated drift, and owner accountability for dFAD use and loss persist in all oceans.

Over time, these and other regulatory gaps in dFAD management have been highlighted by nongovernmental organizations (NGOs), which can play an important role in influencing tuna fisheries governance through their advocacy (49). Their concerns and suggestions for improvement are formally submitted in letter form before each annual RFMO meeting and reflect their organization’s view on the key concerns facing the world’s tuna fisheries. Letters that specifically mention dFADs and advocate relevant mitigation measures only started to be submitted after 2005 (table S7), but the proportion of letters mentioning these devices increased from 2010 to 2016 (Fig. 3C), indicating growing environmental concerns as their use became ubiquitous and effects on the environment were documented. Similar to adopted CMMs, we find an increasing trend, where the topic of dFADs was mentioned in only 17.4% of RFMO observer letters between 1999 and 2011 compared to 56.3% of letters since 2012.

The impact of dFADs on juvenile tuna was the most common concern (64%) mentioned in “early” NGO letters (2007 to 2014), followed by concerns over bycatch and general threats to the marine ecosystem (50%). While some early letters called for a complete prohibition on dFAD use, the most common solutions advocated were improved data collection to better understand and monitor dFAD impacts (64%) and, similarly, science-based dFAD management to limit their use (29%). Notably, while 41% of 2023 letters still emphasized a need for improved data collection, this was now largely in relation to improving industry transparency and, by extension, stricter oversight and monitoring of all dFAD activities. Relatedly, we find that 2023 letters had increased advocacy toward ownership rules and standardized labeling (41%) as well as dFAD recovery policies (37%)—issues that were almost entirely absent from earlier letters and which remain largely unaddressed by all RFMOs (Table 1). Likewise, transitioning to fully biodegradable and nonentangling dFADs was mentioned in 59 and 56% of 2023 letters, respectively. Only once was biodegradable design mentioned before 2015, likely reflecting increased awareness of how these devices contribute to plastic pollution in more recent years (19, 42).

One independent metric of a fishery’s environmental performance is whether it has been eco-certified through a sustainable seafood program. The most common of these is the Marine Stewardship Council (MSC) and 19% of global wild fish catch at present is certified “sustainable” based on their criteria (50). An albacore (Thunnus alalunga) troll association was the first tuna fishing client group to obtain MSC certification in 2007, but it was not until 2011 that free-school purse seine fishing companies began to meet the MSC standard (Fig. 3D). In 2013, one company attempted (but failed) to obtain certification for its dFAD operation and only in 2018 did the same company achieve a passing score. Since there has been a rapid increase with 20 tuna fishing client groups (collectively including more than 90 companies) that use dFADs obtaining MSC certification by mid-2024 (Fig. 3D). This recent trajectory is attributable to fishing companies seeking certification for dFAD use on their first assessment, as well as companies that were originally certified for free-school purse seining and later sought certification for their dFAD operations as well (table S8).

In the context of sustainable fishing and market pressure, it is worth noting that the increased use of dFADs is owed partly to previous pressure from environmental groups and a concerned public over high dolphin bycatch in Eastern Pacific purse seine fisheries before the 1990s. In response, the US government worked with industry “to develop methods to build and deploy relatively inexpensive drifting and anchored dFADs that will attract mature tuna in sufficient abundance to supplement current harvest levels and to decrease fishing activity associated with dolphins” [p. 2 in (51)]. This transition is noted in Fig. 3E, where we provide an illustrative timeline of key events in dFAD knowledge acquisition and scientific advice, as well as at-sea efforts to minimize their impacts since this time (see table S9 for details).

Sustainability shortcomings in certified dFAD fisheries

In addition to observing the number of dFAD fishing companies that are MSC-certified, a more comprehensive means of assessing the impact of dFAD fishing is to investigate the open conditions given to companies during the assessment process (Fig. 4, A and B). These conditions do not prevent a company from obtaining MSC certification but are given when a specific aspect of the fishery does not sufficiently meet the scoring criteria provided by the MSC. Typically, a fishery has until the time of their reassessment (5 years) to address these conditions.

Fig. 4. Sustainability shortcomings in eco-certified dFAD fisheries.

(A) Number of open conditions (unresolved issues to be addressed by the fishing client group) of MSC-certified dFAD fisheries versus free-school purse seine, other tuna, and all other fisheries. (B) Proportion of MSC-certified tuna dFAD fisheries that have open conditions (dark bars) with respect to tuna stock status (MSC Criterion 1), ecosystem impact (Criterion 2), and management (Criterion 3). MSC, Marine Stewardship Council; ETP, endangered, threatened, or protected. See the Supplementary Materials for data sources and methodology.

We find that MSC-certified dFAD fisheries had an average of 15.0 ± 6.5 open conditions during their first assessment cycle (Fig. 4A and table S10), which is higher than free-school purse seine fisheries (mean = 10.7 ± 9.9), tuna fisheries using other gears (mean = 7.6 ± 4.6), and all other MSC-certified fisheries (mean = 10.3 ± 25.4). We find a statistically significant difference between the three tuna groups (Kruskal-Wallis test, H_2,56 = 11.41, P = 0.003), and this trend remains consistent even when accounting for the uneven number of target species and stocks included in each assessment (fig. S3). When looking through the lens of MSC assessment criteria at aspects of dFAD fisheries in most need of improvement, we find all dFAD fisheries had open conditions under sustainability Criterion 2, which relates to their impact on endangered, threatened, or protected (ETP) species as well as their impacts on habitat (Fig. 4B). Almost all certified dFAD fisheries also had outstanding concerns related to tuna harvest strategies and control rules (Criterion 1: Stock health) and more than half had conditions related to decision-making processes, as well as compliance and enforcement (Criterion 3: Management effectiveness). Regarding industry transparency, we note that only 3 of the 24 assessments (12.5%) for MSC-certified dFAD fisheries included sufficient information to determine both the volume of fish the company catches using dFADs and how many devices they deploy annually (table S11).

DISCUSSION

A global footprint

In this study, we sought to better understand the expanding global footprint of dFADs, and how impacts are currently being mitigated through RFMO measures and industry efforts. We show that the freely drifting nature of dFADs extends their reach to a substantially larger area than where fishing takes place (see Fig. 1, B, D, and F) and we estimate a global footprint of 134 million km² for documented dFAD drifts (Fig. 1D), transiting through 157 maritime nations and territories, 104 of which have documented strandings. For context, this footprint equals 37% of global ocean surface, or the combined area of all continents, excluding Antarctica. We further estimate that ~1.41 million dFAD buoys were released into the ocean between 2007 and 2021 (Fig. 2C) and that harvested dFADs incur >90% capture of immature yellowfin and bigeye tuna in all ocean basins (Fig. 2, E and F).

The large increase in dFAD deployments observed between 2007 and 2014 (Fig. 2C) coincides with progressive dFAD technology advancements enabling increased catches (Fig. 2A). This is compounded by steady improvements in fishing efficiency [e.g., 8% per year increase in the Indian Ocean (52)], which may further mask tuna population declines (53). When we trace the 30-year timeline of attempts to understand and regulate the use of dFADs (Fig. 3, A and B), we find that of 13 key conservation and management issues identified by scientists and other stakeholders, only 5 are comprehensibly addressed by any RFMO, and none is addressed comprehensively by all RFMOs (Table 1). Long-standing knowledge and implementation gaps are equally indicated by NGO letters concerning dFAD management at RFMO meetings (Fig. 3C) and by the number of open conditions for eco-certified dFAD fisheries (Fig. 4A), with notable shortcomings in the context of habitat and threatened species interactions, as well as compliance and enforcement (Fig. 4B).

Notably, given their large numbers and ability to freely disperse across vast ocean areas (Fig. 1, B and D), the deployment of dFAD rafts and buoys remains poorly regulated. While existing RFMO measures may limit the number of active dFADs per vessel (Table 1), such limits do not necessary restrict the overall number of dFADs deployed, nor do they account for the cumulative fleet-level impact caused by each vessel’s individual losses. This is because fishers can deactivate a dFAD buoy remotely when they want to deploy another one, while avoiding GPS communication costs associated with devices that have drifted away from accessible fishing grounds (54). Such intentional dFAD abandonment at sea is not only negligent but also violates international maritime law (55). Moreover, many dFADs are also accidentally deactivated when damaged at sea or when stranded. Recent estimates suggest that deactivated buoys may survive for an additional 6 months at sea (56) and, for a single company, losses can amount to thousands of dFADs annually (57). Further, while we estimate a steady increase in buoy deployments since 2005 (Fig. 2C), the number of devices retrieved is declining. In 2021, less than 25% of ~27,000 dFADs deployed in the Eastern Pacific were collected (58) and a comprehensive dFAD tracking program in the Western Pacific estimates a loss rate of ~90% (59). This suggests that companies deploying these devices assume that the vast majority will never be retrieved and, when such annual losses are considered in aggregate, the data suggest a substantial surplus of abandoned dFADs over the last two decades. Notably, a measure to prohibit the abandonment of inactive dFADs was proposed for the Indian Ocean in 2013 (60), but was not adopted. While some retrieval requirements exist in the Pacific around seasonal fishing closures, there are currently no RFMO measures requiring vessels to retrieve all the devices they deploy (Table 1).

Impacts on marine life

It is undeniable that dFADs have rapidly transformed tuna fishing over the last decades. By exploiting the tendency of juvenile pelagic fish to aggregate under floating objects, these devices have substantially widened the size spectrum of purse seine tuna fisheries to include previously lightly fished juveniles and subadults (Fig. 2, D to F). In general, prolonged capture of immature fish in the absence of science-based management measures is unsustainable (61) and concerns over high catches of juvenile bigeye and yellowfin by dFAD fisheries have been raised by scientists (9, 62, 63), NGOs (64), members of the fishing industry (65), and governments for decades. At RFMO meetings, member country governments agreed on the urgent need for understanding and addressing such impacts early on; for example, in 1995, the Spanish delegation at ICCAT (Atlantic) highlighted the importance of obtaining “a reliable scientific base for an eventual regulation of the use of these artificial objects and, if necessary, their prohibition over the medium term” [p. 148 in (66)] while the Brazilian delegation proposed the use of dFADs be “abolished” by 1999 [p. 151 in (67)]. Similarly, in 1998, IATTC (E. Pacific) delegates agreed on the importance of limiting the number of dFADs a vessel could carry (68), but such a limit was never adopted, and only in 2017 did they agree to cap the number of active dFADs permitted for a given vessel (69) (Table 1). Considering the data in Fig. 2, D to F, however, it appears the problem of juvenile mortality remains pertinent in all oceans.

The primary literature (Fig. 3A), NGO advocacy statements (Fig. 3B), and open conditions in MSC-certified dFAD fisheries (Fig. 4B) also provide context for the effects of dFADs on the pelagic ecosystem beyond tuna. Drifting FADs indiscriminately aggregate many pelagic animals, leading to higher bycatch of nontarget species compared to free-school purse seines (36), contributing toward rising extinction threats for pelagic sharks, particularly (34, 70, 71). Beyond bycatch, dFADs, when deployed by the tens of thousands, may fundamentally alter the structure of pelagic open-ocean habitats, and the behavior of many of its inhabitants. The rapid shift toward dFAD fishing during the 2000s may have self-reinforced by progressively associating tuna with floating objects early on in their life cycle, potentially changing their feeding, migration, and other behaviors, such as their tendency to form free schools (10, 29, 32). Given the exceptionally large spatial footprint of dFAD use and loss shown here (Fig. 1), and the large numbers of buoys deployed over time (Fig. 2C), it is conceivable that such ecosystem-wide impacts are already manifest across the world’s tropical oceans, with little action to halt or reverse this trend.

Impacts on coastal nations

We find that active dFADs have drifted through the sovereign waters of 157 countries and territories (Fig 1B). In many of these places, dFAD fishing is permitted by domestic vessels or foreign companies holding an access agreement. In other areas, dFAD fishing is not permitted. However, a legal loophole exists such that devices may not be considered “fishing” while drifting (48) and a vessel can therefore collect any fish that may have aggregated once the dFAD leaves the prohibited area (12, 13). The degree to which this practice occurs intentionally or opportunistically is unclear but given the large volume of fish caught by dFADs, this could contribute substantially to illegal, unreported, and unregulated (IUU) fishing (13). In the Indian Ocean, for example, “illegal” dFAD fishing has been highlighted as an environmental justice issue, with estimates of more than 1400 devices from foreign fleets drifting unauthorized through Somalian waters each year, posing a risk to the sustainability of local fisheries and associated livelihoods (47). Additional spatial management challenges for dFAD fisheries have emerged since 2010 from the ongoing expansion of marine protected areas (MPAs) and other effective conservation measures, established in response to the Convention on Biological Diversity targets (72). Many of these areas are designed to be closed to industrial fishing and/or other extractive industries yet we observe that such areas are among 5 of the top 20 maritime jurisdictions with the highest dFAD strandings relative to the length of their coastline (table S2). Concerns of dFADs entering protected areas or infringing on national sovereignty might in part be addressable though the spatial management of dFAD deployment given the increasing body of research on dFAD movement patterns (16, 73, 74) and by incorporating the extensive knowledge held by fishers regarding dFAD behavior at sea (21).

Our data further show that 104 maritime jurisdictions have reported at least one dFAD stranding event, with certain dFAD hot spots reporting thousands (Fig. 1F and table S1). Variability observed may reflect both real differences in stranding frequencies and differences in reporting, as many regions have only recently started collecting these data and no centralized database of ocean-scale strandings exists, except for the Western Pacific (16). Given the paucity of dedicated dFAD stranding data collection programs and because the dFAD track data we used reflect only a small portion of all dFAD buoy deployments, our global total of 14,787 strandings undoubtedly represents a substantial underestimate given the high loss rates of dFADs in different oceans (74). While we find that most strandings are concentrated in purse seine fishing grounds, dFADs have also been reported in regions far removed (Fig. 1F). Current work to mitigate dFAD strandings in regional hot spots includes the implementation of multisectoral programs that provide the real-time location of buoys that could pose a stranding risk (Fig. 3E). By alerting local government officials when an active dFAD drifts close to shore, the industry-led FADWatch initiative prevented 109 stranding events in the Seychelles in 2016 to 2017 (75), and a similar program has since been started in Palmyra Atoll, a remote Pacific MPA susceptible to dFAD strandings (76). As discussed above in the context of unregulated drift, defined spatial management measures could also prohibit dFAD deployment in areas known to lead to frequent stranding events given prominent current patterns because it is likely that not all countries will have the local capacity to implement this type of program. Prohibiting deployments in areas with low fishing activity in the Atlantic and Indian oceans could reduce regional strandings by 40 and 20%, respectively (17).

Solutions to date

To date, seasonal fishery closures have been the main approach for RFMO member states to address overfishing of juveniles and limit total fishing pressure. Notably, members of industry were among the first proponents of spatiotemporal restrictions and the first Atlantic dFAD closure was voluntarily introduced by French and Spanish companies in 1997 (Fig. 3E), before it was adopted at ICCAT 2 years later (77). In general, seasonal closure effectiveness is contingent on whether closed areas are large enough, closure periods are long enough, measures are taken to limit redistribution of fishing effort, and industry compliance is high (78–80). Well-designed closures should therefore be viewed as a key complement to other RFMO regulations. This type of spatial management may offer particular benefit in the Indian Ocean given the high proportion of juvenile bigeye and yellowfin tuna caught by dFAD fisheries in 2022 (Fig. 2, E and F) and the recent regional over-catch of these species relative to sustainable fishing limits and established quotas (81). This is currently the only region lacking a seasonal closure (Table 1) despite RFMO scientists indicating that a 3-month closure would incur “the most positive impact” on the status of all three tropical tunas over the next decade (82).

Most progress in addressing the potential impact of lost dFADs has been made by changing their design rather than limiting the number of dFADs at sea (Table 1). These include recent moves toward nonentangling and biodegradable rafts, aimed at reducing bycatch, ghost fishing, and plastic pollution impacts (83). Although the testing of nonentangling dFADs began two decades ago (Fig. 3E), this issue became a priority for many NGOs more recently, perhaps due to growing concerns about plastic pollution and ghost fishing. The International Seafood Sustainability Foundation, an industry-focused NGO, was one of the first organizations engaged in dFAD advocacy at RFMO meetings (table S7) and has been actively involved in facilitating improvements to dFAD design and bycatch mitigation (21). In general, fisher attitudes toward lower-impact dFAD designs and best practices for bycatch have improved in recent years (21) and certain fishing companies have also invested substantially into testing and using dFADs with a lower environmental impact (84–86). While this is one area where industry practices and RFMO regulations are rapidly advancing (Table 1 and fig. S1), it is not clear when all dFADs will be required to be fully nonentangling and 100% biodegradable.

Despite these improvements, it is concerning that known problems and poorly quantified risks are not treated with the necessary precaution. This is evident in the context of dFAD eco-certification. MSC certification can be a major driver in the adoption of fisheries management measures more broadly (87). Yet, as more dFAD fisheries are certified with numerous open conditions (Figs. 3D and 4), it remains to be seen whether this will incentivize stronger dFAD oversight, data transparency, and accountability soon. We agree with the assertion by Moreno et al. (88) that “MSC must assess the fishery as a whole as this would force industry to address issues with dFADs.” To this end, we note that when considering potential dFAD ecosystem impacts (MSC Criterion 2), it is concerning that assessments only consider effects at the scale of the company (or client group), rather than the whole dFAD fishery. Thereby, the collective environmental impact of all vessels using these devices in each ocean basin is not considered, although impacts could be substantial. For example, a comprehensive modeling study indicated a 23 to 58% overlap between dFAD deployment trajectories and endangered sea turtle habitat in the Pacific Ocean (89). Further, although Banks and Zaharia (90) suggest that more than 20% of deployed dFADs may wash ashore in the Western Pacific (~282,000 devices since 2007 when extrapolating to global deployments as per Fig. 2C) and ~67% may sink (~945,000 devices since 2007 when extrapolating globally), neither coastal nor seafloor habitats are considered “primary” in MSC dFAD fishery assessments and are hence not thoroughly evaluated. Thus, interactions of sunken dFADs with fragile deep-sea environments such as seamounts have not been quantified. Last, MSC dFAD assessments also fail to consider ecological impacts caused by cases of disproportionately high strandings (Fig. 1F). Over time, these may be especially detrimental to biodiversity hot spots or MPAs (table S2), ecosystems with high local endemism (e.g., the Galápagos Islands; table S1), or those heavily compromised by climate change, as is the case for many coral reefs in the Indian Ocean (91).

Closing regulatory gaps

The development of strong regulatory measures has thus far been challenging, in part because reporting on the use of dFADs is not always straightforward for fishing companies and terminology varies between jurisdictions (92). This challenge is exemplified by the recommendations of RFMO FAD Working Groups: Directives to improve and/or standardize dFAD fishing definitions accounted for 12% of all recommendations while calls for improved data provision and associated research to inform management decisions each accounted for 23% (fig. S2). At the same time, efforts to ascribe, track, and evaluate the full impacts of dFADs are hampered by delays in establishing clear labeling, ownership, and responsibility rules. Currently, labeling requirements are not standardized across RFMOs and dFAD debris often lacks any form of identification (46, 47). Thus, it is often impossible to trace a device that damages coastal habitats back to its owners. Some regional progress has been made, however. In 2022, the first dFAD registry and ownership code of conduct was implemented by the Parties to the Nauru Agreement (93), a group of Small Island Developing States in the Western Pacific whose waters produce nearly half of the global tuna catch. In 2026, the IOTC (Indian Ocean) will be the first RFMO to implement a dFAD registry for all devices used in its waters, with both the buoy and the dFAD raft bearing unique identifiers that can be traced back to the deploying vessel. It is unclear if and when strong ownership and labeling rules will be established through the other RFMOs, but they appear critical for ensuring industry accountability in a fishery that loses between 30 and 90% of all devices deployed.

Given that dFAD fishing was allowed to grow unchecked for so long despite the initial concerns and the complexity of attempting to harmonize dFAD management across regions, we realize the difficulties now facing RFMO member states charged with this task. As was the case for determining annual tuna catch limits, the management of dFADs can be contentious and highly political, which often distracts from scientific advice (94). In 2023, for example, the EU and 10 other IOTC members submitted formal objections to a comprehensive dFAD management measure that had been adopted (Fig. 3E and table S9). When a country objects to an RFMO measure, they are no longer bound to implement the associated provisions. Objections have historically been rare at RFMO meetings because they undermine the democratic decision-making framework of these fora, and this was the first instance of an adopted RFMO management measure being nullified as a result (95). In the Indian Ocean, concerns have been raised that restrictions to dFAD use, especially seasonal closures, would contribute to economic instability for Small Island Developing States, some of which have strong ties to foreign industry (96), and are reliant on the predictability that dFAD fishing provides (97).

While acknowledging these socio-political challenges, we urge the scientific and RFMO community to consider the cumulative impacts of dFAD fishing as a whole and develop an integrated and stepwise management approach that accounts for all dFAD life stages (48). The global harmonization of dFAD fishing terminology reporting requirements is long overdue, as is the implementation of comparable management measures across RFMOs. Clearly, these devices have a near-global impact on one of the largest and least well-understood ecosystems on our planet: the open ocean. Almost every published synthetic review has concluded that numbers and locations of dFAD deployments need to be controlled to meaningfully mitigate documented impacts (7, 98, 99). As eminent tuna scientist Alain Fonteneau concludes in a first global overview in 2000: “It appears that the present massive use of FADs worldwide is perhaps an unsafe fishing mode, which could produce serious overfishing of many stocks. There is then a consensus that the use of FADs needs to be controlled and limited to sustainable biological levels” (9). We note with concern that the global deployment of dFAD buoys and purse seine catches associated with these devices has more than doubled since that statement was published (Fig. 2A). Using a clearly defined dFAD management framework (94), precautionary deployment limits for each ocean basin should be devised based on existing data and knowledge. These limits can be revised accordingly once RFMO dFAD registries are properly established and include sufficient long-term information provided by all fleets. Considering the high number of documented losses and to mitigate data inconsistencies, all dFAD deployments and retrievals should also be subject to 100% observer coverage. Last, we further encourage the development of spatial management measures to minimize further dFAD strandings in regional hot spots combined with measures to disincentivize abandonment and obligate retrieval.

Study caveats and limitations

Our study necessarily carries caveats and limitations that relate to pervasive issues of data availability and accuracy. For example, dFAD strandings are rarely reported and dFAD track data are proprietary and not accessible, except through summary publications. The 88,359 buoys we analyzed between 2007 and 2023 likely represented ~6% of the total number of dFAD buoys we estimated deployed during the same time, i.e., 1.41 million (Fig. 2C). This likely renders our spatial footprint estimates conservative, especially given the paucity of drift data for areas with high reported strandings (e.g., Gulf of Mexico and Caribbean). At the same time, because we only interpreted observed strandings as presence/absence data, our results do not account for variation in the type of dFAD debris that washed ashore or the condition in which it was found, both of which relate to its potential environmental impact.

Regarding our timeline of buoy deployments, we highlight that numbers are increasingly uncertain going back in time, owing to poor or absent reporting at the RFMO level. We also note that the substantial difference between ICCAT reported values and our estimated deployments for the Atlantic Ocean in 2018 to 2021 (table S4) is driven by our deployment reconstruction for the fleet from Ghana, which was responsible for 29% of all reported dFAD catch in the Atlantic during this time. Similarly, disparity between reported and estimated deployments in the Indian Ocean were driven almost entirely by missing deployments for Seychelles-flagged vessels. While we know that supply vessels can play a large role in assisting dFAD fishing fleets with dFAD deployments and retrievals (100), we could not account for their involvement specifically in our study due to a lack of publicly available information. We were also unable to provide context on the degree to which dFAD buoys are stolen, sold, exchanged, or repurposed (rafts redeployed with a fresh buoy) by fishing companies at-sea although these activities are known to be common practice. These topics could be a focus of further research, and a globally integrated dFAD registry would help to trace the total use and spatial footprint of these devices and better understand the impacts of dFAD fishing at the global scale. Last, we highlight that our analysis of MSC open conditions considers observed shortcomings only at the time of a client group’s initial assessment, and therefore does not capture the time and resources companies may have invested to improve their practices. Thus, we encourage further research into how open conditions for dFAD fisheries have changed over time and how holding MSC certification has contributed to fishing companies mitigating dFAD impacts and addressing data transparency shortcomings.

While much of our discussion has focused on the sustainability shortcomings of dFAD use in relation to our results, we note that there are also undeniable economic and practical benefits that continue to drive the widespread use of these devices [reviewed in more detail in (7)]. These include the reduced uncertainty and increase in efficiency that derives from the ability to monitor fish biomass and identify promising fishing locations, resulting in better success rates when fishing dFADs as opposed to unassociated fishing. For example, most fishing on skipjack tuna in the Indian and Atlantic oceans now includes dFADs, while remaining free-school purse seines target adult yellowfin (101).

Future directions

In conclusion, we submit that the global footprint and cumulative impact of dFAD fishing on marine ecosystems require a globally coordinated solution that applies to all relevant industry members and RFMOs. Science-based limits on dFAD deployment, improved spatial management and closures, rapid adoption of fully nonentangling and biodegradable FADs, and standardized labeling, ownership, and responsibility rules are some of the most pressing issues to address concerns that have long been raised about the widespread use of these devices. We note that regulatory efforts are helped by the fact that NGO engagement, market incentives, and science-based management measures have contributed toward addressing other long-standing issues in tuna fishing (87, 102–104) and that the world’s largest dFAD fishing companies have many open conditions to address if they want their MSC certification to remain credible in the eyes of retailers and buyers. As such, we are hopeful that the large footprint of global dFAD fisheries documented here can be limited in a way that helps to conserve and restore the pelagic and deep-water ecosystems occupying more than 70% of our planet.

MATERIALS AND METHODS

Global footprint estimation

Global tuna dFAD purse seine catch data were accessed from published RFMO documents and databases as summarized in (7). These data are reported at a 5° × 5° scale, and total area calculations were made at that scale. We chose to present dFAD purse seine catch here because it is the most direct measure of dFAD fishing.

Because dFAD data are proprietary and not publicly available, we were unable to obtain raw dFAD buoy tracks that would have allowed for a more fine-scale analysis. Instead, we used existing published buoy track density data that were digitized and entered into a Geographic Information System database. As indicated by the original publications, these data represent 9289 dFAD buoys used by 29 French and French-associated purse seiners in the Atlantic and Indian Oceans in 2007 to 2011 (15), 56,263 dFAD buoys used by the same fleets in 2012 to 2018 (74), 10,266 dFAD buoys deployed by 123 vessels of different nationalities fishing in the Western and Central Pacific (WCPO) in 2023 (105), and 12,541 dFAD buoys deployed by 165 vessels in the Eastern Pacific (EPO) in 2022 (58), totaling 88,359 individual dFAD buoys tracked between 2007 and 2023. Although the latter two studies do not mention the specific flags of vessels deploying dFADs in the Pacific Ocean, vessel authorization lists show 16 different countries currently operating a total of 193 tuna purse seiners in the WCPO: China, Ecuador, Federated States of Micronesia, Japan, Kiribati, Korea, Marshall Island, Nicaragua, Papua New Guinea, Philippines, Solomon Islands, El Salvador, Taiwan, Tuvalu, United States of America (USA), and Vanuatu (https://vessels.wcpfc.int/) with 10 countries using 290 purse seines in the EPO: Colombia, Ecuador, Spain, Mexico, Nicaragua, Panama, Peru, El Salvador, USA, and Venezuela (https://iattc.org/en-us/Management/Vessel-register).

Track density data were typically binned in 1° × 1° cells in the source publications, and hence the same resolution was used for our composite (Fig. 1D). We note that all tracking data in the studies used refer to the trajectories of dFAD buoys, not necessarily whole devices (i.e., buoy, raft, and tail). Because a given buoy may be used once or multiple times by the same fleet or stolen/transferred to a different fleet (15, 98), the number of buoy deployments may not always equate to the number of whole dFADs deployed.

We compiled all information available from the literature to determine the location and number of dFAD strandings, which was based on two types of previously published data: physical observations of the device (i.e., buoys or raft material) and inferred strandings based on dFAD tracks (buoys only). The data obtained from all sources (see table S12) were entered into a database binned by the EEZ of the world’s maritime nations and territories. Strandings in nations bordering multiple oceans (USA, Mexico, and Australia), were binned separately by ocean based on the stranding location provided in the source.

Catch and FAD deployment estimates

The composition of tropical tuna catch by gear type was calculated using publicly available data (table S13) as were dFAD buoy deployments (table S3). For deployments, we summarized reported data and estimated total deployments for each RFMO in 2007 to 2021. Annual reported deployments were calculated from published RFMO data and reports (table S3). Because these data are incomplete with certain years and fleets missing, we also report the estimated total number of dFAD buoy deployments using existing time series from the peer-reviewed literature (20, 106) combined with scaled RFMO data.

Our approach to estimating missing dFAD buoy deployments involved scaling deployments from reporting fleets by total fleet size (number of dFAD fishing vessels) in the Indian Ocean and by total fleet dFAD catch in the Atlantic Ocean. This assumes that the number of deployments per vessel and the catch per deployment are similar across fleets, an assumption that is supported by RFMO documents [e.g., (107)]. As it was impossible to discern whether there were gaps in deployment numbers reported for the Eastern Pacific (58), no additional estimates were made here. For the Western Pacific, we used a comprehensive estimate of total deployments based on observer and tracking data available from 2016 to 2019 (20) and estimated the total deployments of other years based on the average proportion of estimated WCPO deployments from this source relative to the total estimated deployments of the three other regions combined. In doing this, we assumed that the WCPO region deploys an approximately stable fraction (32%) of all dFAD buoys deployed worldwide. As noted above, the dFAD deployment data used here generally refer to dFAD buoys, not entire devices. Refer to table S14 for details on deployment estimates.

For calculating the relative catch composition by age for each tropical tuna species, we used L₅₀ as our measure for maturity (i.e., the length at which 50% of the population is mature). Because L₅₀ varies not only by species but also by region, we used the L₅₀ value from each tuna stock’s most recent assessment or a reasonable published alternative (table S15). Catch-at-age estimates from the Atlantic and Western Pacific cover a longer time frame (1990 to 2022 and 1970 to 2021, respectively) while data for the Indian and Eastern Pacific were more recent (2022), thereby providing both a long-term and shorter view of catch-at-age across all three species. Where size composition data were not available directly from the RFMO public databases, they were extracted manually from available RFMO reports (see table S15 for all assumptions and methods).

Trends in scientific publications

We extracted all publications with the keywords “FAD,” “DFAD,” and “Fish aggregating device” from the Open Alex database (https://openalex.org/) which indexes more than 250 million scholarly works from more than 250,000 sources, with extra coverage of humanities, non-English languages, and the Global South. This was important, because a notable fraction of the relevant literature was published in local languages and by authors from low-income countries. The search was completed in May 2023 and yielded 903 papers and publications whose titles and abstracts were manually checked and categorized into papers that covered aFADs, dFADs, or other topics less relevant to this review (e.g., aggregation behavior of fish around oil platforms or fish farms). All papers that dealt exclusively or nonexclusively with dFADs were reviewed in detail, cataloged, and binned by year of publication. This extraction yields mostly journal articles and does not fully capture gray literature publications or RFMO documents, which were reviewed separately by searching the relevant RFMO document repositories.

Conservation and management measures

Using published RFMO compendia (108–111), we amalgamated a time series of RFMO CMMs (also called “Resolutions” or “Recommendations”) to determine how the uptake of dFAD management measures has changed over time (table S6). We chose to include measures specific to dFADs as well as those for target tropical tuna fisheries that have specific dFAD components. While all RFMOs have also adopted measures to prohibit purse seine vessel crew from setting their nets on whale sharks or cetaceans, we did not include these measures because they do not relate to dFAD fishing. We also did not include bycatch-related measures (e.g., shark retention prohibitions) or those related to observer coverage if they were region-wide and did not specifically apply to dFAD fishing.

While we discuss the most current state of dFAD measures in the text and in Table 1 (i.e., up to and including all 2024 RFMO Commission meetings), Fig. 3C includes measures adopted through December 2023. Measures correspond to the year of the meeting in which they were adopted, not the year in which they were implemented. We also performed a content analysis of all measures to track changes in regulatory priorities over time. We used binary scoring (presence; “1”/absence; “0”) to categorize the main topic of dFAD-related provisions within a given CMM (table S16). A single CMM could contain one or more topics. We did not account for incremental changes in dFAD measures (e.g., changes in the type of data required, or moving from “encouraging” to “requiring” states to implement a given provision). We used a similar content analysis to categorize all recommendations (n = 176) provided by RFMO FAD Working Groups over time. Because working groups are mostly tasked to identify research priorities and improve upon weaknesses in dFAD knowledge and regulation, the main topics differed slightly from those in the CMMs (table S17). We categorized each working group recommendation according to its principal action item (e.g., a recommendation to standardize a definition for “biodegradable FAD” was categorized as “dFAD terminology,” while a recommendation for fishing countries to transition to biodegradable dFADs was categorized as “dFAD design”).

RFMO advocacy

To understand how NGO observer attention toward dFADs has changed over time, we updated and expanded an existing list of WCPFC and ICCAT observer advocacy letters (49) with letters submitted to IOTC and IATTC through 2023. Letters for WCPFC (n = 138) were retrieved from “Observer Documents” on the webpage for each Regular Session meeting (https://meetings.wcpfc.int/), and ICCAT letters (n = 138) were found in Vol. 1 of the RFMO’s biennial reports (https://iccat.int/en/pubs_biennial.html), which include annual Commission meeting outcomes. IOTC letters (n = 89) were obtained by searching “NGO Statements” in the “Documents” section of the IOTC website (https://iotc.org/documents). IATTC letters used in our analysis (n = 37) only date back to 2020 as this was the first year when they were provided as part of “Meeting of the IATTC” documents (https://iattc.org/en-US/Event). All NGO observer letters (n = 402; table S7) were coded for the presence (“1”) or absence (“0”) of content related to dFADs. In cases where one NGO submitted multiple letters in a year (or was part of a joint statement), we chose to include only those statements meant for the Commission as a whole (e.g., part of the plenary session) rather than part of a side session or workshop. We included only advocacy statements and did not include submitted reports or peer-reviewed articles in this analysis.

Second, we coded in detail the dFAD priorities identified in each NGO letter for early years and the most recent year to assess how observer attention toward specific dFAD issues has changed over time. Because regional advocacy related to dFADs started at different times, “early” letters analyzed reflect the first 2 to 3 years in which dFAD issues were present at a given RFMO. We analyzed a combined total of 14 early letters for IOTC, ICCAT, and WCPFC between 2007 and 2014 but did not include IATTC due to the absence of letters before 2020. For “recent” letters (n = 27), we coded letters submitted to all four RFMOs in 2023. Our coding was based on topics falling under two key categories: concerns and proposed solutions (table S18), and we used presence/absence scoring to capture whether these topics were mentioned in each letter. Thus, a given letter could include one or more topics.

Certified fisheries and MSC open conditions

Our calculated trajectory of MSC-certified tuna fisheries was based on publicly available data from the MSC website and is current through June 2024. We included all tuna fisheries that had been certified at least once (including those now suspended or withdrawn) but not those currently in assessment for the first time. Any purse seine fisheries originally certified for a free-school component but that later also became certified for a dFAD component were deemed “free school” until the year in which the dFAD component was included (table S8).

To investigate potential sustainability shortcomings of MSC-certified dFAD fisheries, we used the July 2023 version of MSC Conditions Log (https://fisheries.msc.org/en/the-msc-conditions-log/), which included 5478 entries (conditions) across 402 fisheries certified or in assessment. Because the database only included fisheries certified through 23 June 2023, not all tuna fisheries in table S8 were included in the conditions’ analyses. For the purposes of our analysis, we first grouped each fishery in the dataset into one of two categories based on the target species listed: “Tuna” (n = 56) and “Other” (n = 348). Two fisheries that caught both tuna and swordfish were included in Tuna because they were not disaggregated at the species level. Second, we simplified the fishery gear(s) provided in the database into one of three categories for Tuna fisheries: “PS-FS” (free-school purse seine; n = 9), “PS-FAD” (dFAD purse seine; n = 13), and “Other tuna” (all other gears catching tuna; n = 34). Because a fishery that was certified with a dFAD component is also certified for free school but a fishery certified for exclusively free-school fishing is not covered for dFAD fishing, all fisheries using both methods were assigned “dFAD.” For fisheries that were originally certified for free-school components and later for dFADs, we assigned the appropriate gear type for a given condition based on the date of the condition relative to when each component was certified. All species categorized as “Other” above were assigned the gear “Other.”

For all fisheries with at least one MSC assessment (including those suspended or withdrawn), we calculated the total number of open conditions identified during the first assessment cycle. Because of uneven sample sizes and variances, we used a nonparametric Kruskal-Wallis test to determine whether there were significant differences in the number of conditions between the three “Tuna” gear groups. Because tuna fishery certifications often have multiple species included within a single assessment (and therefore potentially more conditions overall), we tested both the total number of conditions for a given fishery and the total number of conditions relative to the number of stocks covered by that assessment (table S10). Last, for dFAD fisheries only, we tabulated the number of conditions (accumulated across all assessment cycles) under each of the MSC assessment criteria (i.e., Performance Indicators) to determine the key areas of improvement required specifically for companies that purse seine with dFADs.

Supplementary Materials

This PDF file includes:

Figs. S1 to S3

Tables S1 to S19

References