Fishing effort, measured in hours of inferred fishing activity. Each asset is the effort for a given flag state and day, with one band for the fishing activity of each gear type. See sample Earth Engine scripts. Also see the main GFW site for program information, fully interactive visualization maps, …

Overview

This dataset contains version 3.0 (March 2025 release) of the Global Fishing Watch apparent fishing effort dataset. Data is available for 2012-2024 and based on positions of >190,000 unique automatic identification system (AIS) devices on fishing vessels, of which up to ~96,000 are active in a given year. Fishing vessels are identified via a machine learning model, vessel registry databases, and manual review by GFW and regional experts. Vessel time is measured in hours, calculated by assigning to each AIS position the amount of time elapsed since the previous AIS position of the vessel. The time is counted as apparent fishing hours if the GFW fishing detection model - a neural network machine learning model - determines the vessel is engaged in fishing behavior during that AIS position.

Data are spatially binned into grid cells that measure 0.01 or 0.1 degrees on a side; the coordinates defining each cell are provided in decimal degrees (WGS84) and correspond to the lower-left corner. Data are available in the following formats:

1. Daily apparent fishing hours by flag state and gear type at 100th degree resolution
2. Monthly apparent fishing hours by flag state and gear type at 10th degree resolution
3. Daily apparent fishing hours by MMSI at 10th degree resolution

The fishing effort dataset is accompanied by a table of vessel information (e.g. gear type, flag state, dimensions).

File structure

Fishing effort and vessel presence data are available as .csv files in daily formats. Files for each year are stored in separate .zip files. A README.txt and schema.json file is provided for each dataset version and contains the table schema and additional information. There is also a README-known-issues-v3.txt file outlining some of the known issues with the version 3 release.

Files are names according to the following convention:

• Daily file format:

• [fleet/mmsi]-daily-csvs-[100/10]-v3-[year].zip

• [fleet/mmsi]-daily-csvs-[100/10]-v3-[date].csv

• Monthly file format:

• fleet-monthly-csvs-10-v3-[year].zip

• fleet-monthly-csvs-10-v3-[date].csv

• Fishing vessel format: fishing-vessels-v3.csv

• README file format: README-[fleet/mmsi/fishing-vessels/known-issues]-v3.txt

File identifiers:

• [fleet/mmsi]: Data by fleet (flag and geartype) or by MMSI

• [100/10]: 100th or 10th degree resolution

• [year]: Year of data included in .zip file

• [date]: Date of data included in .csv files. For monthly data, [date]corresponds to the first date of the month

Examples: fleet-daily-csvs-100-v3-2020.zip; mmsi-daily-csvs-10-v3-2020-01-10.csv; fishing-vessels-v3.csv; README-fleet-v3.txt; fleet-monthly-csvs-10-v3-2024.zip; fleet-monthly-csvs-10-v3-2024-08-01.csv

Key documentation

• For an overview of how GFW turns raw AIS positions into estimates of fishing hours, see this page.

• The models used to produce this dataset were developed as part of this publication: D.A. Kroodsma, J. Mayorga, T. Hochberg, N.A. Miller, K. Boerder, F. Ferretti, A. Wilson, B. Bergman, T.D. White, B.A. Block, P. Woods, B. Sullivan, C. Costello, and B. Worm. "Tracking the global footprint of fisheries." Science 361.6378 (2018). Model details are available in the Supplementary Materials.

• The README-known-issues-v3.txt file describing this dataset's specific caveats can be downloaded from this page. We highly recommend that users read this file in full.

• The README-mmsi-v3.txt file, the README-fleet-v3.txt file, and the README-fishing-vessels-v3.txt files are downloadable from this page and contain the data description for (respectively) the fishing hours by MMSI dataset, the fishing hours by fleet dataset, and the vessel information file. These readmes contain key explanations about the gear types and flag states assigned to vessels in the dataset.

• File name structure for the datafiles are available below on this page and file schema can be downloaded from this page.

• A FAQ describing the updates in this version and the differences between this dataset and the data available from the GFW Map and APIs is available here.

Use Cases

The apparent fishing hours dataset is intended to allow users to analyze patterns of fishing across the world’s oceans at temporal scales as fine as daily and at spatial scales as fine as 0.1 or 0.01 degree cells. Fishing hours can be separated out by gear type, vessel flag and other characteristics of vessels such as tonnage.

Potential applications for this dataset are broad. We offer suggested use cases to illustrate its utility. The dataset can be integrated as a static layer in multi-layered analyses, allowing researchers to investigate relationships between fishing effort and other variables, including biodiversity, tracking, and environmental data, as defined by their research objectives.

A few example questions that these data could be used to answer:

• What flag states have fishing activity in my area of interest?

• Do hotspots of longline fishing overlap with known migration routes of sea turtles?

• How does fishing time by trawlers change by month in my area of interest? Which seasons see the most trawling hours and which see the least?

Caveats

This global dataset estimates apparent fishing hours effort. The dataset is based on publicly available information and statistical classifications which may not fully capture the nuances of local fishing practices. While we manually review the dataset at a global scale and in a select set of smaller test regions to check for issues, given the scale of the dataset we are unable to manually review every fleet in every region. We recognize the potential for inaccuracies and encourage users to approach regional analyses with caution, utilizing their own regional expertise to validate findings. We welcome your feedback on any regional analysis at research@globalfishingwatch.org to enhance the dataset's accuracy.

Caveats relating to known sources of inaccuracy as well as interpretation pitfalls to avoid are described in the README-known-issues-v3.txt file available for download from this page. We highly recommend that users read this file in full. The issues described include:

• Data from 2024 should be considered provisional, as vessel classifications may change as more data from 2025 becomes available.

• MMSI is used in this dataset as the vessel identifier. While MMSI is intended to serve as the unique AIS identifier for an individual vessel, this does not always hold in practice.

• The Maritime Identification Digits (MID), the first 3 digits of MMSI, are the only source of information on vessel flag state when the vessel does not appear on a registry. The MID may be entered incorrectly, obscuring information about an MMSI’s flag state.

• AIS reception is not consistent across all areas and changes over time.

Alternative ways to access

1. Query using SQL in the Global Fishing Watch public BigQuery dataset: global-fishing-watch.fishing_effort_v3

2. Download the entire dataset from the Global Fishing Watch Data Download Portal (https://globalfishingwatch.org/data-download/datasets/public-fishing-effort)

Fishing vessel presence, measured in hours per square km. Each asset is the vessel presence for a given flag state and day, with one band for the presence of each gear type. See sample Earth Engine scripts. Also see the main GFW site for program information, fully interactive visualization maps, …

This map is part of SDGs Today. Please see sdgstoday.orgToday, the health of our ocean is under immense pressure from both intensive human activity and climate change. One-third of global fish stocks are overfished and two-thirds of the ocean has been significantly altered by human actions. Despite the threats it faces, the ocean remains the least observed part of our planet. This lack of visibility allows illicit activity to thrive. Global Fishing Watch (GFW) is advancing ocean governance through increased transparency of human activity at sea. By creating and publicly sharing map visualizations, data and analysis tools, we enable scientific research and drive a transformation in how we manage our ocean.In 2018, GFW published the first-ever global assessment of commercial fishing activity (2012-2016) in Science. The updated 2021 version of this published dataset contains the GFW AIS-based fishing effort and vessel presence from 2012-2020, which includes over 328 million hours of fishing effort across the globe from over 117,000 unique maritime mobile service identity (MMSI) numbers. Fishing vessels are identified via a neural network classifier, vessel registry databases, and a manual review by GFW and regional experts. Data are binned into grid cells 0.01 (or 0.1) degrees on a side and measured in units of hours. The time is calculated by assigning an amount of time to each AIS detection (which is the time to the previous position) and then summing all positions in each grid cell.For more information, contact Global Fishing Watch at research@globalfishingwatch.org.

This is time-series layer of monthly fishing activity for the year 2017. This layer shows the amount of fishing activity recorded by Automatic Identification System (AIS) broken out by month. Each pixel value represents the number of hours per square kilometer of fishing activity. Fishing activity can be described using the two available variables:Hours - Identification of fishing vessels in the AIS data. Fishing Hours (displayed by default) - Detection of fishing activity. Using cloud computing, machine learning and public vessel registry information, Global Fishing Watch (GFW) analyze tens of millions of AIS positions each day to map global apparent fishing effort. Producing such a map involves two key steps: Identification of fishing vessels in the AIS data (Hours) and Detection of fishing activity (Fishing Hours). This annual summary is produced from the daily GFW AIS-based apparent fishing effort data. The daily fishing activity made available by GFW was rasterized at 100th degree resolution then aggregated to produce an annual summary for the given year. The maps show areas where likely fishing activity occurred in the year 2017 and the estimated level of fishing intensity. This information can help understand areas where fishing activity might be considered in a marine spatial planning application. More information: https://globalfishingwatch.org/dataset-and-code-fishing-effort/ Dataset summaryVariable Mapped: Occupancy (hours) and effort (fishing hours) in hours per sq/kmDimension: Time – 12Months (Year 2017)Data Projection: GCS WGS84Service Projection: Web MercatorExtent: GlobalCell Size: (~1km)Source Type: Scientific/DoubleData Source: Global Fishing WatchData Accessed Date: September 23, 2023Version: 2.0, released 18 March 2021 What can you do with this layer?The layer can be used in analysis and visualization. This layer can be used to summarize the values within a polygon (using zonal statistics). Fishing effort can be used to understand the areas impacted by fishing and designate marine protection if needed. This layer allows use of the time dimension to help understand locations and months/seasons when fishing intensity is higher or lower. See companion layers in this ArcGIS Online Group: Global Fishing Watch

This layer shows the amount of fishing activity recorded by Automatic Identification System (AIS) for the year 2020.This layer has two variables:Hours - Identification of fishing vessels in the AIS data.Fishing Hours (displayed by default) - Detection of fishing activity.Using cloud computing, machine learning and public vessel registry information, Global Fishing Watch (GFW) analyze tens of millions of AIS positions each day to map global apparent fishing effort. Producing such a map involves two key steps: Identification of fishing vessels in the AIS data (Hours) and Detection of fishing activity (Fishing Hours).

This annual summary is produced from the daily GFW AIS-based apparent fishing effort data. The daily fishing activity made available by GFW was rasterized at 100th degree resolution then aggregated to produce an annual summary for the given year. The maps show areas where likely fishing activity occurred in the year 2020 and the estimated level of fishing intensity. This information can help understand areas where fishing activity might be considered in a marine spatial planning application.More information: https://globalfishingwatch.org/dataset-and-code-fishing-effort/Dataset SummaryVariable Mapped: Occupancy (hours) and effort (fishing hours) in hours per sq/kmData Projection: GCS WGS84Service Projection: Web MercatorExtent: GlobalCell Size: (~1km)Source Type: Scientific/DoubleData Source: Global Fishing WatchData Accessed Date: September 23, 2023Version: 2.0, released 18 March 2021What can you do with this layer?The layer can be used in analysis and visualization. This layer can be used to summarize the values withing a polygon (using zonal statistics). Fishing effort can be used to understand the areas impacted by fishing and designate marine protection if needed.See companion layers in this ArcGIS Online Group: Global Fishing Watch

This is time-series layer of monthly fishing activity for the year 2019. This layer shows the amount of fishing activity recorded by Automatic Identification System (AIS) broken out by month. Each pixel value represents the number of hours per square kilometer of fishing activity. Fishing activity can be described using the two available variables:Hours - Identification of fishing vessels in the AIS data. Fishing Hours (displayed by default) - Detection of fishing activity. Using cloud computing, machine learning and public vessel registry information, Global Fishing Watch (GFW) analyze tens of millions of AIS positions each day to map global apparent fishing effort. Producing such a map involves two key steps: Identification of fishing vessels in the AIS data (Hours) and Detection of fishing activity (Fishing Hours). This annual summary is produced from the daily GFW AIS-based apparent fishing effort data. The daily fishing activity made available by GFW was rasterized at 100th degree resolution then aggregated to produce an annual summary for the given year. The maps show areas where likely fishing activity occurred in the year 2019 and the estimated level of fishing intensity. This information can help understand areas where fishing activity might be considered in a marine spatial planning application. More information: https://globalfishingwatch.org/dataset-and-code-fishing-effort/ Dataset summaryVariable Mapped: Occupancy (hours) and effort (fishing hours) in hours per sq/kmDimension: Time – 12Months (Year 2019)Data Projection: GCS WGS84Service Projection: Web MercatorExtent: GlobalCell Size: (~1km)Source Type: Scientific/DoubleData Source: Global Fishing WatchData Accessed Date: September 23, 2023Version: 2.0, released 18 March 2021 What can you do with this layer?The layer can be used in analysis and visualization. This layer can be used to summarize the values within a polygon (using zonal statistics). Fishing effort can be used to understand the areas impacted by fishing and designate marine protection if needed. This layer allows use of the time dimension to help understand locations and months/seasons when fishing intensity is higher or lower. See companion layers in this ArcGIS Online Group: Global Fishing Watch

1. MISSION ATLANTIC The MISSION ATLANTIC project is an EU-funded initiative that focuses on understanding the impacts of climate change and human activities on these ecosystems. The project aims to map and assess the current and future status of Atlantic marine ecosystems, develop tools for sustainable management, and support ecosystem-based governance to ensure the resilience and sustainable use of ocean resources. The project brings together experts from 33 partner organizations across 14 countries, including Europe, Africa, North, and South America. MISSION ATLANTIC includes ten work packages. The present published dataset is included in WP3, which focuses on mapping the pelagic ecosystems, resources, and pressures in the Atlantic Ocean. This WP aims to collect extensive spatial and temporal data to create 3D maps of the water column, identify key vertical ecosystem domains, and assess the pressures from climate change and human activities. More specifically, the dataset corresponds to the fishing intensity presented in the Deliverable 3.2, which integrates data from various sources to map the distribution and dynamics of present ecosystem pressures over time, providing crucial insights for sustainable management strategies. 2. Data description 2.1. Data Source Fishing intensity estimates from the Global Fishing Watch initiative (GFW) (Kroodsma et al. 2018), who applies machine learning algorithms to data from Automatic Identification Systems (AIS), Vessel Monitoring Systems (VMS), and vessel registries, has been used for the year 2020. This machine learning approach has been able to distinguish between fishing and routing activity of individual vessels, while using pattern recognition to differentiate seven main fishing gear types at the Atlantic Ocean scale (Taconet et al., 2019). The seven main fishing vessel types considered are: trawlers, purse seiners, drifting longliners, set gillnets, squid jiggers, pots and traps, and other. In this work we have aggregated these into pelagic, seabed and passive fishing activities to align with our grouping of ecosystem components. The GFW data has some limitations: AIS is only required for large vessels. The International Maritime Organization requires AIS use for all vessels of 300 gross tonnage and upward, although some jurisdictions mandate its use in smaller vessels. For example, within the European Union it is required for fishing vessels at least 15m in length. This means that in some areas the fishing intensity estimates will not include the activity of small vessels operating near shore. AIS can be intentionally turned off, for example, when vessels carry out illegal fishing activities (Kurekin et al. 2019). In the GFW dataset, vessels classified as trawlers include both pelagic and bottom trawlers. As trawlers are included in the bottom fishing category, it is highly likely that the data overestimates the effort on the seafloor and underestimates it on the water column. 2.2. Data Processing 1. Data download from the GFW portal. 2. Using R: Add daily files and aggregate fishing hours by fishing gear and coordinates: library(data.table)## Load data fileIdx = list.files(".../fleet-daily-csvs-100-v2-2020/", full.names = T) ## Loop colsIdx = c("geartype", "hours", "fishing_hours", "x", "y") lapply(fileIdx, function(xx) { out = data.table (x = NA_real_, y = NA_real_, geartype = NA_character_) tmp = fread(xx) tmp[, ":=" (y = floor(cell_ll_lat * 10L) / 10L, x = floor(cell_ll_lon * 10L) / 10L)] tmp = tmp[, ..colsIdx] h = tmp[, c(.N, lapply(.SD, sum, na.rm = T)), by = .(x, y, geartype)] outh = data.table::merge.data.table(out, h, by = c("x", "y", "geartype"), all=TRUE) fwrite(outh, ".../GFW_2020_0.1_degrees_and_gear_all.csv", nThread = 14, append = T) }) Group fishing gears into main fishing groups: library(dplyr)library(tidyr)## Load data fishing <- read.csv(".../GFW_2020_0.1_degrees_and_gear_all.csv", sep=",", dec=".", header=T, stringsAsFactors = FALSE) ## Grouping fishing gears (fishing, pelagic, bottom, passive) # unique(fishing$geartype) fishing$group <- NA fishing$group[which(fishing$geartype == "fishing")] = "fishing" # Unknown fishing$group[fishing$geartype %in% c("trollers", "squid_jigger", "pole_and_line", "purse_seines", "tuna_purse_seines", "seiners", "other_purse_seines", "other_seines", "set_longlines", "drifting_longlines")] <- "pelagic" fishing$group[fishing$geartype %in% c("trawlers", "dredge_fishing")] <- "bottom" fishing$group[fishing$geartype %in% c("set_gillnets", "fixed_gear", "pots_and_traps")] <- "passive" ## Total fishing hours (by fishing and position) fish_gr <- fishing %>% group_by(x, y, group) %>% summarise(gfishing_hours = sum(fishing_hours)) Pivot table in order to have fishing groups in columns. Each row corresponds to the coordinates of the left corner of the grid cell (0.1 decimal degrees): ## Pivoting table (fishing groups in columns) fish_gr3 <- fish_gr %>% pivot_wider(names_from = "group", values_from = "gfishing_hours", va

This is time-series layer of monthly fishing activity for the year 2014. This layer shows the amount of fishing activity recorded by Automatic Identification System (AIS) broken out by month. Each pixel value represents the number of hours per square kilometer of fishing activity. Fishing activity can be described using the two available variables:Hours - Identification of fishing vessels in the AIS data. Fishing Hours (displayed by default) - Detection of fishing activity. Using cloud computing, machine learning and public vessel registry information, Global Fishing Watch (GFW) analyze tens of millions of AIS positions each day to map global apparent fishing effort. Producing such a map involves two key steps: Identification of fishing vessels in the AIS data (Hours) and Detection of fishing activity (Fishing Hours). This annual summary is produced from the daily GFW AIS-based apparent fishing effort data. The daily fishing activity made available by GFW was rasterized at 100th degree resolution then aggregated to produce an annual summary for the given year. The maps show areas where likely fishing activity occurred in the year 2014 and the estimated level of fishing intensity. This information can help understand areas where fishing activity might be considered in a marine spatial planning application. More information: https://globalfishingwatch.org/dataset-and-code-fishing-effort/ Dataset summaryVariable Mapped: Occupancy (hours) and effort (fishing hours) in hours per sq/kmDimension: Time – 12Months (Year 2014)Data Projection: GCS WGS84Service Projection: Web MercatorExtent: GlobalCell Size: (~1km)Source Type: Scientific/DoubleData Source: Global Fishing WatchData Accessed Date: September 23, 2023Version: 2.0, released 18 March 2021 What can you do with this layer?The layer can be used in analysis and visualization. This layer can be used to summarize the values within a polygon (using zonal statistics). Fishing effort can be used to understand the areas impacted by fishing and designate marine protection if needed. This layer allows use of the time dimension to help understand locations and months/seasons when fishing intensity is higher or lower. See companion layers in this ArcGIS Online Group: Global Fishing Watch

This map is part of SDGs Today. Please see sdgstoday.orgToday, the health of our ocean is under immense pressure from both intensive human activity and climate change. One-third of global fish stocks are overfished and two-thirds of the ocean has been significantly altered by human actions. Despite the threats it faces, the ocean remains the least observed part of our planet. This lack of visibility allows illicit activity to thrive.Global Fishing Watch (GFW) is advancing ocean governance through increased transparency of human activity at sea. By creating and publicly sharing map visualizations, data and analysis tools, we enable scientific research and drive a transformation in how we manage our ocean.In 2018, GFW published the first-ever global assessment of commercial fishing activity (2012-2016) in Science. The updated 2021 version of this published dataset contains the GFW AIS-based fishing effort and vessel presence from 2012-2020, which includes over 328 million hours of fishing effort across the globe from over 117,000 unique maritime mobile service identity (MMSI) numbers. Fishing vessels are identified via a neural network classifier, vessel registry databases, and a manual review by GFW and regional experts. Data are binned into grid cells 0.01 (or 0.1) degrees on a side and measured in units of hours. The time is calculated by assigning an amount of time to each AIS detection (which is the time to the previous position) and then summing all positions in each grid cell.For more information, contact Global Fishing Watch at research@globalfishingwatch.org.

Fishing effort data publicly accessible thought the Global Fishing Watch (GFW) site (GFW, 2022) have been used to investigate the spatiotemporal distribution of fishing activities in a wide area (Latitude: 30.7°N - 66°N; Longitude: 14.4°W - 41.9°E).

Daily fishing effort data by flag state and vessel class at 0.01° resolution, from 2015 to 2020, have been filtered and aggregated to obtain fishing effort information of six main gear categories. Categories were created by aggregation of the following vessel classes: fixed gears (pots and traps, set longlines, set gillnets and fixed gear), purse seines (purse seines, tuna purse seines and other purse seines), trawlers (trawlers), drifting longlines (drifting longlines), dredge (dredge fishing) and other (pole and line, fishing, trollers, seiners, other seines and squid jigger). Data were aggregated to obtain cumulative (fahs) and average (mfahs) fishing hours by fishing category at 0.1° and 0.5° resolution.

Maps of fishing activity for each gear have been created for eight main areas at 0.1° (Adriatic Sea, Aegean Sea, Balearic Sea, Baltic Sea, Bay of Biscay, Black Sea, Levantine Sea and North Sea) and at Mediterranean and Atlantic level at 0.5°.

The dataset presented includes for each gear spatial layers of the cumulative and average fishing hours at 0.1° and 0.5° resolution (.shp; .csv) and maps of each case study area.

Global Fishing Watch is working across the globe to provide governments and authorities with actionable reports and capacity building to help strengthen fisheries monitoring and compliance. Our global team of experts produce analyses to inform monitoring, control and surveillance of fisheries in five key areas: - Illegal, unreported and unregulated fishing - Transshipment - Port controls - Marine protected areas - Operation support

Collaboration and information sharing are integral to achieving well-managed fisheries. By working with stakeholders and making analyses available to national, regional and intergovernmental partners, Global Fishing Watch is enabling fisheries agencies to make more informed and cost-efficient decisions.

Topics: - Commercial fishing, Global Fishing Watch is harnessing innovative technology to turn transparent data into actionable information and drive tangible change in the way that fisheries are governed. - Transshipment, Through publicly sharing map visualisations and creating data and analysis tools, we seek to inform management and policy efforts and provide a more complete picture of transshipment at sea. - Marine protected areas, Global Fishing Watch is harnessing the data and technology revolution to support the effective design, management and monitoring of marine protected areas.

This data publication contains maps resulting from spatial prioritisations conducted for the iAtlantic D5.3 report on Systematic Conservation Planning of the wider Atlantic Ocean based on results generated by the iAtlantic project. The maps were produced using the prioritizr R package (Hanson et al. 2023), which identifies priority areas for achieving specific conservation goals while minimising costs. The various prioritisations were developed to address multiple research questions related to: (1) identifying priority areas for conservation and restoration, (2) transboundary conservation, (3) climate-smart conservation planning, and (4) protecting 30% of the Atlantic Ocean, including 10% under strict protection. The results are organised into subfolders based on the research questions addressed and further categorised into data-rich and data-poor regions, along with aggregate results for each region. Further, the results are organised into subfolders representing multiple scenarios executed using various cost layers, including area-based, Global Fishing Watch (GFW, 2023) benthic, GFW total fishing, Global Fisheries Landings (GFL, Watson 2019) v4.0 benthic, and GFL v4.0 total landings. Each map filename provides descriptive information about the executed scenario.

Numerical ranking of the global ocean to prioritize conservation actions aimed at safeguarding carbon stored in sediments. The conservation action being prioritized is highly protected Marine Protected Areas (MPAs). Most important pixels are ranked higher (1) and least important pixels are ranked lower (0). Pixels ranked between 0.95-1, correspond to the most important 5% of the ocean. Pixels ranked between 0.9-1, correspond to the most important 10% of the ocean, and so on. Existing highly protected MPAs represent the starting condition of the analyses and are placed at the top of the ranking. Variable mapped: Global priority ranking for safeguarding carbonData Projection: Mollweide Equal Area Mosaic Projection: Mollweide Equal Area Extent: Global Cell Size: 50km Source Type: Thematic Source: National Geographic Society Publication: Sala, Enric, Juan Mayorga, Darcy Bradley, Reniel B. Cabral, Trisha B. Atwood, Arnaud Auber, William Cheung, et al. “Protecting the Global Ocean for Biodiversity, Food and Climate.” Nature, March 17, 2021. https://doi.org/10.1038/s41586-021-03371-z.We define the carbon benefits of a given set of MPAs as the change in carbon content in marine sediments relative to business as usual. We model this benefit by combining a global map of the carbon stored in the first meter of marine sediments, with the global footprint of bottom trawling detected with Global Fishing Watch (2015-2019, 1 km2 resolution), and relevant oceanographic variables that are known to affect carbon loss including: Sediment resettlement ratio – the fraction of disturbed sediment that resettles in the same 1km2 pixel and therefore is trackable by our model.Labile carbon ratio – proportion of carbon that is labile and thus more prone to remineralization after a disturbance. The labile fraction was estimated as a function of the type of sediment. First-order degradation constant—The first-order degradation rate constants were assigned as a function of oceanic region, and were estimated from ranges of values presented in the literature for oxic sediments