We estimated spatially explicit global shark mortality for all true sharks and Rhinopristiformes from 2012-2019 at 1x1 degree grids. Total annual mortality was calculated from Regional Fisheries Management Organization (RFMO) shark catch (purse seine and longlines), coastal shark catch (all gear except purse seine and longline), and high seas catch (non-RFMO and unreported RFMO catch). We used machine learning to predict RFMO shark catch risk globally using self- and observer-reported catch and effort data from RFMOs and a suite of environmental parameters. Machine learning allowed us to predict catch across areas with otherwise low reporting coverage. Detailed reconstructions of shark catch in coastal fisheries and the high seas allocated spatially were provided by the Sea Around Us Project. Mortality was calculated at the taxa, gear type, year, and grid cell level for both RFMO catch and coastal catch using taxa- and gear-specific fate information and post-release mortality estimates. Combining spatially explicit coastal and RFMO data, total shark fishing mortality (M) was calculated in year (y) for each taxa (t) in grid cell (j) as:

$M_{y,t,j} = R + D_d + ( D_a *M_{pr} )$

where $R$ is retained or landed catch, $D_d$ is catch discarded dead, $D_a$ is catch discarded alive, and $M_{pr}$ is the post-release mortality rate. All shark catch was assigned to one of four fate categories – retained/landed, mortality, discarded dead, or discarded alive – with the proportion of catch in each fate category determined using the relevant fate dataset (i.e., coastal or RFMO). We calculated four sets of fate proportions for coastal catch (with various groupings by taxa and gear type) and four sets of fate proportions for RFMO catch. Fates were joined to shark catch on a hierarchy of: (i) taxa, gear type, (RFMO, year), (ii) taxa (RFMO, year), (iii) gear type (RFMO, year), and (iv) total averages (RFMO, year) matching on the highest resolution possible for each observation. Mortality was assigned based on fate. All catch retained/landed or discarded dead was assigned a mortality of 1. Catch discarded alive was assigned a post-release mortality rate using a hierarchy of (i) taxa and gear type; (ii) taxa only; (iii) gear class only; and (iv) a total average.

Please refer to the associated manuscript and Supplemental Methods for more information, available here.

Description of the Data and file structure

Data Inventory:

The mortality estimates provided here resulted from analysis of many datasets both public and private. All datasets used in this project are detailed in the attached Data Inventory. The datasets tab lists all types of data used in the project according to the analysis component. Each data type (e.g. RFMO shark catch) is given an alphabetic dataset id. A description of the dataset is included and one or more associated numeric reference ids. The reference id corresponds to each individual piece of data within a dataset and full citations with links to download publicly available data are provided in the references tab. Data that are not publicly available are noted both under ‘Access Notes’ in the datasets tab and in the url of the references tab. These data must be requested from the appropriate authority.

- data_inventory.xlxs: an inventory of all the datasets, with references, used in the analysis, figures, and tables presented in the associated manuscript 

Please refer to the associated software works for code used to generate the output data.

Literature review data:

The following files used in this analysis were compiled through literature review. Please refer to each file's README for metadata.

- bycatch_fate: data collated through literature review on shark fates by species and gear type from 1992-2013      
- post_release_mortality: data collated through literature review on shark post-release mortality rates and ex-vessel mortality rates by species and gear type from 1986-2018   

Spatial regulations data:

The following files are spatial representations of shark fishing and finning regulations in countries, territories, and RFMOs. Please refer to each file's README for metadata.

- regulations_by_eez: most current shark finning regulation for each country and territory for each maritime boundary in Marine Regions v11     
- regulations_by_rfmo: most current shark finning regulation for four main tuna-RFMOs included in the analysis (IATTC, ICCAT, IOTC, WCPFC)       

Output data:

The following files are shark mortality calculations and represents a global spatially explicit estimate of coastal, high seas, and RFMO shark mortality. Please refer to each file's README for metadata.

- coastal_shark_mortality_1x1.csv: results from the mortality calculation for grid cells within Exclusive Economic Zones; estimated shark mortality at 1x1 degree grids by year, species, gear type, reporting status, sector, and fate    
- high_seas_shark_mortality_1x1.csv: results from the mortality calculation for grid cells within the high seas; estimated shark mortality for non-RFMO catch and unreported RFMO catch at 1x1 degree grids by year, species, gear type, reporting status, and sector   
- rfmo_shark_mortality_1x1.csv: results from the mortality calculation for grid cells within RFMO boundaries; estimated shark mortality at 1x1 degree grids by year, species, and gear type 

The following files are used to test for an association between shark mortality and shark fishing and finning regulations at the EEZ-level. Please refer to each file's README for metadata.

- total_mortality_estimate_eez.csv: results from the mortality calculation for each Exclusive Economic Zone; estimated shark catch and shark mortality by EEZ, year, and gear type    
- eez_predictors_annual.csv: regulations present, total effort, total catch of all species, and World Bank Governance Index estimate by EEZ and year   

Uncertainty disclosure

While the datasets shared here are taxonomically and spatially resolved, we caution over interpretation of these data at this scale. Please refer to the associated manuscript and Supplement Methods Section 5 for discussions of uncertainty surrounding shark mortality estimates.

Sharing/access Information

Links to other publicly accessible locations of the data: None

Was data derived from another source? Yes
If yes, list source(s): Please refer to the attached Data Inventory for a list of all sources