Global monthly catch of tuna, tuna-like and shark species (1950-2023) by 1° or 5° squares (IRD level 2) - and efforts level 0 (1950-2023)

Summary of the treatment for a specific dataset of the Global Tuna Atlas true true May, 2025 Abstract This document provides an overview of the impact of various processing steps on a specific dataset from the Global Tuna Atlas. We begin by presenting the final dataset along with its key characteristics, followed by a global comparison with the initial dataset before any treatment. Next, we outline each step involved in creating the final dataset, detailing the decisions made and their effects on specific data points. If you would like to explore the impact of individual processing steps with more specific filters, feel free to contact us or rerun the process and automated reporting using the project available with attached DOI on GitHub at firms-gta/geoflow-tunaatlas Contents 1 Summary of the final provided data : globalcatchirdlevel219502023 1 1.1 Timecoverage............................................... 2 1.2 Spatialcoverage.............................................. 3 1.3 Coverageforotherdimensions...................................... 5 2 Summary of the treatment done on the data (comparison between the Initial data and the data resulting from treatments) 7 2.1 Maindifferences ............................................. 7 2.2 Main characteristics of the two datasets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 3 Analyse of the processing of the data 22 4 Annexe 22 1 Summary of the final provided data : globalcatchirdlevel219502023 1 1.1 Time coverage Number of fish Tons 1960 1980 2000 2020 0 500000 1000000 1500000 0 200000 400000 600000 time_start Values Dataset global−catch−ird−level2−1950−2023 Figure 1: Evolutions of values for the dimension time-start for global-catch-ird-level2-1950-2023 dataset 2 1.2 Spatial coverage We represent spatial coverage, faceted by geographical category. The geographical category depends on the area of the geographic polygon. In this case there are 2 categories which are 1deg_x_1deg ; 5deg_x_5deg. 1deg_x_1deg5deg_x_5deg global−catch−ird−level2−1950−2023 measurement_value 500,000 1,000,000 1,500,000 2,000,000 2,500,000 3,000,000 Missing Figure 2: Distribution in value for the unit: Tons 3 1deg_x_1deg5deg_x_5deg global−catch−ird−level2−1950−2023 measurement_value 1 mln 2 mln 3 mln 4 mln 5 mln Missing Figure 3: Distribution in value for the unit: Number of fish 4 1.3 Coverage for other dimensions We check the distribution of the value of each dimension and each unit. 5 67% 10% 19% 1% 4% 38% 28% 18% 16% 0% Number of fish Tons SOURCE_AUTHORITY CCSBT IATTC ICCAT IOTC WCPFC (a) (Distribution in value for the dimension: source.authority) 22% 22% 8% 43% 5% 6% 32% 9% 43% 10% Number of fish Tons FISHING_FLEET_LABEL China Japan Korea, Republic of Others Taiwan Province of China Other nei Spain (EU) (b) (Distribution in value for the dimension: fishing.fleet.label) 33% 67% 20% 14% 51% 15% Number of fish Tons FISHING_MODE_LABEL Other set types combined Undefined school Free school Log school (c) (Distribution in value for the dimension: fishing.mode.label) 92% 0%0% 0% 7% 98% 1% 0%0% 1% Number of fish Tons SPECIES_GROUP CARCHARHINIFORMES LAMNIFORMES Others PISCES MISCELLANEA SCOMBROIDEI (d) (Distribution in value for the dimension: species.group) 21% 19% 7% 11% 43% 40% 28% 17% 10% 6% Number of fish Tons SPECIES_LABEL Albacore Bigeye tuna Blue shark Others Yellowfin tuna Skipjack tuna (e) (Distribution in value for the dimension: species.label) 64% 2% 4% 28%3% 53% 10% 8% 11% 18% Number of fish Tons GEAR_TYPE_LABEL Drift gillnets Drifting longlines Longlines (nei) Others Set longlines Handlines and hand−operated pole−and−lines Purse seines (f) (Distribution in value for the dimension: gear.type.label) Figure 4: Other dimensions 6 2 Summary of the treatment done on the data (comparison between the Initial data and the data resulting from treatments) Attention ! In the following document: •All the differences inferior to 0 corresponds to gain in captures. •The initial dataset, referred as, dataset 1 is Initial_data •The final dataset, referred as, dataset 2 is global_catch_ird_level2_1950_2023 2.1 Main differences Table 1: Summary of the difference between the two datasets measurement_unit Initial-data global-catch-ird-level2-19502023 Difference Difference (in %) Number of fish 992,562,800 282,450,720 -710,112,081 -71.54 Tons 148,612,676 213,323,903 64,711,227 43.54 68% 32% 86.8% 13.2% 0% 25% 50% 75% 100% global−catch−ird−level2−1950−2023 Initial−data Dataset Percentage Measurement unit Number of fish Tons Figure 5: Distribution of number strata for each measurement_unit by dataset The strata differences (completely lost or appearing) between the first one and the second one (representing 4 % of the total number of strata) are : 7 Table 2: Disappearing or appearing strata between Initial-data and global-catch-ird-level2-1950-2023 Dimension measurement_unit Loss / Gain Precision Difference in millions fishing-fleetlabel Number of fish Loss Norway -2,297.0 Bulgaria (EU) -5,934.0 Tanzania, United Rep. of -6,418.0 Cyprus (EU) -8,802.0 Thailand -9,586.0 Greece (EU) -11,312.0 Malta (EU) -13,138.0 Libya -21,663.0 Italy (EU) -48,003.0 Tunisia -120,805.0 gear-type-label Number of fish Loss Gear nei -182,792.0 gear-type-label Tons Gain Longlines (nei) 511,108.7 species-label Number of fish Loss Narrow-barred Spanish mackerel -32.0 Mediterranean spearfish -35.0 Longtail tuna -6,400.0 Table 3: Comparison of number of stratas between the two datasets Initial-data global-catch-ird-level2-19502023 Difference Number of fishing-fleet-label 114 114 0 Number of fishing-mode-label 4 4 0 Number of gear-type-label 27 27 0 Number of geographic-identifier 12,266 12,266 0 8 Table 3: Comparison of number of stratas between the two datasets Initial-data global-catch-ird-level2-19502023 Difference Number of gridtype 2 2 0 Number of measurement-unit 2 2 0 Number of source-authority 5 5 0 Number of species-group 6 6 0 Number of species-label 65 65 0 Number of time-start 880 864 -16 2.1.1 Differences in temporal data Representing the differences in percent for each year. Number of fish Tons 1960 1980 2000 2020 −100 −75 −50 −25 −100 0 100 200 time_start Difference (in %) Figure 6: Difference in percent of value for the dimension time-start for Initial-data and global-catch-ird-level21950-2023 dataset 9 2.2.2 Spatial coverage We represent spatial coverage, faceted by geographical category. The geographical category depends on the area of the geographic polygon. In this case there are 2 categories which are 1deg_x_1deg ; 5deg_x_5deg. 1deg_x_1deg5deg_x_5deg global−catch−ird−level2−1950−2023 Initial−data measurement_value 500,000 1,000,000 1,500,000 2,000,000 2,500,000 3,000,000 Missing Figure 9: Distribution in value for the unit: Tons 16 1deg_x_1deg5deg_x_5deg global−catch−ird−level2−1950−2023 Initial−data measurement_value 10 mln 20 mln 30 mln 40 mln 50 mln 60 mln 70 mln 80 mln 90 mln 100 mln 110 mln Missing Figure 10: Distribution in value for the unit: Number of fish 17 2.2.3 Coverage for other dimensions We check the distribution of the value of each dimension and each unit. 50% 22% 26%0% 2% 52% 17% 14% 16% 0% Number of fish Tons Initial−data 67% 10% 19% 1% 4% 38% 28% 18% 16% 0% Number of fish Tons global−catch−ird−level2−1950−2023 SOURCE_AUTHORITY CCSBT IATTC ICCAT IOTC WCPFC Figure 11: Distribution in value for the dimension: source.authority 18 28% 13% 18% 7% 34% 46% 4% 5% 7% 37% Number of fish Tons Initial−data 22% 22% 8% 43% 5% 6% 32% 9% 43% 10% Number of fish Tons global−catch−ird−level2−1950−2023 FISHING_FLEET_LABEL Japan Korea, Republic of Others Spain (EU) Taiwan Province of China Ecuador Other nei China Figure 12: Distribution in value for the dimension: fishing.fleet.label 50%50% 27% 20% 35% 18% Number of fish Tons Initial−data 33% 67% 20% 14% 51% 15% Number of fish Tons global−catch−ird−level2−1950−2023 FISHING_MODE_LABEL Other set types combined Undefined school Free school Log school Figure 13: Distribution in value for the dimension: fishing.mode.label 19 97% 0%0%0% 3% 99% 0%0% 0%0% Number of fish Tons Initial−data 92% 0%0% 0% 7% 98% 1% 0%0% 1% Number of fish Tons global−catch−ird−level2−1950−2023 SPECIES_GROUP CARCHARHINIFORMES LAMNIFORMES Others PISCES MISCELLANEA SCOMBROIDEI Figure 14: Distribution in value for the dimension: species.group 23% 4% 9% 17% 47% 52% 29% 6% 8% 5% Number of fish Tons Initial−data 21% 19% 7% 11% 43% 40% 28% 17% 10% 6% Number of fish Tons global−catch−ird−level2−1950−2023 SPECIES_LABEL Albacore Bigeye tuna Others Swordfish Yellowfin tuna Skipjack tuna Blue shark Figure 15: Distribution in value for the dimension: species.label 20 5% 49% 4% 6% 35% 67% 8% 6% 13% 6% Number of fish Tons Initial−data 64% 2% 4% 28%3% 53% 10% 8% 11% 18% Number of fish Tons global−catch−ird−level2−1950−2023 GEAR_TYPE_LABEL Drifting longlines Handlines and hand−operated pole−and−lines Others Set longlines Trolling lines Purse seines Drift gillnets Longlines (nei) Figure 16: Distribution in value for the dimension: gear.type.label 21 3 Analyse of the processing of the data The following table recap all the treatment done on the mapped and standardized rawdata provided by tRFMOs. Table 5: Evolution of captures in tons and number of fish during the process Step Millions of tons Millions of fish Difference (in % of tons) Difference (in % of fish) Percentage of nominal Conversion factors (kg) Step number rawdata 148.61 992.56 0.00 0.00 60.5 750.00 1 iattc raised for billfish and shark 148.64 992.64 0.02 0.01 60.5 291.02 2 overlap-iattcwcpfc 148.64 992.39 0.00 -0.03 60.5 3 overlap-iotcwcpfc 148.63 992.29 0.00 -0.01 60.5 49.26 4 Removing-datawith-no-nominal 148.50 991.87 -0.09 -0.04 60.5 306.04 5 Conv-NOnominal1 164.50 425.17 10.77 -57.13 67.0 28.23 6 RF-pass-1 176.10 425.17 7.05 0.00 71.7 7 Conv-NOnominal2 176.10 425.17 0.00 0.00 71.7 8 RF-pass-2 176.27 425.17 0.10 0.00 71.8 9 Conv-NOnominal3 176.27 425.16 0.00 0.00 71.8 64.04 10 RF-pass-3 176.41 425.16 0.08 0.00 71.8 11 Conv-NOnominal4 176.41 425.16 0.00 0.00 71.8 7.37 12 RF-pass-4 177.10 425.16 0.39 0.00 72.1 13 Conv-NOnominal5 180.90 393.02 2.15 -7.56 73.6 118.33 14 RF-pass-5 186.06 393.02 2.85 0.00 75.7 15 Conv-NOnominal6 195.59 290.67 5.12 -26.04 79.6 93.15 16 RF-pass-6 198.50 290.67 1.49 0.00 80.8 17 Conv-NOnominal7 199.22 290.66 0.36 0.00 81.1 204,170.49 18 RF-pass-7 205.59 290.66 3.19 0.00 83.7 19 Conv-NOnominal8 206.40 282.45 0.40 -2.83 84.0 99.66 20 RF-pass-8 213.32 282.45 3.35 0.00 86.8 21 4 Annexe 22 0 250 500 750 1000 rawdata iattc raised for billfish and shark overlap_iattc_wcpfc overlap_iotc_wcpfc Removing_data_with_no_nominal Conv_NO_nominal1 RF_pass_1 Conv_NO_nominal2 RF_pass_2 Conv_NO_nominal3 RF_pass_3 Conv_NO_nominal4 RF_pass_4 Conv_NO_nominal5 RF_pass_5 Conv_NO_nominal6 RF_pass_6 Conv_NO_nominal7 RF_pass_7 Conv_NO_nominal8 RF_pass_8 Step Millions of fish 0 50 100 150 200 rawdata iattc raised for billfish and shark overlap_iattc_wcpfc overlap_iotc_wcpfc Removing_data_with_no_nominal Conv_NO_nominal1 RF_pass_1 Conv_NO_nominal2 RF_pass_2 Conv_NO_nominal3 RF_pass_3 Conv_NO_nominal4 RF_pass_4 Conv_NO_nominal5 RF_pass_5 Conv_NO_nominal6 RF_pass_6 Conv_NO_nominal7 RF_pass_7 Conv_NO_nominal8 RF_pass_8 Step Millions of tons Figure 17: Evolution of captures in tons and number of fish during the process Table 6: Review of all the impact and purpose of every the treatment done Step Explanation Functions rawdata The georeferenced data of the included tRFMOS, are binded. get_rfmos_datasets_level0 iattc raised for billfish and shark The effort is expressed here in terms of number of sets. This means in the case of the EPO get_rfmos_datasets_level0 overlap_iattc_wcpfc The georeferenced data present on the overlapping zone between IATTC and WCPFC is handled.The option for the strata overlapping allows handling the maximum similarities allowed between two data to keep both. In the case the data is identical on the stratas provided, the remaining data is from IATTC function_overlapped overlap_iotc_wcpfc The georeferenced data present on the overlapping zone between IOTC and WCPFC is handled.The option for the strata overlapping allows handling the maximum similarities allowed between two data to keep both. In the case the data is identical on the stratas provided, the remaining data is from IOTC function_overlapped Removing_data_with_no_nominal Since the nominal catch dataset does not cover every year, and the georeferenced data for the first years are not complete, raising would only apply to certain years and/or raising would be not accurate for first years. To avoid mixing raised and unraised data, we prefer to remove records for years that have no equivalent in the nominal dataset. 23 Table 6: Review of all the impact and purpose of every the treatment done Step Explanation Functions Conv_NO_nominal1 The data that remains in Number of Fish, for which the entirety of the strata with the following dimensions:gear_type, species, year, source_authority, fishing_fleet, geographic_identifier_nom, fishing_modecontaining catch information in tons, is converted and raised using the nominal dataset global_nominal_catch_firms_level0_2025.csv. The key identifier for this operation is: global_nominal_catch_firms_level0_2025.csv. This process relies on the fact that for a strata reported in both number and tons, the spatial footprint of the data in number is more often containing the spatial footprint of the data in tons.Then, as there is need to choose one of the measurment_unit for the raising and with limited conversion factors, we choose to keep raise the data in ’Number of fish’ on the basis of the equivalent nominal strata downgraded by the corresponding georeferenced catch in tons. RF_pass_1 Pass 1/8 on gear_type, species, year, source_authority, fishing_fleet, geographic_identifier_nom, fishing_mode. Tons raised 11599281.844, decreased 0.000. Numbers raised 0.000, decreased 0.000. No down-scaling was applied on this pass. iterative_raising Conv_NO_nominal2 The data that remains in Number of Fish, for which the entirety of the strata with the following dimensions:gear_type, species, year, source_authority, fishing_fleet, geographic_identifier_nomcontaining catch information in tons, is converted and raised using the nominal dataset global_nominal_catch_firms_level0_2025.csv. The key identifier for this operation is: global_nominal_catch_firms_level0_2025.csv. This process relies on the fact that for a strata reported in both number and tons, the spatial footprint of the data in number is more often containing the spatial footprint of the data in tons.Then, as there is need to choose one of the measurment_unit for the raising and with limited conversion factors, we choose to keep raise the data in ’Number of fish’ on the basis of the equivalent nominal strata downgraded by the corresponding georeferenced catch in tons. RF_pass_2 Pass 2/8 on gear_type, species, year, source_authority, fishing_fleet, geographic_identifier_nom. Tons raised 169448.342, decreased 0.000. Numbers raised 0.000, decreased 0.000. As one or multiple dimensions are removed in this pass for the raising, the raising is only done on the corresponding unspecified data of the stratas. The strata having a ’perfect’ match in the nominal is not included in this raising.No down-scaling was applied on this pass. iterative_raising 24 Table 6: Review of all the impact and purpose of every the treatment done Step Explanation Functions Conv_NO_nominal3 The data that remains in Number of Fish, for which the entirety of the strata with the following dimensions:species, year, source_authority, fishing_fleet, geographic_identifier_nom, fishing_modecontaining catch information in tons, is converted and raised using the nominal dataset global_nominal_catch_firms_level0_2025.csv. The key identifier for this operation is: global_nominal_catch_firms_level0_2025.csv. This process relies on the fact that for a strata reported in both number and tons, the spatial footprint of the data in number is more often containing the spatial footprint of the data in tons.Then, as there is need to choose one of the measurment_unit for the raising and with limited conversion factors, we choose to keep raise the data in ’Number of fish’ on the basis of the equivalent nominal strata downgraded by the corresponding georeferenced catch in tons. RF_pass_3 Pass 3/8 on species, year, source_authority, fishing_fleet, geographic_identifier_nom, fishing_mode. Tons raised 136345.900, decreased 0.000. Numbers raised 0.000, decreased 0.000. As one or multiple dimensions are removed in this pass for the raising, the raising is only done on the corresponding unspecified data of the stratas. The strata having a ’perfect’ match in the nominal is not included in this raising.No down-scaling was applied on this pass. iterative_raising Conv_NO_nominal4 The data that remains in Number of Fish, for which the entirety of the strata with the following dimensions:gear_type, species, year, source_authority, geographic_identifier_nom, fishing_modecontaining catch information in tons, is converted and raised using the nominal dataset global_nominal_catch_firms_level0_2025.csv. The key identifier for this operation is: global_nominal_catch_firms_level0_2025.csv. This process relies on the fact that for a strata reported in both number and tons, the spatial footprint of the data in number is more often containing the spatial footprint of the data in tons.Then, as there is need to choose one of the measurment_unit for the raising and with limited conversion factors, we choose to keep raise the data in ’Number of fish’ on the basis of the equivalent nominal strata downgraded by the corresponding georeferenced catch in tons. RF_pass_4 Pass 4/8 on gear_type, species, year, source_authority, geographic_identifier_nom, fishing_mode. Tons raised 693159.520, decreased 0.000. Numbers raised 0.000, decreased 0.000. As one or multiple dimensions are removed in this pass for the raising, the raising is only done on the corresponding unspecified data of the stratas. The strata having a ’perfect’ match in the nominal is not included in this raising.No down-scaling was applied on this pass. iterative_raising 25