User menu
This report presents a stock assessment for Indian Ocean swordfish (Xiphias gladius) using Stock Synthesis 3 (SS3). The assessment uses a spatially disaggregated, sex explicit, and age structured model that includes several sources of fisheries and biological data. The assessment model covers the period 1950–2021 and represents an update and revision of the 2018 assessment model with the inclusion of updated longline CPUE indices and length composition data. A range of sensitivity models are also presented to explore the impact of key data sets and model assumptions.
The assessment assumed the Indian Ocean swordfish constitute a single spawning stock partitioned into 4 spatial areas (NW, NE, SW, and SE), to account for differential abundance and depletion levels among regions. The assessment model defined 15 fisheries, based on fleet and region. Standardised CPUE series (as relative abundance indices) are available from the Japanese, Taiwanese，Portuguese, Spanish, South African, and Indonesia longline fleets. The reference model focused more on the Japanese (for 4 regions), Portuguese (SW), and South African (SW) indices, assuming that they are proportional to the regional abundance of swordfish. Final models also include alternative configurations on more CPUE index, replacing more recent index suing CPUE indies from fleets in each region. Another key difference made to the previous assessment model is that the selectivities for the Japanese fleet were estimated for each region separately given the spatial variability in the length compositions, rather than assumed to be common for all regions, as in the previous assessment.
The final assessment model options correspond to a combination of model configurations, including alternative CPUE configurations for recent years (Japanese/Portuguese/South Africa index, or Japanese/Taiwanese/Indonesia index) , assumptions on growth (otolith or spine based age estimates), alternative values of SRR steepness (0.7, 0.8, or 0.9), recruitment variability σR (0.2 or 0.4), and the alternative sample size of length composition data (20 or 5). The model ensemble (a total of 24 models) encompass a range of stock trajectories. Estimates of stock status were combined across from the 24 models and incorporated uncertainty estimates from both within and across the model ensemble.
The overall stock status estimates do not differ substantially from the previous assessment. Fishing mortality rates have remained relatively stable in recent years and are well below the FMSY. Biomass was estimated to have been increasing since 2010. Spawning stock biomass in 2018 was estimated to be 42% of the unfished levels and 170% of the level that can support MSY (SSB2018/SSBMSY = 1.75). With high likelihood, current fishing mortality was estimated to be lower than FMSY (F2018/FMSY = 0.6). The probability of the stock being currently in the green Kobe quadrant is estimated to over 95%. Considering the quantified uncertainty, the stock is considered not to be overfished and is not subject to overfishing in 2018. The retrospective analysis provided some confidence on the robustness of the model with respect to recent data, yet the uncertainly on levels of most recent recruitment may undermine the predictive capabilities of the model. Annual catches since the 1990s have been within the range of the estimated MSY and the current catch is projected to be sustainable over the longer term. the estimated stock status is summarized as below:
• Catch in 2021
23237
• Average catch 2017–2021
30 809
• MSY (80% CI)
29 856 (26 319–33 393)
• FMSY
0.16 (0.12–0.20)
• SB0
224 673 (200 328–249 019)
• SB2021:
75 891 (58 019–93 764)
• SBMSY
55 055 (40 243 –69 866)
• SB2021/SB0
0.35 (0.32–0.37)
• SB2021 / SSBMSY
1.39 (1.01–1.77)
• F2021 / FMSY
0.60 (0.43–0.77)
