Data
The data used for the assessment included total catch which was collected monthly by the governments from all processing plants. The whitebelly shrimp, which was part of the catch and which formed part of the ‘broken’ category in the catches, or the small sized seabob, was estimated and removed from the catch. Catch and effort data were used from spreadsheet forms used by Noble House Seafoods (Guyana) and Heiploeg Suriname (Suriname) to record landings and processing operations. It was found that number of trips was a better indicator of a trip’s catch than days-at-sea. Days at sea may not measure trawling effort well, since it included other activities, such as travelling to the fishing ground and search for the shrimp. A standardization method also used in 2013 was used to define effort and this resulted in a better relationship between landings (catch) and effort, but the measurement of fishing effort remained unsatisfactory.
Assess Models
Type: Age-structured
Bayesian size structured
OVERVIEW OF THE ASSESSMENT APPROACH - The stock assessment used an integrated assessment that utilized all available information in a single model to estimate the past stock dynamics and current status for seabob within the separate national boundaries of Suriname and Guyana. For the purposes of the present assessment, the CSWG has recommended to keep the assumption of separate stocks for Suriname and Guyana. The assumption of separate stocks will be kept under review and re-examined when results become available following the completion of doctoral research work on the population genetics of both stocks that was currently being undertaken by a PhD candidate from Ghent University, Belgium. The assessment consisted of a population model that described the dynamics, an observation model that calculated what would be the expected observed values of total catch, catch and effort and size composition derived from the population model and a likelihood model that linked the observation model to the data. The stock assessment was applied to both Guyana and Suriname fisheries with the same priors and model structure, but independent data. This allowed direct comparison between the assessments and should be informative on differences, if any, between the fisheries. The stock assessment was fully implemented in a Bayesian MCMC using Stan (mc-stan.org). Growth was modelled using a size transition matrix, which avoided maintaining the population in age categories and the excessive smoothing that occurred in converting weight to age or vice versa. A full summary of the changes made to the 2013 assessment approach is given in the assessment method summary report, the final version of which is given in Annex 1 to this report. - EXPLANATION OF USE OF DATA ON SEX AND MATURITY - A significant improvement in the model fit, compared to the 2013 assessment, was achieved by modelling not only maturity, but sex allocation as applied in the random sampling. The basis for this was that immature males were easily confused with immature females as external parts may not have been present (e.g. lost through damage or not yet formed through a moult). The maturity model was therefore included in the stock assessment model and data were provided in three categories: immature females, mature females and males. There was now a considerable data set linking female size (tail weight in grams) to maturity (presence of a “green vein”) in females. This allowed the maturity ogive to be estimated, which could be used to estimate spawning stock biomass within the stock assessment model. For the 2013 assessment, this was done externally to the stock assessment model. It was now done within the model because it was found that the size composition data could be best explained if some (estimated) proportion of immature females were actually males. In this context, maturity also increased the correct sex identification and so there was a need for maturity to be estimated within the model.
Results
RESULTS SUMMARY - RECOMMENDATIONS:
• The stock assessment model and HCR are heavily dependent on measures of fishing effort. New data are available from VMS to evaluate this. This evaluation should be carried out with a matter of urgency.
• Combine data for 2018 to generate complete trip information.
• An important part of the assessment is estimation of selectivity and catchability. More information on selectivity and catchability would be useful to better determine how to represent changes in the time series. Although flexible, the cubic spline currently used is inherently unstable. Preliminary attempts to change selectivity failed to obtain a fit, suggesting that parameters will be difficult to estimate if selectivity is modelled in this form. Although there may be some justification for changing selectivity with a changing inner line and introduction of BRD, it is not strongly supported by residual patterns. This would need more time and resources to explore properly.
Sci Advice
RECOMMENDATIONS:
1. Logbook data for 2018 to be digitized. The digitized logbook data along with the VMS and CCTV data should assist in improving understanding of the catch rates, selectivity patterns and CPUE, as well as the spatial fishing patterns in the two countries. A Sub Working Group (P. Medley, D. Maison, R, Bumbury, T. Willems & Y. Babb-Echteld) was established to undertake the following tasks: (i) look at samples of the logbooks and develop a standardized spreadsheet for logbook data to be digitized; (ii) examine available data and develop list of trips to be included in the 2018 data set; (iii) make arrangements for digitization of the data in agreed format. The Sub Working Group will commence its work the week of 30 September 2019. VMS and CCTV data for 2018 should be transmitted to Dr. Medley.
2. The agreed HCR be implemented and performance monitored for a year then reviewed and adjustments made for practicality purposes, as necessary. Dr. Medley will prepare a first draft of a standardized spreadsheet for tracking HCR performance by 01 October 2019, and share with CSWG for review.
3. The data limited method presented should be further developed so that an alternate method of analysis would be available for comparison, which would serve to broaden understanding and appreciation of the results and possibly support better convergence of ideas and hence results.
4. Consideration to be given to the impacts of Sargassum and climate change on fisheries and how these impacts could be incorporated into management decision-making in the future.
5. Local scientific sampling should be done to inform development of a suitable maturity ogive and to determine if the ‘Castilho’ ogive (Castilho et al., 2015) as currently used in the stock assessment remains applicable. In this regard, the value of M should also be kept under review in the immediate future.
6. A chronological table should be developed (both for Suriname and Guyana) providing information on changes in fishing gear and/or fishing practices (e.g. introduction or changes to BRDs and TEDs) that might influence catchability and selectivity, in order to better understand the observed differences in selectivity between Suriname and Guyana.
7. Similarly, management actions should be documented (and preferably also reconstructed from the past), including dates of implementation. This could be very useful in explaining changes in CPUE patterns.
8. The exact composition of the ‘broken shrimp’ category as reported by the peeling plants should be assessed as it is currently unclear what shrimp sizes are contained within this group.
9. Dr. Medley and the CSWG acknowledged that future stock assessment efforts should preferably be done less remotely and would benefit from increased interaction and input from local stakeholders during the assessment process.
to turn layers on and off 
