A summary of the results of the AMSY workshop held at GEOMAR in Kiel, Germany on 23-25 April 2019, as reported by Rainer Froese, can be found below.
The workshop was kindly sponsored by the Minderoo Foundation. Participants were Rainer Froese, GEOMAR (Chair); Nina Garilao, GEOMAR (Org.); Daniel Pauly, UBC, Canada (remote participation); Gianpaolo Coro, CNR, Italy (remote participation); Henning Winker, DAFF, South Africa; Nazli Demirel, IU, Turkey; Athanassios Tsikliras, AUTH, Greece; Donna Dimarchopoulou, AUTH, Greece (remote participation); Giuseppe Scarcella, CNR-IRBIM, Italy; Deng Palomares, UBC, Canada; Manuel Dureuil, DU, Canada.
The goal of the workshop was to further develop and test a new method (AMSY) for assessment of data-limited stocks. AMSY estimates fisheries reference points (Fmsy, F/Fmsy, B/Bmsy) from abundance index data (such as catch-per-unit-of-effort = CPUE) plus priors for resilience (available in FishBase and SealifeBase for most commercial species) plus a prior for stock status (B/B0) anywhere in the time series (from expert knowledge or estimated from length frequency data with LBB). AMSY is the first method to only work with these inputs. It is expected to provide first-ever assessments for more than 200 stocks globally.
The workshop started with a review of the logic behind AMSY and the results of previous simulations and evaluations against 128 real stocks. These were deemed satisfactory. Several improvements were then discussed and implemented:
- a better prior for carrying capacity for stocks with very low resilience — a better implementation of process and observation error
- better handling of stocks where CPUE and catches were close to zero for a number of years
- output of a Kobe plot such as used by managers
- addition of a retrospective analysis to test robustness of results (i.e., how do results change if fewer years with CPUE are available)
The new AMSY version was tested against regional data provided by the participants. For those stocks where independent assessment with more data as inputs were available, the data-limited AMSY results were similar. But more excitingly, from the regions represented by the participants, there will be first-ever assessments for 50-100 stocks. These are typically bycatch species or other overlooked stocks and quite a number were severely depleted. It was decided to include these in the planned AMSY paper, so that these assessments can be directly used for MSFD (= mandatory determination of good environmental status in the EU) and for IUCN Red-List assessments.
Participants agreed to publish AMSY in the ICES Journal of Marine Science (fast processing, good acceptance by managers), aiming for submission before the 2019 summer break (= latest mid July). AMSY will also be presented (together with CMSY and LBB) at a data-poor methods section at the ICES Annual Conference in September this year. Participants agreed on the need to develop a more user-friendly graphical interface (GUI) to the suite of methods (CMSY, LBB, AMSY) for use in training courses and by others, and on the need for better integration, version control, and distribution.
An unexpected but highly welcome distraction was the development of a new stand-alone tool (BCrumb) that uses a state-space Bayesian approach
to fill gaps (missing years) in CPUE time series and to combine CPUEs from different areas or gears and of different length, a major problem in stock assessment in general. Users can specify their degree of trust in the different CPUE sources and BCrumb produces a harmonized CPUE time series with confidence limits. BCrumb will be very useful not only for AMSY but also for CMSY, which combines catch with CPUE data if available.
Participants agreed on a need for a follow-up workshop, to migrate new developments to CMSY and to follow up on some of the other issues raised. This workshop is planned for 6-8 November 2019 in Thessaloniki, preceding the summer school on CMSY for managers from the region planned for the subsequent week by AUTH.
Click here to download Froese’s presentation.
[Report written by Rainer Froese, GEOMAR]
