Workshop%20on%20use%20of%20Ageing%20and%20Maturity%20Staging%20Error%20Matrices%20in%20Stock%20Assessment%20(WKAMEMSA).pdf (1.69 MB)Download file
Workshop on use of Ageing and Maturity Staging Error Matrices in Stock Assessment (WKAMEMSA; outputs from 2021 meeting).
reportposted on 2022-02-17, 00:00 authored by ICESICES
The Workshop on use of Ageing and Maturity Staging Error Matrices in Stock Assessment (WKAMEMSA) aimed to compile previous experiences with incorporating age and maturity error information in the assessment and scientific advice process within ICES, create protocols to quantify the observation error in age and maturity data used as input data in the assessment models, identify the lines of development required to incorporate that information, and select candidate stocks and models to be used as case studies and progress in each of the steps related to age and maturity error that affect the final uncertainty in the assessment. The approaches followed by three of the main assessment models used in ICES (SAM, SS3 and Gadget) to incorporate ageing errors in the assessment, data requirements, as well as the steps to consider in the organization of age reading and maturity staging exchange events were addressed during the workshop. The revision of previous research studies showed that not accounting for ageing error in stock assessment produces bias in the estimation of the Spawning-Stock Biomass (SSB) and smoothed estimates of annual recruitment, which can lead to under/overestimation of fishing mortality in the short-term forecast. Another group of studies proposed different approaches to incorporate ageing errors in stock assessment by using use the information from exchange events to calculate model-based Ageing Error Matrices (AEM) that could be incorporated into stock assessment models.SAM, SS3 and Gadget can use and incorporate ageing error information in the stock assessment, however none of these assessment models include the possibility of considering errors in the maturity process at this stage. These models can use the output from age exchange events to produce internally model-based AEMs during the model optimization. The results of preliminary studies with these assessment models were presented. Alternative approaches to use the exchange event information and produce AEM inside and outside the assessment models were discussed. The need of increasing the communication between coordinators of calibration exercises and stock assessors was highlighted; in order to assess if a given stock requires a tailored design of exchange events, data input format and model structure. The organization of regular calibration exercises with adequate sample size is required to assess the potential changes in ageing errors over time. Finally, a selection of stocks was suggested to be used as case studies for SAM, SS3 and gadget. It is expected that these case studies serve to set guidance on the best practices throughout the entire process, from the organization of exchange events to the use of their outputs in the stock assessment during the benchmark process.
Published under the auspices of the following ICES Steering Group or Committee