The WKD3C3THRESHOLDS meeting provided a platform for experts from the EU member states to meet and progress the assessment methodology on Criteria 3 of Descriptor 3 upon re-quest by EC (DGENV). WKD3C3THRESHOLDS is the second of a series of three workshops (WKD3C3SCOPE, WKD3C3THRESHOLDS and WKD3SIMUL) to identify operational indica-tors for MSFD D3C3.

The workshop was organised as a series of presentations of results with intermittent group dis-cussions.

The D3C3 indicators agreed at WKD3C3SCOPE were estimated and documented for a selection of stocks representing different life-histories (Tor a). Plots comparing indicators were investi-gated for stocks with all estimated indicators. The age structure indicators ABI, ASA, POS and SSB/R generally followed the temporal development of SSB and react similarly to F. A gap of up to 10 years was observed between changes in F and subsequent changes in age structure indica-tors for long-lived species while SSB responded quickly to changes in F. For medium-lived stocks, the four age structure indicators exhibited similar temporal patterns, with SSB divided by R tending to be more variable. Recruitment and mean weight at age documented shifts in productivity, impacting age structure indicators differently when changes occur. Plots of F, re-cruitment, weight at age and SSB are considered useful for understanding cases where changes in F do not impact SSB as expected (e.g. rebuilding does not occur or stock remains high in spite of high F). Higher proportions of older fish as measured by ABI/ASA/POS or SSB/R did not ap-pear linked to an immediate increase recruitment. A comparison of length-based and age-based indicators for Mediterranean stocks was also conducted. Recruitment detection from survey time series showed uneven patterns over stocks and time series, and in some cases depended on sur-vey timing. Length-based indicators exhibited weak consistency information from stock assess-ments, and confounding effects of biological variability and sampling timing on observed re-cruitment pulses were noted. The indicator L90R, calculated from the length-frequency distribu-tion of fish larger than recruiting length, seemed to perform well among those inspected.

Thresholds for the D3C3 indicators for stocks representing different life-histories, data availabil-ity and MSFD (sub)regions (when possible) were discussed (ToR b and c). The suggested thresh-olds covered all approaches identified by WKD3C3SCOPE. Clear thresholds where the indicator signifies stock productivity declines could not be identified from the data as none of the age structure indicators showed a positive correlation with stock productivity. As a result, threshold levels cannot be determined based on levels at which stock productivity is either impaired or enhanced. In the absence of clear relationships between the indicators and stock health, the work-shop used varying percentages (10th percentile, median/50th percentile) of the simulated or ob-served distributions of indicators to determine good status of the indicator. The analyses pre-sented emphasized the direct influence of recruitment and growth on fisheries yield and precau-tionary fishing mortality limits. Finally, a decision tree to choose a threshold setting method was proposed for further testing in WKSIMULD3 on the basis of listed pros and cons discussed by WKD3C3THRESHOLD participants.

The SSB/R indicator responded to recruitment in an undesirable manner but there was insuffi-cient evidence to determine which of the three remaining age structure indicators provided a higher signal to noise due to recruitment variability. Selectivity indicators under D3C3 were re-tained despite unclear guidance in the MSFD guidance document. The retained indicators for medium-lived stocks with age-based assessment data include ABI, POS, ASA, R, ASW, and Fjuv/Fapical. The value of age structure indicators as management indicators was unclear for short-lived and long-lived species. For short-lived species, no strong link was found between age-structure indicators and F or SSB, and high age at spawning may lead to senescence rather than increased viability of spawning products. For long-lived species, age structure indicators appeared to react substantially later than F and SSB, making their added value for management unclear. The definition of thresholds for these indicators will be further investigated in WKSIM-ULD3.

The assessment of stock health under D3C3 relies on crucial data such as recruitment, weight at age, and size/age distribution (ToR d). In the absence of this information, D3C3 assessments can-not be conducted, and Member States were encouraged to enhance data collection efforts. For stocks with age-based assessments, these data are considered essential input and/or output for the assessment, and assessments based on age data are preferred over those based solely on length distributions for the estimation of age structure indicators.

Finally, the group drafted a list of actions to be completed for the reparation of WKSIMULD3.