Report of the Workshop on Age Reading of European Atlantic Sardine (WKARAS)

The Workshop on Age reading of European Atlantic Sardine (WKARAS) was held in Lisbon, Portugal, from 14 to 18 February 2011.
The terms of reference of this Workshop were addressed as specified in Sub-section 1.1.
There were 11 participants in the Workshop from five institutes located in France, Spain and Portugal (Sub-section 1.2). Seven participants had more than 4 years of experience in sardine age reading and are responsible of providing age determinations for the assessment of the Iberian sardine stock (Sardine in ICES Sub-Areas VIIIc and IXa) and for the description of the catch and survey age structure of sardine from the Bay of Biscay within the ICES Working Group on Anchovy and Sardine (WGAN-SA). The final Agenda of the Workshop is presented in Annex 1. The activities carried out in the different laboratories regarding otolith sampling, storing and reading methods are presented in Section 3.
This Workshop was preceded by an otolith exchange which was undertaken from September to December 2010 in order to prepare material and data for the workshop (Section 4, Annex 3). Eight readers participated in this exchange, one of them as a guest from the UK with no experience in sardine otolith age reading and therefore whose results were not included in the final analysis of the exchange data.
The otolith exchange included a total of 300 otoliths from samples collected in 2008 spanning the area from ICES Sub-Area VIIIa (Bay of Biscay North) to Sub-Division IXa-South-Cadiz (Gulf of Cadiz). They were prepared and interpreted according to the guidelines in Soares et al. (2007).
The relative accuracy of sardine age determination was generally good: the average percentage of agreement with modal age was 77.0% and 75.2% for the Iberian Stock and the Bay of Biscay respectively and average bias was ca. 0.03 years. Precision was substantially higher in the Bay of Biscay (CV= 14.1%) than that in the Iberian stock area (CV=32.8%), although the latter is strongly influenced by high CVs at age 0 in the Gulf of Cadiz. By excluding age 0, the precision in the Iberian Stock area becomes similar to that in the Bay of Biscay (14.7%), being still above the 5% threshold recommended in the PGCCDBS Guidelines for Workshops on Age Calibration.
Compared to the previous Workshop, relative accuracy of sardine age determination within the Iberian Stock area (Cantabrian Sea and South Iberian peninsula areas) has improved substantially and there is less evidence of bias (% of agreement increased ca. 20%) while precision increased in the Cantabrian Sea (CV declined from 17.2 to 15.7%) but decreased in south Iberia (CV increased from 17.6 to 22.0%) (Section 8). In the Bay of Biscay, the relative accuracy was maintained but the precision declined (CV increased from 10.3 to 18.6%).
The identification of the otolith edge and of the first annual ring were the main discrepancies between readers in sardine age determination, accounting for 25-53% and 32-43% of the total number of discrepancies in otolith interpretation, respectively(Section 6). These are recurrent issues in sardine age determination (e.g. ICES, 1997; Soares et al., 2007).
The occurrence of a first wide clear translucent ring around the whole otolith slightly closer than typically to the nucleus is generally a source of disagreement among readers, leading part of them to consider it as the first annual ring and the others not.
The results of an on-going study of otolith seasonal growth showed that the seasonality of the otolith edge varies with fish age, with older individuals forming an opaque edge later in the year for a shorter period as observed in several other clupeids (e.g. anchovy, ICES, 2009a). Marginal growth was also different between the Bay of Biscay and the Portuguese waters, with a longer period of opaque growth in Age 1 individuals and more similar seasonal patterns for Ages 2-4 years in the former area.
The attribution of age in years of young individuals was corroborated by preliminary results from the analysis of daily growth rings of sardine juveniles in northern Portugal. The diameter of the first annual ring was positively correlated with the fish length at the time of formation. Measurements of ring diameter indicate that the first translucent ring forms at a distance around 1.1 mm from the otolith nucleus, corresponding to a diameter around 2 mm. The radius of the potential check did not over-lap  with  the  radius  of  the  true  ring  (as  considered  by  some  readers)  but  measurements were very close and certainly impossible to discern just by observation of the otolith.
Guidelines related to the age reading problems were modified and their impact on age reading was evaluated by a second reading of a sub-set of the exchange otoliths (Section 7). Changes in relative accuracy and precision between the first and second readings differ among areas: agreement was similar while precision declined substantially in the Gulf of Biscay, generally worse results were obtained for western Portugal and generally better results were obtained for the Gulf of Cadiz.
Although the results available to the workshop did not fully clarify the major age reading uncertainties, the participants considered that sardine age reading would be improved by taking into account age related differences in edge seasonality and by using a flexible gauge to guide the identification of the first annual ring. These changes were accommodated in the age reading protocol (Section 10). Conversely, the present birthdate assumption, 1st of January, was considered to be well adapted to the spawning seasonality and growth pattern of sardine across the area.
A reference collection of annotated (catch area and date, fish length and location of modal rings) otolith digital images was assembled (see Section 9 and image examples in Annex 4). The collection contains 139 otoliths covering the area from northern Gulf of Biscay to the Gulf of Cadiz and Age groups 0-8 years.
Recommendations and follow up actions are presented in Section 11. The Workshop recommends that Age reading Workshops take place regularly at 4-5 years intervals, with participants experienced in the Iberian Stock area, in the Bay of Biscay and in ICES Div. VIIe.
Also the revision of procedures to calculate CV’s by age group, taking into account the problem of inflation observed for age group 0, is recommended.
The Workshop also supports the recommendation of WKARA 2009 to set up of a Workshop on daily ring methodology and interpretation and suggests this is ex-tended to sardine and eventually, to other small pelagic fish (e.g. herring, sprat).
In section 13 a list of the working documents and presentations to WKARAS can be found.