Maritime Forum

Maritime Forum Themes


Study 2006-15 lot 9: Improved methodology for Cod age estimation

Published on: Fri, 21/05/2010 - 17:26
Table of Contents

    Assessing the Eastern Baltic Cod stock is difficult. One of the reasons is the problem of correct age-determination. This leads to problems with the catch-at-age and abundance indices of the surveys at sea.


    Why this study?

    The current way of determining the age of cod ( by interpreting the annual rings in the fish’ otoliths) is rather subjective and usually without proper validation. Especially interpreting the rings in Eastern Baltic Cod otoliths is difficult since cod in this area does not have clearly defined rings.

    Without a proper determination of the age of the fish in for example sea surveys has an effect on the estimate on both the age structure and the size of the stock.

    Therefore this study will look for alternative methods to determine age. For example in other stocks with age reading problems the weight of the otoliths has proven to be a very good alternative to age reading.


    1. Based on historic data form both commercial catches and surveys put together length distributions, several measurable characteristics of otoliths (such as weight) and biological facts and figures.
    2. Use these data to develop an alternative approach to determining the age of cod. Of course this new approach should not only work but also should be  robust.
    3. Use the new method to reconstruct the historic age composition of the commercial catches and surveys done in the past


    The background for this project is the longstanding problem with age-determination of Baltic Cod. The species’ prolonged spawning period, the heterogeneous hydrography of the Baltic Sea, coupled with the cod’s pronounced migration patterns, affect the otolith microstructure in such a way that the contrast between opaque and transparent zones (the counting of which is used for age determination) becomes rather weak.

    It was suggested to use the Length-mediated Mixture-model based on otolith

    biometrics, particularly otolith weight. The Mixture-model approach has been developed for Faroese cod. This approach aims at estimating the age distribution in a population, rather than individual ages. It involves two types of samples, a production sample, where fish length and otolith weight is known, and a calibration sample, where the fish’s age is also known.

    Overall the conclusion of this study is that this new method for age determination is working. Comparison of traditionally read ages with the fish’s true age indicated that the error in age determination may be up to 80 %. Error seems to be attributable to under-estimation with the traditional method, but the error is not consistent over ages or year classes. It became evident that in Eastern Baltic cod, transparent zones, which are counted as “annuli” are not formed synchronously but occur at widely different times of the year in individual fish. This shows that in this stock reliable age determination is not possible with the traditional ageing method.

    For the production sample data is available from all Baltic countries back to 2000, for some even to 1995. All otolith weight data have bee quality controlled and are ready to be used. For the calibration sample a new method for estimating the fish’s true age using otoliths micro-increment patters as indicator of annual temperature cycles was used. This method seems promising, even though the preparation procedure is time-consuming. A drawback with this method is that validation of annual increment patterns has to be validated before implementation is possible.

    The Length-Mediated Mixture Analysis proved to be easy to apply to Baltic cod, and to result in proportions at age and precisions of estimates that are very similar to those obtained by traditional age-length-key method. The approach makes better use of

    the population’s length distribution and may require a smaller calibration sample than the traditional method. From a economic perspective, the LMMA approach looks promising, but its full potential needs to be evaluated based on samples of the entire size range, and evaluated with a detailed cost benefit analysis.


    Full title: DECODE ImproveD mEthodology for Baltic COD age Estimation

    Organisations: DTU Aqua, Institute for Baltic Sea Fisheries, Swedish Fishery Board, Sea Fisheries Institute, Latvian Fish Resources Agency, Centre for Environment, Fisheries & Aquaculture Science.

    Full report: