Report of the Atlantic bluefin tuna otolith ageing workshop Workshop time and place

Report of the Atlantic bluefin tuna otolith ageing workshop

The Atlantic bluefin tuna otolith ageing workshop was held April 7 and 8, 2011 at the Large Pelagics Research Center in Gloucester, MA.

The goal of this workshop was to standardizing ageing methods for Atlantic bluefin tuna (ABFT) otoliths and to establish methods to ensure reader precision.

Robert Allman, NOAA Fisheries, Panama City, FL

Dheeraj Busawon, Fisheries and Oceans Canada, St. Andrews, New Brunswick, Canada

Ben Gahagan, Chesapeake Biological Laboratory, University of Maryland, Solomons, MD

Walter Golet, University of Maine School of Marine Sciences, Portland, ME

John Hoolihan, NOAA Fisheries, Miami, FL

Jessie Knapp, Large Pelagic Research Center, and University of New Hampshire, Durham, NH

Peter Abdu, Large Pelagic Research Center and University of New Hampshire, Durham, NH

Patricia Lastra, Spanish Institute of Oceanography, Santander, Spain

Molly Lutcavage, Large Pelagics Research Center, University of Massachusetts Amherst, Gloucester, MA

John Neilson, Fisheries and Oceans Canada, St. Andrews, New Brunswick, Canada

Enrique Rodriguez-Marin, Spanish Institute of Oceanography, Santander, Spain

Dave Secor, Chesapeake Biological Laboratory, University of Maryland, Solomons, MD

Thursday April 7

1. Otolith processing methods

After introductions, the workshop began with a discussion of the otolith preparation methods employed by the U.S. and Canada.

Robert Allman gave a short presentation describing the methods used by NOAA Fisheries Panama City derived from approaches occurring in John Neilson’s and Dave Secor’s laboratories. The following steps were described.

1. 
   Prior to sectioning an image is taken of the whole sagittal otolith (hereafter referred to as otolith) sulcus (proximal) side down using reflected light on a black background. The longest and widest axes are measured and a weight recorded.
   
2. 
   A small portion of epoxy mixed with hardener is poured into petroleum jelly lined silicone molds for small otoliths or ice cube trays for large otoliths and allowed to dry for approximately 5 hours in a vacuum aspirator.
   
3. 
   The otolith sulcus side down is placed on the dry epoxy and covered with another mixture of epoxy and hardener and allowed to dry.
   
4. 
   The core is marked on the embedded otolith with a pencil and a transverse section is removed using a low-speed saw equipped with two 4 inch blades with a 0.7-1.0 mm spacer.
   
5. 
   The single section is mounted on a geological slide using thermoplastic cement, ground down using 240-600 grit sandpaper to obtain a 0.5 to 0.3 mm section and polished with 0.3 micron polishing compound. Care should be taken not to grind down the section too much since the overall pattern can be lost among sub-annual increments.
   

This method is slightly modified to obtain a sample for micro-constituent analysis prior to ageing. Dave Secor said that for this purpose, a relatively thick section (1.5-2 mm) is cut and the core milled to obtain the sample, then the remaining section is ground down for ageing as described above. Approximately 700 microns thickness within the core region are required to supply sufficient material for elemental analysis.

Dheeraj Busawon described the processing method employed by Fisheries and Oceans Canada. The otolith is embedded in resin similar to the description above and the embedded otolith is sectioned using four blades with spacers. Three 0.5 mm serial sections are polished using 30 and 0.3 micron lapidary film. These methods are summarized in detail in an otolith processing protocol document produced by the Campana and Neilson laboratories.

There was discussion of which area of the otolith was optimum to section to obtain the core. In the past, the Secor lab sectioned using the tip of the antirostrum as a guide, while the Neilson lab sectioned slightly posterior to this region. Dave Secor pointed out the concentric daily increments in the core of his lab’s sections. These sections were roughly V shaped while the sections more posterior were Y shaped (Figure 1). John Neilson commented that for the serial sections, one section could be used for ageing and one for extraction of the core material for elemental analysis.

In order to determine if annuli counts from the different sectioning planes (i.e., anterior and posterior of the antirostrum) are consistent, it was recommended that annuli counts be compared on the ventral arm of the otolith from different serial sections. Dheeraj Busawon and Ben Gahagan offered to compare counts from approximately forty otoliths using a paired t-test, and circulate the results to participants.

2. Microscopic techniques

The group discussed the advantages and disadvantages of transmitted versus reflected light and viewed otoliths using both methods. In the past the Secor lab has used transmitted light while the Neilson lab has used reflected light. It was agreed that both methods could be useful, but reflected light might allowed for better manipulation of the light direction for viewing the otolith. John Neilson mentioned that software was currently available that allowed for multiple digital images from different focal planes to be merged into a single image, and such new technologies may have application for our work.

Images of otolith sections were compared using transmitted and reflected light. Dheeraj Busawon demonstrated methods using Photoshop software to enhance annuli. These methods are described in the Campana and Neilson lab’s otolith processing protocol document. Walter Golet mentioned that Image J software was available online for free and could be used to make measurements. Walt also commented that the group could share images and documents through an ftp type server or some other method to share large files.

It was agreed that the ventral (long) arm of the sectioned otolith was the most reliable path for counting annuli and that the dorsal (short) arm could sometimes be used for verification of counts made on the ventral arm, but should not be relied on exclusively. Crenulations along the ventral margins may be useful in locating annuli (particularly at younger ages), but should not be relied upon solely since they do not occur in all otoliths. The region most difficult to distinguish annuli was after the first inflection point, generally comprising the first five annuli.

There was a discussion of using the first and second inflection points as guides for locating annuli as described in: A manual for age determination of southern bluefin tuna Thunnus maccoyii (Anon., 2002). This manual also described false annuli. John Neilson commented on the danger of strict application of descriptions in the Southern bluefin tuna manual to ABFT, noting that while the two species are congeners, parallels in life history have not been established. Thus, procedures for age determination of southern bluefin tuna should be used as guides but not as absolute protocols for ABFT.

Enrique Rodriguez-Marin gave a presentation title: Validated age and growth analysis of Atlantic bluefin tuna, which described the methodology for dorsal fin spine preparation and the age interpretation criteria using spine sections. This presentation highlighted that the major limitation of using dorsal fin spines for ageing is the nucleus vascularisation or “resorption” resulting in an age underestimation or overestimation of growth (Figure 2). Thus, resorption was quantified by analyzing the relationship between the number of translucent annuli lost by age based on the first annulus that was observed or measured. Preliminary results showed that nearly half of the translucent annuli (yearly considered) were lost, due to vascularisation, in relation to total number of translucent bands that the spine should have had.

Otoliths represent an advantage in relation to other calcified structures because all ages can be interpreted since there is no nucleus resorption. Otoliths can be used to age giants, while other hard parts are difficult to interpret from 10 years upwards. Conversely, spines are easier to collect and prepare than otoliths. Spines are the quickest and easiest structure to collect since their sampling does not require significant damage to the specimen and may be the only option in certain fisheries. See pros and con of different calcified structures in Table 5.1 from Rodríguez-Marín et al. (2007).

Since it is not always possible to sample otoliths because in certain fisheries ABFT is not processed or because of difficulties in obtaining them, spines are an alternative due to the simplicity of their collection.

The first annulus was often difficult to indentify. Dave Secor has images of young-of-the-year otolith sections with measurements to use as a guide for identifying the location of the first annulus (Figure 3).

Dave Secor also presented annotated images of otoliths from the Gulf of Mexico and Canada and pointed out crenulations and the area of first annulus formation. Generally, the distance from the outer edge of the first inflection point (elbow) on the ventral arm to the first increment is approximately 1 mm (Figure 4). Currently Dave Secor has a small set of images of young-of-the-year otoliths (n= 10). The group decided that in order to better identify the region in which the first annulus was formed, it would be beneficial to photograph and measure additional young-of-the-year bluefin tuna otolith sections to use as templates. In particular, it would be useful to obtain measurements of young-of-the-year bluefin tuna otoliths collected during different years to assess interannual variation. The importance of comparability of sectioning plane in the young-of-the-year samples and adults was also discussed.

4.2 Otolith edge identification and timing of annulus formation in otoliths and spines

The occurrence of a peripheral translucent band in ABFT otoliths throughout the year and the criteria of counting or not counting this translucent band were discussed. These issues were also a matter of discussion in the last ABFT direct ageing workshop (Rodríguez-Marín et al., 2007). Different calcified structures were used for ABFT age interpretation in this direct ageing workshop and the conclusions were:

- “Translucent zones (bright under transmitted light) are laid down from November to May and opaque growth zones (dark under transmitted light) from June to October (Mather & Schuck, 1960; Butler et al., 1977; Farrugio, 1980; Hurley & Iles, 1983, Lee et al., 1983; Cort 1991; Foreman, 1996)”.

- “An Atlantic bluefin tuna with a translucent band (bright under transmitted light) formed at the edge of the hard part and caught at the beginning of the year would be interpreted as having one year more, despite there still being five or six months before its true date of birth. Consequently, when the peripheral translucent band is present and the fish was caught in autumn, this band should not be considered as one year more (Mather & Schuck, 1960; Butler et al., 1977; Hurley & Iles, 1983)”.

The percentage of edge type classifications was plotted monthly to estimate the timing and periodicity of the translucent annulus formation in spine sections. Results show the appearance of this translucent band from September to June with 50% occurrence between November and April.

In general the type of edge and the capture date have not been taken into account for age interpretation in previous otoliths readings. Interpreting age in otoliths is based on counting summers (i.e. dark zones in transmitted light and bright zones under reflected light), whereas winters (i.e. bright zones in transmitted light) are counted in spines age interpretations. Until now, an opaque zone (dark zone in transmitted light) on the edge of an otolith is counted as annuli no matter the time of capture of the specimen.

If we are going to take into account the type of edge of the structure it is recommended to apply the criteria adopted in the 2006 direct ageing workshop (Rodríguez-Marín et al., 2007) for both structures, particularly for specimens caught at the beginning of the year. Otherwise, if a bright edge (in transmitted light) is observed in any calcified structure of an ABFT caught at the beginning of the year, it will be counted as one year more for spines but not for otoliths.

4.3 Readability codes for otoliths

The group discussed the utility of assigning readability codes to otolith sections. The suggestion was made to adapt the otolith confidence scale described on page 22 of: A manual for age determination of southern bluefin tuna Thunnus maccoyii (Anon., 2002). In order to better delineate the codes and eliminate redundancy the following four codes were suggested for inclusion:

1. 
   Pattern present-no meaning
   
2. 
   Pattern present-unsure with age estimate
   
3. 
   Good pattern present-slightly unsure in some areas
   
4. 
   Good pattern-confident with age estimate.
   

Dave Secor favored the use of a single reader making multiple reads on different occasions. Enrique Rodriguez-Marin mentioned that spines are read by two independent readers. For those spines with differences in the estimated age, a consensus between readers is used to address any discrepancies. The Panama City Laboratory uses two readers on difficult to age species, with consensus used to resolve differences. The group discussed exploring how other ageing centers dealt with repeated age estimates. Robert Allman mentioned that he would check on this for the NEFSC and NWFSC ageing programs.

Individuals were suggested to serve as expert readers for bluefin tuna otolith reference material. Steve Campana (Canada) and Naomi Clear (Australia) were suggested as potential candidates

5. Spine and otolith annuli calibration

The group suggested that a summary of all biological samples for bluefin tuna be made available to workshop participants. John Hoolihan volunteered to collect and compile these data from each laboratory and distribute to the group.

Molly Lutcavage discussed the importance of collecting as many biological samples (otoliths, spines, vertebra and gonads) as possible and that communication of sampling needs among the group was key.

The group agreed that it was important to calibrate spines and otoliths. For almost all datasets there were more spines than otoliths available. Enrique mentioned that ICCAT considers spines to be acceptable ageing structures and spines can be collected much easier and faster than otoliths and with less damage to the fish carcass. It was noted that some participants were convening next week in Canada to review ages obtained from otoliths and spine sections collected from the same individuals.

Molly Lutcavage discussed the importance of obtaining otolith based ages for Walt Golet’s research project in the Gulf of Maine. Walt has spine based ages but no otolith ages. In addition, Molly’s graduate student Jessie Knapp needs otolith based ages to complete her dissertation on bluefin tuna reproduction. The most important years to complete otoliths ageing are 2007-2009. Dave Secor has the 2007 otoliths and thought that Jay Rooker had the 2008 samples (he would contact Jay). The Panama City Laboratory has the 2009 otoliths.

Friday April 8

6. Targeted sampling needs

The group discussed future sampling needs. Dave Secor mentioned that samples are needed from North Carolina and the Mid-Atlantic in particular.

Walter Golet commented that he has been extensively sampling the Gulf of Maine and contiguous areas.

Robert Allman mentioned that starting May 5^th the Gulf of Mexico pelagic long-liners will be required to use “weak” hooks and that this would likely reduce numbers of bluefin caught. Panama City has samples collected from the recreational fishery (mainly tournament) in the Atlantic by the contractor Quantec, the commercial fishery in the Northeast and in the Gulf of Mexico from the pelagic observer program.

Jay Rooker is working with French, Spanish and Italian scientists and will be sampling Mediterranean bluefin looking for Western stock origin fish. Jay will be targeting 100 otoliths during each of the next two years. Enrique Rodriguez-Marin commented that he would also be working with Jay Rooker.

John Neilson mentioned that during discussions at ICCAT meetings with Clay Porch, it was recommended to keep sampling proportional to catches. Dave Secor was concerned that this approach would not sample enough length classes and suggested targeting length classes in order to establish a good length at age relationship. John Neilson commented that we should try to sample according to effort at a broad level of geographic aggregation (i.e., Gulf of St. Lawrence versus Scotian Shelf fisheries) but recognize that within those areas, sampling may be somewhat opportunistic. His program, starting this year, is targeting about 400 otoliths per year. Enrique suggested length-stratified sampling over random sampling because there is no ABFT fishery with all age classes fully represented.

Walt Golet said that sampling most often is opportunistic and that he obtained samples from collecting himself and from fish houses by retrieving data from reference numbers. Walt collected about 300 otoliths last year and mentioned that the strong 2003 year class was beginning to appear in the fishery and that he expected even more 2003 fish to be collected this year. Dave Secor asked that Walt examine his dataset for underrepresented sizes. Enrique Rodriguez-Marin asked if sampling 300 fish would cover the whole range of sizes. Dave Secor commented that this moderately large sample would likely have good length coverage for fisheries sampled. Molly Lutcavage noted that Walt Golet’s age and growth samples comprised an important part of the Large Pelagic Research Center’s long term study of the ecology of ABFT in the NW Atlantic. Dave Secor mentioned that the there was a need to sample the high-seas fishery in the Central North Atlantic. John Neilson replied that Canada was sampling this fishery, but sampling was limited due to the expense of sending samplers out to these vessels.

Dave Secor has developed an equation to convert snout length (i.e., distance from tip of upper jaw to anterior part of eye orbit) to curved fork length. He requested that others collect these measurements as well.

1. 
   The group agreed that the objectives for future sampling should concentrate on ageing and determining stock origin.
   
2. 
   Everyone agreed that better communication was needed regarding sampling needs and coverage.
   
3. 
   There is a need for better coverage of the Mid-Atlantic, Dave Secor could possibly concentrate on this.
   
4. 
   There is a need for sampling both the reproductive and non-reproductive segments of the population in order to develop a more rigorous maturity ogive for this species.
   

The group discussed the advantages and disadvantages of exchanging slides versus images and decided the images would be most practical for training purposes. An annotated set of about 20 images was suggested. There was some discussion about whether to include exceptionally clear (i.e., relatively easy to age) otoliths or to include otoliths of varying difficulty. It was decided that the training set should include the range of ages observed in the fishery, as well as images of varying difficulty.

The group discussed assembling a set of otoliths that have a corresponding spine age in order to calibrate the two ageing structures. Enrique Rodriguez-Marin offered to provide 40 otoliths that had corresponding spines and ranged in age from 8 to 14 years.

The group agreed that a reference set should include at least 50 otolith images. Robert Allman mentioned that for reef fish reference collections in the Gulf of Mexico region, each laboratory contributes material. John Neilson mentioned that his lab could contribute older individuals and that he would ask Steve Campana to review the images.

Robert Allman commented that for production ageing programs an average percent error (APE) ≤ 5% is generally considered acceptable for moderately difficult to age otoliths. Dave Secor commented that for ageing ABFT otoliths a CV of 10 (APE of 7) may be acceptable.

The group agreed that due to the difficulty in locating the first annuli that a scale bar should be included on all images as a reference and that each image should be photographed using reflected and transmitted light.

The group agreed that reflected light needs to be explored further. Ben Gahagan described how he uses a single fiber optic filament taped with electrical tape to a fiber optic light source to direct light into the otolith.

John Neilson commented that perhaps funding could be obtained to hire experts to augment the existing collection of annotated images, given that the current collection from Canada only includes larger fish. It would be helpful to include experts involved in the previous exercise (Naomi Clear, Persephone Megalafonou, and Steve Campana) possibly adding others currently involved in the age determination initiative.

Enrique Rodriguez-Marin showed some of the patterns found in ABFT spine sections. Spines are the main ageing structure for the eastern stock. Double or triple (i.e. false annuli) rings grouped together are common and generally are counted as a single annulus (Figure 5). It is important to examine the overall pattern to establish a final age. Walt Golet wondered if the second (false) ring was associated with Atlantic crossing. Enrique reported that the oldest individual spine age was 24 years. Annuli start to reabsorb at age 3.

The group examined otolith sections from Robert Allman, Ben Gahagan and Dheeraj Busawon using transmitted and reflected light. Light type did not appear to make a difference in final age agreed upon by the group. The group took turns ageing otolith images from Dave Secor’s collection. The annuli were pointed out using a laser pointer and then compared to annotated layers in Photoshop. Overall the group had good agreement with Becca (D. Secor’s former assistant tuned to S. Campana), however compared to Steve Campana’s ages the group consistently under-aged. This further illustrated the importance of location of the first annuli, using the reference of reliable measurements of otoliths from young of the year fish.

The group discussed the importance of tracking the strong 2003 cohort as a method of age validation. This cohort should be highly visible this sampling year.

8. Recommendations

1. 
   Compare annuli counts from serial sections of otoliths to determine if counts differ by distance from the core. Dheeraj Busawon and Ben Gahagan will compare annuli counts from two sections (anterior and posterior to the anti-rostrum) from about 40 otoliths. A paired t-test will then be used to compare counts.
   
2. 
   Include a readability code (1-4) for all otolith sections. Examples of an otolith section for each readability code will be provided by Dheeraj Busawon and Ben Gahagan.
   
3. 
   Expand the set of young-of-the-year otolith images to use as templates to establish where the difficult to identify first annulus occurs. Dave Secor and Ben Gahagan will provide these.
   
4. 
   Create a training set of 20 annotated otolith images using both reflected and transmitted light. Dave Secor has a collection of annotated images that could be used for some of these.
   
5. 
   Assemble a reference collection of 50+ otolith images and provide an expert age determination. Perhaps the Secor and Neilson labs can provide the bulk of these otoliths with Panama City contributing the material it has available. A reference scale needs to be included on these images.
   
6. 
   Continue to collect otolith material for micro-constituent analysis. If only one otolith is available the core material must be removed prior to ageing.
   
7. 
   Age otolith sections which have a corresponding spine age to calibrate the two structures. Enrique Rodriguez-Marin (n= 40) will provide these otolith sections.
   

Anon. 2002. A manual for age determination of southern bluefin tuna Thunnus maccoyii. Otolith sampling, preparation and interpretation. In: Report of the Direct Age Estimation Workshop. 11-14 June 2002. Victoria, Australia. Meeting Report of the Commission for the Conservation of Southern Bluefin Tuna (CCSBT). 36p.

Butler, M.J.A., J.F. Accy, C.A. Dickson, J.J. Hunt & C.D. Burnett. 1977. Apparent age and growth, based on otoliths analysis, of giant bluefin tuna (Thunnus thynnus) in the 1975-76 Canadian catch. Collective Volume of Scientific Papers, ICCAT, 6(2): 318-330.

Cort, J.L. 1991. Age and Growth of the Bluefin Tuna, Thunnus thynnus (L.) of the Northwest Atlantic. Collective Volume of Scientific Papers, ICCAT, 35 (2): 213-230.

Farrugio, H. 1980. Age et croissance du thon rouge (Thunnus thynnus) dans la pecherie francaise de surface en Mediterranee. Cybium 3e série (9): 45-59.

Foreman, T. 1996. Estimates of age and growth, and an assessment of ageing techniques, for northern bluefin tuna, Thunnus thynnus, in the Pacific ocean. IATTC Bulletin, 21: 2

Mather, F. J. & H. A. Shuck. 1960. Growth of bluefin tuna of the western North Atlantic. U.S. Fish and Wildlife Service, Fishery Bulletin, 179(61): 39-52.