link to AFSC home page
Mobile users can use the Site Map to access the principal pages

link to AFSC home page link to NMFS home page link to NOAA home page

Age & Growth Program

Age Validation of Northern Rockfish and Yellowfin Sole with Bomb-Produced 14C

Research Reports
Summer 2014
Contents
Feature
ABL Reports
FMA Reports
NMML Reports
REFM Reports
Complete Rpt (pdf)
Complete Rpt (flash)
Quarterly Index
Quarterly Home
refer to caption
Figure 1. Northern rockfish otolith cross sections with presumed annual growth zones indicated (A). Yellowfin sole cross sections with presumed annual growth zones indicated (B).

Accurately ageing fish is often a difficult task. The routine method used to estimate fish age is by counting presumed annual growth zones in cross sections of fishes’ otoliths (Fig. 1). Otoliths are small, calcium carbonate stone-like structures in the inner ear of fish. The interpretation and counting of otolith growth zones is not always clear and can require subjective decisions. The impacts of systematic errors in counting growth zones include difficulty in estimating stock-recruitment relationships, unrealistic fish growth estimates, and population forecasting inaccuracy. Therefore, validating the accuracy of estimated fish ages is a critical step in determining the reliability of age data used in stock assessments.

Northern rockfish (Sebastes polyspinis) has one of the most northerly distributions among the 60+ species of Sebastes in the North Pacific Ocean and are important to the commercial fishery in the Gulf of Alaska, where about 4,000 metric tons are caught by trawlers each year. Yellowfin sole (Limanda aspera) is one of the most abundant flatfish species in the eastern Bering Sea and is the target of the largest flatfish fishery in the United States. The Center’s Age and Growth program routinely estimates age for these species by growth zone counts from an otolith cross section (Fig. 1). The maximum age is about 71 and 31 years of age for northern rockfish and yellowfin sole, respectively. Both species are assessed using age-structured population models. But how accurate are the age data?

refer to caption
Figure 2. Map of collection locations for northern rockfish, yellowfin sole, eastern Bering Sea halibut reference, and Gulf of Alaska halibut reference.

To validate the ageing methods of growth-zone counts, we used a technique known as the bomb-radiocarbon assay. This method is based on the increase of in the atmosphere and surface layer of the ocean caused by above-ground testing of nuclear bombs during the late 1950s and 1960s. To apply this validation method, first a 14C “reference chronology” (a time series of 14C measurements from otoliths of known-age fish) is developed from 1-year-old juvenile fish collected during the era of marine 14C increase (1950s to 1960s). The 14C reference chronology is then compared to14C measured in otolith cores from adult test fish.

The otolith core is material deposited in the first year of life. If the bomb-produced increase of ∆14C in both the reference and test chronology is synchronous, the growth zone ages are considered validated. This comparison assumes that the test and reference chronologies are based on biologically similar species that are from the same geographical area. Our goals were to 1) validate the accuracy of ages determined from otolith growth zone counts in northern rockfish and yellowfin sole and 2) expand our fish age validation abilities to species from the eastern Bering Sea (EBS) by developing a new reference chronology from EBS juvenile Pacific halibut.

refer to caption
Figure 3. Eastern Bering Sea halibut reference, individual samples and predicted model, and eastern Bering Sea yellowfin sole test samples, individual samples and predicted model (A). Gulf of Alaska halibut reference, individual samples and predicted model, and Gulf of Alaska northern rockfish test samples, individual samples and predicted model (B).  
refer to caption
Figure 4. Eastern Bering Sea halibut reference, individual samples and predicted model, and eastern Bering Sea yellowfin sole test samples, individual samples and predicted model (A). Gulf of Alaska halibut reference, individual samples and predicted model, and Gulf of Alaska northern rockfish test samples, individual samples and predicted model (B).
refer to caption
Figure 5. Probility of under or over ageing by one or more years for northern rockfish is about 10% and 1% respectively (A). Probility of under or over ageing by one or more years for yellowfin sole is about 1% respectively (B).

Our test samples, northern rockfish and yellowfin sole otoliths, were collected during AFSC scientific trawl surveys or by NMFS fishery observers aboard commercial vessels. Specific specimens were chosen such that their posted birth years (based on catch year and estimated growth-zone age) were during the era of ∆14C increase. The otolith’s cores were removed, analyzed for 14C, and reported here as ∆14C ‰ (a standardized method of presenting radiocarbon results).

Two reference chronologies were used, one developed prior to our study using juvenile Pacific halibut collected in 1954–81 from the Gulf of Alaska (GOA) by the International Pacific Halibut Commission (IPHC), and second, our new reference from EBS juvenile Pacific halibut otoliths collected by the IPHC between 1956 and 1980 in our study (Fig. 2).

To validate the northern rockfish and yellowfin sole ages and estimate any potential ageing bias, we quantitatively compared the test ∆14C ‰ datasets to the reference curves from each basin using a 4-parameter logistic function. The logistic function was modified to account for a post-peak decay in ∆14C ‰ levels after 1970 to more realistically reflect the Bering Sea data (Figs. 3 and 4). Bayesian inference and Markov Chain Monte Carlo (MCMC) simulations were used to estimate all parameters and derive a probablistic framework for estimating bias. Bias was approximated as the difference between the parameter that describes the year at 50% rise of ∆14C curve.

We found a difference in the rate of increase and amount of bomb-produced ∆14C in the GOA and EBS; also we found that ages determined for the two species were accurate. The estimated year at 50% rise in the GOA reference chronology was 1962.6, while in the new EBS reference chronology it was 1962.0 (Fig. 3). The difference between these parameters appears small; however, the other estimated parameters clearly indicated that the two references were different, as demonstrated by the two predicted references chronologies seen in Figure 3.

In the EBS, the ∆14C increased at a faster rate, rose slightly earlier, increased to a greater level, and decreased faster than in the GOA. The validation of accurate estimated ages for yellowfin sole was demonstrated by a year at 50% rise of 1962.2, which is comparable to the EBS reference of 1962.0. This was further demonstrated by the similarity of the other parameters, as seen in the nearly identical predicted model fits (Fig. 4).

The validation of accurate estimated ages for northern rockfish was demonstrated by a year at 50% rise of 1963.0, which is comparable the GOA reference of 1962.6. Again, this was further demonstrated by the similarity of the other parameters, as seen in the predicted model fits (Fig. 3).

For northern rockfish, the probability of under-ageing by 1 or more years (ageing bias) is about 10%, and the probability of over-ageing by 1 or more years is only about 1% (Fig 5).

For yellowfin sole, the probability of under-ageing by 1 or more years is about 1% (Fig. 5). For yellowfin sole, and especially for northern rockfish where the average age in the population exceeds several decades, these results indicate a very small likelihood of age inaccuracy.

In addition, the comparison of the two reference chronologies indicate that different geographic areas and unique oceanographic conditions may led to differences in the assimilation and incorporation of 14C in fish otoliths. Therefore, future age validation studies need to make careful choices in the reference chronology used. Historically, there has been a lack of available reference chronologies; here we developed a new one for a region where none existed before.

By Craig Kastelle, Thomas Helser,
and Steve Wischniowski












<<< previous

next >>>


            | Home | Site Map | Contact Us | FOIA | Privacy | Disclaimer | USA.gov | Accessibility | Print |           doc logo