Discriminating among Alaska’s Pacific herring stocks using chemometry of heart tissue fatty acids 
and otolith microchemistry

Principal Investigators

Ted Otis, Alaska Department of Fish and Game, Commercial Fisheries Division, Homer
Ron Heintz, NOAA-National Marine Fisheries Service, Auke Bay Lab, Juneau  
Dr. Nathan Bickford, Advanced Instrumentation Lab, University of Alaska, Fairbanks

INTRODUCTION

Herring are an important component of the marine ecosystem providing a trophic pathway for energy flowing from secondary producers to apex predators, including humans. Despite decades of study and well over a hundred years of commercial exploitation, considerable uncertainty continues to exist regarding: 1) the scale at which herring population structure exists within large geographic areas and, 2) the degree to which herring return to natal areas to spawn.  These fundamental life history traits are directly relevant to how exploited herring stocks should be assessed and managed. 

Cape Ann herring fishermen landing their gill-nets after a night's fishing.  (Etching from a photograph by T. W. Smillie, circa late 1800’s; credit NOAA-National Marine Fisheries Service HistoricArchives).

Alaska’s fishery managers require a tool that can identify ecologically significant population structuring among adjacent spawning aggregations that are exploited during spring sac-roe herring fisheries. They also require a mixed stock analysis tool that allows them to investigate whether winter herring fisheries (e.g., food/bait fisheries) target only the local spawning stock or a mixture of nearby stocks that aggregate during winter.  The ability to manage stocks discretely is a principal component of sustainable fisheries management- one that requires the ability to accurately apportion the catch from mixed stock fisheries.

Reported Atlantic and Pacific herring homing rates range from 66-94 percent (Tester 1949; Cushing and Burd 1957; Hourston 1982; Wheeler and Winters 1984).  The corresponding stray rates of 6-34% indicate there is more than sufficient gene flow between neighboring spawning areas to compromise the ability of allozyme markers to discriminate between putative stocks (Smith and Jamieson 1986; Bembo et al. 1996; Waples 1998).  Waples (1998) warned, “because the amount of migration necessary to obscure most genetic evidence of stock structure (only a handful of individuals per generation) is generally inconsequential as a force for rebuilding depleted populations on a time scale of interest to humans, there is no guarantee that genetic methods alone will provide sufficient precision for key management decisions involving marine species”.   Thus, herring fishery managers have continued to seek a tool that allows them to identify population structure within and among their respective management areas. 

Researchers from the Alaska Department of Fish and Game, the National Marine Fisheries Service-Auke Bay Lab, and the University of Alaska-Fairbanks have teamed up to investigate whether fatty acid signature analysis and otolith chemistry can be used to discriminate among spawning aggregations of Pacific herring Clupea pallasi at relatively fine spatial scales (i.e., > 100 km) 

Fatty acid compositions of fish lipids have been investigated for decades (Ackman et al. 1963); however, much of the early lipid research was directed at determining the commercial value of fish oils (e.g. Ackman and Eaton 1966) and understanding how fat content relates to various life history functions (e.g. Rajasilta 1992).  Because the composition of certain lipids can be closely related to the types of food recently ingested (Navarro et al. 1995; Kirsch et al. 1998), recent investigations have been directed at diet analysis and foraging distribution (e.g. Iverson et al 1997; Iverson et al. 2001; Budge et al 2002). 

Many studies have established that fatty acid analysis (FAA) also has utility as a stock identification tool. As early as the 1930’s it was demonstrated that different stocks of fin whale Balaenoptera physalus could be distinguished by the degree of unsaturation of their oils (measured as iodine value: Lund 1934, as cited in Grahl-Nielsen et al. 1993).  More recent research suggests that the composition of phospholipid fatty acids prominent in some body tissues (e.g., heart tissue, brain, eggs) have a genetic basis that makes analysis of these tissues appropriate for stock identification studies (Joensen and Grahl-Nielsen 2000, Joensen et al. 2000).   Chemometry of fatty acids from heart tissue have been used to discriminate stocks of striped bass Morone saxatilis (Grahl-Nielsen and Mjaavatten 1992), Atlantic herring Clupea harengus harengus (Grahl-Nielsen and Ulvund 1990), and Atlantic cod Gadus morhua (Joensen et al. 2000).  Fatty acid analysis of eggs has been used to discriminate between American lobster Homarus americanus populations (Castell et al. 1995), Baltic cod Gadus morhua stocks (Pickova et al. 1997), and even the wild/domestic origin of sturgeon ova (Czesny et al. 2000).  Chemometry of fatty acids have also been used to distinguish between closely related species of the genus Sebastes from the Faroe Islands (Joensen and Grahl-Nielsen 2000) and juvenile chinook and coho salmon from the Fraser River (Mjaavatten et al. 1998). 

Before fatty acid analysis can proceed as a stock discriminator, it will be important to demonstrate the temporal stability of the differences within and among a priori stocks. In a pilot study (Otis and Heintz 2003), we targeted heart tissues based on the understanding that heart phospholipids are less subject to environmental influences than other tissues or lipid classes (Grahl-Nielsen and Ulvund 1990, Czesny et al. 2000, McKenzie 2001). This is important because the fatty acid composition of different tissues and lipid classes can be very sensitive indicators of environmental change. For example, numerous studies have shown the effect of diet on the fatty acid composition of the triacylglycerols (e.g., Henderson and Tocher 1987). Thus, tissues whose lipids comprise triacylglycerols will have fatty acid compositions that vary over time. The heart lipids studied here were primarily phosphatidylcholine and phosphatidylethanolamine (data not shown). As a class, these phospholipids are generally less sensitive to diet effects, but dietary influences on the phospholipid content of Atlantic salmon hearts has been demonstrated (Grisdale-Helland et al. 2002). In addition, phospholipids can be influenced by temperature (Hazel 1984; Henderson and Tocher 1987), salinity (Cordier et al. 2002), and developmental stage (Kreps et al. 1969). Therefore, demonstrating that the variation in heart tissue fatty acid composition observed between stocks exceeds that imposed by the environment on a given stock will be a key element in the development of this method (Begg et al. 1999).

Otolith microchemistry has also been used  for stock discrimination in a variety of fish species including: pink snapper, (Edmonds et al. 1989), orange roughy (Edmonds et al. 1991), yellow-eye mullet (Edmonds et al. 1992), Atlantic cod (Campana and Gagne 1995, Campana et al. 1995), walleye (Bickford and Hannigan 1990) and salmonids (Kalish 1990) among others.  Thresher (1999) provides a comprehensive review of the use of otolith elemental composition as stock discriminators and offers some cautionary suggestions for researchers interested in employing this promising technique. 

Successful application of otolith elemental analysis for stock discrimination is likely dependent on the extent of the differences in water chemistry between the environments inhabited by each stock and the precision of the instruments used to measure trace elements. Laser ablation- inductively coupled plasma- mass spectrometry (LA-ICP-MS) can be used to analyze trace elements at specific loci (30 mm) on the otolith (Gray 1985, Denoyer et al. 1991).  Electron microprobes (EM) also allow analysis of specific loci (5-7 mm), albeit at reduced resolution (parts per thousand, pers. comm. K. Severin, UAF Dept. of Geology and Geophysics).  Techniques that target specific loci, such as EM and LA-ICPMS, are most appropriate for identifying stocks that spawn in different environments but later reside in similar environments (Coutant and Chen 1993).

Study Objectives:

The goals of this research are to a) evaluate the temporal stability and biological variability of the heart tissue fatty acid compositions that have already been used to discriminate Alaska herring stocks (Otis and Heintz 2003), and b) to determine whether otolith microchemistry is useful for discriminating population sub-units within traditionally recognized stocks of Pacific herring in Alaska (e.g., Sitka, PWS, Kamishak, Kodiak, Togiak, Dutch Harbor, Bering Sea).  Because otolith samples were collected from the same individual fish used to evaluate the temporal stability of fatty acid compositions (Otis and Heintz 2004), this project offers the unique opportunity to directly compare two cutting edge herring stock identification techniques. 
Accurate knowledge of stock structure is relevant to the manner in which state officials assess and manage this commercially and ecologically important resource.  The ability to identify the stock of origin for herring collected away from their natal spawning areas would also have tremendous utility to managers of fisheries that may be harvesting mixed stocks (e.g. herring food/bait fisheries).  For these purposes, we proposed the following objectives for the fatty acid component of our work: 

Objective 1) Assess the temporal stability and biological variability of stock discrimination criteria derived from fatty acid analysis of cardiac tissues. 

This objective addresses three hypotheses: 

1. At spawning, the variation in fatty acid composition within a spawning stock is equal to the variation observed between that stock and other spawning stocks.
2. At spawning, the variation in fatty acid composition of a spawning aggregation is equal to that of a similar aggregate spawning in the same general area, but later during the spawning period.
3. The variation in fatty acid composition of a spawning stock in a given year is equal to the variation between that stock and a stock using the same spawning area in a different year.

The first of these hypotheses is an attempt to re-create the results described in Otis and Heintz (2003), without controlling for age, sex, and gonad maturity, as was done in their pilot study.  Evaluation of this hypothesis will establish the extent to which heart fatty acid composition naturally varies across all contributing members (i.e., sexes, cohorts) of a putative spawning stock. The second examines temporal variation within a putative stock over the course of a protracted spawning period, within a given spawning year. The third hypothesis, undertaken in year two of the proposed study, examines the temporal variation in heart fatty acid composition across successive years. In addition, since we are proposing to resample the stocks examined in Otis and Heintz (2003), hypothesis 3 can be examined over a 6-year period (i.e. 2001 – 2006).

Objective 1) Assess whether the stock(s) of origin for herring harvested in winter food/fisheries can be determined by comparing their heart fatty acid composition to those of local area spawning aggregations.

This objective addresses one additional hypothesis:

4. The variation in fatty acid composition within herring schools aggregating during winter is equal to the variation observed between herring schools using the same general area for spring spawning.

This final hypothesis evaluates whether or not fatty acid compositions from spawning herring can be used to determine the stock(s) of origin for herring harvested during winter food/bait fisheries.

The otolith chemistry component of our stock identification work includes two additional objectives that are very similar to those used for our fatty acid work:
Objective 1) Using samples from the same individual fish, assess whether population sub-unit boundaries derived from otolith chemistry match those derived by fatty acid analysis.

This objective addresses the following hypothesis:

1). At spawning, the variation in otolith chemistry within a spawning stock is equal to the variation observed between that stock and other spawning stocks.

This hypothesis attempts to corroborate and expand upon the results described in Otis and Heintz (2003), as well as the results forthcoming from EVOS Project 050769.  Evaluation of this hypothesis will also establish the extent to which otolith microchemistry naturally varies across all contributing members (i.e., sexes, cohorts) of a putative spawning stock.

Objective 2) Assess whether the stock(s) of origin for herring collected during fall/winter can be determined by comparing their otolith chemistry to those of local area spawning aggregations.

This objective addresses the following hypothesis:

2).  The variation in otolith microchemistry within herring schools aggregating during fall/winter is equal to the variation observed between herring schools using the same general area for spring spawning.

This hypothesis evaluates whether or not otolith chemistry from spawning herring can be used to determine the stock(s) of origin for herring sampled/harvested during other times of the year.

View project details on the EVOS-TC website

This project was funded by the Exxon Valdez Oil Spill Trustee Council (EVOS-TC).  For more information on this herring stock identification study, the successful pilot study that preceded it, selected references, or the principal investigators, please follow the links provided on this home page.