Coastwide Otolith Signatures of Juvenile Atlantic Menhaden, 2009–2011

Abstract

The Atlantic Menhaden Brevoortia tyrannus is a clupeid that plays a critical role in the marine food web and supports one of the largest fisheries on the U.S. East Coast. Along with a decrease in overall numbers and spawning stock biomass, recruitment levels have remained low since the 1990s. Atlantic Menhaden use numerous estuaries along the Atlantic coast for juvenile development before recruiting to the adult population, but the contribution of each of these nursery grounds is currently unknown. Chesapeake Bay is thought to contribute 70% of the total recruits, although this estimate is over 20 years old and predates current low recruitment levels. We investigated the potential of trace element (Li, Mg, Mn, Rb, Sr, Y, Ba, and Pb) and stable isotope ratio (δ13C and δ18O) signatures in otoliths to distinguish among Atlantic Menhaden collected from various nursery grounds along the U.S. Atlantic coast (Connecticut to South Carolina) during 2009–2011. Juveniles were classified to four regional nursery areas with nearly 90% accuracy. Due to significant interannual variation in the chemical signatures, our attempts to classify juveniles from adjacent year‐classes or combined year‐classes resulted in lower accuracy. However, this study provides a 3‐year library of geochemical fingerprints for assigning adults to their regions of origin. This research builds the foundation for a comprehensive estimate of Atlantic Menhaden recruitment rates from each of the major nursery areas along the U.S. Atlantic coast for 2009–2011.Received May 28, 2014; accepted September 16, 2014

BeyondZar: the use and abuse of classification statistics for otolith chemistry

“…Otolith chemistry is one of the most widely used methods. Provided that water chemistry is reflected in the chemistry of the otolith layer being deposited and remains unchanged with fish growth (Campana, ), the rationale for estimating adult natal origin is to compare the microchemistry of the inner part of the adult otolith with those of juveniles from all putative sources (Elsdon et al ., ; Anstead et al ., ). This, however, is often a challenging task (Catalan et al ., ; Morales‐Nin et al ., ) and, among other technical problems, scientists face the dilemma of which classification methods for assignment to choose.…”

Contribution of Nursery Areas to the Adult Population of Atlantic Menhaden

“…Proceeding with four clusters in the adult data based on previous studies (Ahrenholz et al 1989;ASMFC 2004;Anstead et al 2015Anstead et al , 2016 and supported by the small pseudo t 2 statistics and large pseudo F statistic, we assigned adults of unknown origin to their nursery ground using the QDA. We found that, on average, 40% of age-2 Atlantic Menhaden originated in the northeastern nursery region, 8% in Delaware Bay, 22% in Chesapeake Bay, and 30% in the southeastern region for the 2008-2010 yearclasses (Table 4).…”

“…Juvenile Atlantic Menhaden were collected in 2009-2011 from multiple sites along the U.S. Atlantic coast, prepared, and analyzed as described in Anstead et al (2015). Adult Atlantic Menhaden (ages 2 and older) were collected in 2010-2012 from the coastal waters from Cape Cod to the Hudson River, Delaware Bay, Chesapeake Bay, Pamlico Sound, and Charleston Sound ( Figure 1).…”

“…Classification functions specific to 2009 and 2010 were used to assign adults that belonged to those cohorts to their nursery region. For the 2008 adults, a combined signature from the 2009-2011 juvenile data were applied, as its utility in correctly classifying adjacent year-classes has been shown when year-specific data are not available (Anstead et al 2015).…”

“…We previously established the utility of otolith chemistry for building geochemical signatures that distinguished between coastwide nursery grounds for Atlantic Menhaden with nearly 90% accuracy (Anstead et al 2015). We showed that in addition to using year-specific signatures for classification in 2009-2011, a combined signature could be employed with little decrease in accuracy (from 90% to 86%) for data-poor years or when juvenile data have not been collected for adjacent year-classes.…”