MEPS 267:253-265 (2004)  -  doi:10.3354/meps267253

ABSTRACT: The spatial relationships and relative contributions of known juvenile gag Mycteroperca microlepis habitats to specific fishery grounds and populations along the Florida west coast are virtually unknown. To determine if otolith composition is a valid tracer of specific nursery sites and can be used to classify adult fish to their nursery area, chemical concentrations in juvenile gag otoliths (Li, Na, K, Mg, Ca, Sr, Ba, Mn, Cu, Pb, δ¹³C and δ¹⁸O,) were measured for 4 nursery areas along the Florida west coast in 1992, 1995 and 1996. Classification of fish to nursery area was by parametric discriminant function analysis and neural network simulation; both gave similar results in the spatial and temporal patterns of classification error and in identification of important classification variables (Mn, Sr and δ¹³C). Classification success rates ranged from 66 to 100%. Interannual variability in otolith composition had a negative effect on classification success rate at the spatial scale of site separation (<200 km). At regional spatial scales (>300 km) the year effect was reduced and classification success remained high without consideration of year class. Analysis of classification errors supports the separation of sites into northern and southern groups. The observed negative trend in otolith Sr and positive trend in otolith Mn with increasing latitude are positively correlated with corresponding trends in groundwater Sr and sediment Mn, respectively, which are hypothesized to be the proximal causes for the otolith trends. Otolith composition is a valid technique for classifying juvenile gag to estuarine habitats along the west coast of Florida. Classification error rates of less than 10% are obtained when the year class of the fish is considered. Similar error rates were achieved across multi-year data at regional spatial scales.

KEY WORDS: Otolith chemistry · Juvenile habitat · Gag · Manganese · Strontium · Carbon isotopes