Document Type

Dissertation

Degree

Ph.D.

Degree Granting Department

Marine Science

Major Professor

Ernst B. Peebles

Keywords

estuarine habitat, natural tags, otolith chemistry

Abstract

Knowledge of fish habitat use and connectivity is critical for understanding the structure and dynamics of fish populations and, therefore, necessary for the implementation of successful fisheries management strategies. Tagging is an effective means of providing such information, and the elemental composition contained within fish otoliths is increasingly being used as a natural tag. The chemical composition of otoliths reflects the incorporation of elements from different water bodies and can thus be used to understand the habitat use, movements, and life history patterns of fishes. To assess the applicability of otolith elemental composition as a tagging technique within the Tampa Bay estuary, Florida, laser ablation-inductively coupled-plasma mass spectrometry (LA-ICP-MS) was used to analyze the elemental composition of otoliths from two estuarine-dependent fish species, Snook (Centropomus undecimalis) and Red Drum (Sciaenops ocellatus).

Otolith elemental fingerprints can be used to quantify the proportion of juveniles from different nurseries that survive to join the sub-adult and adult fisheries, thus, providing resource managers with quantitative criteria to prioritize the most productive areas for conservation and restoration. To evaluate the use of otolith elemental fingerprinting in the Tampa Bay estuary, the spatial and temporal scales of chemical variation in otoliths collected from throughout Tampa Bay were examined by performing permutation-based multivariate analyses of variance (MANOVA) on the elemental data at several spatial (individual tributary, two-region, and three-region) and temporal (annual and seasonal) scales. Canonical Analysis of Principal Coordinates (CAP) was used to generate classifiers based on the otolith elemental fingerprints of juvenile fish, and `leave-one-out' cross-validation procedures indicated that the greatest classification accuracy was obtained by using the two-region model (upper vs. lower Tampa Bay) for both species (for Snook F=45.8, p=0.001, CAP cross-validation success=76%; for Red Drum F=9.7, p=0.001, CAP cross-validation success=87%). For both species, all temporal analyses at the inter-annual scale indicated that otolith elemental fingerprints were significantly different across years (two-way MANOVA; p<0.05). The lack of significance for the site-by-year interaction term for all models, except the two-region model for Snook, indicated that the degree of variation between sites was consistent across years, and that samples could potentially be pooled across years. However, in all cases, combining cohorts weakened the classifiers; therefore, classification models for both species were based on the elemental fingerprints of juveniles from a single cohort (2007 cohort for Snook and 2008 cohort for Red Drum). Seasonal differences in the elemental composition of Snook otoliths were not found for the elements that contributed to the discrimination between groups in the spatial models, except for a single element, manganese (ANOVA; p=0.051). Insufficient sample sizes of juvenile Red Drum across seasons precluded inter-seasonal tests for this species.

Several environmental factors which may have contributed to the regional differences in otolith chemistry were examined, including physico-chemical parameters (surface measurements of salinity, dissolved oxygen, pH, and temperature taken at the time of sample collection), surficial geologic stratigraphy, and land development. Weak, but significant correlations were identified between some elements and physico-chemical parameters; however, instantaneous measurements taken at the time of fish collection may not have provided an accurate representation of the overall conditions experienced by the fish during the period in which the otolith material used in analyses had been deposited (2 - 4 weeks). A significant correlation between latitude and otolith Sr/Ca was found, likely corresponding to an increasing ambient gradient that occurs from the upper to lower bay (for Red Drum F=77.1, p=0.001; for Snook F=69.2, p=0.001). The Land Development Intensity metric was negatively correlated with otolith Li/Ca and Sr/Ca. While surficial geologic inputs may have also contributed to the elemental composition of otoliths, the relationships revealed by redundancy analyses (RDA) were somewhat unclear or contradictory.

Once the appropriate chemical characterization of the study area was identified (the two-region models for both species), elemental fingerprints from the core portions of sub-adult and adult otoliths were assigned to their most probable juvenile habitat region using a maximum likelihood estimator based on the posterior probabilities generated by CAP analyses (CAP-MLE). Application of the two-region model revealed that the majority of Red Drum (83%) was determined to have originated from juvenile habitats in the upper Tampa Bay region, while most Snook (60%) originated from juvenile habitats in the lower Tampa Bay region. The majority of sub-adult/adult Snook and Red Drum were collected from the same region in which they were determined to have originated (for Snook, 36 out of 55 = 65%; for Red Drum, 58 out of 78 = 74%), indicating some level of site fidelity to juvenile habitat areas.

The use of otolith elemental profiling to reconstruct specific environmental and physiological experiences has the potential to provide unique insights regarding the life histories of Snook, a species with unpredictable spawning and movement characteristics. Otoliths from Snook maintained in captivity at the Mote Aquaculture Park (MAP) were analyzed to elucidate the degree to which various factors, including otolith growth (macrostructure features), spawning events, handling stress, and salinity influence otolith chemistry. Cross-correlation analyses of otolith elemental profiles and quantified macrostructure features (including annuli and checks) demonstrated that interpretations of elemental patterns should not be confounded by changes associated with otolith crystallography. An elemental marker for known spawning events was not identified (ANOVA spawners vs. non-spawners, p>0.05); however, because the physiological costs and alterations in blood chemistry associated with gonadal maturation (rather than the singular act of spawning) could affect otolith chemistry, additional studies which more thoroughly track maturation stages may be able to identify a suite of elements that can be used to discern the reproductive histories of Snook. Significantly elevated Zn:Ca (ANOVA: F=5.64, p=0.012) and decreased Fe:Ca (ANOVA: F=25.02, p<0.001) were observed in otolith elemental profiles of Snook exposed to higher levels of handling stress, indicating that these elements may be useful for retrospectively identifying stressful periods within a fish's life. Despite the Snook having been maintained in a constant, high salinity environment within the MAP tanks, the magnitudes and variations of otolith Sr:Ca observed in otolith profiles were much larger than those observed in wild otoliths, warranting caution in the sole use of this element for reconstructing salinity histories of Snook, and perhaps other species. In contrast, otolith Ba:Ca remained low and stable, despite exposure to various physiological perturbations, and may provide a very useful indicator of the occupation of low salinity habitats.

Continuous life history Ba:Ca and Sr:Ca profiles of 56 wild Snook collected from throughout Tampa Bay revealed significant plasticity in the types of juvenile habitats settled, as well as in the timing of ontogenetic movements from these habitats. Of the profiles examined, 55% exhibited otolith core signatures characterized by an opposing Sr:Ca and Ba:Ca pattern, followed by an inverted pattern, providing an indicator of the movement of larvae from high salinity, pre-settlement environments into mesohaline, tidally-influenced juvenile habitats. In contrast, nearly half (45%) the Sr:Ca and Ba:Ca profiles indicated settlement in higher salinity environments, suggesting a high degree of habitat plasticity for juveniles of this species. For fish that settled into mesohaline habitats, decreases in Ba:Ca and/or increases in Sr:Ca over the first several years of life signaled the ontogenetic transition out of the juvenile habitat, with the timing of emergence ranging from within the first year to age-3. Because conditions during early life may propagate into divergent behaviors in subsequent life stages, information on the experiences of early life and juvenile stages could help to inform whether the occupation of different juvenile habitat types, or the precocious or delayed emergence from those habitats, explain the peculiar spawning and movement habits that occur in this species.

Share

COinS