Bottlenose Dolphins as Marine Ecosystem Sentinels: Developing a Health Monitoring System
- First Online:
- Cite this article as:
- Wells, R.S., Rhinehart, H.L., Hansen, L.J. et al. EcoHealth (2004) 1: 246. doi:10.1007/s10393-004-0094-6
- 1.2k Downloads
Bottlenose dolphins (Tursiops truncatus), as long-lived, long-term residents of bays, sounds, and estuaries, can serve as important sentinels of the health of coastal marine ecosystems. As top-level predators on a wide variety of fishes and squids, they concentrate contaminants through bioaccumulation and integrate broadly across the ecosystem in terms of exposure to environmental impacts. A series of recent large-scale bottlenose dolphin mortality events prompted an effort to develop a proactive approach to evaluating risks by monitoring living dolphin populations rather than waiting for large numbers of carcasses to wash up on the beach. A team of marine mammal veterinarians and biologists worked together to develop an objective, quantitative, replicable means of scoring the health of dolphins, based on comparison of 19 clinically diagnostic blood parameters to normal baseline values. Though the scoring system appears to roughly reflect dolphin health, its general applicability is hampered by interlaboratory variability, a lack of independence between some of the variables, and the possible effects of weighting variables. High score variance seems to indicate that the approach may lack the sensitivity to identify trends over time at the population level. Potential solutions to this problem include adding or replacing health parameters, incorporating only the most sensitive measures, and supplementing these with additional measures of health, body condition, contaminant loads, or biomarkers of contaminants or their effects that can also be replicated from site to site. Other quantitative approaches are also being explored.
Keywordsbottlenose dolphinecosystem healthsentinel speciesrisk assessment
Bottlenose dolphins (Tursiops truncatus), can serve as important barometers of the health of marine ecosystems. They are long-lived, long-term coastal residents in tropical and temperate regions throughout the world (Wells and Scott, 1999; Reynolds et al., 2000). Long-term research on such a species allows one to document the history of exposure to ecosystem perturbations and their effects. They are top-level predators on a wide variety of fishes and squids, and thus concentrate contaminants through bioaccumulation and integrate broadly across the ecosystem in terms of exposure to environmental impacts. Dolphin health and population status not only reflect the effects of natural and anthropogenic stressors on the species, but they serve as sentinels of the health and status of lower trophic levels in the marine ecosystem.
Over the last 17 years, another reason to monitor bottlenose dolphin health and population status has emerged due to the occurrence of large-scale dolphin mortality events. During 1987–1988, it was estimated that half of the putative coastal migratory stock of bottlenose dolphins died along the mid-Atlantic coast of the U.S. (Scott et al., 1988), leading to a designation of this stock as “depleted.” Other unusual dolphin mortality events occurred in the northern Gulf of Mexico in 1990, 1992, and 2000, in which hundreds of bottlenose dolphins died (Hansen, 1992). A variety of factors, such as environmental contaminants, natural biotoxins (Geraci, 1989), and morbillivirus (Duignan et al., 1996; Lipscomb et al., 1996), have been suggested as probable agents responsible for the mortalities. In each of these cases, conclusive determination of the cause of death was hampered by a shortage of fresh carcasses to allow all desired tests to be performed. Evaluation of the impacts of the mortalities was hampered by inadequate background information on dolphin stock structure, abundance, life history, and vital rates prior to the mortalities. The lesson reinforced by each subsequent investigation is that a proactive approach to evaluating risks by monitoring living dolphin populations would be preferable to waiting for large numbers of carcasses to wash up on the beach.
The authors have been developing methods for assessing the population status and health of coastal bottlenose dolphins, not only to monitor the risks to the populations themselves, but also to be able to use them as sentinels of the health of marine ecosystems. Much of this work has occurred in Sarasota Bay, Florida, where research on the resident dolphin community has been ongoing since 1970, and where four generations of identifiable individuals of known gender, age, and genetic relationships are currently under study (Irvine and Wells, 1972; Irvine et al., 1981; Scott et al., 1990a; Wells, 1991). Population monitoring efforts benefit from the fact that at least 60% of inshore dolphins on the west coast of Florida are individually identifiable from dorsal fin features, facilitating direct counts and mark-recapture estimates (Wells, 2002; Wells and Scott, 1990; Wells et al., 1996a, 1996b, 1997). Using photographic identification techniques (Scott et al., 1990b; Würsig and Jefferson, 1990) it is possible to define individual ranges (e.g., relative to contaminant sources) and measure female reproductive success as well as monitor population-level trends in abundance, losses, and other vital rates (Wells and Scott, 1990). In addition, the shallow waters of much of the habitat of inshore bottlenose dolphins facilitate safe capture and release operations, in which dolphins can be examined by veterinarians and sampled for subsequent health-related analyses. This article focuses on the information for health assessment that can be derived from directly examining and sampling dolphins.
Dolphins are captured for examination and sampling by encircling them with a 500 m × 4 m seine net in shallow waters where handlers can safely stand and support dolphins as necessary. One at a time, each dolphin is transferred to foam pads on the shaded deck of a boat, where it is weighed and a standard series of length and girth measurements is collected. Adult females are first given an ultrasound examination for pregnancy before a decision is made to bring them aboard the vessel. Throughout the examination, behavior and respiratory patterns are closely monitored, and water is sponged over the animals. Blubber depth is measured ultrasonically at standard sites. Abdominal and thoracic organs are evaluated via ultrasound examination. Core body temperature is measured through a colonic probe. Blood samples (up to 320 ml) are collected through venipuncture from a vessel in the fluke. Blood samples are: 1) analyzed for standard chemistry, hematology, and reproductive hormones; 2) used for immunological studies; 3) applied to genetic studies including paternity analyses; 4) examined for circulating levels of environmental contaminants; and 5) stored for retrospective investigations of disease processes. Urine is obtained through sterile catheterization. Milk is expressed into a custom suction collection system for compositional analyses and measurements of environmental contaminants. Samples for evaluation of the presence or absence of a suite of microorganisms are collected from the blowhole, feces, and genital tracts. Small blubber wedges (4-cm long × 3-cm wide × up to 1.5-cm deep) are obtained under local anesthesia from a standard location below the dorsal fin, for contaminant analyses. Most of the Sarasota Bay residents are of known age through observation, but a tooth is collected under local anesthesia from individuals of unknown age for sectioning and counting of growth layer groups (Hohn et al., 1989). A full examination and sampling program typically requires about 1 hour, and then the dolphin is returned to the water and released. More than 180 individuals have been examined in Sarasota Bay to date. Similar sampling was conducted by the National Marine Fisheries Service (NMFS) in Matagorda Bay, Texas in 1992 (n = 36 dolphins) and near Beaufort, North Carolina in 1995 (n = 31 dolphins) (Hansen and Wells, 1996).
Develop and refine a “grading system” for the health of individual animals examined and sampled during the annual Sarasota bottlenose dolphin capture–release project, based on consideration of the full range of health and condition parameters measured. The grading system should be based on defensible, objective criteria;
Develop and refine a system to extrapolate the individual animal grades to a seasonal grade for the population as a whole based on the animals sampled during a given season;
Identify the indicators of greatest value in assessing the overall health of an individual or population; and
Relate trends in population health grades to trends in population parameters such as mortality or loss rates, natality, and fecundity.
Data Sources for Development of Health Assessment Protocol of Bottlenose Dolphins (Tursiops truncatus) in Sarasota Bay, Florida
Collected from Sarasota Bay dolphins since
Blood chemistry and hematology
Results of physical examinations
Sex, age, morphometrics, life history
Body condition: weight, girth, blubber depth
Diagnostic ultrasound examination
Results of necropsies performed by Mote Marine Laboratory
Population vital rates
Behavior: ranging and social patterns
Blubber biopsy wedges
Blood/milk for contaminant analyses
At the workshop, the biologists worked with the veterinarians to evaluate their grading system, and to identify the most informative health parameters of those that have been or could be measured regularly and reliably in the field. The authors evaluated a variety of potential assessment parameters, including 56 blood hematology and chemistry measures, weight, girths, blubber depths, and the presence or absence of 80 microorganisms. These parameters were evaluated relative to the sex, age, and reproductive condition of each individual, and the type of sample (serum vs. plasma) and the analytical laboratory used (Specialty Veterinary Laboratory Services [SVLS], Santa Cruz, CA vs. SmithKline-Beecham, locations throughout the US). From this list, we selected a set of the most indicative parameters that allowed us to develop a quantitative measure of the health of each individual. A mean annual health score for the population as a whole was calculated. The scores provided a set of tentative baseline values for comparison with other populations.
HEALTH ASSESSMENT SCORING SYSTEM
Baseline Blood Values, and Points scored if Outside of the Baseline Values of Bottlenose Dolphins (Tursiops truncatus) in Sarasota Bay, Floridaa
≥7.0 (if adult)
Calf/juvenile (0–4.9 years)
Subadult (5–11.9 years)
Adult (12 + years)
Alanine aminotransferase—ALT (or GGT or AST)
Aspartate aminotransferase—AST (or GGT or ALT)
Gamma glutamyl transferase—GGT (or ALT or AST)
Blood urea nitrogen—BUN
Hemoglobin—Hb (or hematocrit)
Hematocrit—HCT (or hemoglobin)
Erythrocyte sedimentation rate—ESRc
A (0–4 points)—The dolphin is apparently in good health, with no obvious medical problems or need for follow-up medical attention.
B (5–9 points)—The dolphin would benefit from a follow-up veterinary examination.
C (10–19 points)—The dolphin would benefit from medical treatment.
D (>20 points)—The dolphin has a serious medical problem that requires treatment.
Several parameters were considered to be duplicative and therefore interchangeable if necessary. For example, only one of alanine amino transferase (ALT), gamma glutamyl transferase (GGT), and aspartate aminotransferase (AST) was used for any given score (the highest value of the three was used), but in the absence of any two of these, the other was acceptable for scoring. The higher of the values for hemoglobin and hematocrit was selected. Similarly, the total leukocyte count (WBC) would have duplicated scoring of the leukocyte differential, so only the differential was scored.
The reproductive condition, age class, and analytical laboratory affected some variables. Hemoglobin and hematocrit were scored differently for lactating vs. nonlactating females. Alkaline phosphatase (AP) was scored differently for calves/juveniles, subadults, and adults. Comparisons of health scores between age classes within genders found a significant difference (P < 0.01) only between adult and subadult females, with subadults exhibiting better health (mean = 3.0, SD = 2.70, n = 25 vs. adult female mean = 5.6, SD = 4.64, n = 38).
Consistent differences between the two laboratories used prior to the workshop (SVLS and SmithKline-Beecham) led to different scoring for albumin, globulin, and bilirubin. SmithKline-Beecham was selected as our standard laboratory due to widespread availability of linked laboratories throughout the U.S. However, after 1995 SmithKline-Beecham no longer performed analyses on veterinary samples. Subsequent analyses were performed by a variety of laboratories (University of Miami, Sea World, Cornell University) in an attempt to find one that was both easily accessible and would perform reliably all 19 of the analyses required by the scoring system. For example, some laboratories would not perform analyses for iron or erythrocyte sedimentation rate. As a result, the data presented here reflect only the 1990–1995 period during which a consistent laboratory (SmithKline-Beecham) performed all 19 analyses.
EVALUATION OF THE HEALTH SCORING SYSTEM
Summer Health Scores of the Bottlenose Dolphin (Tursiops truncatus) Population in Sarasota Bay, Florida, 1990–1995
Possible relationships between dolphin health scores and mortality patterns, as one extreme of health, were examined using three measures: 1) total number of Tursiops strandings recovered by Mote Marine Laboratory, as an indication of health conditions across a broad geographic area; 2) losses of known Sarasota residents of all ages; and 3) losses of known Sarasota residents of >1 year of age (the category most directly comparable to the set of animals sampled for the health assessment scoring). No correlation was found between the total number of strandings and health scores for any given year. Positive correlations were obtained between annual health scores and losses of known Sarasota dolphins when 19 blood parameters were considered, but the correlations were not statistically significant. When only 16 blood parameters were considered (reflecting limitations provided by some laboratories), no correlation was found. These results suggest that the health index might benefit from further refinement through the inclusion of additional parameters.
The validity of the scoring system approach was further supported anecdotally by several case histories. For example, FB19, an adult female, maintained a “B” grade during five samplings from 1988 through 1991. She declined to a “C” in June 1993, and died, at 50 years of age, in January 1994. FB97, a dependent female calf, received a “C” score in June 1989, and disappeared several months later. FB206, a stranded adult male, arrived at Mote Marine Laboratory in June 1993 with a score of 42 (“D”). He underwent medical treatment for 107 days, before being reintroduced to the wild. Upon release, his score had improved to 9 (“B”) and he was observed in the wild 6 months later in apparently good condition. Thus, the scoring system appears to accurately track individual medical conditions in at least some cases.
FUTURE OF THE HEALTH SCORING SYSTEM
Though the scoring system in its current form appears to roughly reflect dolphin health, its general applicability is hampered by several factors. Interlaboratory variability is perhaps the most serious problem associated with the blood parameter health scoring method. The requirement of using the same laboratory/analytical techniques in order to be able to identify trends over time within a site/individual or to make valid comparisons across sites has been problematic. Variability due to different analytical techniques across laboratories, changes in laboratory availability, and the occasional lack of availability of analyses of some of the critical blood parameters can negate the utility of this additive approach. Missing values potentially bias the health scores downward. The blood parameter algorithm also suffers from a lack of independence between some of the variables, and the possible effects of weighting variables remain to be evaluated.
The high variance associated with the scores resulting from the current health assessment scoring system seems to indicate that the approach may lack the sensitivity to identify trends over time at the population level. We are examining several potential solutions to this problem. One refinement might be to consider adding or replacing health parameters, incorporating only the most sensitive measures. The initial approach of limiting the scoring system to blood parameters was in part an attempt to obtain objective, quantitative, comparable measures regardless of the specific composition of the investigating team or the location of the research site. However, it may be advisable to identify a subset of blood parameters of clinical significance that can be measured most reliably across laboratories, and supplement these with additional health or body condition measures that can also be replicated from site to site. For example, age/sex/season-specific weights, girths, and blubber depths could provide important measures of body condition. Measures of thermal characteristics from colonic temperatures, surface temperatures, heat flux, or thermal imaging could provide indications of health status (Meagher, 2001). Measures of immunocompetency could also be incorporated (Lahvis et al., 1993; Erickson et al., 1995). Integration of data about skin lesions could be useful as well (Wilson et al., 1999). Ultrasonic imaging of organs and microbiological cultures give insights into health, but the means to quantify these data for inclusion in a scoring system remain to be developed.
As we begin to understand more about the effects of environmental contaminants on dolphin health (O’Shea, 1999), incorporating measures of contaminant loads or biomarkers of contaminants or their effects becomes increasingly important. Even in a relatively nonindustrialized area such as Sarasota Bay, inorganic and organic pollutants may be health factors. Rawson et al. (1993, 1995) related mercury-associated pigment granules and liver disease to concentrations of mercury in stranded dolphins from the Sarasota Bay area. Lahvis et al. (1995) reported an apparent relationship between increasing concentrations of organochlorine contaminants (PCBs and DDT metabolites) and decreasing Sarasota Bay male immune system function as indicated by lymphocyte proliferation, but their sample size was small. Though as yet unconfirmed, one possible explanation for the high rate of first-born calf mortality in Sarasota Bay (Wells et al., 2001, 2003) may relate to transfer of environmental contaminants from mother to calf, as described by Cockcroft et al. (1989) for dolphins in South Africa. Vedder (1996) provided some support for this hypothesis in her findings from analyses of Sarasota Bay dolphin milk from females of different ages and reproductive histories, in which she noted apparent depuration of contaminants. Similarly, Küss (1998) found elevated PCB and chlorinated pesticide concentrations in carcasses of young Sarasota Bay resident dolphins followed by a decline in maturing individuals, and subsequent increase as males aged, and decline or steady state at lower concentrations as females reared offspring. Schwacke et al. (2001) and Wells et al. (2003) identified a similar pattern from analysis of blubber samples from living Sarasota Bay dolphins. Significant challenges remain in identifying the specific chemicals, congeners, concentrations, interactive effects, and biomarkers (e.g., DNA damage: Gauthier et al., 1999) of greatest importance to dolphin health, and the most appropriate and accessible sites for sample collection, recognizing that contaminants are not distributed uniformly through blubber (O’Shea, 1999).
The authors are also beginning to approach the concept of population health assessment from a different quantitative perspective. Given that large-scale health problems typically do not affect all members of a population to the same degree or at the same time, it may be more appropriate to examine health scores by looking in detail at individuals that are outliers, rather than at population means. Changes in the proportions of outliers over time may provide a more meaningful and sensitive evaluation of population trends and risks, as the outliers themselves may serve as potential sentinels of environmental problems. This approach would also eliminate the need to compare specific parameter values directly from year to year, as the most important parameter comparisons would be within-year. It should be possible to test this approach both within the Sarasota Bay reference population and across study sites as expanded bottlenose dolphin health assessment programs are planned for the next several years.
Veterinary examinations and sampling were conducted through the support of the National Marine Fisheries Service, the U.S. Environmental Protection Agency, the National Science Foundation, Earthwatch Institute, Dolphin Quest, and the Chicago Zoological Society. W. Jarman, J. (Vedder) Greene, and associates conducted preliminary analyses of contaminant levels in blood and milk samples. The participation of J. Buck, C. Driscoll, J. Reif, and G. Worthy in the 1993 Bottlenose Dolphin Health Assessment Workshop provided valuable input for development and evaluation of the approach. The field sampling in Sarasota Bay benefited greatly from the efforts of B. Irvine, L. Fulford, K. Urian, S. Hofmann, K. Hull, and S. Nowacek, along with a host of collaborating researchers and volunteer dolphin handlers. This research was conducted under Scientific Research Permits Nos. 417, 655, 945, and 522-1569 issued by the National Marine Fisheries Service.