CHANGE IN HABITAT USE AND POPULATION PARAMETERS FOR THE BOTTLENOSE DOLPHINS (Tursiops truncatus) OCCUPYING BOCA CIEGA BAY, FLORIDA

Allen, J.B., Eide, S.D., Forys, E.A., and J.E. Reynolds, III.

    Between June 1993 and August 1998, the Eckerd College Dolphin Project conducted ad libitum and
focal animal surveys of bottlenose dolphins (Tursiops truncatus) in Boca Ciega Bay, Florida, and its peripheral waters. Surveys were conducted primarily during the summer months and occurred over 274 days and 937 hours. A significant redistribution (p<<0.001) of the bottlenose dolphin was seen after mid-summer, 1995. In addition, mark-resight analysis using a Jolly-Seber model for open populations (assuming immigration emigration) provided the following population estimates: 1994-211 dolphins (95% CI=154-270 dolphins); 1995-211 dolphins (95% CI=150-271); 1996-287 dolphins (95% CI=224-351 dolphins); 1997-290 dolphins (95% CI=237-342 dolphins). Thus a 36% increase in population size occurred from 1995 to 1996. In addition, significant increases in mean herd size (p=0.001), sightings with feeding activity (8.2%; p=0.003), and sightings with calves (10.1%; p<<0.001) have been noted. Although causality cannot be proven, these findings are strongly correlated with an increase in some inshore fish stocks due to the Florida State ban on commercial netting of the inshore waters (Article 10, section 16) initiated on July 1, 1995. Although not assessed, other factors that may have also contributed to the observed changes in the dolphins biology include: changes in temperature, salinity, rainfall, incidence of red tides, or other physical/chemical changes; recovery of the area from a large oil spill in 1993; degradation of habitat elsewhere, leading to immigration of dolphins into Boca Ciega Bay; differences in observers; changes in food availability unrelated to the net ban; and differences in boat traffic amount or distribution. These results suggest the need for regular assessment of dolphin populations occupying coastal and inshore waters where combinations of human and natural factors can alter habitat, and thereby dolphin populations status, quickly.
 
 

EFFICIENT CORRECTION OF DIFFERENCE IN PERSPECTIVE, ORIENTATION AND SCALE IN DARWIN: DIGITAL ANALYSIS TO RECOGNIZE WHALE IMAGES ON A NETWORK

Roberts, Z.W., Wilkin, D.J., Debure, K.R. and J.E. Reynolds, III.

    Researchers engaged in the study of various aspects of dolphin biology (e.g., migrations, ranging patterns and social association patterns) frequently employ photo-identification techniques to identify individual dolphins in a geographic region of interest. Although quite effective, photo-identification can be a tedious and visually stressful task, particularly when large catalogs of known dolphins are employed. The quantity of image data and the somewhat constrained properties of dolphin dorsal fin images make the photo-identification process an appealing and very feasible task for automated processing. DARWIN is a computer vision system which allows a user to compare a digital image of a dolphin dorsal fin against a database of images of previously identified dolphins. Weighted mean squared error is used as a measure of similarity between the outline of the image of unknown fin and each fin in the database. This system presents the user with the images of the most similar fins from the database in rank order, significantly reducing the number of images the biologist need consider to verify the identity of the dolphin. The current software is fully-functional and provides a graphical user interface to facilitate the matching of individual images against the database. The difficulties associated with the task of comparing digital representations of dolphin dorsal fins include difference in the scale and orientation of the dorsal fin. Distortions which result from perspective differences are among the most difficult and computationally expensive to correct. Three techniques for representation and subsequent matching are compared in terms of their accuracy and efficiency.
 
 

ASSOCIATION PATTERNS OF BOTTLENOSE DOLPHINS (Tursiops truncatus) IN BOCA CIEGA BAY, FLORIDA

Viezbicke, J.B., Eide, S.D., Reynolds, J.E., and E.A. Forys.

    From 1993 through 1998, ad libitum and focal animal surveys of bottlenose dolphins occupying Boca Ciega Bay, Florida, and its peripheral waters was conducted by the Eckerd College Dolphin Project. Surveys were performed primarily in the summer months and occupied 274 days and 937 hours. Taking repeated photographs of distinctly-marked dolphins allowed mark-resight techniques to be used. The photo-identification catalog includes 662 dolphins to date. These data along with the appropriate software (in our case, SOCPROG, developed by H. Whitehead, and MATLAB) allowed association indices to be calculated. Seventeen individuals were selected for analysis due to their high frequency of sighting (more than 10 times each). Association indices, ranging from 0.0 to 0.62 +/- .149 (SD), were similar to those found by Brager et al. (1994) for dolphins in Galveston Bay, Texas. An index of 1.0 indicates animals seen together 100 percent of the time, while an index of 0.0 denotes no affiliation. The Galveston Bay dolphin population is open, in contrast to the closed population studied by Wells, Scott, and colleagues in Sarasota Bay, Florida. The low or moderate association indices for Boca Ciega Bay dolphins provide evidence for an open dolphin population. Additional evidence for an open population stems from the frequency of new dolphins that are cataloged during each survey year. Further, the Bay's dolphin population expanded from 210 dolphins in 1994 and 1995 to approximately 290 dolphins in 1996 and 1997 (Jolly-Seber model for open populations). Interestingly, Sarasota Bay and Boca Ciega Bay are approximately the same size and are within 25 kilometers of one another, but the dynamics of the dolphin populations in the two areas vary considerably. This raises questions about assumptions one can make for management purposes.
 
 

ASSESSMENT OF RANGING PATTERNS AND HABITAT PREFERENCE OF THE BOTTLENOSE DOLPHINS, Tursiops truncatus, UTILIZING A GEOGRAPHIC INFORMATION SYSTEM (GIS)

Wilkin, S.M., Forys, E.A., and J.E. Reynolds, III

    Home ranges for 19 individually identifiable bottlenose dolphins, Tursiops truncatus, were calculated using the Minimum Convex Polygon (MCP) method. Sighting data for these individuals were collected over a five-year study (1993-1998) by Eckerd College (EC) personnel in Boca Ciega Bay, parts of Tampa Bay and the Gulf of Mexico, Florida. Data were also used from studies conducted by the Sarasota Dolphin Research Program (SDRP) in the lower and coastal portions of Tampa Bay, and from the Clearwater Marine Laboratory in the Intracoastal Waterway north of Boca Ciega Bay. All dolphins analyzed were sighted five times or greater by EC to ensure a significant presence in Boca Ciega Bay. GIS software (ARC/Info and ArcView) was used to calculate the areas of all polygons and to exclude land from the MCP's. In addition, GIS was used to calculate differential habitat use with regard to bathymetry and seagrasses. A variety of ranging patterns was observed, including a transient individual and a potential home range shift. The majority of the animals ranged throughout the EC study area, with a high degree of overlap between individual's ranges. These individuals were not observed in other parts of Tampa Bay, including the southeast shore, despite high survey effort by SDRP in this area. Differential use of bathymetry was analyzed by comparing the proportion of the MCP at various depths with the proportions available in the entire study area, and was found to vary significantly from expected use (Chi-squared test, x²=55.1, p<0.0001, df=4). Analysis using Bonferroni confidence intervals indicated preference for waters shallower than 12 feet, and avoidance of water deeper than 12 feet. Dolphin use of seagrass was also examined, and found to be significantly higher than expected (Chi-squared test, x²=61.8, p<0.001, df=1).

Top of page

Doty, S.M., E.A. Forys, J.E. Reynolds, III, and R.S. Wells. 1998. Analysis using GIS of home range characteristics and habitat use by bottlenose dolphins, Tursiops truncatus, in Boca Ciega Bay, Florida, and surrounding waters. Proceedings of the Sixth Annual Atlantic Coastal Dolphin Conference, 1-3 May 1998, Sarasota, FL.

Forys, E.A., E.L. Gallizzi, J.E. Reynolds, III, S.M. Doty, and S.D. Eide. 1998. Field use of technology to assist photo-identification research on bottlenose dolphins in the Tampa Bay area, Florida. Proceedings of the Sixth Annual Atlantic Coastal Dolphin Conference, 1-3 May 1998, Sarasota, FL.

Top of page
 

Eide, S.D. and J.E. Reynolds, III.
Correlations between bottlenose dolphin (Tursiops truncatus) and presence of calves in Boca Ciega Bay, Florida and the Florida State ban on commercial netting. Proceedings of the World Mammal Science Conference, Monaco, 20-24 January 1998

Thoms, K.A., S.L. Carlson, M.E. Bolen, S.M. Doty, S.D. Eide, E.A. Forys, J.L. Odell, J.E. Reynolds, III, and J.B. Viezbicke.
The use of Geographic Information Systems to relate behavior of bottlenose dolphins (Tursiops truncatus) and environmental parameters in Boca Ciega Bay, Florida. Proceedings of the World Mammal Science Conference, Monaco, 20-24 January 1998.

Wilkin, D.J., J.H. Stewman, M.C. Allen, E.L. Gallizzi, and J.E. Reynolds, III.
DARWIN - Software to identify dolphins in digital images. Proceedings of the World Mammal Science Conference, Monaco, 20-24 January 1998.

Top of page

Eide, S.D., M.E. Bolen, S. Carlson, S.M. Doty, E.A. Forys, J.M. Odell, K.A. Thoms, and
J.E. Reynolds, III.
Assessment and habitat use of bottlenose dolphins, Tursiops truncatus, in Boca Ciega Bay, Florida. Proceedings of the 5th Annual Atlantic Coast Dolphin Conference, 4-6 April, 1997, Wilmington, N.C.

Top of page

Allen, M.A., D.J. Wilkin, A.P. Cameron, J.H. Stewman, and J.E. Reynolds, III.
DARWIN - Software to identify dolphins in digital images. Proceedings of the Eleventh Biennial Conference on the Biology of Marine Mammals, 14-18 December 1995, Orlando, Florida.

Kostka, J., J. Jeffers, J. Odell, J. Hood, M. Bolen, L. Tomaselli, M. Allen, E. McMahon, J. Reynolds, and B. Weigle.
Use of the Boca Ciega Bay, Florida, area by bottlenose dolphins, Tursiops truncatus. Proceedings of the Eleventh Biennial Conference on the Biology of Marine Mammals, 14-18 December 1995, Orlando, Florida.

Top of page

    Although a few studies have addressed levels of noise associated with certain human activities (e.g., oil and gas drilling, seismic surveying, shipping traffic, recreational boat traffic), and some studies have attempted to describe the effects of anthropogenic sounds on marine mammals, virtually no work has also included an assessment of "normal" sound types and levels experienced by these animals. Yet, studies of ambient noise in local habitats are an important part of assessing the impact of anthropogenic sounds on marine mammals. This study has two parts: (1) to design an acoustical data acquisition system and establish a methodology to assess ambient noise characteristic of the inshore habitat of Boca Ciega Bay, and (2) to digitally sample the acoustical contribution of the Pinellas Bayway drawbridge, Charlie, located in the Bay. However, the design and implementation of an acoustical data acquisition system at Eckerd College was not completed in time for part (2) of this study to be addressed. Anticipated results included the characterization of low frequency sounds being contributed to the local marine environment by Charlie and analyses suggesting that the contribution be substantial. Studies of ambient noise help scientists establish typical noise levels with which local marine mammals must contend. These data can facilitate our understanding of marine mammal behavior and habitat use, as well as conservation and management.

    Home ranges for 19 individually identifiable bottlenose dolphins, Tursiops truncatus, were calculated using a minimum convex polygon (mcp) method. Data for these individuals were collected over a five year study (1993-present) by Eckerd College (EC) personnel in Boca Ciega Bay, parts of Tampa Bay, and the Gulf of Mexico, Florida. Data were also used from studies conducted by the Sarasota Dolphin Research Program (SDRP) in the lower and coastal portions of Tampa Bay, and from the Clearwater Marine Laboratory, Clearwater, Florida, in the intra-coastal waterway north of Boca Ciega Bay. All dolphins analyzed were sighted 5 or greater times by EC, to ensure a significant presence in Boca Ciega Bay. GIS software (ARC/INFO and ArcView) was used to calculate the areas of all polygons and to exclude land from the MCPs. In addition, GIS was used to calculate differential habitat use with regard to bathymetry and seagrasses. An area-observation curve indicated that these animals did not have enough locations to adequately determine home ranges, so ranging patterns were analyzed. A variety of ranging patterns, including a transient individual and an animal having potentially undergone a home range shift, were observed. The majority of the animals ranged throughout the entire EC study area, with a high degree of overlap between individuals' ranges. However, animals were not observed in other parts of Tampa Bay, including the south east shore despite high survey effort by SDRP in this area. Differential use of bathymetry by the dolphins was examined by comparing the proportion of the MCP at various depth with the proportions of the total study area, and was found to vary significantly from the expected use (Chi squared test, x²=55.1, p<0.0001, df=4). Dolphin use of seagrass beds was also examined, and was found to be significantly different than expected (x²=61.8, p<0.001, df=1). Analysis using Bonferroni confidence intervals indicated a significant preference for waters shallower than 12 feet, and avoidance of waters deeper than 12 feet, especially waters deeper than 18 feet, including channels and spoils.

    On July 1, 1995 the state of Florida initiated a ban on commercial netting in inshore waters less than three miles offshore (Article 10, section 16). The goal of the ban was to protect and replenish depleted fish stocks. Since its initiation, decreased commercial catches of some of the fish stocks (e.g., Mugil cephalus) have resulted. It is possible that the net ban has also had beneficial impacts to inshore ecosystems of Florida. Often it is cost effective to monitor status of a single species as an ecosystem indicator, rather than to assess a variety of parameters of species. Due to its behavioral plasticity, including feeding habits and certain other biological or life history characteristics, the bottlenose dolphin (Tursiops truncatus) may be useful as an ecosystem indicator. Data obtained through ad libitum boat surveys of Boca Ciega Bay, Florida, and adjacent waters were used to compare of bottlenose dolphins' distribution, feeding frequency, and presence of calves during pre-net ban and post-net ban periods from June 1993 through August 1997. Maps produced using Geographic Information Systems and statistical analyses (G-test) were used to demonstrate differences. A significant redistribution (p<0.001) of the bottlenose dolphin (e.g., a 31% increase in sightings in sector BCN) occurred in inshore waters after the net ban. A significant increase (0.01<p<0.025) in sightings with feeding and probable feeding (7%), as well as sightings with calves (17%;0.05<p<0.1) was also found post-net ban.  These correlations show changes have occurred in a bottlenose dolphin community since the initiation of the net ban; however, the numbers do not prove a cause and effect relationship between the net ban and abundance of fish or abundance of fish and changes in the bottlenose dolphin community.

    Computer vision and image processing techniques can be used to solve a variety of problems and perform tasks that humans find difficult, tedious, or time consuming. Recently, several software applications that apply these techniques to the photo-identification of various marine mammals have been developed. The subject of this thesis is the design and development of a computer vision application which can assist biologists with the process of identifying individual Atlantic bottlenose dolphins, Tursiops truncatus. Many marine mammalogists currently use a system of photo-identification to gather data about bottlenose dolphins. They take hundreds of photographs of dorsal fins each year and manually identify the dolphins by the comparison of the characteristic nicks, notches, bumps, and scars on the fins with those in previously identified photographs. The DARWIN system matches newly photographed fins with previously identified dolphins and organizes the images in rank order based on the similarity between fins. The software reduces the tediousness of photo-identification. Image processing routines produce a list of 500 points that approximate the fin outline, and then calculate the center point, known as the centroid. The distance from the centroid to the outline is calculated as a function of angle in degrees to create a one dimensional graph known as a signature. The mean squared error between two signatures determines the rank of database fin. Thorough testing of the system revealed that matching performance was degraded by: 1) misplacement of the centroid because of variations in the user designation of the beginning and end of each fin outline; 2) sensitivity of signature scaling to these same terminal points; and 3) distortion of the fin shape due to out-of-plane rotations. The positions fo the centroids are corrected during the matching process by shifting them and recalculating the signatures repeatedly. The scaling problem is minimized by scaling the signatures based only on the region of overlap between the new and database signatures. Additionally, the outlines are rotated three dimensionally to correct the fin shape for distortion due to perspective. A mammalogist using DARWIN would have to search only a small subset of the database to find a correct match, thus saving valuable time.

    The ability to identify individuals within a group is essential to the study of animal social systems and association patterns. Photo-identification is a method commonly used to identify marine mammals based on distinctive natural and anthropogenic markings. Boat-based surveys were conducted in Boca Ciega Bay, Florida, to study the association patterns of bottlenose dolphins, Tursiops truncatus. Analyses of the half-weight indices of association showed that the dolphins form mostly low-level associations and often have more associates than the mean group-size. The fluidity of the social system in Boca Ciega Bay may be due, in part, to its use as a passage between surrounding areas.

    Photo-identification is a method of using photographs to recognize individuals of a particular species and, thereby, to obtain information about those individuals and their population. Photo-identification is tremendously useful but also problematic in that it is difficult to efficiently process and analyze the number of photos taken on an annual basis by dedicated research programs. The DARWIN (Digitally Automated Recognition of Whale Images on a Network) System was developed to assist marine mammalogists who utilize photo-identification to study the bottlenose dolphin, Tursiops truncatus. The DARWIN System ideally will increase the speed and accuracy of the process by which marine mammalogists can analyze their data. The conceptual design of the DARWIN System links an automated dolphin dorsal fin matching scheme with a client-server database architecture. The current version of DARWIN, however, is a stand-alone system which utilizes computer vision to semi-automatically analyze and match dolphin dorsal fin images. Specifically, the system attempts to extract meaningful features from two-dimensional curves (or active contours) which approximate the outline of the dorsal fin, and then match these features. The large-scale shape characteristics of dolphin fins are represented through a one-dimensional signature, and these signatures are compared using a mean squared error calculation. Results from testing indicate that the signature can effectively match fins with highly unique fin shape. The signature descriptor is unreliable with substantial changes in scale or variability in the start- and end-points of the fin boundary.

Top of page

The Bottlenose Dolphin: Biology and Conservation

The Bottlenose Dolphin Biology and Conservation John E. Reynolds III, Randall S. Wells, and Samantha D. Eide "This excellent introduction on the biology of the bottlenose dolphin also provides a review of conservation issues and outlines current knowledge of dolphins in general.... for students, professionals, and anyone interested in the bottlenose dolphin." -Charles W. Potter, National Museum of Natural History    The Bottlenose Dolphin presents for the first time a comprehensive, colorfully illustrated, and concise overview of a species that has fascinated humans for at least 3,000 years.    After reviewing historical myths and legends of the dolphin back to the ancient Greeks and discussing current human attitudes and interactions, the author replaces myths with facts--up-to-date scientific assessment of dolphin evolution, behavior, ecology, morphology, reproduction, and genetics--while also tackling the difficult issues of dolphin conservation and management.    Although comprehensive enough to be of great value to professionals, educators, and students, the book is written in a manner that all dolphin lovers will enjoy. Randall Wells's anecdotes interspersed throughout the work offer a firsthand view of dolphin encounters and research based on three decades working with them. Color photographs and nearly 100 black and white illustrations, including many by National Geographic photographer Flip Nicklin, beautifully enhance the text.

Readers of The Bottlenose Dolphin will better appreciate what dolphins truly are and do, as well as understand some of the controversies surrounding them. While raising compelling questions, the book provides a wealth of information on a legendary species that is loved and admired by many people.  John E. Reynolds, professor of marine science at Eckerd College, St. Petersburg, Florida, is chair of the U.S. Marine Mammal Comrnission. He has written over 100 articles on marine mammal biology and conservation and is coauthor with Daniel K. Odell of Manatees and Dugongs and coeditor of Biology of Marine Mammals. Randall S. Wells is a behavioral ecologist with the Conservation Biology Department of the Chicago Zoological Society and adjunct associate professor of ocean sciences at the University of California, Santa Cruz. He also serves as director of the Center for Marine Mammal and Sea Turtle Research at Mote Marine Laboratory, Sarasota, Florida, where he conducts the world's longest running study of wild dolphins. Samantha D. Eide, a graduate student at the University of South Florida, is field leader for the Eckerd College Dolphin Project, St. Petersburg, Florida. August.  304 pp.  6 X 9. Jjacket.  13 color plates, 59 b&w photos, 36 figures, 4 tables, glossary, notes, bibliography, index. ISBN 0-8130-1775-0  Cloth,  $24.95

Top of page