Behavior of Loggerhead Sea Turtles on an Urban Beach. I. Correlates of Nest Placement

Transcription

1 lourno1 of Herpelology, Vol. 29, No. 4, pp , 1995 Copyright 1995 Society for the Study of Amphibians and Reptiles Behavior of Loggerhead Sea Turtles on an Urban Beach. I. Correlates of Nest Placement Department of Biological Sciences, Florida Atlantic University, Box 3091, Boca Raton, Florida , USA ABSTRACT. - Loggerhead sea turtles nesting in Florida sometimes deposit their clutches on urban beaches. This study was undertaken at a city beach to determine correlations between physical variables and where nests were placed. Over a four year period, the distribution of nests on the beach was statistically identical. Nesting density variation at particular sites was unrelated to offshore depth profiles.or to beach width, but was strongly correlated with the presence of tall objects (clusters of mature Australian pine trees and rows of multi-storied condominiums) located between the beach and the city. There are no reports that females nest preferentially in front of tall objects (dune or vegetation) at natural rookeries. The response may be unique to urban rookeries where the nesting habitat is exposed to artificial lighting. Tall buildings and trees shielded the beach from city light, with the magnitude of the effect (and the number of nests) positively related to object elevation. Planting vegetation and reestablishing dunes on urban beaches may be effective methods for attracting nesting turtles to these sites. Most of the loggerhead sea turtles (Caretta caretta) found in the western Atlantic Ocean nest in the southeastern United States (Pritchard, 1989), with 90 % of all nests deposited on Florida's beaches (Ehrhart, 1989). In Florida, nesting densities are highest in southern Brevard County, Hutchinson Island, Jupiter Island- Hobe Sound, and Juno Beach (Murphy and Hopkins, unpubl. data, Conley and Hoffman, 1987). The beaches in these regions are adjacent to areas which are relatively undeveloped compared to the locations in south Florida (Dade, Broward, and southern parts of Palm Beach counties). Sea turtles prefer to nest on dark beaches (Mann, 1978; Witherington, 1986) and will avoid areas of an otherwise attractive nesting beach exposed to luminaries (Witherington, 1992). In highly urbanized south Florida, beaches are chronically exposed to city lighting. Loggerhead turtles nest at these sites but at reduced levels (Conley and Hoffman, 1987). We know of no studies at urban sites designed to determine where and why nests are placed in particular locations, or how placement affects the ability of hatchlings to locate the sea. In this paper, we examine correlates of nest placement; in the accompanying paper, we document how hatchling orientation is affected by nest location. There are several reasons why Boca Raton beach is ideal for such a study. First, the city protects all nests and maps their beach location. We had access to data for four consecutive nesting seasons, allowing us to determine if females consistently preferred some nesting sites over others. Second, nesting densities at Boca Raton are relatively high for an urban beach, decreasing the probability that any spatial pattern we discerned could occur by chance. Third, the beach is almost equally divided into regions backed by heavily vegetated public parks, and those backed by on-the-beach condominiums. We therefore had an opportunity to compare how females responded to each area. MATERIALS AND METHODS The study site was 7.32 km in length. It was bordered to the north by the Village of High-

2 M. SALMON ET AL. Ocean Depth Profile FIG. 2. View of the beach showing the variables measured in this study. The nesting zone spans the region between the highest and lowest nests on the beach. Other measurements include: (1) Object elevation for a condominium (foreground) and (2) a stand of Australian pine trees (background); (3) object width (for a condominium) and beach width (between objects behind the beach and the high tide [HT] strand line). Ocean depth profile (lower right) was measured from shore to deep water. See text for other measurements. buildings, low bushes growing on the dune (sea grape, Coccoloba uvifera), or tree clusters (Australian pine, Casuarina equisetifolia). (3) Silhouette width: The horizontal extent of buildings or of clustered vegetation, as measured from scaled county maps (Fig. 2). Measurements were made exclusively at the South beach where the borders of condominiums and tree stands were distinct. (4) Light intensity: A stellar photometer (Optec, model SSP; angle of acceptance = 16"), mounted on a small tripod, was used to measure light intensity on moonless nights. Each measurement was integrated over a 10 sec sampling period. An internal (Schott "V") filter narrowed system response to wavelengths between nm. Maximum transmission (74%) was at 520 nm, which closely matched wavelengths detected with greatest sensitivity by darkadapted juvenile green turtles (510 nm; Granda and O'Shea, 1972). The instrument was previously calibrated to give absolute measurements of light intensity at 500 nm. Measurements were made within 13 sites: four zones at the South beach and 9 zones at the North beach. At each site, light readings were made in four directions: (a) out to sea, just above the oceansky horizon; (b) toward land, just above the silhouette of any object on shore; (c) at a lower elevation in the same direction (centered within the object on shore); and (d) perpendicular to the beach (photometer 1.0 m above the sand at mid-beach level). (5) Depth profiles: In July, 1992, we made 20 transects (9 spaced evenly within the South beach; 11 within the North beach) oriented perpendicular to the shoreline. An initial sonar reading was made within 10 m of the surf zone (depths ranged between m). Six additional readings were made at 0.05 nautical mile (NM) intervals (as determined by Loran) to a distance 0.3 NM from shore (depths ranged from 9-10 m). Description of the Beach. -The South beach was bordered almost exclusively by condominiums (Fig. 1). One small park (South Inlet) was present. The North beach was almost equally divided between areas dominated by buildings (condominiums) to the south, and by three city parks (South Beach, Red Reef, and Spanish River) to the north. In a few locations, features near shore could limit access to the beach by gravid females. These include: boating activity at the Boca Raton inlet, a 70 m wide rock outcrop in shallow water within one zone located in-south Beach park, and a formation of three underwater concrete modules (an artificial "reef," about 100 m in length) located within one zone inside Red Reef Park. Nest Distributions.-Total numbers of nests were 834,964,725, and 970 for the 1989 to 1992 nesting seasons, respectively. Nest densities within zones during each year were not uniform (Kolmogorov-Smirnov "D values over

3 NESTING BY SEA TURTLES land Beach and to the south by the Browardl Palm Beach County line. The region was subdivided into 25 zones (#723-#747) of equal length (approx. 293 m). One zone (#728) was bisected by the Boca Raton Inlet (entrance into the intracoastal waterway) which divided the site into two regions: a North beach (19 zones) and a South beach (six zones, with one zone bisected by the inlet; Fig. 1). Recent (1987) county aerial maps showed the location of parks, buildings, and vegetation within all zones. We updated these maps as necessary. City workers used them to plot nest position with an estimated accuracy of k4 m. Throughout the nesting season (March to October; 1989 to 1992), morning surveys were made to locate eggs deposited the previous night. Each new nest was covered with a screen cage bearing a number, and its position was plotted on a map. Nest Distribution.-After nesting ended, the number of nests deposited within each zone was tallied. To determine if annual nest distributions within zones were uniform, observed distributions were compared to a uniform distribution using Kolmogorov-Smirnov tests. To determine if density patterns between years were statistically similar, zone totals were converted into percentages, then compared to one another using a Friedman x2 test. Correlates of Density Patterns.-Analyses showed that within years nests were not deposited uniformly and that between years, distributions were statistically indistinguishable. To determine which variables were correlated with differences in nesting density, features on the beach (beach width, the height and width of objects behind the beach, beach lighting conditions) and adjacent to shore (depths at set intervals from the beach) were measured. Spearman rank correlation coefficient analyses were used to determine which terrestrial features were significantly correlated with nesting density. Correlation coefficients for the depth profiles were compared within and between the North and South beaches using standard techniques (Zar, 1984). Specific procedures were as follows (Fig. 2). (1) Beach width: Distance (in m) between the spring high tide line and objects located at the upper margin of the beach. At different locations these were dunes, trees, bushes, or the sea walls of buildings. All measurements were made during July and August when beaches contained the largest numbers of nests. In 1991, we measured beach width in the center of 9 of the 14 zones located on the North beach. In 1992, similar data were obtained at 11 sites within six zones located on the South beach. Sites were selected to include the full range of FIG. 1. Boca Raton beach. The site is divided by an inlet into a short South and a longer North beach. Black squares are sites bordered by condominiums; regions filled by horizontal lines are city parks behind the beach (SR, Spanish River; RR, Red Reef; SB, South Beach; SI, South Inlet). State map (above) shows area location on Florida's east coast. variation in beach topography and nesting density which occurred along both beach areas. (2) Silhouette height: Elevation (in degrees) from the horizontal to the top of the tallest object immediately behind the beach. Measurements were made with an inclinometer (Edmund Scientific), positioned on the beach surface and centered within the "nesting zone" (the region between the highest and lowest nests on the beach; Fig. 2). The objects were either

4 NESTING BY SEA TURTLES Zones FIG. 3. Mean (+ SD) number of nests in 25 zones within the study site over the four year study. Most nesting occurred in the southern (left) two-thirds of the area. The Boca Raton Inlet is located in zone 728 (see Fig. 1). four years ranged from to ; in all aged by zone, nest density was higher in the cases, P < 0.01). Zones which contained rela- southern (predominantly condominium) than tively high (or low) numbers of nests in one in the northern (mostly park) beach zones (Fig. year also did so in other years. Thus there were 3). no statistical differences in the distribution of Along the east coast of the United States nests between years (xz = 1.0, P = 0.75, Fried- loggerheads nest about four times per season man test). When data for all years were aver- (Murphy and Hopkins, unpubl. data), most TABLE 1. Light intensity measurements on clear, moonless nights at 13 Boca Raton sites. Values are given in Optec units (1 unit = 1.3 x lo9 photons/s/cm2 at 500 nm). Sites directly bordering the beach are: C, condominium; AP, Australian Pines; SG, Sea Grape bushes. See Fig. 1 for a map of the park (Pk.) locations. Sector Site Stratford Arms Villa del Mar So. Inlet Park So. Inlet Park Sable Point Boca Mar So. Beach Pk. Red Reef Pk. Red Reef Pk. Ocean View San Remo Span. Riv. Pk. Span. Riv. Pk. Above East Beach Object object

5 564 M. SALMON ET AL. TABLE 2. Spearman rank correlation coefficient analysis of the relationships between sector nesting density and the light intensity measurements given in Table 1. Sample size = 13. None was significantly correlated with nesting density. Variable Rho Probability Object Above object Beach surface Out to sea commonly on a two or three year cycle (Richardson et al., 1978; Bjorndal et al., 1983). Thus over the duration of our study (four years), we estimate our data reflected the nesting preferences of no fewer than 500 individuals. Light Measurements. -Light intensity was lowest when measured toward the sea, from the beach, or from the center of objects behind the beach (Table 1). Sky measurements captured the glow of city lights and consistently yielded the highest light intensities. There were no significant correlations between average nesting densities (Fig. 3) and any of these light intensity measurements (Table 2). Beach Measurements.-At the North beach, site width in 14 zones (Table 3) was unrelated to ina, m e, E? 40 C $ 20 m > TABLE 3. Beach width, silhouette height and nesting densities within sectors of the north beach (1991 season). Abbreviations, as in Table 1. Silhou- No. ette Beach of height width Sector Site nests (9 (m) Sable Point Sable Ridge Marabella So. Beach Pk. So. Beach Pk. Golf Course Red Reef Pk. Red Reef Pk. Brighton Residence Ocean View San Remo Span. Riv. Pk. Span. Riv. Pk. (C) (C) (C) (AP) (SG) (SG) (SG) (SG) (C) (SG) (C) (C) (SG) (SG) nesting density during the 1991 season (Rho = , P = 0.98). Silhouette height was positively correlated with the number of nests in each zone (Rho = 2.809, P = 0.005). At the South beach (Table 4), there was no correlation between beach width (Rho = 0.34, * - - m 0' W meters i Whitehall South Stratford Arms FIG. 4. Placement of all nests (1992 season) in front of several condominiums located on the South beach. Lower diagram: aerial view of the beach, traced from county maps. Each dot is one nest. Upper diagram: elevation of the buildings. There are few nests between buildings, but many in front of the buildings. Note that more nests are found in front of taller condominiums.

6 NESTING BY SEA TURTLES 565 TABLE 4. Silhouette height, silhouette width, and beach width at 11 sites on the south beach. Abbreviations, as in Table 1. Nest numbers are those present on 27 July, 1992 (less than the season total shown in Fig. 4). Sector Site 3000 South Whitheall So. Whitheall So. Stratford Arms Villa del Mar Arvida Whitheall Placide Addison So. Inlet Pk. So. Inlet Pk. No. Silh. Beach of height Silh. Width nests c') Width* (m) ' Width was measured in mm from scaled aerial (county) photographs. ** Whitehall South consists of two adjacent buildings of identical elevation and width (see Fig. 4) P = 0.74) or silhouette width (Rho = 1.80, P = 0.07) and nest density. Silhouette height again was positively correlated with nest density (Rho = 2.49, P = 0.013). Thus throughout the beach females most often placed their nests in front of the tallest trees found in city parks, and in front of the tallest condominiums (Fig. 4). Nest Density and Underwater Profiles.-There were no statistical differences between the bottom profiles measured within the North and within the South beach (North x2 = 5.62,10 df, P > 0.9; South x2 = 5.28, 8 df, P > 0.5). Correlation coefficients for each beach were r2 = 0.96 (North) and r2 = 0.98 (South). These differences were statistically indistinguishable, indicating that depth increase with distance from shore was similar along the entire study site. Boca Raton turtles nest more often at some sites than at others. They nevertheless show a consistent nesting pattern. Most nests are placed at sites backed by tall objects (multi-storied condominiums and mature Australian pine trees) behind the beach. Nest Site Selection by Sea Turtles.-Several features are characteristic of the most productive sea turtle rookeries (Mortimer, 1982). Beaches are easily accessible from the sea; they contain sands which provide an optimal environment for embryological development; they present elevated platforms which minimize the probability that the nest will be destroyed by sformgenerated high tides and wave action (Provancha and Ehrhart, 1987); they are located where turtles (females, their eggs, and their hatchlings) will encounter fewer terrestrial predators (Mortimer, 1982); and for most species, they are located near to offshore currents which transport hatchlings to oceanic nursery habitats (Collard and Ogren, 1990). Few nesting beaches excel in all of these characteristics; most beaches may represent compromises between several physical and biological attributes (Mortimer, 1982). The green turtle rookery at Ascension Is., for example, is located where there are few terrestrial predators but where beach sands are less than optimal for egg survival (Mortimer, 1990). These considerations mean that features of importance at one site (such as Boca Raton) may not be important at other sites. However, if a feature is important it should be found at some rookeries. No other studies in the literature report that females place nests in front of tall objects. In fact, this feature is usually unimportant. The Melbourne nesting beach, for example, is a major rookery in Florida (25% of all nests in the state; Ehrhart, 1989). A low dune is located behind some portions of the beach but its presence is unrelated to variation in loggerhead nesting density (Witherington, 1986). In South Carolina, the beach is often backed by a low, rounded dune, but the most attractive nesting sites correlate with beach width and slope characteristics, not the presence of a dune (Caldwell, 1959). In Australia, loggerheads often nest on small, flat, and sparsely vegetated islands and cays (Gyuris and Limpus, 1988). Thus the tendency of turtles to nest in front of tall objects may represent a response which is unique to urban nesting beaches. Its importance may be related to another feature associated with urban habitats: high levels of artificial lighting. Artificial Lighting, Tall Objects, and Nesting at Boca Raton.-At Boca Raton, city lights illuminate the overhead sky (Table 1). Condominiums (most of which are unoccupied and dark during the summer months) and tall Australian pine trees (located in parks behind the beach) are interposed between the city and the ocean. When viewed from the beach, these objects pro-

7 566 M. SALMON ET AL. vide an obvious barrier which shadows the foreground, and thereby emphasize the sharp light intensity contrast between the sky glow and the much darker beach. We hypothesize that when turtles seek a suitable nesting site, this photic contrast may be important. Supporting evidence comes from plots of nest distributions in front of buildings. At such locations light intensity contrasts in the vertical plane (a function of building elevation) and in the horizontal plane (gaps or spaces between adjacent buildings) are evident. Both appear to influence the location of nests (Fig. 4) Further support for this hypothesis comes from our light intensity measurements. The data indicate that levels of light intensity vary from the beach with location (Table 1) and that nest density is unrelated to this variation (Table 2). These results are consistent with a hypothesis that sites are selected on the basis of light intensity contrasts, but not with a hypothesis that site selection depends upon absolute differences in light intensity. Management Implications.-From a management perspective, consistent nesting preferences are important as they suggest any change in the relative distribution of nests from ~ one ~- season to the next is probably related to a modification of the local area. The first priority of any management program should be to determine what has changed. The second should be to determine whether its effect has a positive or negative impact on nesting density. The third priority should be to use this information to make urban rookeries more attractive to nesting turtles. For example if the proportion of females nesting in certain areas increases immediately after reducing light levels or reestablishing dune vegetation, then it is likely that the effect is important. This study shows that nesting density patterns are a sensitive assay. For these reasons recording nest numbers and location should be a routine component of proper management procedure. These records should prove invaluable for determining trends and eventually, for predicting which changes degrade or enhance urban nesting beaches. While nesting densities at Boca Raton are low (mean of 99 nests/km) compared to those observed on relatively dark beaches (> 600 nests/ km at Melbourne beach; L. M. Ehrhart, pers. comm.), they nevertheless are high compared to other urban sites on the east coast. At Broward County's beaches immediately to the south, densities varied from 45 to 63 nestslkm be- darker area so that hatchlings are not attracted to the lights. We hypothesize that tall objects act as a light barrier, making urban sites otherwise more attractive to nesting turtles. If our hypothesis is correct, reducing light levels by the use of natural barriers (vegetation or reestablished dune) could increase nest densities at other urban beaches. We believe that such efforts would benefit Florida's nesting loggerhead population, which increasingly finds fewer dark beaches. Those few regions which are still relatively dark now receive many nests. It is unclear whether this pattern is a consequence of artificial lighting conditions associated with coastal development, or if it represents a historically natural pattern. Since there are no data which could be used to distinguish between these alternatives, the most conservative management strategy is to provide the nesting population with a large number of suitable and spatially separated nesting sites, including some urban beaches. An advantage of such a policy is that the impact of risks (natural disasters such as storms, or unnatural disasters such as oil spills or continued development behind presently dark beaches) is spread among fewer members of the breeding population. Attracting females is only half of an effective strategy for managing urban rookeries. Sites must also provide an environment which allows the hatchlings to reach the sea. In the accompanying paper, we describe when and where at Boca Raton lights interfere with hatchling seafinding behavior. The results provide further insight into how city beaches might be suitably modified to achieve these management goals. Acknowledgments.-This project was funded through a cooperative agreement from the National Oceanic and Atmospheric Administration, by Florida Atlantic University, and by personal funds. We are grateful to Alan Huff of the Florida DEP for his encouragement and steadfast support. Ann Broadwell, Barbara Squillante and the City of Boca Raton's Sea Turtle Program (administered through the Gumbo Limbo Nature Complex) provided the nest distribution data. Without their assistance, this project would not have been possible. Cathy F. Lohmann, Ken Lohmann, Blair Witherington, and two referees made suggestions which improved the manuscript. tween 1989 and 1992 (Burney and Margolis, un- LITERATURE CITED publ. data). It is likely that beach lighting is an BJORNDAL, K. A., A, B, MEYLAN, AND B.,, TURNER. important determining factor in this difference sea turtles nesting at ~ ~ ~ ~ l ~ So much artificial light reaches Br0wa1-d county Florida. I. Size, growth and reproductive biology. beaches that most nests must be relocated to a Biol. Conserv. 26:65-77.

8 NESTING BY SEA TURTLES 567 CALDWELL, D. K The loggerhead turtles of Cape Romain, South Carolina. Bull. Florida State Mus. Biol. Sci. 4: CONLEY, W. J., AND B. A. HOFFMAN Nesting activity of sea turtles in Florida, Florida Sci. 50: COLLARD, S. B., AND L. H. OGREN Dispersal scenarios for pelagic post-hatchling sea turtles. Bull. Mar. Sci. 47: EHRHART, L. M Status report of the Loggerhead turtle. In L. Ogren (ed.), Proceedings of the Second Western Atlantic Turtle Symposium, pp NOAA Technical Memorandum NMFS- SEFC-226. GRANDA, A. M., AND P. J. O'SHEA Spectral sensitivity of the green turtle (Chelonia mydas mydas) determined by electrical responses to heterochrornatic light. Brain, Behav. Evol. 5: GYURIS, E., AND C. J. LIMPUS The loggerhead turtle, Caretta caretta, in Queensland: population breeding structure. Aust. Wildl. Res. 15: MANN, T. M Impact of developed coastline on nesting and hatchling sea turtles in southeastern Florida. Florida Mar. Res. Publ. 33: MORTIMER, J. A Factors affecting beach selection by nesting sea turtles. In K. A. Bjorndal (ed.), Biology and Conservation of Sea Turtles, pp Smithsonian Institution Press, Washington, D.C The influence of beach sand characteristics on the nesting behavior and clutch sur- vival of green turtles (Chelonia mydas). Copeia 1990: PRITCHARD, P. C. H A Summary of the Distribution and Biology of Sea Turtles in the Western Atlantic. Proceedings of the Second Western Atlantic Turtle Symposium. NOAA Technical Memorandum NMFS-SEFC-226. PROVANCHA, J. A., AND L. M. EHRHART Sea turtle nesting trends at Kennedy Space Center and Cape Canaveral Air Force Station, Florida, and relationships with factors influencing nest site selection. In W. N. Witzell (ed.), Ecology of East Florida Sea Turtles, pp NOAA Technical Report NMFS 53. RICHARDSON, J. I., T. H. RICHARDSON, AND M. W. Dm Population estimates for nesting female loggerhead sea turtles (Caretta caretta) in the St. Andrew Sound area of southeastern Georgia, U.S.A. Florida Mar. Res. Publ. 33: WITHUIINGTON, B. E Human and natural causes of marine turtle clutch and hatchling mortality and their relationship to hatchling production on an important Florida nesting beach. M.S. Thesis, Univ. Central Florida, Orlando Behavioral responses of nesting sea turtles to artificial lighting. Herpetologica 48: ZAR, J. H Biostatistical Analysis. Prentice-Hall, New Jersey. Accepted: 19 July 1995.