Using Strip-transect Aerial Surveys to Estimate Manatee Abundance in the Ten Thousand Islands Region of Southwest Florida

Lynn W. Lefebvre¹, Dean E. Easton¹ and Terry J. Doyle<sup>2

1) U.S. Geological Survey, Florida Integrated Science Center,
Sirenia Project, 412 NE 16th Ave., Room 250, Gainesville, FL 32601

2) U.S. Fish and Wildlife Service, Ten Thousand Islands National Wildlife Refuge,
3860 Tollgate Blvd., Suite 300, Naples, FL 34114</sup>
 

Poster presentation at the 16th Biennial Conference of the Estuarine Research Federation, St Petersburg Beach, November  4 - 8, 2001.

Strip-transect aerial surveys have been used extensively in Australia to estimate trends in offshore dugong populations (Marsh and Sinclair 1989; Marsh 1995). The use of strip–transect methods in estimating manatee population size and trend, however, has been limited (Miller et al. 1998).   Manatee surveys have typically not been designed to sample quantified survey areas, or to produce estimates of abundance. While useful in obtaining minimum manatee counts and distribution information, the latter surveys do not permit statistical comparison of survey results over time (Lefebvre et al. 1995).

Figure 1. Distribution of manatee groups during  17 aerial surveys in the Ten Thousand Islands. Data were collected by the U.S. Fish and Wildlife Service between August 1999 and February 2001.

Our ultimate objective is to determine if manatee density and distribution in the nearshore waters of the Ten Thousand Islands and the Everglades National Park change in response to restoration of natural hydrologic patterns in southwestern Florida.  The Ten Thousand Islands region is of particular interest because of proposed changes to the Southern Golden Gate Estates and Faka Union Canal drainage. We want to statistically compare pre- and post-restoration indices of manatee abundance. We also believe that strip-transect methods are likely to be successful in the Ten Thousand Islands region, unlike many other regions of Florida, in which manatees may be highly aggregated at winter sites, or their density may be too low and distribution too linear to permit this approach.

Six surveys were conducted between 25 July and 22 October 2000 and eight surveys were conducted between 15 July and 30 August 2001.  We established parallel transects, 1 km apart, with a survey strip width of approximately 250 m. Transects flown during July – October 2000 were oriented perpendicular to shore, between Palm Bay and the Ferguson River (Fig. 2).  Based on results from these surveys, we omitted 5 transects (26 – 30, Fig. 2) and established 5 new transects (31 – 35, Fig. 2) near Cape Romano prior to the 2001 surveys. Transect lengths ranged from 6.6 to 8.4 km and 3.4 to 8.4 km in 2000 and 2001, respectively; water area surveyed ranged from 0.79 to 1.53 km² per transect in 2000 and 0.83 to 1.53 km² per transect in 2001 (Table 1).  Two of the transects, number 31 and 32, were completely over water (Fig. 3).

Manatee locations were plotted on topographic maps, and flight paths were recorded on a Trimble Basic Plus GPS.  Surveys were conducted from a Cessna 172 at an altitude of 153 m, traveling at approximately 120-140 km per hour. Perception bias, which occurs when some of the manatees visible within a strip transect are missed by an observer, was estimated by applying a Petersen mark-recapture model to counts made by two observers (Pollock and Kendall 1987; Marsh and Sinclair 1989).

Figure 2. Spatial arrangement of 30 manatee aerial survey strip-transect polygons in the Ten Thousand Islands during flights July-October 2000 (top) and July-August 2001 (bottom).

Table 1. Length of and area covered by 30 aerial survey strip transects in the Ten Thousand Islands.  Surveys were conducted between 22 August - 22 October 2000 and 15 July – 30 August 2001.

Manatee group locations for all survey dates are shown in (Fig. 4).  The corrected number of manatee groups (a group = 1 or more individuals in the same location) sighted on transects ranged from 7.0 to 25.7 and 12.9 to 27.1 per survey during 2000 and 2001, respectively (Table 2).  The corrected number of individuals counted ranged from 10.0 to 39.8 in 2000, and from 15.1 to 61.7 per survey in 2001 (Table 2).  Mean group size per survey ranged from 1.0 to 2.0 during 2000 and 1.1 to 2.3 during 2001. Survey-specific population estimates in this study ranged from 39 to 164, or 1.09 to 4.57 per km² during 2000 and 58 to 237, or 1.62 to 6.64 per km² in 2001. Annual population estimates from these surveys were 9927 (S.E.) in 2000 (n=5 flights) and 12619 in 2001 (8 flights) (Table 3).

Excluding the Cape Romano transects, the overall distribution of sightings was somewhat bimodal, with average (= 0.44 groups per transect) or higher than average number of groups sighted on transects 1-9 and 17-21 during 2000 (Fig. 4).  Transect 6 starts near the mouth of the Barron River, and Transects 19 and 20 start near the mouth of the Faka Union Canal.  Virtually no manatee sightings were made on transects 25-30, at the western end of the study area during 2000.  The replacement of these five transects with five transects near Cape Romano (31-35: Fig. 2) in 2001 produced a somewhat higher estimate of manatee abundance in the region. The average number of groups was also higher (0.62 groups per transect) in 2001.

The Faka Union Canal is known to attract large numbers of manatees, particularly in the winter, presumably because of the availability of freshwater at its head and thermal buffering provided by its depth. In this study, we considered the canal to be a separate, high-density stratum, analogous to the "hot spots" described by Miller et al. (1998).  When manatee counts from this stratum were added to the transect-based estimates, estimates for the whole study area on all dates ranged from 47 to 187 in 2000 and 59 to 247 in 2001.

Figure 3. Locations of manatee groups on 30 strip transects during 6 aerial surveys in 2000 (top) and 8 surveys in 2001(bottom) in the Ten Thousand Islands.

Table 2. Manatee counts corrected for perception bias and multiplied by mean group size to obtain population estimates in the Ten Thousand Islands during 2 survey years, July through October 2000 and July through August 2001.

Table 3. Annual manatee population estimates using low and combined low-high density strata, Ten Thousand Islands, Florida.

To assess the potential for detecting statistically significant trends in the Ten Thousand Islands population, we used TRENDS software (Gerrodette 1993). We used CVs of 0.30 and 0.15, based on results from the surveys in 2000 and 2001, respectively (Table 3).

The following parameters were also selected: = 0.05; 1-tailed test; linear model of rate of change; CV proportional to the square root of the abundance estimate; standard normal distribution.

If number of sampling periods is 6 per year, sampling is continued for a minimum of 4 years, and CV = 0.30, then power = 0.32.

If number of sampling periods is 8 per year, sampling is continued for a minimum of 4 years, and CV = 0.15, then power = 0.86.

We are likely to need a minimum of 8 surveys per year for a minimum of 4 years to detect an annual rate of change of 10% per year.

Population estimates and densities in this study were similar to those for the Banana River, an important area for manatees on the Atlantic coast in the warm season. The latter estimates ranged from 112 to 209, or approximately 0.67 to 1.26 per km² (Miller et al. 1998). Mean group size per survey in Ten Thousand Islands (1.62) was lower than in Banana River surveys (2.19).  Group size was 2.00 in 13 of the 14 Ten Thousand Island surveys and 2.00 in 13 of 15 Banana River surveys (Miller et al. 1998). These findings suggest that poorer water clarity in the Ten Thousand Islands than in the Banana River, where the bottom can be seen in most of the survey area, may contribute to greater variability and smaller observed group size in our surveys.

The number of sediment plumes observed during the Ten Thousand Island surveys (32 in 2000 and 72 in 2001; Table 2 & Fig. 5) suggests that many manatees may not be directly observable.  Observations from distribution surveys suggest that approximately 0.75 of plumes represent an actual manatee group.  A correction factor should be developed to incorporate the number of observed plumes in the total estimated population size.

Variation in group size and population estimates is a reflection of the challenging survey conditions presented by the Ten Thousand Islands, as well as additional variability caused by weather. Nevertheless, the strip-transect approach shows promise for monitoring the manatee population using this region during the warm season, if weather-related variability can be minimized.

• A minimum of 8 surveys should be conducted every year.
• Survey conditions must be good to excellent to minimize variation among surveys.
• Surveys should be repeated every year for a minimum of 4-5 years.
• Correction for availability bias (manatees present but invisible) should be explored. 
• Similar strip-transect surveys should be designed for other areas of the southwest coast.

Gerrodette, T. 1993.  TRENDS: software for a power analysis of linear regression. Wildlife Society Bulletin 21:515-516.

Lefebvre, L.W., B.B. Ackerman, K.M. Porter, and K.H. Pollock.  1995. Aerial survey as a technique for estimating trends in manatee population size–problems and prospects. Pages 63-74 in T.J. O'Shea, B.B. Ackerman, and H.F. Percival, editors. Population biology of the Florida manatee. U.S. Department of the Interior, National Biological Service Information and Technology Report 1.

Marsh, H.  1995. Fixed-width aerial transects for determining dugong population sizes and distribution patterns. Pages 56-62 in T.J. O'Shea, B.B. Ackerman, and H.F. Percival, editors. Population biology of the Florida manatee. U.S. Department of the Interior, National Biological Service Information and Technology Report 1.

Marsh, H., and D.F. Sinclair.  1989. Correcting for visibility bias in strip transect aerial surveys of aquatic fauna.  Journal of Wildlife Management 53:1017-1024.

Miller, K.E., B.B. Ackerman, L.W. Lefebvre, and K.B. Clifton.  1998. An evaluation of strip-transect aerial survey methods for monitoring manatee populations in Florida. Wildlife Society Bulletin 26(3):561-570.

Pollock, K.H., and W.L. Kendall. 1987.  Visibility bias in aerial surveys: a review of estimation procedures. Journal of Wildlife Management 51:502-510.

U.S. Fish and Wildlife Service and the U.S. Geological Survey, through the Resource Partnership Program, the Big Cypress National Preserve, and Columbus Zoo supported this research.  We thank Skip Snow, Everglades National Park, for his guidance and enthusiasm, which helped to launch this study. Completing these surveys was truly a team effort.  Backseat observers were Margie Barlas and Lucy Keith of the Florida Fish and Wildlife Conservation Commission - Manatee Rescue Program, Steve Schulze of Big Cypress National Preserve, Doug Suitor of Collier County Natural Resources Department, and Khabira Al-MuhyeeEttaji of Everglades National Park. Special thanks to Deborah Jansen of Big Cypress National Preserve for securing funding for several flights, loaning us a GPS to record flight lines, and allowing Steve Schulze to help with the flights.  Last but not least, thanks to Emilio Echeverria, Gary Roam, and Ken Poteet of Speed Aviation for piloting us safely through the flights.