Use of aerial survey and aerophotogrammetry methods in monitoring manatee populations

MISSING IMAGE

Material Information

Title:
Use of aerial survey and aerophotogrammetry methods in monitoring manatee populations
Series Title:
Florida Cooperative Fish and Wildlife Research Unit Research Work Order 116
Alternate Title:
RWO 116
Physical Description:
Book
Language:
English
Creator:
Miller, Karl E.
Ackerman, Bruce B.
Lefebvre, Lynn W.
Clifton, Kari B.
Affiliation:
University of Florida -- Florida Cooperative Fish and Wildlife Research Unit -- Department of Wildlife Ecology and Conservation -- Institute of Food and Agricultural Sciences
Publisher:
Florida Cooperative Fish and Wildlife Research Unit, University of Florida
Publication Date:

Subjects

Subjects / Keywords:
University of Florida.   ( lcsh )
Aerial photography   ( lcsh )
Manatees   ( lcsh )
Biotic communities -- Florida   ( lcsh )
Natural history -- Florida   ( lcsh )
Manatees -- Florida
Spatial Coverage:
North America -- United States of America -- Florida

Notes

Funding:
This collection includes items related to Florida’s environments, ecosystems, and species. It includes the subcollections of Florida Cooperative Fish and Wildlife Research Unit project documents, the Sea Grant technical series, the Florida Geological Survey series, the Coastal Engineering Department series, the Howard T. Odum Center for Wetland technical reports, and other entities devoted to the study and preservation of Florida's natural resources.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
All rights reserved, Board of Trustees of the University of Florida
System ID:
UF00073806:00001

Table of Contents
    Title Page
        Title page
    Abstract
        i
        ii
        iii
    List of Tables
        iv
        v
    Introduction
        Page 1
        Page 2
    Objectives
        Page 3
    Study area
        Page 4
    Methods
        Page 5
        Page 6
        Page 7
        Page 8
    Results
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
    Discussion
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
    Literature cited
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
    Tables
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
    Figures
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
    Appendix A: Daily record of aerial surveys
        Page 41
        Page 42
    Appendix B: Intensive aerial surveys for manatees in the northern Banana River
        Page 43
        Page 44
        Page 45
Full Text







USE OF AERIAL SURVEY
AND AEROPHOTOGRAMMETRY METHODS
IN MONITORING MANATEE POPULATIONS



FINAL REPORT
to the U.S. Department of the Interior
National Biological Service
(RWO-116: Aerial Survey objective)





Prepared by:

Karl E. Miller'
Bruce B. Ackerman2
Lynn W. Lefebvre3
Kari B. Clifton2

'Florida Cooperative Fish and Wildlife Research Unit
and
Department of Wildlife Ecology and Conservation
University of Florida, Gainesville, FL 32611

2Florida Department of Environmental Protection
Florida Marine Research Institute, St. Petersburg, FL 33701

National Biological Service, Sirenia Project, Gainesville, FL 32601


June 1996











EXECUTIVE SUMMARY/ ABSTRACT

Aerial surveys have been used to obtain information on the distribution and
relative abundance of manatees (Trichechus manatus) in Florida for more than 25 years.
Because managers and researchers believe surveys underestimate manatee population
size, survey effort traditionally has been geared to maximize counts rather than to
standardize counts. Consequently, surveys have not produced population estimates with
associated confidence intervals, making them inadequate for estimating population size
or determining trends in abundance over time.
We evaluated the use of strip-transect survey methods for manatees through a
series of replicate aerial surveys in the Banana River, Brevard County, Florida, during
summer 1993 and summer 1994. Transect methods sample a representative portion of
the total study area, thus allowing for statistical extrapolation to the total area. Other
advantages of transect methods are less flight time and less cost than total coverage, ease
of navigation, and reduced likelihood of double-counting.
Our objectives were: (1) to identify visibility biases associated with the transect
survey method and to adjust the counts accordingly; (2) to derive a population estimate
with known variance for the Banana River during summer; and (3) to evaluate the
potential value of this survey method for monitoring trends in manatee population size
over time.
We selected the Banana River as a candidate site for transect sampling because it
is a relatively homogeneous, shallow water body with good visibility and it is known to be
an important summer area for manatees. Surveys were flown in a Cessna 172 at an
altitude of 183 m (600 ft). Forty 250m-wide strip transects were established at 1 km
intervals, resulting in 25% coverage of a 160 km2 area. A team of independent
observers in the right-front and right-rear seats simultaneously scanned the same
transect. Their independent observations provided counts which were used in a Petersen
mark-recapture model to estimate perception bias (the number of groups that are visible
in the transect but are not observed). The number of observed groups was multiplied by
the perception correction factor to obtain the corrected number of groups per transect,
then multiplied by the mean group size to obtain the corrected number of individuals per
transect. Finally, a ratio estimator was used to estimate the total population size and its
variance.
Eight replicate surveys were conducted between August and September each year.
A total of 775 individuals belonging to 374 groups was observed during the 16 surveys,









for a mean group size of 2.07. Total survey time was 25.0 hrs, for a sighting rate of 31
manatees per hr (0.52 per min).
Double counts by 2 right-side observers produced survey-specific correction factors
for perception bias averaging 1.13 (range: 1.00 1.60). A perception correction factor
(Cp) of 1.13 indicates that 88% (the reciprocal) of the groups estimated to be available
were observed by at least one of the observers. The C, was highest on days with poor
(1.60 on 25 Aug 1993) or fair (1.42 on 15 Aug 1994) survey conditions. Observers
performed similarly in their ability to recognize manatee groups. Although the
proportion of groups detected varied from survey to survey, within a survey date it was
relatively similar between the 2 observers. Single animals were more likely to be
overlooked by at least one observer than were larger groups.
Mean group size per survey for the right-side observer team ranged from 1.65 to
3.10. Most of the groups sighted were single animals (51% of groups) or groups of two
(26% of groups), but a few large (>6) groups also were seen. These statistical outliers
resulted in larger variance in mean group size and, consequently, larger variance in the
overall population estimate (9).
Manatees were not evenly distributed throughout the study area. Four
aggregation sites ("hot-spots") along the Banana River shoreline were identified and
manatees were counted by repeated circling over these areas. Hot-spots were treated as
a separate sampling stratum.
Population estimates were moderately consistent among survey dates, with the
exception of 2 surveys (25 Aug 1993, 28 Sep 1993) conducted under extreme weather
conditions. Excluding these surveys, mean population size was 125 4 (SE) in 1993
(N= 6 flights) and 179 8 (SE) in 1994 (N= 8 flights). We added the daily counts of
manatees in the hot-spot stratum to each day's population estimate to obtain an estimate
of the total population size (9+) in the survey area. Mean ", was 159 7 (SE) in 1993
(N=6 flights) and 238 10 (SE) in 1994 (N=8 flights).
Although mean population size differed between years (P < 0.05), we cannot
infer an increasing trend in population size based on 2 annual estimates. To assess the
possibility of detecting statistically significant trends in the Banana River population, we
used power analysis software for linear regression (TRENDS). Manatee density in the
Banana River is relatively high (> 1/km2) and our sampling intensity is high (25% of the
total survey area is sampled, 7-8 replicate flights are performed each year), resulting in
annual population estimates with high levels of precision. With the stringent
assumptions of cv = .04 and power = 0.75, we would need 4 sampling periods (i.e.,








iii
years) to detect an annual rate of change (r) = 0.05. An increase in r or a decrease in
the cv would increase the power of our test.
The strip-transect survey technique presented here is an improvement over past
attempts to estimate absolute manatee abundance, because it is a repeatable,
standardized survey design that produces population estimates with known precision.
The technique should be useful in trend analysis in the Banana River. However,
application of the survey technique to other areas in Florida during summer appears to
be limited because of excessive water depths or water turbidity. In addition, we have not
measured the availability bias associated with these surveys.
Despite limitations, we feel managers would benefit by continuing warm-season
transect surveys in the Banana River for the following reasons: (1) Banana River is
clearly important as a manatee summer "sanctuary"; (2) Complementary long-term
datasets are needed to corroborate other analyses that suggest a manatee population
increase on the east coast of Florida.










LIST OF TABLES, FIGURES, AND APPENDICES


Tables

Table 1. Manatees observed on strip-transect aerial surveys, Banana River, Brevard
County, Florida, summer 1993-94.

Table 2. Manatee counts corrected for perception bias (Petersen estimator), multiplied
by mean group size, and adjusted for area sampled (ratio estimator) to obtain population
estimates for Banana River, Brevard County, Florida, summer 1993-94.

Table 3. Detection probabilities for 3 right-side observers on strip-transect aerial
surveys, Banana River, Brevard County, Florida, summer 1993-94. Proportions indicate
the number of manatee groups seen by the observer divided by the number of groups
estimated to be available on transects.

Table 4. Total groups of each size class observed by both right-side observers (RB)
versus those observed by only one right-side observer (RF or RR), strip-transect aerial
surveys, Banana River, Brevard County, Florida, summer 1993-94.

Table 5. Manatee concentrations observed at "hot-spots" in Banana River during strip-
transect aerial surveys, summer 1993-94. Hot-spots were counted by 2 right-side
observers; highest count is presented.

Table 6. Observed distribution of manatee groups in 8 zones (each zone = 5
consecutive transects) and the expected distribution of manatee groups in relation to the
area surveyed in each zone, Banana River, Brevard County, Florida.

Table 7. Mean annual population estimates with and without hot-spot stratum,
Banana River, Brevard County, Florida, summer 1993-94.


Figures

Figure 1. Area surveyed for manatees in Banana River lagoon, Brevard County, Florida,
1993-94. Parallel strip-transects were flown from Jack Davis Cut (start) to Highway 404
(stop). High-density stratum included (a) Hangar AF, (b) Cape Canaveral Sewage Plant,
(c) 520 Causeway, and (d) Cocoa Beach Golf Course.

Figure 2. Manatee observability at different distance intervals from the aircraft on 7
survey flights.









Figure 3. Size-class distribution of all manatee groups (N=374) observed by right- and
left-side observers on 16 strip-transect aerial surveys, Banana River, Florida, summer
1993-94.

Figure 4. Annual difference in size class distribution of manatee groups observed by
right-side observers in summer 1993 (N= 101) and 1994 (N= 152).

Figure 5. Total number of manatee groups observed on each transect in (a) 1993 and
(b) 1994, Banana River, Brevard County, Florida.


Appendices

Appendix A. DAILY RECORD OF AERIAL SURVEYS.
Table A-1. Record of personnel, survey time, and flight conditions for strip-
transect aerial surveys of manatees in Banana River, Brevard County, Florida,
1993-94.

Appendix B. INTENSIVE AERIAL SURVEYS FOR MANATEES IN THE
NORTHERN BANANA RIVER.
Table B-1. Four years of summer (June-September) counts of manatees in the
northern Banana River, Brevard County, Florida. Counts were obtained through
intensive-search of a 78.5 km2 study area by 3-4 observers in a NASA helicopter
(Source: Jane Provancha, Bionetics Corp., unpubl. data).











I. INTRODUCTION

Aerial surveys have been used to identify seasonal patterns in distribution and

relative abundance of manatees (Trichechus manatus) in Florida for more than 25 years.

In recent years, state-wide "synoptic" surveys have been conducted during the winter

months to derive a minimum estimate of the size of Florida's manatee population

(Ackerman 1995). Because aerial surveys are believed to underestimate manatee

population size (Hartman 1974, Irvine and Campbell 1978, Eberhardt 1982, Kinnaird

1985, Packard 1985, Packard et al. 1985), survey effort traditionally has been geared to

maximize counts rather than to standardize counts (Lefebvre et al. 1995). Consequently,

state-wide "extended-area" surveys produce composite counts with unknown variance,

which cannot be compared statistically across years (Lefebvre et al. 1995). Most

manatee surveys also have not been corrected for sources of visibility bias (Lefebvre et

al. 1995). Thus, current aerial surveys may provide important information on the

distribution and relative abundance of manatees, but they are inadequate for estimating

population size or determining trends in abundance over time (Ackerman 1995, Lefebvre

et al. 1995).

Researchers at the 1992 Technical Workshop on Manatee Population Biology

(O'Shea et al. 1992, Ackerman 1995, Lefebvre et al. 1995) recommended that the testing

of transect survey methods should be a high priority for future manatee research. In

Australia, Helene Marsh and her colleagues have developed fixed-width transect surveys

for dugongs (Dugong dugon) to provide standardized population estimates and density-










distribution maps for management purposes (Marsh and Saalfeld 1989; Marsh and

Sinclair 1989a,b; Marsh et al. 1994; Marsh 1995). Transect surveys are superior to

counts that attempt total coverage of an extended area because a complete count is

rarely possible; extensive counts do not quantify the area actually surveyed or produce an

estimate of absolute abundance. In contrast, transect methods sample representative

portions of the study area and allow statistical extrapolation to the total area (Burnham

et al. 1980). An additional advantage of transect surveys is that they require less flight

time and less cost than total coverage. For example, Johnson et al. (1991) conducted

aerial surveys for pronghorn (Antilocapra americana) with line transect methods and

with "trend" counts that attempted total areal coverage, and found that adequate line

transect samples were obtained in 19-51% of the flight time required for "trend" counts.

Transect sampling methods also are superior to quadrat or block sampling in terms of

lower cost, ease of navigation, reduced likelihood of double-counting, and less fatigue for

observers (Norton-Griffiths 1978, Firchow et al. 1990).

The goal of this research project was to evaluate the use of transect survey

methods for manatees through a series of replicate aerial surveys in the Banana River,

Brevard County, Florida, during the summers of 1993 and 1994. The Banana River

lagoon was selected as a candidate site because of its importance as a warm-season

refuge for manatees and because it is a shallow body of water with good visibility.

Previous aerial surveys in Brevard County (e.g., ShanfieT981, 1983; Bonde, unpubl. data

from mid-1980s; Provancha and Provancha 1988, 1989; Ackerman, unpubl. data from

1992) have indicated that large numbers of manatees use the area throughout the year.









The study was conducted during the warm season in the belief that summer

counts would be less subject to the influence of fluctuating environmental factors (e.g.,

water temperature) on visibility bias. Problems associated with counting large

aggregations of animals also were avoided because of the more even dispersion of

manatees throughout coastal water bodies during summer.

We are indebted to our pilots, J.T. Lundstrom and V. Renaud, for skillfully and

safely piloting flights in 1993 and 1994, respectively. Thanks are due C. Deutsch, A.

Perry, and A. Spellman for filling in as left-side observers on survey flights. We also

wish to extend our gratitude to J. Provancha and the Bionetics Corporation for

coordinating their helicopter surveys with our fixed-wing surveys and for making their

data available to us. J. Harrison of the University of Florida's Institute of Food and

Agricultural Sciences (IFAS) Statistics is acknowledged for assistance with the derivation

of variance estimators. H.F. Percival and B. Fessler of the U.S. Fish and Wildlife

Service (USFWS) provided administrative support.



II. OBJECTIVES

(1) To identify visibility biases associated with counting manatees with the transect

aerial survey method and to adjust the counts accordingly.

(2) To derive a population estimate with known confidence intervals of the manatee

population in the Banana River in summer 1993 and summer 1994.

(3) To evaluate the potential value of this survey method for monitoring trends in

manatee population size over time.












III. STUDY AREA

The Banana River is an estuarine lagoon in Brevard County on the east coast of

central Florida. It is a shallow water body (average water depth outside of dredged

canals = 1.2 m, Provancha and Provancha 1988) with submerged aquatic vegetation

(SAV) throughout most of the river bottom. Four seagrasses, manatee grass

(Syringodium filiforme), shoal grass (Halodule wrightii), widgeon grass (Ruppia

maritima), and star grass (Halophila engelmannii), and several species of algae dominate

the SAV (Provancha and Provancha 1988; L. Lefebvre, unpubl. data).

Aerial surveys have indicated that the Banana River has the highest late winter-

early spring concentration of manatees along Florida's east coast (B. Wiegle unpubl. data

cited in Provancha and Provancha 1989). Numbers peak in the spring, and then decline

to a stable population that persists throughout the summer (Provancha and Provancha

1989).

We surveyed a 160 km2 area extending from Jack Davis Cut at the northern end

of the Banana River south to state Highway 404 (Fig. 1). The study area incorporates

lands administered by several government agencies, including Merritt Island National

Wildlife Refuge (USFWS), Kennedy Space Center (NASA), and Patrick Air Force Base.

Recreational boating and fishing are common throughout the river except for the portion

north of the NASA Causeway which is closed to the public.









IV. METHODS

Strip-transect Surveys

Line transect sampling is often used to survey marine fauna from ships (e.g.,

Barlow 1988), where the observers have sufficient time to measure perpendicular

distances from the center line. However, in a rapidly moving aircraft there is usually

inadequate time to measure distances correctly while thoroughly searching the entire

field-of-view and counting the number of animals in each group (Pollock and Kendall

1987), and, thus, strip-transects are usually used in aerial surveys (Leatherwood et al.

1978; Gasaway et al. 1985; Barlow et al. 1988; Conroy et al. 1988; DeYoung et al. 1989;

Firchow et al. 1990; Marsh and Sinclair 1989a,b; Marsh 1995). We used strip-transect

sampling because we believed that it would be more efficient to focus on a narrow

"window" of view. We selected a 250-m strip-width in accordance with other aerial

surveys for sirenians (Bayliss 1986; Marsh and Sinclair 1989a,b). A fundamental

assumption of strip-transect sampling is that all objects present in the strip have an equal

probability of being observed (Eberhardt et al. 1979, Seber 1982). We tested this

assumption by estimating perpendicular distance from the plane to each manatee group

on a subset of our surveys. No significant decrease in sightability occurred beyond the

outer edge of the strip, although appreciably more groups were seen in the middle of the

strip (Fig. 2).

Forty parallel transects were established at 1 km intervals in the Banana River,

with the 250-m strip-width enabling 25% coverage of the study area. We spaced

transects evenly at 1 km intervals both for ease of navigation and to ensure that animals









were not counted twice, as often happens in closely-spaced transects (Caughley

1977a:32). Transects were oriented on an east-west axis perpendicular to the linear

Banana River to maintain homogeneity among transects, thereby minimizing sampling

error (Norton-Griffiths 1978), as well as to reduce glare from the morning sun (Marsh

and Sinclair 1989a).

Eight surveys per year were flown in a Cessna 172 (Merritt Island Air Service) at

70-80 knots along these predetermined transects. Transects were drawn and labelled on

a small-scale map, and reference to topographic features and landmarks were used

effectively by the pilot to navigate along transects. Survey altitude was standardized at

183 m (600 ft); restrictions on flights over inhabited areas precluded flights at lower

altitudes. Calibration of strip-width followed Pennycuick and Western (1972) and

Norton-Griffiths (1978). The 250m strip-width was measured beginning from the outer

edge of the wheel of the plane. Lines were drawn with wax pencil on the windows of the

aircraft and on the plane's struts to delineate inner and outer strip-width boundaries and

to divide the strip into 5 distance bands for estimating perpendicular distances. Due to

federal aviation restrictions, plastic streamers, rods, or other attachments to the outside

of the aircraft were not used to mark strip-width boundaries. All surveys were

completed before 1200 h Eastern Standard Time.

Visibility Bias

Aerial surveys are biased because sighting probabilities often are no better than --

50-60% (Caughley 1977). Two common types of bias in aerial survey for marine

mammals are perception bias and availability bias (Marsh and Sinclair 1989b, Lefebvre









et al. 1995). Perception bias in manatee surveys results when a proportion of manatees

visible within the strip-transect is missed by the observerss. Availability bias in manatee

surveys occurs when a proportion of manatees within the strip-transect area is not

"available" to be counted because it is submerged under turbid water or hidden by some

other habitat feature.

We estimated perception bias using a double count by two independent observers

(Seber 1982, Pollock and Kendall 1987). Observers in the right-front and right-rear seats

of the aircraft formed a tandem team searching the same 250m-wide transect

independently. Right-side observers recorded their sightings on small-scale maps, then

conferred afterwards to determine whether each group was seen (a) by the front

observer only, (b) by the rear observer only, or (c) by both observers. Their independent

observations provided counts which were used in a modified mark-recapture model

(Petersen estimator) to determine the number of animals missed by both observers. This

technique has been used successfully in surveys for a variety of species, including

crocodiles (Crocodylus porosus; Magnusson et al. 1978), emus (Dromaeous

novaehollandiae; Caughley and Grice 1982), feral horses (Equus caballus; Graham and

Bell 1989), and dugongs (Marsh and Sinclair 1989b, Marsh et al. 1994, Marsh 1995).

Three different biologists rotated among seating positions so that different right-side

combinations could be used to calibrate observer performance.

Because members of a group of animals are not sighted independently, groups,

rather than individuals, were considered appropriate sample units. Matching of groups

seen by right-side observers occurred during the flight via intercom (11 Aug 1993 10









Aug 1994) or after the survey flight was completed (11 Aug 1994 23 Aug 1994).

Conversations during the flight were tape-recorded directly from the intercom as a back-

up for later verification if needed. The following information was recorded for each

manatee group sighted: location, time, behavior (feeding, resting, travelling, cavorting),

total number in group, and number of calves. When observers differed as to the number

of manatees in a group, the higher number was used in analyses. Although the right-

front and right-rear observers formed the primary survey team, on most flights a left-rear

observer was also present. However, data collected by left-side observers could not be

adjusted for perception bias and so are not used in population estimates.

Availability bias was intentionally minimized by the selection of a study area with

shallow water and good visibility. We did not use an availability correction factor to

correct our counts because we found it was not feasible to establish different sighting

probabilities for manatees based on their position in the water column (but see Marsh

and Sinclair 1989b).

Population Estimates

The equations used to calculate survey-specific correction factors for perception

bias follow that of Marsh and Sinclair (1989b). Despite small sample sizes, the Chapman

modification of the Petersen estimator (Seber 1982) was not used because it did not

significantly reduce the variance of overall population estimates. The corrected number

of manatees per transect was obtained by multiplying these components: (a) the number

of groups on the transect sighted by right-side observers, (b) the survey-specific

correction factor, and (c) the mean group size for that survey (Marsh and Sinclair








9
1989b:1021-1022). Because transects were of variable length, the ratio method (Norton-

Griffiths 1978, Seber 1982, Marsh and Sinclair 1989b) was used to estimate the total

population size in the 160 km2 study area. Population estimates were averaged to obtain

a mean and standard error for each year.

Trend Analysis

The utility of these surveys for trend analysis was explored using power analysis

software for linear regression (TRENDS, T. Gerrodette, Southwest Fish. Sci. Center, La

Jolla, Cal.). A power analysis allows estimation of the probability of being able to detect

an upward or downward trend in population size given the number of replicate samples

per year and sample variability (Gerrodette 1987, 1991, 1993).



V. RESULTS

Survey Counts

Eight surveys were conducted each year in August-September (Table 1). A total

of 775 individuals belonging to 374 groups was counted during the 16 surveys, for a mean

group size of 2.07 1.74 (SD). Flights were conducted under standardized survey

conditions with good visibility except on 25 Aug 1993, when overcast, rainy weather

associated with a stationary front produced poor visibility; the low count of 11

undoubtedly underestimated the number of manatees actually present. Daily survey

conditions and personnel are summarized in Appendix A. Total survey time was 25.0

hrs, with a sighting rate of 31 manatees per hr (0.52 per min) by all observers and 22

manatees per hr (0.36 per min) by right-side observers.









Observer Bias

Right-front observers sighted a total of 206 manatee groups, whereas right-front

and right-rear observers combined for a total of 253 manatee groups. Compared with a

conventional single-observer aerial survey, the double-observer technique increased the

number of manatee groups by 23% due to the presence of the second observer and by

36% when the double counts were corrected by the Petersen model.

Double counts by 2 right-side observers produced survey-specific correction factors

for perception bias averaging 1.13 (range: 1.00 1.60) (Table 2). A perception correction

factor (Cp) of 1.13 indicates that 88% (the reciprocal) of the groups estimated to be

available (N) were observed by at least one of the two right-side observers. The Cp was

highest on days with poor (1.60 on 25 Aug 1993) or fair (1.42 on 15 Aug 1994) survey

conditions. Although the proportion of groups detected varied from survey to survey,

within a survey it was relatively similar between observers (Table 3). Mean detection

probabilities for each biologist ranged from 75-78% in 1993 and 70-78% in 1994.

Group Size

Mean group size per survey for the right-side observer team ranged from 1.65 to

3.10 (Table 2). Most of the groups sighted were single animals (51% of groups) or

groups of two (26% of groups), but a few large (> 6) groups also were seen (Fig. 3).

These statistical outliers resulted in larger variance in mean group size and,

consequently, larger variance of the overall population estimate (Table 2). Although

more single manatees were seen in 1994 than in 1993 (Fig. 4), mean group size did not

differ between years (Wilcoxon rank sum test, Z = 1.058, P = 0.2902).








11
Combining all data collected by right-side observers, sighting probability was 0.87

for single manatees (N= 123), 0.96 for groups of 2 manatees (N= 62 groups), and 0.99 for

groups of 3 or more manatees (N= 68 groups). Single animals were more likely to be

seen by only one observer than were larger groups (X2 = 18.737, 2 df,

P < 0.0005, Table 4).

Of the 253 groups sighted by right-side tandem observers, 151 (60%) were seen by

both members (Table 4). Observers differed in their estimates of group size for 26

(17%) of these 151 groups. Discrepancies were somewhat more frequent with larger

groups, suggesting that observers may have had greater difficulty counting larger groups

of manatees.

Clumped Distributions

Manatees formed summertime aggregations at a few sites along the Banana River

shoreline (hereafter referred to as "hot-spots"). These areas were attractive to manatees

primarily because they were sources of freshwater. Counts at 4 hot-spots were made by

repeated circling by the aircraft over each area (Table 5). On 12 of 16 (75%) surveys

the total count at hot-spots exceeded the total number of manatees counted on transects.

More manatees were counted at hot-spots in 1994 (mean = 59.4) than in 1993 (mean =

45.7) due to the increased numbers at the 520 Causeway (520C) and Cape Canaveral

Sewage Plant (CCSP) (Table 5).

The clumped distribution of manatees is further demonstrated by comparing the

number of manatee groups observed on each transect (Fig. 5). We collapsed the 40

transects into 8 zones of 5 adjacent transects and performed a Chi-square Goodness-of-









Fit test to see if groups were distributed evenly across the study area (Table 6).

Manatee distribution did not fit the expected distribution in either year (1993: P < 0.005;

1994: P < 0.001), indicating that the number of manatees observed in each of the 8

zones was not proportional to the area surveyed in each zone. Manatee density was

highest between transects 41-45 (the area north and south of Hwy 528 Causeway); more

manatees occupied this zone than expected in both years. Note also that some of the

highest manatee densities were observed south of transect #43 in the area outside the

manatee "sanctuary" (Fig. 5, Table 6).

Population Estimates

Population estimates (') ranged from 64 to 240 in 1993 and 148 to 208 in 1994

(Table 2). Population estimates were moderately consistent among surveys, with the

exception of 2 surveys conducted under extreme weather conditions. Results from 25

Aug 1993 (' = 64) are suspect because of poor visibility caused by bad weather, and we

excluded that survey from all analyses of population size. Results from 28 Sep 1993

(' = 240) also may be unreliable because they were obtained late in the season

coincident with the first fall cold-front. We suspect that cool weather probably changed

manatee behavior and distribution, but because we cannot verify this, we conducted

analyses of population size both with and without this survey (hereafter referred to as

"late" survey).

There are several components of the variance estimator for '9, including variance

from the correction factor, variance from group sizes, and variance from sampling.







13
Generally, high variance was mostly due to group size variability and sampling variability

in number of manatees per transect.

Mean population size in 1993 was 142 17 (SE) with the late survey (N=7), and

125 4 (SE) without the late survey (N=6). Mean population size in 1994 was 179 8

(SE) (N=8). The benefit of replicate sampling is clear; although the coefficient of

variation (cv) of I from a single survey was large (generally 30-40%, Table 2), the cv of

mean I was smaller (8% in 1993 without the late survey, 13% in 1994; Table 7).

By adding the daily counts of manatees in the high-density hot-spot stratum to I,

we obtained an estimate of the total population size in the survey area (hereafter

referred to as +.) (Table 7). Mean 'I was 27% and 33% greater than in 1993 and

1994, respectively, indicating that the proportion of the population using hot-spots was

similar between years.

Trend Analysis

Mean population size (f) differed between years (late 1993 survey excluded:

Wilcoxon rank sum test, Z = -3.034, P = 0.0024; late 1993 survey included: Wilcoxon

rank sum test, Z = -2.257, P = 0.0240). Obviously, we cannot infer an increasing trend

in population size based on 2 annual estimates. To assess the possibility of detecting

statistically significant trends in the Banana River population, we used TRENDS

software (Gerrodette 1993) for a power analysis of linear regression. We used a cv of

0.04 and 0.05 in our power analysis, because the cv of mean was in this range in both

years. The following parameters also were selected: alpha = .05; 1-tailed test; linear

model of rate of change; cv is proportional to the square root of the abundance estimate










(appropriate when strip-transect or line-transect methods are used for estimating

abundance; Gerrodette 1987); standard normal (Z) distribution (because an estimate of

variance of mean is produced each year; Gerrodette 1991, 1993).


When number of sampling periods (n)

For r = .05 and power = .75:



For r = .10 and power = .75:


When power

For r


(1 p) is parameter to be

= .05 and n = 5:


For r = .10 and n = 5:


s parameter to be computed:

If cv = 0.04, n = 4

If cv = 0.05, n = 4

If cv = 0.04 n = 3

If cv = 0.05 n = 3

computed:

If cv = 0.04 power = 0.96

If cv = 0.05 power = 0.86

If cv = 0.04 power = 1.00

If cv = 0.05 power = 1.00


With the stringent assumptions of high precision (cv = 0.04) and a high level of

power (power = 0.75), we would need 4 sampling periods (i.e., years) to detect an

annual rate of change (r) = 0.05. An increase in r or a decrease in the cv will increase

the power of our test. Note that the above analyses are for detecting an increasing trend

(r > 0) with a one-tailed test; to detect a decreasing trend (r is negative) of the same

magnitude with a one-tailed test, the power will be reduced slightly (Gerrodette 1993).









VI. DISCUSSION

Observability Bias

Sighting probabilities in this study were consistent among observers and averaged

nearly 80% (Table 3). In contrast, mean sighting probability of manatees was <50% in

areas in the St. Johns River with turbid water and/or overhanging tree branches

(Packard et al. 1985). Sighting probabilities in aerial surveys for terrestrial mammals are

generally less than 60-70% (Floyd et al. 1979, Bartmann et al. 1986, Beasom et al. 1986,

DeYoung et al. 1989, Firchow et al. 1990, Fuller 1990, Potvin et al. 1992), due in part to

environmental factors such as tree cover and broken terrain that help to obscure animals.

We believe that low perception bias in our study may be due primarily to the

conspicuous nature of the manatees in the predominately clear, shallow water of the

Banana River. Water turbidity and depth influence both availability bias and perception

bias (Lefebvre et al. 1995). Therefore, we may have reduced perception bias by selecting

a study area with optimal survey conditions that minimize availability bias (but see

Marsh and Sinclair 1989b).

Low correction factors for perception bias in this study also could be reflective of

observer skill level, as each of the 3 principal observers had prior multi-year experience

in the aerial survey of ungulates or marine fauna or both. Packard et al. (1985) reported

that observer experience influenced the observability of manatees in aerial surveys.

Aerial surveys conducted by Holt and Cologne (1987) indicated that experienced

observers saw more dolphins than did inexperienced observers but the difference were

not statistically significant.








16
For the Petersen estimator to be valid, 4 main assumptions must be met: (1) the

population is closed during the time between the 2 sampling periods; (2) animals do not

lose their marks between the 2 sampling periods; (3) all marks are correctly identified,

noted, and recorded; and (4) each animal has the same probability of capture in the first

sample, and marking does not affect the capture probability of an animal in the second

sample (Otis et al. 1978, Seber 1982).

The first assumption is valid because observers see manatees almost at the same

instant. Although aircraft noise can cause ungulates to flee from the flight path (e.g.,

Bartmann et al. 1986, Johnson et al. 1991), we did not observe a flight response by

manatees. Manatees are typically slow-moving creatures, and it is unlikely that they can

move laterally outside of the transect in the short period of time between observation by

front and rear observers.

The second and third assumptions are linked. Each manatee group must be

uniquely identified so that groups can be matched properly, and recorded as seen by the

front observer only, the rear observer only, or by both observers (Seber 1982, Pollock

and Kendall 1987). Separate intercom systems can be used for the 2 right-side observers

to communicate independently with a third person acting as recorder (see Marsh and

Sinclair 1989b, Marsh 1995). Such an intercom system was unavailable in our aircraft,

thus, we used 2 other systems for recording data. Initially, right-side observers recorded

their sightings on small-scale maps, then conferred immediately afterwards to determine

whether each group was seen by one or by both observers. Later (after 10 Aug 1994),

observers recorded their sightings in silence and conferred after the flight to determine








17
group identity. Matching of groups after a double-count is facilitated when the density of

the species surveyed is low, because observations are usually separated by enough time

that no confusion arises (Graham and Bell 1989, Potvin et al. 1992). Criteria used to

describe a group, such as size, movement, and behavior also helped to discriminate

between groups in cases where they occurred close together (Potvin et al. 1992). (In the

few cases where it was difficult to interpret the observer's sightings we erred toward

matching groups together rather than separating them, thus reducing the correction

factor and making a conservative estimate of population size.) At any rate, correction

factors were similar whether we conferred during the flight or afterwards, leading us to

believe that observers operated independently without taking cues from each other as to

the timing and nature of their observations. We also note that the problem of matching

groups is not solved by using separate intercoms, as simultaneous registration of a group

by 2 observers does not necessarily prove they are one and the same---the only sure

documentation would be a high-quality photographic record of the strip as it is surveyed.

The fourth assumption is perhaps the most difficult assumption to meet. Some

have argued that marking and recapturing in the aerial survey double-count technique

are not truly independent because observers have essentially the same vantage point

from the aircraft (Caughley and Grice 1982, Pollock and Kendall 1987), thus, the second

sample is not random. Also, animals are not likely to have equal capture probability

(i.e., probability of being sighted) because some may be inherently more difficult to see.

Larger groups of animals were more easily detected than smaller groups in aerial surveys

of white-tailed deer (Odocoileus virginianus; Fuller 1990), mule deer (Q. hemionus;








18
Ackerman 1988), elk (Cervus canadensis; Samuel et al. 1987), and feral horses (Graham

and Bell 1989), and in shoreline surveys of sea otters (Enhydra lutris; Estes and Jameson

1988). In contrast, group size was not a factor in the sightability of dugongs in aerial

surveys (Marsh and Sinclair 1989a). Our data suggest that single manatees or manatees

in groups of 2 are more likely to be missed than larger (03) manatee groups, but the bias

is not as significant as in surveys of the above-mentioned species. Our sample sizes were

inadequate to calculate group-size-specific correction factors for each survey. However,

our results are bolstered by the fact that simulations have shown the technique to be

robust to unequal capture probability (Magnusson et al. 1978) unless the sighting

probability is below 0.45 (Caughley and Grice 1982). Sighting probability in this study

far exceeds the minimum recommended by Graham and Bell (1989) in their survey of

feral horses, or that recommended by Caughley and Grice (1982) in their simulations of

the mark-recapture model for emus.

Spatial Distribution of Manatees

Contrary to our expectations, manatees tended to aggregate in large groups

(usually >10 manatees) at a few shoreline locations we labelled "hot-spots". We stratified

these groups into a separate "high-density" stratum to improve the precision of our

population estimates (Norton-Griffiths 1978). Marsh (1995) stratified out groups of 10

or more dugongs from transect estimates by interrupting the flight path whenever they

were sighted inside or outside transects, and circling and photographing the group to

obtain a total count. We found groups of > 10 manatees were restricted almost entirely

to shoreline hot-spots (<1% of groups on strip-transects were > 10 manatees).








19
The pattern of concentration at hot-spots changed between years (Table 5). The

520C was more heavily occupied by manatees in 1994 than in 1993, while the area west

of Cocoa Beach Golf Course (CBGC) was essentially unused by manatees in 1994. The

ephemeral nature of the CBGC hot-spot underscores the fact that manatee distribution

changes over time (see Fig. 5). Observers need to adequately search the study area each

year to identify changes in the location and size of hot-spots if they are to be counted

accurately and added to the strip-transect population estimate. Knowledge of the

proportion of the population that is aggregated is necessary to assess population trends.

Despite stratification, variance in mean group size among transects and variance

in the number of groups sighted per transect were still high; these components were

responsible for high variance of each population estimate (Table 2). Similarly, variability

in moose (Alces alces) distribution was the most important component of corrected

population estimates in aerial surveys of moose in winter (Crete et al. 1986).

Population Estimates and Population Trends

Manatee population size in the Banana River was higher in 1994 (mean ". =

238) than in 1993 (mean P = 159) (P = 0.0024; late 1993 survey excluded), an increase

of nearly 50%. The survey design, survey conditions, flight-path, type of aircraft, and

observers used were the same in both years. Summer surveys of manatees conducted

since 1977 (Shane 1983; Provancha and Provancha 1988; J. Provancha, NASA, unpubl.

data) indicate a steady increase in the manatee population occupying the northern

portion of the Banana River. The portion of the river surveyed by Provancha is closed

to boating and protected as a manatee sanctuary. Provancha and Provancha (1988)









suggested the observed increase could be due to an increase in manatee use of the

sanctuary rather than an actual increase in the manatee population in Brevard County.

However, some of the highest densities of manatees we observed were south of the

protected portion of the river. Although we conducted only 2 years of surveys, our data

support the hypothesis that the manatee population of the entire Banana River is

increasing.

Biologists generally have difficulty detecting trends in vertebrate species that

occur at low densities (Forney et al. 1991, Taylor and Gerrodette 1993). Manatee

density in the Banana River is relatively high (> 1/km2) and our sampling intensity is

high (25% of the total survey area is sampled, 7-8 replicate flights are performed each

year). Together, these factors produce annual population estimates with high levels of

precision. We should be able to use the current survey protocol to detect annual rates of

change of 5-10% within only a few years, with alpha = 0.05 and power z 0.75. Trend

analysis will be valid only if the same survey effort (i.e., 7-8 flights per year with a cv of

4-5%) is maintained from year to year (Gerrodette 1987, 1993). Surveys should be

performed at regular intervals (e.g., every year, or every second year) if possible.

Fortunately, TRENDS software is not very sensitive to missing data points as long as

they occur at fairly regular intervals (T. Gerrodette, pers. comm.). Thus, the fact that

surveys were not performed in 1995 is not a detriment.

We also assume that biases not directly measured (i.e., availability bias, absence

bias) remain constant from year to year. Availability bias is likely to remain constant

from year to year as long as water quality and other environmental conditions in the








21
river remain unchanged. We also must assume that the proportion of animals using the

survey area remains reasonably constant if we are to extrapolate survey trends to the

Brevard County region. If the proportion of manatees using creeks and canals

contiguous to the river changes each year then we are limited in our ability to infer

trends about the region. It is recommended that analysis of telemetry data for radio-

tagged manatees in Brevard County waterways be conducted to explore diel, seasonal,

and annual patterns of movement in and out of the Banana River study area.



VII. SUMMARY AND MANAGEMENT RECOMMENDATIONS

The strip-transect survey technique represents an improvement over past attempts

to estimate absolute manatee abundance because it is a repeatable, standardized survey

design that produces population estimates with associated variance and confidence

intervals. Observer bias in our surveys was adjusted for with survey-specific correction

factors derived from double counts (Petersen estimator). Although our data indicate

perception bias is low relative to other studies, it is still beneficial to include double-

counts in the survey protocol. Compared to a conventional single-observer aerial survey,

the double-observer technique increased the number of manatee groups by 23% due to

the presence of the second observer and by 36% when the double counts were corrected

by the Petersen model. Survey-specific correction factors are advantageous because they

adjust for variable weather conditions and different observer skill levels.

The strip-transect surveys also allow for the detection of trends in population size

with reasonable statistical power, which makes them useful for monitoring purposes in









the Banana River. Because of the assumption that availability bias and absence bias

remain relatively constant among years, transect surveys perhaps are best used for

monitoring in conjunction with other datasets.

An important, still unresolved, issue is whether the strip-transect survey method

can be used successfully in other areas in Florida during summer. Transect surveys for

manatees are most appropriate in large, homogeneous areas (Ackerman 1995).

Researchers recently selected 3 areas to evaluate warm-season strip-transect surveys for

manatees: Banana River, Charlotte Harbor, and the Ten Thousand Islands (O'Shea et

al. 1992, Ackerman 1995). Trials in the Ten Thousand Island region in Collier County

during June-August 1993 were hampered by excessive water turbidity and other

environmental features (B. Ackerman, unpubl. data). Because manatees commonly

inhabit coastal waters that are turbid or partially obscured by overhanging vegetation

(Lefebvre et al. 1995), it may be very difficult to find other areas like the Banana River.

Marsh (1995) questioned the applicability of dugong transect survey methods to manatee

habitats because the spatial configuration of manatee habitats tends to be linear,

precluding the use of parallel transects.

Despite these limitations, we feel managers would benefit by continuing warm-

season transect surveys in the Banana River for the following reasons: (1) Banana River

is clearly important as a manatee summer sanctuary and we have a need to continue to

monitor the abundance, survival, and movements of manatees in this area. Our data

support the hypothesis that manatees are increasing throughout the Banana River and

not just within the northern portion protected from development and disturbance, but








23
additional years of data are needed to verify this; (2) We need complementary datasets

to confirm recent analysis that suggests manatee population size on the east coast of

Florida is increasing (e.g., Garrott et al. 1994). Garrott et al. (1994) developed a

population index by using a temperature covariate to model a linear trend in aerial

survey data from the winters of 1977-78 through 1991-92. Their analyses showed an

increasing trend in the temperature-adjusted counts of 7-12% annually on the Atlantic

coast, but the degree to which these increases are related to true population growth is

unknown. Changes in manatee behavior may have altered the proportion of manatees in

winter aggregations (Garrott et al. 1995). Different aerial surveys (including NASA

funded aerial surveys in Banana River; see Appendix B) and different demographic

datasets should be used to corroborate evidence of population trends in manatee

populations.











LITERATURE CITED


Ackerman, B.B. 1988. Visibility bias of mule deer aerial census procedures in southeast
Idaho. Ph.D. thesis, Univ. of Idaho.

1995. Aerial surveys of manatees: a summary and progress report.
Pages 13-33 in T.J. O'Shea, B.B. Ackerman, and H.F. Percival, eds. Population
biology of the Florida Manatee. Natl. Biological Serv. Information and
Technology Rep. 1, U.S. Dep. of Interior, Wash., D.C.

Barlow, J. 1988. Harbor porpoise, Phocoena phocoena, abundance estimation for
California, Oregon, and Washington: I. ship surveys. Fish. Bull. 86:417-432.

C.W. Oliver, T.D. Jackson, and B.L. Taylor. 1988. Harbor porpoise,
Phocoena phocoena, abundance estimation for California, Oregon, and
Washington: II. aerial surveys. Fish. Bull. 86:433-444.

Bartmann, R.M., L.H. Carpenter, R.A. Garrott, and D.C. Bowden. 1986. Accuracy of
helicopter counts of mule deer in pinyon-juniper woodland. Wildl. Soc. Bull.
14:356-363.

G.C. White, L.H. Carpenter, and R.A. Garrott. 1987. Aerial mark-
recapture estimates of confined mule deer in pinyon-juniper woodland. J. Wildl.
Manage. 51:41-46.

Bayliss, P. 1986. Factors affecting aerial surveys of marine fauna, and their relationship
to a census of dugongs in the coastal waters of northern Australia. Aust. Wildl.
Res. 13:27-37.

Beasom, S.L., F.G. Leon III, and D.R. Synatzke. 1986. Accuracy and precision of
counting white-tailed deer with helicopters at different sampling intensities.
Wildl. Soc. Bull. 14:364-368.

Burnham, K.P., and D.R. Anderson. 1984. The need for distance data in transect
counts. J. Wildl. Manage. 48:1248-1254.

_, ,_ and J.L. Laake. 1980. Estimation of density from line
transect sampling of biological populations. Wildl. Monogr. 72. 202 pp.

Caughley, G. 1974. Bias in aerial survey. J. Wildl. Manage. 38:921-933.









1977. Analysis of vertebrate populations. John Wiley and Sons, New
York. 234 pp.

and D. Grice. 1982. A correction factor for counting emus from the air,
and its application to counts in Western Australia. Aust. Wildl. Res. 9:253-259.

Conroy, M.J., J.R. Goldsberry, J.E. Hines, and D.B. Stotts. 1988. Evaluation of aerial
transect surveys for wintering American black ducks. J. Wildl. Manage. 52:694-
703.

Cook, R.D., and J.O. Jacobson. 1979. A design for estimating visibility bias in aerial
surveys. Biometrics 35:735-742.

Crate, M., L.-P. Rivest, H. Jolicoeur, J.-M. Brassard, and F. Messier. 1986. Predicting
and correcting helicopter counts of moose with observations made from fixed-wing
aircraft in southern Quebec. J. Applied Ecol. 23:751-761.

DeYoung, C.A, F.S. Guthery, S.L. Beasom, S.P. Coughlin, and J.R. Heffelfinger. 1989.
Improving estimates of white-tailed deer abundance from helicopter surveys.
Wildl. Soc. Bull. 17:275-279.

Eberhardt, L.L. 1982. Censusing manatees. Manatee Population Res. Rept. No. 1,
Technical Rept. 81, Fla. Cooperative Fish and Wildl. Res. Unit, Univ. Fla,
Gainesville. 18 pp.

D.G. Chapman, and J.R. Gilbert. 1979. A review of marine mammal
census methods. Wildl. Monogr. 63. 46 pp.

Estes, J.A., and R.J. Jameson. 1988. A double-survey estimate for sighting probability of
sea otters. J. Wildl. Manage. 52:70-76.

Firchow, K.M., M.R. Vaughan, and W.R. Mytton. 1990. Comparison of aerial survey
techniques for pronghorns. Wildl. Soc. Bull. 18:18-23.

Floyd, T.J., L.D. Mech, and M.E. Nelson. 1979. An improved method of censusing deer
in deciduous-coniferous forests. J. Wildl. Manage. 43:258-261.

Forney, K.A., D.A. Hanan, and J. Barlow. 1991. Detecting trends in harbor porpoise
abundance from aerial surveys using analysis of covariance. Fish. Bull. 89:367-
377.

Fuller, T.K. 1990. Dynamics of a declining white-tailed deer population in north-central
Minnesota. Wildl. Monogr. 110. 37pp.










Garrott, R.A., B.B. Ackerman, J.R. Cary, D.M. Heisey, J.E. Reynolds III, P.M. Rose, J.R.
Wilcox. 1994. Trends in counts of Florida manatees at winter aggregation sites.
J. Wildl. Manage. 58:642-654.

Gasaway, W.C., S.D. Dubois, and S.J. Harbo. 1985. Biases in aerial transect surveys for
moose during May and June. J. Wildl. Manage. 49:777-784.

Gerrodette, T. 1987. A power analysis for detecting trends. Ecology 68:1364-1372.

__. 1991. Models for power of detecting trends---a reply to Link and
Hatfield. Ecology 72:1889-1892.

1993. Program TRENDS: user's guide. Natl. Marine Fish. Serv.,
Southwest Fish. Sci. Center, La Jolla, Calif.

Graham, A., and R. Bell. 1989. Investigating observer bias in aerial survey by
simultaneous double-counts. J. Wildl. Manage. 53:1009-1016.

Hartman, D.S. 1974. Distribution, status and conservation of the manatee in the United
States. Natl. Tech. Information Serv. Publ. PB81-140725, Springfield, Va. 246 pp.

Holt, R.S., and J. Cologne. 1987. Factors affecting line transect estimates of dolphin
school density. J. Wildl. Manage. 51:836-843.

Irvine, A.B., and H.W. Campbell. 1978. Aerial census of the West Indian manatee,
Trichechus manatus, in the Southeastern United States. J. Mamm. 59:613-617.

Johnson, B.K., F.G. Lindzey, and R.J. Guenzel. 1991. Use of aerial line-transect surveys
to estimate pronghorn populations in Wyoming. Wildl. Soc. Bull. 19:315-321.

Kinnaird, M.F. 1985. Aerial census of manatees in northeastern Florida. Biol. Conserv.
32:59-79.

Leatherwood, S.R., J.R. Gilbert, and D.G. Chapman. 1978. An evaluation of some
techniques for aerial censuses of bottlenosed dolphins. J. Wildl. Manage. 42:239-
250.

Lefebvre, L.W., B.B. Ackerman, K.M. Portier, 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, eds.
Population biology of the Florida Manatee. Natl. Biological Serv. Information
and Technology Rep. 1, U.S. Dep. of Interior, Wash., D.C.









Magnusson, W.E., G.J. Caughley, and G.C. Grigg. 1978. A double-survey estimate of
population size from incomplete counts. J. Wildl. Manage. 42:174-176.

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, eds. Population biology of the Florida Manatee. Natl. Biological Serv.
Information and Technology Rep. 1, U.S. Dep. of Interior, Wash., D.C.

and W.K. Saalfeld. 1989. The distribution and abundance of dugongs in
the northern Great Barrier Reef Marine Park. Aust. Wildl. Res. 16:429-440.

and D.F. Sinclair. 1989a. An experimental evaluation of dugong and sea
turtle aerial survey techniques. Aust. Wildl. Res. 16:639-650.

and 1989b. Correcting for visibility bias in strip transect
aerial surveys of aquatic fauna. J. Wildl. Manage. 53:1017-1024.

R.I.T. Prince, W.K. Saalfeld, and R. Shepherd. 1994. The distribution
and abundance of the dugong in Shark Bay, Western Australia. Wildl. Res.
21:149-161.

Norton-Griffiths, M. 1978. Counting animals. Handbook No. 1. 2nd ed. African Wildl.
Leadership Found., Nairobi. 22pp.

O'Shea, T.J., B.B. Ackerman, and H.F. Percival, eds. 1992. Interim report of the
Technical Workshop on Manatee Population Biology. Manatee Population Res.
Rep. 10. Fla. Coop. Fish and Wildl. Res. Unit, Univ. Florida, Gainesville. 83pp.

Otis, D.L., K.P. Burnham, G.C. White, and D.R. Anderson. 1978. Statistical inference
from capture data on closed animal populations. Wildl. Monogr. 62. 135 pp.

Packard, J.M. 1985. Development of manatee aerial survey techniques. Manatee
Population Res. Rept. No. 7, Technical Rept 8-7. Fla. Cooperative Fish and
Wildl. Res. Unit, Univ. Fla., Gainesville. 68 pp.

R.C. Summers, and L.B. Barnes. 1985. Variation of visibility bias during
aerial surveys of manatees. J. Wildl. Manage. 49:347-351.

Pennycuick, C.J., and D. Western. 1972. An investigation of some sources of bias in
aerial transect sampling of large mammal populations. E. Afr. Wildl. J. 10:175-
191.

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










Potvin, F., L. Breton, L-P. Rivest, and A. Gingras. 1992. Application of a double-count
aerial survey technique for white-tailed deer, Odocoileus virginianus, on Anticosti
Island, Quebec. Can. Field-Naturalist 106:435-442.

Provancha, J.A., and M.J. Provancha. 1988. Long-term trends in abundance and
distribution of manatees (Trichechus manatus) in the northern Banana River,
Brevard County, Florida. Marine Mamm. Sci. 4:323-338.

and 1989. Summary of 1987 and 1988 manatee aerial
surveys at Kennedy Space Center. NASA Technical Memo. TM-102783, Kennedy
Space Ctr.

Samuel, M.D., E.O. Garton, M.W. Schlegel, and R.G. Carson. 1987. Visibility bias
during aerial surveys of elk in northcentral Idaho. J. Wildl. Manage. 51:622-630.

Seber, G.A.F. 1982. The estimation of animal abundance and related parameters. 2nd
ed. Macmillan Publ. Co. Inc., New York. 654 pp.

Shane, S.H. 1981. Abundance, distribution and use of power plant effluents by
manatees (Trichechus manatus) in Brevard County, Florida. U.S. Fish and Wildl.
Lab., Gainesville, Fla. Unpubl. rept. prepared for Fla. Power and Light Co.
Contract No. 61552-86540, NTIS Publication No. PB81-147019. 240 pp.

S1983. Abundance, distribution and movements of manatees (Trichechus
manatus) in Brevard County, Florida. Bull. Marine Sci. 33:1-9.

Taylor, B.L., and T. Gerrodette. 1993. The uses of statistical power in conservation
biology: the vaquita and the Northern Spotted Owl. Conservation Biol. 7:489-500.










Table 1. Manatees observed on strip-transect aerial surveys, Banana River, Brevard
County, Florida, summer 1993-94.

Right observers Left observer All observers
Date Groups Individuals Groups Individuals Groups Individuals

1993
11 Aug 11 22 9 22 20 44
25 Auga 5 10 1 1 6 11
26 Aug 13 31 --b -- 13 31
31Aug 12 24 2 2 14 26
01 Sep 12 30 6 9 18 39
02 Sep 11 28 5 13 16 41
16 Sep 10 31 --b -- 10 31
28 SepC 27 54 9 18 36 72
Total 101 230 32 65 133 295

1994
08 Aug 20 40 12 25 32 65
09 Aug 21 45 13 28 34 73
10 Aug 20 49 4 6 24 55
11 Aug 18 36 6 7 24 43
15 Augd 17 28 12 30 29 58
18 Aug 19 34 8 15 27 49
19 Aug 19 42 13 22 32 64
23 Aug 18 37 21 36 39 73
Total 152 311 89 169 241 480

a Very poor survey conditions---overcast with scattered rain.
b No left observer.
c Survey coincided with first cold front of the fall season.
d Fair survey conditions---partly cloudy with moderate wind.













6%o6666m tt66
tttttt-t


U4







42








rA
~I










ci



la
'3-4

.i
SI3





Scu
E'^'






U) U
e5a
ia


8s
& ^2



CO

cil
, is




I.)
&|


60
0
*a




U
SMa



% c








04

00
II
CO











A
0
V













Qc
*- C




Ei












0z
U)
8
*a
















0

&
Zt


00 c0 N 0
---n --t i


- W)MC NN- o r-


r, -e oor oo, e
66666666csmc


cr-
t-
v


0c
..


or-

a- M


- 4 r- ~ N T -



0 o0I 0 0


0000-000
2--2 -- .B2.


W)0> 0
4 C4


0 -l- (\ \

-4 -0- -- i-4
(S w-ti-n o in -'c^ 'r
VI T-^1-^ ^^13^<4


Tr cn cmcn "c =W


< < (o( < < < << <<
000o0 -(o o o-'ryooo m
) (
T^ '4


Se00
-4


mt tn en
>. T I"l:
-.4 C4d


Cu


43 u o r


Sa
'-8

o

4-
S s


o





mM
II





S*3


a
0



2 E





8a D
U U
.5 8

fp

8

II +






o 3
oi0
2i
C e~












U3
*C _.


s (^










IIc
tzl a
n


"
"llP1


1-z -1- a z ll I-



0V en o v n 0








Ci o -
8n ono8l 0- 8


(S' OC-O 4I


0

3





I_




'5;
co



.-
^ 0
m o




I-s




aQ.
0
" 00
c a

0 *
0 vt








Table 3. Detection probabilities for 3 right-side observers on strip-transect aerial
surveys, Banana River, Brevard County, Florida, summer 1993-94. Proportions
indicate the number of manatee groups seen by the observer divided by the number
of groups estimated to be available on transects.

Observer A Observer B Observer C
(Ackerman) (Miller) (Clifton)

1993
11 Aug .50 .57
25 Auga
26 Aug -- .90 .75
31 Aug .63 .56
01 Sep .82 .90
02 Sep 1.00 .82
16 Sep 1.00 -- .90
28 Sep .74 -- .64
Mean .78 .75 .76

1994
08 Aug .80 .71
09 Aug .79 .88
10 Aug .76 .81
11 Aug .57 .67
15 Aug -- .45 .45
18 Aug -- .79 .69
19 Aug .86 -- .71
23 Aug .88 -- .94
Mean .78 .72 .70

a Poor survey conditions produced insufficient data.








Table 4. Total groups of each size class observed by both right-side observers (RB)
versus those observed by only one right-side observer (RF or RR), strip-transect
aerial surveys, Banana River, Brevard County, Florida, summer 1993-94.

Group size
Observer 1 2 > 3 Total

RB 57 42 52 151

RF or RR 66 20 16 102

Combined 123 62 68 253








Table 5. Manatee concentrations observed at "hot-spots" in Banana River during strip-
transect aerial surveys, summer 1993-94. Hot-spots were counted by 2 right-side
observers; highest count is presented.

Cape Canaveral South of Cocoa Beach
Date Hangar AF Sewage Plant 520 Causeway Golf Course Total


1993
11 Aug
25 Aug
26 Aug
31 Aug
01 Sep
02 Sep
16 Sep
28 Sep

Total
Mean

1994
08 Aug
09 Aug
10 Aug
11 Aug
15 Aug
18 Aug
19 Aug
23 Aug

Total
Mean


14
29
15
16
10
17
22
15

138
17.3



11
20
18
26
9
11
11
11

117
14.6


33
14

21
4
6
7
6

91
13.0



18
20
14
20
15
23
29
16

155
19.4


--a


10
6
17

33
11.0



14
28
11
18
17
38
35
31

192
24.0


3
15
8
1
0
7
0
1

35
4.4



0
5
0
0
0
0
0
6

11
1.4


50b
58b
23
38b
14
40b
35b
39

297
45.7c



43b
73b
43
64b
41b
72b
75b
64b

475
59.4c


a No data.
b "Hot-spot" total was greater than total number of manatees
during same survey.
C Sum of column means.


counted on transects








Table 6. Observed distribution of manatee groups in 8 zones (each zone = 5
consecutive transects) and the expected distribution of manatee groups in relation to the
area surveyed in each zone, Banana River, Brevard County, Florida.

1993a 1994b
Transect Observed Area Expected Observed Area Expected
Zone groups km2 groups groups km2 groups

56-60 2 67.1 5.5 5 58.6 7.3

51-55 20 173.4 14.2 13 161.9 20.2

46-50 12 214.8 17.6 37 214.8 26.8

41-45 24 177.8 14.6 34 177.8 22.2

36-40 18 196.0 16.1 28 196.0 24.4

31-35 14 100.8 8.3 2 89.5 11.2

26-30 7 203.2 16.7 16 203.2 25.3

21-25 4 97.8 8.0 17 116.8 14.6

Total 101 1230.9 152 1218.6


a X2
b X2


= 24.275, 7 df, P < 0.005.
= 25.368, 7 df, P < 0.001.








Table 7. Mean annual population estimates with and without hot-spot stratum,
Banana River, Brevard County, Florida, summer 1993-94.


ja


1993


Mean
SE
(cv)


1994


130.7
128.0
117.7
123.0
112.7
139.5

125.3
3.9
(.031)


172.7
187.5
208.3
198.1
161.9
148.0
203.8
151.4

179.0
8.4
(.047)


Mean
SE
(cv)


180.7
151.0
155.7
137.0
152.7
174.5

158.6
6.6
(.042)


215.7
260.5
251.3
262.1
202.9
220.0
278.8
215.4

238.3
9.9
(.042)


a Population estimate derived from strip-transects (excludes manatees counted at
aggregation sites.
b Population estimate derived from strip-transects, added to the number of manatees
counted at aggregation sites.














































Stop
(*) ''




5 1k



Figure 1. Area surveyed for manatees in Banana River lagoon, Brevard County, Florida,
1993-94. Parallel strip-transects were flown from Jack Davis Cut (start) to Highway 404
(stop). High-density stratum included (a) Hangar AF, (b) Cape Canaveral Sewage Plant,
(c) 520 Causeway, and (d) Cocoa Beach Golf Course.








<--- Strip-Transect Width --


S25-

0. 20-
0
.-i 15-
O
0

E o
2 10

5-


---~
i


- ---- ....... .. F ..


<100 101-150 151-200 201-250 251-300 301-350 350-400
Distance from Center Line
m
Figure 2. Manatee observability at different distance intervals from the aircraft on 7
survey flights.


4;


I


301-


rlllllll V


prr








0.6

191
0.5-
0
C 0.4-
0
I-
0.3-
0 96

0
c 0.2-

o 43
o 0.1
I12 12


1 2 3 4 5 6 7 8 9 10 11 12 13
Group Size

Figure 3. Size class distribution of all manatee groups (N=374) observed by right-
and left-side observers on 16 strip-transect aerial surveys, Banana River, Florida,
summer 1993-94.














0.5- | 1994 (N=152)
0)

S0.4-
0

m0.3-
02 -

S0.2-
0
0
S0.1-



1 2 3 4 5 6 7 8 9 10 11 12 13
Group Size



Figure 4. Annual difference in size class distribution of manatee groups observed by
right-side observers in summer 1993 (N= 101) and 1994 (N= 152).







-" 40

LO

z -E


LOC
n


C\l.
o-
0w Oa



E 3




0O CO O O C0 ^- C O CO O (0 CO 0 i 0 CO o 1 O L LO LO LOLO Tr I cO- ` O co Im co o c moN C C N


a,


S
jeqwnu IoesueJ1





o o
CO




"o



0
C\i "
LO 0


u-)



C o o -ti- co ( N co 3 N


(CoLOLO LLOLO Ict It'lO"L CO r CO)CJC'MC'mJCmN J( N


ciqwnu ioasusji







41











APPENDIX A

DAILY RECORD OF AERIAL SURVEYS












ft^
U













4-4
0
c/1


U
CU
I-





0
g






o
*4-4

0


t'


0
*a




'o

CU

4-ON





o '~



B2 -
o
0
0


.0

1
a &<


Il


(U o


11
EioQ


= 1




m ON in o ..0 C



f--- -4-


- oo


ON 0
0
00 OS~


Cn co n I oo









--. ---
ooooo^^nor-'
in c cnt)i~
r^(N ^^( (ic ^ (


----
^l-b m C
0000


- -


e mO
o- a
os -


N 0
() 0-


MONO
- -
00 00


0 0
E --







CO
23 2


a _








04 00
o 04 a U
cW cO .ooe










000000 O OUOOOOOU


o
<
E



= o
O
2 .1
- 0

IUm

II s U








IS.0
o,
Icf
.a -1 ,>

.U U/









II t0

y o
ac
s 0
^1

F^S s



II



Q &a



| U



-c c h-, 1-


en n

-0 0
0000


u U4 u < --: m 2U
UU U<




<;<<22240
MecC==MM u
"t u


Biiaiu








43











APPENDIX B

INTENSIVE AERIAL SURVEYS FOR MANATEES
IN THE NORTHERN BANANA RIVER











Summary

A cost-benefit analysis of different survey methods is beyond the scope of this
project. However, we present the following summary (Table B-l) of warm-season
manatee surveys conducted by the Bionetics Corporation (now Dynamac), NASA, in the
Banana River to illustrate their potential use in trend analysis. For more than 15 years,
Bionetics has conducted intensive searches of the northern portion of our study area with
the goal of a complete count of the manatees in that area (for details on methodology
see Provancha and Provancha 1989). The data suggest a gradual but steady increase in
the number of manatees that summer in the northern Banana River (Table B-l).
However, the small number of replicates per year, as well as variability in the number
each year, limit our ability to infer a population increase. A consistent sample of 26
flights per summer should provide the precision necessary to detect a statistical trend in
population size in the Banana River (see Population Estimates and Population Trends
above). Provancha and colleagues have assembled an excellent long-term database in
the Banana River. These data can be used effectively to monitor population trends with
only a small increase in the number of replicates during the summer months (late June -
early September).

Acknowledgments

These data were graciously provided for comparison purposes by Jane
Provancha and her co-workers.










Table B-1. Four years of summer (June-September) counts of manatees in the northern
Banana River, Brevard County, Florida. Counts were obtained through intensive-search
of a 78.5 km2 study area by 3-4 observers in a NASA helicopter (Source: Jane
Provancha, Bionetics Corp., unpubl. data).

Total Manatees
Date Time Conditiona Areas 1-7 Areas 3-7b


1990 04
09
17
25


1991 14
25
15
30
29
25


1992 19
21
20
17


Jun
Jul
Aug
Sep


Jun
Jun
Jul
Jul
Aug
Sep


Jun
Jul
Aug
Sep


1993 15 Jun
17 Aug
16 Sep
27 Sep


S 1
1
2
2
1
2


1-2
2
NAd
2


40
55
41
32
42.0 (4.8)

104
91
95
68
101
80
89.8 (5.6)

67
139
83
109
99.5 (15.8)

64
118
127
120
107.3 (14.5)


37
46
34
30
36.8 (3.4)

99
63
86
50
77
72
74.5 (7.0)

45
126
67
85
80.8 (17.2)

57
110
71
87
81.3 (11.4)


a Surveys conducted in conditions >3 were excluded from analysis.
b Areas 3-7 also were surveyed by the USFWS/FL-DEP transect method in 1993.
c Annual mean (SE).
d Data not available.