ArticlePDF Available

The Effects of Hurricanes on Birds, with Special Reference to Caribbean Islands

Authors:

Abstract

Cyclonic storms, variously called typhoons, cyclones, or hurricanes (henceforth, hurricanes), are common in many parts of the world, where their frequent occurrence can have both direct and indirect effects on bird populations. Direct effects of hurricanes include mortality from exposure to hurricane winds, rains, and storm surges, and geographic displacement of individuals by storm winds. Indirect effects become apparent in the storm's aftermath and include loss of food supplies or foraging substrates; loss of nests and nest or roost sites; increased vulnerability to predation; microclimate changes; and increased conflict with humans. The short-term response of bird populations to hurricane damage, before changes in plant succession, includes shifts in diet, foraging sites or habitats, and reproductive changes. Bird populations may show long-term responses to changes in plant succession as second-growth vegetation increases in storm-damaged old-growth forests. The greatest stress of a hurricane to most upland terrestrial bird populations occurs after its passage rather than during its impact. The most important effect of a hurricane s i the destruction of vegetation, which secondarily affects wildlife in the storm's aftermath. The most vulnerable terrestrial wildlife populations have a diet of nectar, fruit, or seeds; nest, roost, or forage on large old trees; require a closed forest canopy; have special microclimate requirements and/or live in a habitat in which vegetation has a slow recovery rate. Small populations with these traits are at greatest risk to hurricane-induced extinction, particularly if they exist in small isolated habitat fragments. Recovery of avian populations from hurricane effects is partially dependent on the extent and degree of vegetation damage as well as its rate of recovery. Also, the reproductive rate of the remnant local population and recruitment from undisturbed habitat patches influence the rate at which wildlife populations recover from damage.
Bird Conservation International (1993) 3:319-349
The effects
of
hurricanes
on
birds, with
special reference
to
Caribbean islands
JAMES W. WILEY and JOSEPH M. WUNDERLE, JR.
Summary
Cyclonic storms, variously called typhoons, cyclones,
or
hurricanes (henceforth,
hurricanes),
are
common
in
many parts
of the
world, where their frequent occurrence
can have both direct
and
indirect effects
on
bird populations. Direct effects
of
hurricanes
include mortality from exposure
to
hurricane winds, rains,
and
storm surges,
and geo-
graphic displacement
of
individuals
by
storm winds. Indirect effects become apparent
in
the
storm's aftermath
and
include loss
of
food supplies
or
foraging substrates; loss
of
nests
and
nest
or
roost sites; increased vulnerability
to
predation; microclimate changes;
and increased conflict with humans.
The
short-term response
of
bird populations
to
hurricane damage, before changes
in
plant succession, includes shifts
in
diet, foraging
sites
or
habitats,
and
reproductive changes. Bird populations
may
show long-term
responses to changes in plant succession as second-growth vegetation increases
in
storm-
damaged old-growth forests.
The greatest stress
of a
hurricane
to
most upland terrestrial bird populations occurs
after its passage rather than during its impact. The most important effect
of a
hurricane
is
the
destruction
of
vegetation, which secondarily affects wildlife
in the
storm's after-
math. The most vulnerable terrestrial wildlife populations have
a
diet
of
nectar, fruit,
or
seeds;
nest, roost,
or
forage
on
large
old
trees; require
a
closed forest canopy; have
special microclimate requirements and/or live
in a
habitat
in
which vegetation has
a
slow
recovery
rate.
Small populations with these traits are
at
greatest risk to hurricane-induced
extinction, particularly
if
they exist
in
small isolated habitat fragments.
Recovery
of
avian populations from hurricane effects
is
partially dependent
on the
extent
and
degree
of
vegetation damage
as
well
as its
rate
of
recovery. Also,
the
repro-
ductive rate
of the
remnant local population
and
recruitment from undisturbed habitat
patches influence
the
rate
at
which wildlife populations recover from damage.
Introduction
Environmental catastrophes are of particular interest and concern to conserva-
tion biologists because of their potential impact on small populations. These
concerns arise, in part, from models of population viability which indicate that
catastrophes can cause extinction of small populations (Ewens et
al.
1987). Inter-
est is likely to increase given the increasing loss and fragmentation of habitat
in much of the world, where more species will exist in small isolated popula-
tions,
thereby becoming more vulnerable to catastrophic events. Survival of
small populations in habitat fragments becomes more precarious if the destruct-
ive potential of catastrophes increases, as predicted for cyclonic storms assum-
ing increased global warming (Emanuel 1987).
James
W.
Wiley and Joseph
M.
Wunderle,
Jr. 320
Cyclonic storms, variously called typhoons, cyclones, or hurricanes
(henceforth hurricanes), are found throughout the world (Nalivkin 1983),
with most frequent occurrence in the Caribbean and south-eastern United
States; eastern Pacific along the Mexican coast; western Pacific along the
north coast of Australia, throughout Indonesia, Malaysia, South-East Asia
north to Japan; northern Indian Ocean, along the coast of Bangladesh and
India; and western Pacific Ocean including Madagascar and the east coast
of Africa. Even in some parts of the north temperate zone, hurricanes can
sweep across continents causing extensive destruction, as recorded through-
out Europe (Nalivkin 1983).
Although the frequency of hurricane occurrence varies within a region, the
re-occurrence rate can be high as illustrated by the return rates of hurricanes in
the south-eastern United States and the Caribbean: once every 15 years for
Jamaica (Neumann et al. 1990), once every 16 years to coastal South Carolina
(Hooper et
al.
1990), once every 20 years in south Florida (Lugo et
al.
1976),
and
once every 22 years for Puerto Rico (Weaver 1986). The swath of destruction
can also be wide, as substantial damage has been detected 43 to 60 km from
the track of the hurricane eye (Wadsworth and Englerth 1959, Thompson 1983).
In some of these regions, hurricanes occur with sufficient frequency and
destructive force to be important factors in determining the structure and spe-
cies composition of biotic communities (e.g. Wadsworth and Englerth 1959,
Odum 1970, Crow 1980, Doyle 1981, Weaver 1986). The ecological importance
of hurricanes in influencing biotic communities is consistent with the prevailing
view that natural systems are largely organized by disturbance (Pickett and
White 1985, Denslow 1987).
Given their frequency and potential power, it is not surprising that the effects
of hurricanes on human life and property have been well documented (e.g.
Anthes 1982, Saffir 1991, Sparks 1991). Descriptions of the effects of hurricanes
on wildlife have been widely scattered throughout the literature for some time.
Many of the major effects of hurricanes on bird populations were recognized
by Clark (1906) and are evident in his vivid description of a storm's impact on
bird populations:
On September
11,
1898, St Vincent experienced one of the most destructive
hurricanes that has ever occurred in the West Indies. The center of the
storm passed directly over the island, and the interior forest as well as the
fruit trees on the cultivated areas were almost entirely destroyed. On the
next day the island appeared as if it had been swept by fire: there was not
a leaf nor any green thing in sight. Everything was brown. The number of
birds was very sensibly diminished, those of the "high woods", especially
the [St Vincent] parrots
[Amazona
guildingii],
appearing to have suffered the
most. Hundreds, if not thousands of birds were killed on the island, and
quantities were driven out to sea and lost.
Allenia albiventris
[Scaly-breasted
Thrasher,
Margarops fuscus]
after the storm became a common resident on
Union Island and Carriacou (possibly on some of the other Grenadines
also),
places where previously it had been unknown. It has since, however,
died out at both places. One or two [St Vincent] parrots were picked up
dead on the shores of St Lucia. The effects of the storm were not felt all at
Effects
of
hurricanes
on
bird populations
321
once.
For days afterwards parrots and "Ramier" [Scaly-naped Pigeon
Col-
umba squantosa]
would stray into the smaller towns in so helpless a condition
that many fell prey to the negroes. It is possible that starvation was the
cause of this, as every green thing had been destroyed, and it was several
days before the trees began to put forth buds. When the vegetation did
begin to recover from the shock, the whole island, I was told, presented
much the appearance of a rugged New England landscape in the spring.
A number of the parrots were obtained alive at this time, and some of them
are living in captivity yet.
In this passage, Clark describes many of the direct and indirect effects of
hurricanes on avian populations, as well as some of the responses of birds to
habitat destruction. Direct effects of hurricanes include mortality from exposure
to hurricane winds, rains and storm surges, and geographic displacement of
individuals by storm winds. In the aftermath, the indirect effects become appar-
ent and include loss of food supplies or foraging substrates; loss of nests and
nest- or roost-sites; increased vulnerability to predation; microclimate changes;
and increased conflict with humans. The short-term response of bird popula-
tions to hurricane damage, before changes in plant succession, includes shifts
in diet, foraging sites or habitats, and reproductive responses. Finally, had Clark
remained on St Vincent for a longer time period, he would undoubtedly have
documented the long-term response of wildlife populations to changes in plant
succession as second-growth vegetation increased in the damaged old-growth
forests.
Until recently, descriptions of the effects of hurricanes on wildlife have mostly
been anecdotal and lacked prior baseline data for quantitative comparison after
the storm's impact (e.g. Huntington and Barbour 1936). However, recent hurri-
canes have swept across regions where investigators have had adequate pre-
hurricane baseline data for comparison with post-hurricane effects (as reviewed
in Woodley et
al.
1981, Benito-Espinal and Benito-Espinal
1991,
Finkl and Pilkey
1991 and Walker et
al.
1991).
The purpose of our review is to summarize the effects of hurricanes on bird
populations, concentrating on the adverse effects on terrestrial species in an
effort to identify the characteristics which make a bird population vulnerable to
hurricane damage. Many of our examples come from Caribbean islands where
we have personally studied the effects of hurricanes on bird populations. These
island examples are especially appropriate to continental areas where habitat
fragmentation increases the vulnerability of some species to hurricane impacts.
Limitations of hurricane studies
Care should be taken when evaluating wildlife population changes attributed
to hurricanes, even when excellent pre-hurricane baseline censuses are available
for comparison. Pre- and post-hurricane census results may not be comparable
for a variety of reasons Gegg° and Taynton 1981). As a result of hurricane
damage and lessened food resources, some species may disperse to non-
traditional habitats or wander over a greater geographic area. Defoliated habitats
may increase the likelihood of detection, whereas extensive treefalls could
James
W.
Wiley
and
Joseph
M.
Wunderle,
Jr. 322
obscure detection of understorey- or ground-dwelling species. Wunderle et al.
(1992) compared potential detectability changes associated with a hurricane,
based on detectability ratios derived from bird census data. They found that
the number of species that were more detectable after the hurricane was not
statistically different from the number of species that were less detectable. Even
though some species showed statistically significant changes in detectability
after the storm, no obvious relationship was found between changes in a spe-
cies's detectability ratio and changes in its population estimate. They concluded
that although statistically significant changes in detectability ratios were found
in some species after the storm, it is unlikely that these changes cause a consist-
ent bias in census results.
Direct effects of hurricanes
Mortality attributed to rain,
wind,
and flooding
Hurricane winds, rain, and wave action are commonly assumed to cause bird
mortality directly, although this has often been difficult or impossible to docu-
ment. Indeed birds, with their advantage of flight, may be able to escape the
direct ravages of a severe storm by moving out of its path, provided they have
some way of sensing the need to do so. The literature on wildlife mortality from
hurricanes is sparse and occasionally contradictory (Cely 1991). Accounts of
avian mortality are confined primarily to waterbirds rather than landbirds,
although the former group's larger size could bias mortality surveys (Wayne in
Tomkins 1934, Cely 1991).
Kennedy (1970) reviewed studies relating to the direct effects of rain on birds,
noting that the most important consequence is the wetting of plumage and
possible death from hypothermia. Chicks probably suffered to a greater extent
than adult birds.
Strong winds and rain together can substantially reduce bird populations.
Species commonly found in open and exposed sites, such as seabirds, are espe-
cially susceptible to severe storms (Threlfall et
al.
1974, Cely 1991). High winds
alone can cause mortality in birds (Helmuth 1954, Byrd and Tobish 1978,
McNicholl and Hogan 1979). Birds may be killed outright by treefalls or flying
debris, or they may be thrown against objects, or high winds might blow them
far out to sea where they die of exhaustion and drowning (e.g. Clark 1906,
Jeggo and Taynton
1981,
Pierce 1990). Jones (1987) suggested that cyclones prob-
ably killed some Echo Parakeets
Psittacula (eques) echo
directly by damaging the
hollow limbs in which they seek shelter. Deshayes (in Buffon 1779) suggested
that the flight of the Smooth-billed Ani
Crotophaga
ani was weak, so that many
were killed during hurricanes. J.W.W. found an adult Red-shouldered Hawk
Buteo lineatus
that apparently died from head injuries during Hurricane Agnes.
It is likely that 200-400 dead Brown Pelicans
Pelecanus occidentalis
reported by
Cely (1991) following Hurricane Hugo (1989) died as a result of the high winds,
although it is not explicitly stated. However, Cely (1991) concluded that most
South Carolina birds, except roosting pelicans and other waterbirds, moved out
of the path of Hurricane Hugo or sought shelter on the ground. During the
height of Hurricane Hugo, J.M.W. observed several species of Puerto Rican
Effects
of
hurricanes
on
bird populations
323
landbird (Zenaida Dove
Zenaida
aurita, Common Ground-dove
Columbina
pas-
serina,
Black-faced Grassquit
Tiaris
bicolor,
Bananaquit
Coereba
flaveola)
perching
or foraging on the ground on the leeward side of buildings and large trees.
Similarly, Sutton (1945) noted that many landbirds moved to the ground during
a major hurricane in Florida; he reported only one avian fatality.
Although only 10 dead Red-cockaded Woodpeckers
Picoides borealis
were
found after Hurricane Hugo, Hooper et al. (1990) estimated that 63% of these
woodpeckers in the Francis Marion National Forest were killed or missing,
based on the mean number of surviving birds per colony (1.5) and mean number
of birds per colony prior to Hugo (4.0). Loss of cavity trees undoubtedly contrib-
uted to some of this mortality, because woodpeckers forced to roost outside of
cavities are more exposed to inclement weather and predators (Engstrom and
Evans 1990).
Storm surges or flooding in lowland areas can cause mortality in some popula-
tions.
"Heavy mortality" of roosting waterbirds was attributed to a 3.9 m storm
surge, rather than to high winds associated with Hurricane Donna in Florida
Bay (Robertson and Paulson 1961). Clapper Rail Rallus
longirostris
populations
suffered an estimated 10% mortality rate (20,300 birds, estimated from dead
birds at sample sites) across 31,000 ha of New Jersey saltmarsh, after flooding
caused by Hurricane Bell (Ferrigno and Widjeskow 1977).
Change in geographic distribution and migratory pattern
Storms have been cited as responsible for affecting local movements of animals
(Darlington 1938). Birds are particularly vulnerable to displacement by high
winds (Murphy 1934, Cruickshank 1939, Hill 1945, McClure 1945, Lowery 1946,
Pangburn 1948, Peterson 1948, Mayhew 1949). However, Robertson and Kush-
lan (1974) cautioned that "hurricane transport, particularly of small land birds,
seems to have been accepted rather uncritically . . . During hurricanes most
land birds seek shelter on or near the ground . . . and, despite the enormous
energy of hurricanes, the common vision of such birds being swept away by
the winds is probably incorrect". Nevertheless, severe storms over the North
Atlantic are credited with bringing a spectacular influx of Northern Lapwings
Vanellus vanellus
from Europe to North America (Bagg 1967) and Dovekie (Little
Auk) Alle
alle
to Cuba (Murphy and Vogt 1933). Seabirds are commonly blown
inland, and southern shorebirds are often displaced northward with the north-
ward movement of tropical storms. This frequently occurs on the east coast of
North America where southern shorebirds, including terns, skimmers, pelicans
and frigatebirds, as well as pelagic species such as shearwaters, petrels,
tropicbirds and Sooty Terns
Sterna
fuscata,
have been displaced by storms well
to the north of their normal range (Ciferri in Wetmore 1931, Newman and
Andrle 1961, Pettingill 1970, Hall 1981, Legrand 1985, DeBenedictis 1986).
Although these birds may remain inland for a few days in the aftermath of a
storm, many individuals are found dead or emaciated, suggesting that the mor-
tality rate for such storm-driven waifs may be high.
Hurricanes have the potential to disrupt migration, as documented by several
observers. For example, displacements of migrant shorebirds to West Indian
islands have also been attributed to the effects of hurricanes (Potts 1927, Murphy
James
W.
Wiley and Joseph
M.
Wunderle,
Jr. 324
and Vogt 1933). Fisk (1979) attributed the exceptional wave of fall migrants
detected at her banding station in south Florida to Hurricane Gerda. Thurber
(1980) suggested that Hurricane Fifi (September 1974) was responsible for that
fall's unusual bird migration in El Salvador and that such storms affect migration
patterns in Middle America. Hurricane Fifi had four major influences on bird
migration: the storm caused local concentrations of migrants, deflected indi-
viduals from their normal routes, brought about at least temporary extensions
of winter ranges, and delayed passage of certain species. In contrast, Fletcher
(1989) reported that arrival dates of migrating warblers in parts of Jamaica were
not delayed by the passage of Hurricane Gilbert (10 September 1988). Yunick
(1977) reported nocturnal migration, exhaustion and death of the normally
diurnal Barn Swallow
Hirundo rustica
in the Caribbean during a hurricane.
Thurber (1980) suggested that the extension of the winter ranges of a few
Nearctic species into El Salvador, should these prove permanent, would show
that cyclonic storms can have a lasting effect on the distribution of migrants
by introducing species into new areas that become suitable for winter habitat.
Darlington (1938) believed that hurricanes were an important dispersal agent of
fauna in the Greater Antilles. For instance, the Lesser Antillean Bullfinch
Loxig-
illa noctis
may have been introduced to the Virgin Islands by Hurricane Donna
(Raffaele and Roby 1977). The European Starling Sturnus vulgaris was first
sighted at St Ann, Jamaica, just after a hurricane in 1944 (Jeffrey-Smith 1953).
The now-common American Kestrel
Falco sparverius
was first sighted on Mona
Island in 1935 (Danforth 1936b) and its arrival may have been associated with
a 1930 hurricane which struck Hispaniola, where it is common (Bond 1946).
Similarly, the Smooth-billed Ani is believed to have first arrived on Mona follow-
ing the same hurricane, although it has remained rare (Raffaele 1973).
Conversely, hurricanes may further alter fragments of remaining wintering
habitat thereby affecting survival of wintering residents, such as Bachman's
Warbler Vermivora
bachmani
in Cuba (Amadon 1953, Arendt 1992a; but see
Hamel 1986).
Indirect effects of hurricanes on birds
Loss of food supplies
Defoliation is a conspicuous effect of major storms and, although variable in
both degree and extent, it is the commonest type of damage caused by hurri-
canes (Brokaw and Walker 1991). The high winds which strip leaves from trees
also remove flowers, fruits, and seeds, thereby eliminating the food supply for
many species. Following a major hurricane on Dominica, Snyder and Snyder
(1979) commented on the relative absence of most fruit. Even after more than
one year following Hurricane David, production of flowers and fruits was still
less than 50% of pre-hurricane levels on Dominica (Zamore 1981). One month
after the passage of Hurricane Hugo, J.M.W. found substantial declines in
flowers and fruits (relative to 2.5 years of baseline), including their complete
absence at one site, on long-term phenology trails in the Luquillo Mountains of
Puerto Rico. Lynch (1991) noted the absence of flowers and fruits in the after-
math of Hurricane Gilbert in the defoliated forests of Yucatan. Swindell (1950)
Effects
of
hurricanes
on
bird populations
325
described extensive hurricane damage to acorn and other seed crops in Florida,
and noted that the larger canopy trees, which are usually superior mast pro-
ducers, were those most affected by hurricane damage, as also observed by
others (Gunter and Eleuterius 1971, Platt et
al.
1986, Cely 1991).
In the aftermath of hurricanes, following major loss of flowers, fruits and
seeds,
observers have found birds concentrated around remnant food sources,
including some species that show superior fruit or seed retention. For example,
Snyder and Snyder (1979) found that royal palms
Roystonea
sp. retained their
fruits in spite of high winds, and that frugivorous birds of all sorts were concen-
trated around these food sources on Dominica following a major hurricane.
Two weeks after the passage of Hurricane Hugo, J.M.W. observed three species
of nectarivores feeding abnormally close to the ground (<i.o m) on the flowers
of windthrown
Inga vera
trees in the Luquillo Mountains of Puerto Rico. In the
same forest four months after the passage of Hurricane Hugo, fruit-eating birds
concentrated in fruit-rich but small (120-150 m2) pre-existing treefall gaps,
where mist-net capture rates were exceptionally high relative to those in nearby
pre-existing forest understorey, where fruit was absent (Wunderle 1990).
Given that hurricane winds strip flowers, fruits and seeds from plants, it is
reasonable to expect that populations of nectarivores, frugivores or seed-eaters
are more likely to decline in the aftermath of a hurricane than insectivores or
raptors. Indeed, recent hurricane studies of birds have documented this pattern
in sites with varying degrees of vegetation damage in a diversity of habitats,
including lowland rainforests of Nicaragua (Boucher 1990), subtropical dry for-
ests in Yucatan (Lynch 1991), montane habitats in Jamaica (Wunderle et al.
1992),
subtropical wet forest in Puerto Rico (Waide 1991a), subtropical dry forest
on Guadeloupe (Be"nito-Espinal and Benito-Espinal 1991), wet montane forest
in Dominica (J.W.W., unpubl. data), subtropical dry woodland and subtropical
moist forest on both St John (Askins and Ewert 1991) and St Croix (Wauer and
Wunderle 1992), and mature evergreen forest in Mauritius (Cheke 1987). These
declines in primary consumer populations may, in part, reflect site or habitat
shifts (see section below entitled
Foraging sites
and
habitat
shifts).
It is not surprising that raptor populations remain unchanged after a major
hurricane (e.g. Wauer and Wunderle 1992), given the potential increase in
detectability of their prey. For many insectivores, arthropod populations might
be initially reduced by hurricanes, but because of their shorter life cycles and
rapid reproductive rates, they may recover more quickly than vegetation. In
addition, some insect populations (e.g. Coleoptera, Diptera, Lepidoptera) may
survive storms as larvae or pupae in protected sites in the soil, in leaf-litter
and under bark (Wolcott 1932, B. Freeman verbally). Also, some arthropod
populations have shown major outbreaks following hurricanes, possibly as a
result of decreased predation and increased food resources (Thompson 1983,
Waide 1991b, Torres in press). Thus, many insectivore populations might be
buffered from hurricane-induced food shortages by relying on a food source
characterized by a high level of diversity, resiliency and turnover.
Population declines of waterbirds related to changed food supplies following
hurricanes have been reported. Shepherd et al. (1991) discovered extensive
damage to freshwater habitats through saltwater intrusion in South Carolina
following Hurricane Hugo's passing. They related the decline (>io,ooo pairs in
James
W.
Wiley
and
Joseph
M.
Wunderle,
Jr.
326
1989 to o in 1990) in the breeding population of White Ibis
Eudocimus albus
to
this disturbance of local freshwater feeding sites and the subsequent reduced
availability of freshwater crayfish. Schmitz and Baldassarre (1992) related a
sharply elevated feeding rate of Greater Flamingos
Phoenicopterus ruber
to Hurri-
cane Gilbert, which probably influenced a change in food abundance and availa-
bility in their Yucatan study area.
Loss of foraging substrate
Whereas post-hurricane food loss can explain many population declines, factors
such as changes in foliage distribution and vegetation structure can also affect
certain species. Loss of high-canopy foraging substrates undoubtedly explains
the absence of Black-and-white Warblers Mniotilta varia from montane pine
plantations and their decline in lowland coffee plantations on Jamaica following
Hurricane Gilbert (Wunderle et
al.
1992). The overstorey trees in these habitats
suffered substantial loss of branches and twigs, where this wintering Nearctic
migrant normally gleans insects. Similarly, the extensive hurricane damage to
mimosaceous trees in the overstorey of lowland shade coffee, lowland pasture
and ruinate scrub in dry limestone probably contributed to the decline of win-
tering migrant Prairie Warblers
Dendroica
discolor,
which glean insects from their
leaves (Wunderle et al. 1992). Thus, it appears that gleaning insectivores are
especially vulnerable to the loss of foraging substrates resulting from hurricane
impacts.
Hurricane damage to a Jamaican forest may have contributed to the low return
rate in October of banded Black-throated Blue Warblers
Dendroica caerulescens
to
their winter territories (Holmes and Sherry 1992). Although banded Black-
throated Blue Warblers showed a high return rate in October to their winter
territories in a hurricane-damaged forest in Puerto Rico, a higher proportion
than normal disappeared in the next several months (J.M.W. unpubl. obs.).
Perhaps this Nearctic migrant abandoned traditional wintering sites because of
loss of foliage from which it normally gleans insects.
Hurricanes are important to mangrove ecosystems as mechanisms for "flush-
ing" accumulated organic materials from estuarine systems. However, estuary
substrate changes caused by hurricanes can affect foraging by dependent
waders. Schmitz et al. (1990) suggested that estuarine habitat changes caused
by Hurricane Gilbert resulted in increased foraging times by a Yucatan popula-
tion of Greater Flamingos, which had switched to feeding at abnormal substrate
depths. Furthermore, they suggested that ingestion of accumulated lead shot
deposited at those depths before the hurricane contributed to lead poisoning in
the birds.
Loss of nests, nesting sites or roosting sites
Nests,
nesting sites, nesting material and roosting sites can be destroyed by
high winds, rainfall or flooding. For example, McNicholl (1979) reported the
effects of a wind and rain storm on the breeding marsh birds in Manitoba: 59%
Effects
of
hurricanes
on
bird populations
327
(17) of 29 nests under observation did not survive the storm, and survival of
eggs or chicks was even lower - 76%
(37/75)
succumbed.
Shepherd et
al.
(1991) reported substantial declines in some species of breed-
ing wading birds (White Ibis, Great Egret
Casmerodius
albus,
Tricoloured Heron
Egretta
tricolor),
whereas other species (Snowy Egret
E.
thula,
Glossy Ibis
Plegadis
falcinellus)
showed less change in breeding populations following the passage
of Hurricane Hugo in South Carolina. They suggested the decline in some spe-
cies was related to habitat changes, whereas other species (Snowy Egret) may
have benefited from these changes. Alleng (1990) noted that despite extensive
habitat damage caused by Hurricane Gilbert in the mangroves of Jamaica's south
coast, a small population of waterbirds continued to use the site for breeding,
showed no change in population size, and exhibited only minor behaviour
adjustments.
Species requiring large old trees for nesting or roosting are particularly sus-
ceptible to hurricane effects. For example, tall trees and trees with larger dia-
meters at breast height were more likely than shorter trees to snap or be
uprooted by Hurricane Hugo in Puerto Rico (Walker 1991, You and Petty 1991)
and in South Carolina (Gresham et
al.
1991, Putz and Sharitz 1991). Similarly,
tall trees were especially prone to direct wind damage on St John, Virgin Islands
(Reilly 1991). Trees which had sustained previous mechanical damage were
more damaged during Hurricane Hugo than trees with no previous damage in
the Congaree Swamp, South Carolina (Putz and Sharitz 1991). This suggests
that wildlife using storm-damaged old trees for nests or roost-sites runs the risk
of losing these sites in future storms.
Bald Eagles
Haliaeetus leucocephalus
in the south-eastern United States typically
nest in large exposed pines which are susceptible to hurricane wind damage
(Broley 1951, 1957). This was evident following Hurricane Hugo in South
Carolina where 44% (24/54) or the state's Bald Eagle nests were destroyed (Cely
1991).
Although no eagles were reported dead, 18 nest-trees were blown down.
Many of the eagle territories had few suitable large pines remaining which
might serve as alternative nest-sites. Yet despite major habitat changes, includ-
ing disturbance from timber salvage operations, eagles showed a high degree
of site fidelity and rebuilt nests in 21 of the 24 damaged territories. In the
breeding season following the hurricane, nest-building got under way too late
in the season to produce young on some territories. Nonetheless, hurricane
destruction of eagle nests caused only a temporary set-back to the Bald Eagle
recovery in South Carolina (Cely 1991).
Temple (1978) reported that the majority of known nest-trees of the cavity-
nesting Mauritius Parakeet were toppled by Cyclone Gervaise (1975). Although
nest-sites were not considered to be limiting to the population, the loss of sites
through storm damage may have increased the importance of nest-site competi-
tion with other species (Jones 1980).
Jeggo and Taynton (1981) believed that the effect Hurricane Allen had on
the availability of St Lucia Parrot Amazona
versicolor
nest-sites was of greater
significance to the population than temporary depression of food supplies. Of
the four known nest-sites, three were still standing after the storm, although
other large trees around them had been blown down. The fourth tree was
James
W.
Wiley
and
Joseph
M.
Wunderle,
Jr.
328
completely destroyed. Nichols (1981) reported that two known nest cavities
used by parrots on Dominica before Hurricane David were destroyed by the
storm. All three remaining known sites had been changed and only two were
reused.
Snyder (1977), Wiley (1980, 1985a), and Snyder et al. (1987) discussed the
probable role of hurricanes in reducing availability of nesting cavities suitable
for the Puerto Rican Parrot
Amazona vittata
in certain parts of the Luquillo Forest.
The rarity of cavities in the western forest relative to other parts of the Luquillo
Mountains and the presence of many dead trees on that side of the forest may
be the result of the two hurricanes that hit that forest in 1928 and 1932. Con-
versely, trees that suffer limb breakage, but remain standing, might be suscept-