SEI
The Montserrat Oriole - PVA
Population Viability Analysis for the Montserrat Oriole,
November 1997
by
Deborah M. Brosnan PhD
Steven P. Courtney PhD
Sustainable Ecosytems Institute
0605 SW Taylors Ferry Road,
Portland, Oregon, 97219 USA
In consultation with
Resit Akcakaya PhD
Applied Biomathematics
Setauket, New York, USA
This manuscript may not be cited or distributed without prior permission
of the authors.
Executive Summary
This manuscript presents a Population Viability Analysis for the Montserrat
Oriole and a risk assessment of management options. Reviews and evaluations
of volcanic impacts on birds and associated species, probability of continued
eruptions on Montserrat, issues relating to Icterids in captivity, and answers
to frequently asked questions are presented in an appendix at the end of
the analysis.
The Montserrat Oriole, endemic to the island of Montserrat, was previously
regarded as threatened by loss of habitat to agriculture. Hurricane Hugo
and the ongoing eruption of the Soufriere volcano have further reduced Oriole
numbers and habitat.
The current best estimate is that 57% of habitat has been destroyed. Unfortunately,
habitat destruction is concentrated in prime habitat areas, and in ghauts
(gulleys) which are essential to population survival. As of November 1997,
76% of prime habitat has been completely destroyed.
Of the three main population areas, only one (Centre Hills) currently maintains
substantial habitat.
Quantitative population viability analysis allows us to predict the future
of the Montserrat Oriole. Under optimistic scenarios (if eruptions cease
or do not affect Orioles) there is a high risk (1 in 10) that the population
will crash to low levels. Under realistic scenarios (continuing eruptions),
there is essentially no likelihood that a viable population will persist
for 50 years. There is a 50:50 chance of total extinction.
The predicted outcome for the Montserrat Oriole is effective extinction
in the next 10-15 years.
Under impartial scientific criteria, the Montserrat Oriole would be classified
as either endangered or critically endangered.
We examined the effect of removing birds from the population, for development
of a captive breeding program. Removal of birds at this point will not harm
the population. However later removal, when the population has crashed,
may increase extinction rates.
We conclude that the Montserrat Oriole is at great risk of extinction in
the short term. There is no scientific or ethical justification for delaying
implementation of a captive breeding program leading to reintroduction of
the bird to its native habitat.
Population Viability Analysis for
the Montserrat Oriole
The Montserrat Oriole (Icterus oberi) is an endemic bird, restricted
to the island of Montserrat, British West Indies. It is currently threatened
by agricultural development of its habitat, montane forest, and by the action
of the Soufriere volcano. Eruptions have increased in frequency since 1995,
and the best quality habitat has already been destroyed. The purpose of
this Population Viability Analysis (PVA) is to present an impartial analysis
of the current situation, and of conservation options.
Biology of the Montserrat Oriole
The Montserrat Oriole is found only on the island of Montserrat. Its
black and orange plumage once graced forests over the whole island. The
arrival of humans and the ensuing deforestation reduced the Oriole's habitat.
It is now restricted to central and southern forests. In 1984 the population
density of this species was estimated at 500 pairs (Arendt and Arendt, 1984).
Hurricane Hugo in 1989 likely reduced their numbers (Arendt 1990). Prior
to the volcano, the Oriole could be found in three main areas: the bamboo
forest east of Galways Soufriere, the leeward slopes of Chances Peak Mountain,
and the Center Hills (especially in the area of Runaway Ghaut). In 1990
Arendt concluded that the forests of the Soufriere Hills offered the best
habitat for Orioles.
Orioles begin breeding around April. Pairs weave a basket-shaped nest
of vines that they suspend and sew to a lower horizontal leaf, usually of
a forest or banana palm, or in Heliconia vines. Preliminary studies indicated
that the Montserrat Oriole is particularly choosy in selecting a nest site
(Siegel 1983). Two spotted eggs are incubated by the female and both parents
feed the young. The Montserrat Oriole is primarily an insect feeder: In
146 feeding observations, only insects were taken as food (Siegel 1983).
95% of the time these insects are obtained from the underside of tree and
shrub foliage. However, the Montserrat Oriole is also a frugivore. (Wayne
Arendt USFS, personal communication 1997).
First recognized as a new species in 1880 by Fred Oberi of the Smithsonian
Institute US. The Oriole is a member of the Troupial family (which includes
91 species distributed in North, Central and South America, and the West
Indies). Recent mitochondrial DNA (mtDNA) genetic analysis confirmed that
the Montserrat Oriole is a distinct species. It is more divergent in mtDNA
from the other Antillian Orioles than many North American congeneric species
are from one another. (Dr. Irby Lovette and Dr. Eldredge Bermingham (Smithsonian
Institute), and Dr. Robert Ricklefs (University of Missouri, St Lewis) personal
communication, and in prep). Lovette (1997) concluded that the Montserrat
Oriole represents an Evolutionary Significant Unit.
The lack of suitable habitat currently limits the range of the Montserrat
Oriole (Blankenship 1990). The IUCN listed the Montserrat Oriole as a threatened
bird, under stress from human-induced habitat loss. As a result of work
by Blankenship (1980), and Siegel (1983) the Oriole was thought to be threatened
with extinction due to the continued destruction of its habitat by local
agriculturists. A comprehensive study of the Oriole was carried out in 1984
(Arendt and Arendt 1984), and results classified the Oriole as vulnerable
at best or more likely threatened by habitat destruction (Arendt 1990).
Faaborg and Arendt (1985) regarded it as threatened. The International Council
for Bird Preservation and the Smithsonian Institution regarded the Montserrat
Oriole as 'threatened by unchecked habitat destruction' (Collar 1988) and
included the species among its world check-list of threatened birds. Only
one other Oriole (the Martinique Oriole) was listed, indicating that the
Montserrat Oriole was among the most threatened species in its Family. The
Government of Montserrat began implementing habitat protection measures,
and the prognosis for the Oriole was improved until Hurricane Hugo devastated
the island (Arendt 1990).
In a post-hurricane Hugo assessment of birdlife, Arendt (1990) concluded
that bird numbers were in the hundreds, and recommended that to maintain
species :
¨ The government should maintain and enforce the law prohibiting deforestation
above 1500ft to safeguard prime habitat for the Oriole and other unique
species.
¨ Preserve the vegetation and insularity of the major ghauts, as they
serve as refuge areas and corridors, especially during disturbances. Arendt
concluded that ghauts were important in Oriole viability and noted that
"The preservation of the ghauts guarantees preservation of Oriole habitat"
¨ Prohibit land alteration along the proposed road leading to Chances
Peak
¨ Enact a National Park for habitat protection.
The South Soufriere Hills volcano has devastated much of the Oriole's
habitat. The vegetation on Chances Peak has been completely destroyed by
pyroclastic flows and toxic gases (Figure 1). Tuitts Ghaut, where 24 Orioles
were observed in 1990 (the highest number in the survey), and Mosquito Ghaut
where 13 Orioles were observed in 1990 (the third highest number in 1990)
(Arendt 1990), have been completely destroyed by pyroclastic flows. If the
birds were threatened before the eruption, they are now in serious jeopardy
from the extensive habitat destruction from the eruption. The ongoing volcano
poses a major threat to the viability of the species.
Effects of volcanic eruptions on
bird populations
Previous studies show that volcanic eruptions have drastic effects
on birds (see for example, Butcher 1981, Anderson and McMahon 1985, Lambert
1985, Miller and Hill 1985 Manuwal et al 1987, and references below). Pyroclastic
flows kill birds outright. Habitat destruction and reductions in food supply
lead to further declines in bird populations. Insectivorous birds are most
strongly impacted. Ashing has significant effects on nesting success. During
the Mount St Helens eruption many species, including the Northern Oriole
abandoned nests during ashfalls, and that there was an overall reduction
in nesting success of many species (Butcher 1981, Miller and Hill 1985).
Large numbers of St Vincent parrots were killed by toxic gas during the
eruption in 1979 (Lambert 1985). Full details of the impacts of volcanic
eruptions on bird populations are presented in the Appendix at the end of
this manuscript.
Volcanic activity in Montserrat
Since the eruption began in 1995, the overall tendency has been for
activity to increase in a series of successive escalations. Each new peak
of activity is higher than the previous one (MVO report August 1997). Pyroclastic
flows are common and ash eruptions frequently reach heights in excess of
30,000ft. In 1997 the volcano entered an explosive phase. In a recent report,
scientists at MVO unanimously agreed that the eruption is likely to be long-lived
and continue for many more years. Already two thirds of the island has been
evacuated and designated a "danger zone" More detailed on past
and predicted volcanic activity is presented in the Appendix.
Population Viability Analysis
Population Viability Analysis (PVA) is a way to predict the future of
a population. The output of the analysis is usually a set of probabilities
of alternative outcomes ('the risk of extinction', ' the risk of major decline'
etc.) and should be used by managers as a guide to risk. The technique has
been widely applied in conservation biology to many different animal and
plant species (Akcakaya 1992; Akcakaya et al 1994; Boyce 1992; Boyce et
al 1994; Haig et al 1993; La Haye et al 1994; Lindenmayer and Lacy 1995;
Murphy et al 1990; Noon et al in press; Possingham et al 1994; Shaffer 1990).
Formal PVA has been applied in management situations by the US government
(e.g. President Clinton's Northwest Forest Plan; Thomas et al 1995). More
recently the IUCN has adopted PVA as a regional planning tool for many species
at risk.
There are many possible factors that put populations at risk of extinction.
PVA is a technique which allows us to examine the magnitudes of these risks,
and to develop a more formal understanding of how factors interact to determine
the probability of population persistence. While most PVAs have a more or
less explicit mathematical formulation (model), there is no set format,
nor a single 'best' approach. The structure of any given PVA is usually
dictated by data availability, and the goals of management.
PVA is an assessment of the likelihood of future population trends and
events, based on our current understanding of a species' ecology. When there
is uncertainty about essential information, the typical PVA assumes either
the values that are best supported by available data, or examines the range
of future outcomes under alternative assumptions. PVA thus provides a way
to identify those factors that have the greatest impact on the future of
a population (this can be formalized in a sensitivity analysis).
All successful PVAs have certain elements in common (Noon et al, in
press). The most important of these are:
* A fundamental understanding of the species' ecology, including what
constitutes suitable habitat, and the ability of the species to disperse
to distant patches of habitat.
* An understanding of the environmental disturbances that constitute a threat
to the species, their probabilities, and impacts on the populations. This
should include the effects of both deterministic threats (e.g. timber harvest),
and stochastic threats (e.g. hurricanes)
* An understanding of the response of the species to the threats (e.g. density
dependence).
* An assessment of the current state of the population.
* An assessment of the probabilities of future risks.
* An evaluation of the habitat as well as the population itself.
* PVAs are valuable to the extent that they are useful: i.e. that they help
managers make decisions.
The Model
The demography of Icterus oberi was modeled using RAMAS-GIS software.
This modeling approach allows us to examine the probability that the species
will survive under different scenarios. This model has been successfully
applied to many bird conservation situations, e.g. Northern Spotted Owl,
California Gnatcatcher, etc.
The basic structure of the model is a stage-based Leslie matrix projection.
Age-specific fecundities and survival are used to predict the abundance
of birds in different populations. The model also allows for population
structure, with dispersal between spatially defined populations. The effects
of weather on demographic change can be included through the correlation
of population changes across the region. The model is sophisticated in that
it allows the experimenter to select among numerous types of population
dynamics. Density dependence of several forms can be incorporated, as can
changes in the carrying capacity of the environment.
In short, the model is a powerful tool. It lets us examine alternative
scenarios, and to evaluate the risks associated with each.
We have decided to examine five situations:
Scenario 1. Recovery: The 'best case scenario' .
Under this model, we have assumed that the volcano will immediately cease
activity, and that the population and habitat of Orioles will begin to recover.
While this scenario is unrealistic (given the ongoing activity of the volcano)
the results of this analysis will give us a benchmark for comparison to
more realistic scenarios.
Scenario 2. Continued eruptions with small effect: The 'optimistic scenario'
Under this approach, we assume that the volcano will continue to erupt but
will not affect the survival or reproduction of Orioles, and that there
will be no further loss of habitat to the volcano. However recovery of habitat
will not occur.
Scenario 3. Continued eruptions with large effect: The 'realistic scenario'
In this scenario, Orioles are subject to increased mortality, and decreased
reproduction as a result of volcanic effects. We have attempted in this
model to give our best estimate of the existing conditions.
Scenario 4. Continued eruptions with major effects: The 'pessimistic
scenario'
In this approach, we assume the worst conditions for survival, reproduction
and other demographic parameters. In effect, this is an analysis to determine
what will occur if the Oriole is unlucky - for instance if a major hurricane
damages the little remaining habitat. It should be emphasized that this
scenario is a real possibility.
Scenario 5. Management
Under this scenario, we have modeled the effects of removing birds from
the population, as would occur if a captive breeding or rescue program was
attempted. Other model parameters are set as in the 'realistic' scenario.
The worst case scenario
We have not modeled the 'worst case scenario'. Simply put, this would entail
the catastrophic eruption of the Soufriere volcano, and the extinction of
the species. This possibility is addressed in the Appendix. It should also
be made clear that, should the volcano enter a highly explosive phase, there
may be little warning, and little opportunity for intervention and removal
of Orioles.
Model parameters
Several parameters of the populations must be set in each analysis. These
parameter values determine the exact predictions of the model. It is therefore
important to make these values as reasonable as possible. We have attempted
to estimate parameter values from the known biology and status of the Montserrat
Oriole (e.g. Arendt and Arendt 1984, Arendt 1990). Sometimes information
was lacking on some points. In this case we have 1. examined the available
data for similar species, and 2. sought outside expert opinion from scientists
familiar with the biology of Icterids. All parameter values are justified
with source.
Other parameters in the model require estimation of habitat loss, volcanic
activity, hurricane frequency etc. Again, the parameter values used are
justified with source.
We carried out 400 simulations for each scenario, and projected populations
over the next 50 years. Results are presented as an average of these 400
data for each scenario.
Oriole life-history
Following Arendt and Arendt (1984), Bond (1993), Blankenship (1990)
and others, we have assumed that Montserrat Orioles are monogamous, with
each pair producing a maximum of 2 eggs per year. Monogamy is typical of
many but not all passerines. Some isolated populations with little dispersal
opportunity (e.g. Florida Scrub Jays) show cooperative breeding, with only
two birds in the group actually reproducing. The Montserrat Oriole occupies
a restricted habitat, where new breeding opportunities may be limited -
in other species such conditions favor cooperative breeding. By assuming
monogamy we have made an optimistic assumption, which will maximize the
reproductive rate and recovery of the population. The fecundity of the Montserrat
Oriole, at only two eggs per year, is low compared to other Icterids, which
typically lay 3 to 5 eggs. However the low reproductive rate of I.oberi
is well established. It is probable that the species has evolved lower fecundities
in association with its unusual forest habitat. Most Caribbean Icterids
are associated with early successional or edge habitats, where a higher
fecundity is to be expected. The low fecundity of I.oberi is a major factor
increasing its vulnerability in the current crisis.
In all models we have assumed that Orioles begin breeding at one year
of age. We have no data from I.oberi to support this conclusion; instead
we have based this assumption on the biology of similar species (J.Seltz
pers.comm). If we are incorrect, and I.oberi begins breeding at a later
age, then we have over-estimated the recovery rate of Orioles, and our models
are too optimistic.
We have initially set survival and reproductive rates of the Montserrat
Oriole as tabulated here:
Survival Rate Fecundity
Adults .75 .50
Subadults (1 year old) .45 .25
These values are justified as follows:
Survivorship in non-migratory small passerines is typically midrange for
birds as a whole (Karr et al 1990). No long-term studies have been carried
out on survivorship of I.oberi or similar Caribbean Orioles; however we
have no reason to expect that they suffer abnormally high mortality. Under
an annual adult survival rate of .75, we can anticipate 18% of the population
living to 7 years of age or greater. J.Seltz (pers.comm.) indicates that
this is the maximum reproductive lifespan expected for captive Icterids.
Other estimates of survival of tropical forest passerines are given by Karr
et al (1990) who reported that 25 species of Panamanian passerine had an
average yearly survival of .56. Ralph and Fancy (1994 a,b,1995, 1996) reported
adult survival of Akepa, Hawaii Creeper, Omao, Apapane, Iiwi, and Akiapolaau
on Hawaii at .70, .73, .66, .72 ,.55, and .71 respectively. Ricklefs (1997)
reporting on nine species of Turdus, including those studied by Snow and
Snow (1963) and Skutch (1985), showed that survival in lowland tropical
areas averages .761.
Our estimates of adult survival are therefore plausible, but are on
the high range of values seen in tropical forest passerines. Our model therefore
probably overestimates the likelihood of Oriole survival.
We have set subadult survival slightly lower than adult survival. This
pattern is typical of most bird species, where inexperienced subadults must
disperse to new habitats, find nest sites and mates, and succeed in evading
predators. Field studies showing this phenomenon in tropical forest passerines
include: Karr et al (1990); Ralph and Fancy (1994, a, b; 1996); Ricklefs
(1997); Ralph et al (in press). The values reported by these authors are
much lower than we have assumed - i.e. we are probably again over-estimating
the probability that Orioles will survive.
Average fecundity of adults is set at .5 chicks per adult per year.
Fecundity in this sense is the number of young surviving to adulthood (compounded
of egg hatching, chick fledging and first year survival rates). The maximum
fecundity of I.oberi is set by the low number of eggs produced: 2 eggs per
pair, or 1 egg per mated bird. Hence if all eggs hatch, all chicks fledge,
and all survive to year 1, fecundity would be 1. We have therefore assumed
that Montserrat Orioles have a very high survival rate of eggs, chicks and
subadults. Other tropical forest birds have hatching and fledging rates
much lower than 50% (Skutch 1985). For instance, four Turdus species show
fledging rates of eggs from 17 to 28 % (Snow and Snow 1963, Skutch 1983,
Ricklefs 1997). However we feel that the optimistic assumption of high survival
is justified because of the relative scarcity of nest predators on the island.
Wayne Arendt (personal communication 1997) considered that nest predators
were not a significant source of mortality for the Montserrat Oriole. (Snakes
and juvenile Iguanas are known nest predators on Montserrat). High nest
survival is also predicted from the observed low fecundity of I.oberi. We
have however assumed lower reproductive success for inexperienced subadult
birds (as above).
The effect of our assumption of high nest survival will be to be over-estimate
the probability of persistence of I.oberi . The net effect of all our assumptions
on Oriole demography is to predict a maximum growth rate R = 1.04. Thus
our basic model assumes that, in the absence of volcanic effects, or of
other factors reducing survival and reproduction, the Oriole population
will grow at a rate of nearly 4% a year, until it reaches the ceiling set
by the habitat. Thus our basic model, before adding in volcanic or other
effects, or variability in survival rates, predicts an increasing population.
The overall rate of increase is low, which fits our expectations given low
reproductive rates. Our deterministic model predicts that a population would
grow by 43% in ten years.
Oriole distribution and abundance
Prior to Hurricane Hugo, the Montserrat Oriole was found in forested
habitat in three areas: the northern Central Hills, the Chances Peak volcano
area, and the southern Bamboo Forest (south Soufriere) (Arendt and Arendt
1984) (Figure 1). Approximately 1000 birds were found distributed over these
three areas. Hurricane Hugo destroyed large areas of habitat, particularly
in the southern area, where Arendt (1990) found few surviving birds. However
Arendt (1990) argued that birds from this area were 1. persisting in the
'ghauts' or steep volcanic gulleys, which had refugial habitat, and 2. had
emigrated to Chances Peak and other locations. He stated that the population
after the Hurricane was 'still in the hundreds' (Arendt 1990). However Arendt
was unable to give a more accurate estimate of the effect of the Hurricane,
how many birds had been killed, or how many had survived. Arendt (personal
communication) considered that there had been a reduction in density due
to Hurricane Hugo, but noted that the birds had a 5 year recovery period
prior to the volcanic activity.
The status of the Oriole prior to the volcanic eruptions (beginning
1995) is hard to evaluate. It is probable that the population had recovered
somewhat from Hurricane Hugo, but the exact number of birds cannot be reasonably
estimated. In our models we have decided to be optimistic, and to assume
that the species had fully recovered, so that there were 1000 Orioles in
1995.
The number of Orioles presently surviving is similarly uncertain. In
our models we have followed a simple assumption, that the number of Orioles
is affected by the availability of habitat. This is a standard assumption
and result in studies of birds and other wildlife in many environments (e.g.
Thomas, 1986; Larson et al 1986; Raphael and Marcot 1986; Rotenberry 1986;
Dobkin and Wilcox 1986;Akcakaya 1992). We have then estimated that the carrying
capacity and hence the numbers of Orioles will depend upon the extent of
habitat remaining.
Current estimates of the destruction of vegetation on Montserrat are
that less than 40% of Oriole habitat remains. This estimate is based on
1. Documented extent of prior forest cover and of Oriole habitat (provided
by Blankenship 1990, following Arendt and Arendt 1984, which included a
map of concentrated and sparse Oriole distributions), and 2. estimates of
the extent of pyroclastic flows which have been mapped by Montserrat Volcano
Observatory (MVO 1997), and 3. direct observations of vegetation loss to
acid rain, ash and fire. We overlaid the map of Oriole distribution with
the MVO map of habitat destruction (Habitat destruction was classified as
only areas destroyed by pyroclastic flows and with burnt and denuded vegetation
areas.) The table below shows the results of these calculations (see Table
1 for additional information). We calculated that 57% of Oriole habitat
has been completely destroyed. Unfortunately however this damage is concentrated
in higher quality "concentrated" habitat, of which 76% is totally
destroyed. Moreover, pyroclastic flows have tended to follow the course
of the valleys or ghauts, which are the main refuge for the Oriole population.
The amount of habitat in the three population centers is estimated as
follows:
Percentage of Oriole habitat, pre-eruption Percentage of habitat lost by
Nov. 1997 Proportion of Oriole habitat remaining
Centre Hills 37 0 37
Chances Peak 47 94 03
South Soufriere 16 81 03
Our calculations only concern habitat that has been destroyed. A far
larger area has been affected by persistent ashing, acid rain (throughout
the island), and volcanic bombs and fires (especially the southern habitat
area). Our best estimate of this damage is that the remaining southern Soufriere
area is heavily damaged; the Centre Hills less so.
Based on these calculations, we estimate that in the range of 57% to
76% of Orioles have been lost (from total or prime habitat loss respectively).
However it is possible that birds have moved out of these affected areas,
and are still alive (e.g. Arendt 1990). It is unclear that these animals
could continue to survive, or to breed in saturated forest habitat in the
remaining Centre Hills area, or the small pocket in the South Soufriere.
Our considered opinion is that a large majority of Oriole habitat is
either destroyed or damaged. There appears to be little opportunity for
colonization of other habitat in the arid north of the island.
In the PVA model, we have decided to make the optimistic assumption
that many Orioles have so far survived.
In developing the PVA model, we factored in a low mortality from pyroclastic
flows. This was based on previous studies from Mount St Helens volcano on
the effect of pyroclastic flows on bird survival (see Appendix). We therefore
set initial population conditions as follows. We estimate the remaining
population of I.oberi at 600 birds, distributed with 400 birds in the Centre
Hills population, 200 in the South Soufriere. We have also set carrying
capacity in the two areas at the same numbers - that is we have made the
large assumption that there is sufficient habitat for all remaining 600
birds to survive and breed.
Other demographic factors
We have elected to model the remaining Oriole populations using the
following assumptions:
1. The natural variability in reproduction and survival is small.
2. There are no density-dependent factors (such as food shortage) that operate
to limit the population below its carrying capacity (ceiling model).
3. There is no minimum number of birds necessary for breeding (Allee effect).
4. There are no local catastrophes that wipe out one of the remaining populations.
5. The extinction threshold, below which the population cannot recover,
is very low (10 birds).
6. Most surviving birds are adults, with higher breeding success and survival
than subadults.
7. Factors affecting success of one population (e.g. weather) will also
affect the other population (high correlation).
8. There are no factors causing either variation in carrying capacity, or
long-term deterioration of habitat, other than those specified in the model.
9.The distribution of variation in vital rates follows a lognormal rather
than normal distribution.
These assumptions are reasonable; however if incorrect they will all
tend to over-estimate the chances of the survival of I.oberi and to under-estimate
the risk to the species.
Migration and dispersal
We have decided to include different levels of dispersal in our 'optimistic',
'realistic', and 'pessimistic' scenarios. The rationale for these different
values is that, in some species, the overall persistence of the species
depends upon recolonization of abandoned or unoccupied sites from distant
breeding areas (Noon et al in press). The behaviour and persistence of such
species can best be understood in terms of 'metapopulation dynamics'. Broadly
speaking, local populations that go extinct can be recolonized. It is therefore
important to understand the probability of recolonization, which depends
upon dispersal.
Montserrat Orioles appear to move in response to habitat modification.
Arendt (1990) notes that, following the loss of southern habitat to Hurricane
Hugo, birds had moved to local ghauts, and to the Chances Peak area. Recently,
Ministry of Forestry officials have noted birds in the Centre Hills area
that may have moved from the Chances Peak area (Mr. G.Gray pers.comm.).
These observations suggest that Orioles can move considerable distances.
However, in the model, we need to estimate the probability of movement between
the Centre Hills and southern South Soufriere forest populations - i.e.
through the heavily destroyed area of Plymouth and Chances Peak. We believe
that the probability of movement of Orioles through this area is lower than
would be expected through forest, and may indeed be impossible. We have
therefore specified two levels of dispersal: in the ' best case', 'optimistic'
and 'realistic' scenarios we allow 5% of birds to move between populations.
In the 'pessimistic' scenario, we allow no movement or rescue effects.
Hurricanes
Hurricanes are an important feature of Caribbean ecology. Bird biogeography
throughout the Antilles shows a patchy distribution, which can be interpreted
as the results of colonization and extinction. Hurricanes are known to negatively
affect bird persistence on heavily impacted islands. For instance Arendt
(1984) records that the Green-Throated Carib (Sericotes holosericeus), a
hummingbird, was significantly reduced by Hurricane Hugo on Montserrat,
although other species were less affected (the species had recovered, and
was reasonably common in 1995-7 S.P.Courtney pers.obs.). The Antillean Crested
Hummingbird (Orthorhychus cristatus) was driven locally extinct on St.Barths
in 1995 as a result of Hurricane Luis (Jean-Claude Plessey personal communication).
Reintroduction of this species has been proposed. Hurricane Hugo dramatically
reduced the numbers of Puerto Rican Parrots, Yellow-Shouldered Blackbirds,
Guadeloupe Woodpeckers and several other rare or endemic species (Haney
et al 1991). Hurricanes David and Frederick (1979) nearly extirpated the
two parrot species of Dominica (Haney et al 1991).
Hurricane Hugo dramatically altered the habitat of I.oberi, with 100%
of trees in the Bamboo Forest and Chances Peak areas having major damage,
and 72% of trees in the Centre Hills (Arendt 1990). 63% of trees in the
Bamboo Forest were either uprooted or had their main trunk snapped. However
an exact count of the birds remaining was not possible. Other Hurricanes
that have hit Montserrat include Luis and Marilyn (1996), and two major
events in the 1920's.
Hurricanes have the potential to be a major catastrophe for I.oberi.
However the modeling of such large but rare events is problematic. We have
decided to set two levels of catastrophe in our models of the Montserrat
Oriole. Under the 'best case', and 'optimistic' models we allow Hurricanes
to occur with annual frequency of .02. Under the 'realistic' and 'pessimistic'
models, we allow Hurricanes to occur with an annual frequency of .05 (5%
chance annually). This frequency is lower than observed on Montserrat in
the past ten years.
The effects of a catastrophic Hurricane were modeled by reducing population
size by 25%. This seems conservative but reasonable given Arendt's observations
of moderately reduced populations after Hurricane Hugo, and other observations
on Caribbean birds (e.g. 50% of Puerto Rican Parrots were lost (Haney et
al 1991)). Larger catastrophes are possible, but we elected not to model
these; hence all our models are conservative. We ignored additional effects
of Hurricanes on the habitat rather than the Orioles themselves. Given the
scale of the effects of Hugo, Luis, and Marilyn, this may be unreasonable.
However we feel that the loss of habitat is temporary, and will be rapidly
offset by vegetative growth of host trees, and by succession. Loss of habitat
to Hurricanes can thus be seen as only a temporary setback, unlike permanent
losses to agriculture, or long-term damage due to pyroclastic flows.
Effects of volcanic activity on Oriole
survival and breeding
It is reasonable to predict that volcanic activity will continue to
affect I.oberi. This impact could take several forms. Pyroclastic
flows to the north of Chances Peak could harm the Centre Hills population.
These are likely to be limited in impact, in that a substantial part of
the population is in areas sheltered from flows. However the southern population
is extremely vulnerable. Volcanic ballistics regularly drop into the Bamboo
Forest, which is already heavily damaged. There have been extensive pyroclastic
flows down the Whites River, (leading to a formation of a new delta), the
entire Waterfall area has been destroyed by pyroclastic flows, and the Galways
Wall is considerably weakened.
More pervasive effects will be felt by Orioles throughout the island.
Birds can be expected to show reduced survival because of both direct and
indirect effects of ashing and acidic rain (see Appendix). Fine volcanic
ash is breathed in by all vertebrates; birds are especially vulnerable to
respiratory diseases. The effects of the eruption at Mount St. Helens are
well-studied, and have shown dramatic effects on bird populations. A full
discussion of the health effects of ash is given in the Appendix.
Indirect effects on Orioles are likely to include reduced food availability.
Insectivorous birds suffered greatest reductions during and after the Mount
St Helens volcanic eruption. Most affected were birds that fed on tree,
and foliage dwelling insects, and the Oriole tends to feed on foliage-dwelling
insects. Volcanic ash causes high levels of insect mortality by abrading
the cuticle, clogging spiracles and guts. Volcanic impacts on insects will
reduce food supply for Orioles.
We have elected to model ongoing volcanic activity as follows: under
the 'best case' and 'optimistic' scenarios, there is no continuing effect.
Under the 'realistic' scenario we reduce survival by 5%, adult reproduction
by 10%, and juvenile reproduction (already low) by 5%. Similar changes are
also incorporated in the 'pessimistic' scenario which also includes a 50%
reduction in the carrying capacity of the southern population.
Management
We have decided to model management of the species by simply reducing
the number of birds in the model populations, to mimic removal of birds
for a captive breeding and release program. We have elected to remove 50
birds in the model, from either the northern or southern population. 50
is a large group, and probably exceeds the number that will be removed in
any captive breeding effort. In our models we have not included the rescue
effect that is to be expected from reintroducing birds to recovered habitat.
The models are conservative in predicting effects of removal, in that we
have assumed a large number of birds have survived the recent volcanic events.
Hence our models consider only the negative effects of management, and
should be seen as advising on the increased risk that will result from removing
birds.
Summary of Results
Comparison of scenarios
Table 2 gives the overall probability of the Montserrat Oriole surviving
for another 50 years .
Under the 'realistic' and 'pessimistic' scenarios there are appreciable
risks of extinction. Even in the 'optimistic' case (assuming that the volcano
destroys no more habitat and that volcanic ash does not affect Oriole survival
and reproduction) there is an appreciable risk (1 in 10) that the species
will be reduced to 10% of its present numbers. This result is a reflection
of the already low numbers of Orioles remaining - the population is already
so small (at 600 birds) that it is at immediate risk of extinction from
random events. Even under the 'best case' scenario, where new habitat grows,
and the population has the opportunity of recovery, the risk of a population
crash is 7% (1 in 14).
Under the 'realistic' scenario, the risk of Oriole extinction is much
higher. There is a 50:50 chance of the population becoming extinct in the
next 50 years. Under this scenario we expect the population to be greatly
reduced, because of reduced breeding success and survival. Either the population
will go extinct, or it will be reduced to a small remnant of survivors.
Under the 'pessimistic' scenario, the outlook for the Oriole is similarly
bleak. There is a 54% probability of extinction; under the remaining conditions,
a few stragglers will persist for 50 years.
It should be pointed out that any prediction of the future of the Montserrat
Oriole is quite uncertain. The results of the models depend critically on
the assumptions we have used. For instance, lower fledging rates would dramatically
reduce the probability of survival; similarly, it is difficult to incorporate
rare events (such as major hurricanes) in any predictive sense.
Removal of Birds for a Captive Breeding
Program
The effect of removing birds from the population for a captive breeding
program is shown in Table 3. In all situations the effect appears to be
minimal. Removing birds from the population does not lead to detectable
increases in risk, provided that the number is kept small. If large numbers
of birds are removed, of course the population will be damaged. Removal
of up to 50 birds appears safe under the present model.
We have not modeled the effects of waiting several years before removing
birds from the population. Under the optimistic scenario, we expect risks
to be essentially unaltered. Conversely, under the realistic and pessimistic
scenarios, the population will be so reduced after three years (to approximately
half of the current population) that removal of birds at that point will
probably push the wild population close to extinction.
Note that it is possible that many of the birds now seen in the Centre
Hills may have been displaced from the other habitat areas. It is probable
that some of these birds will die, or fail to breed in the areas they have
been forced into. If these birds are removed from the population now, there
will be little effect on population persistence. If however we wait several
years before taking action, the emigrant birds will be dead, and we will
remove more breeders. Our models show clearly that a 'wait and see' approach
is inappropriate - if birds are going to be removed from the population,
this should be done now, when there are emigrant birds present that might
otherwise die.
We have not modeled the effects of reintroduction: we can reasonably
expect extinction risks to drop following reintroduction. For instance the
Puerto Rican Plain Pigeon (Columba inornata wetmorei) has been reduced
to c.200 birds because of hunting, habitat destruction, and Hurricane Hugo,
but is buffered by the captive population of > 50 birds. A similar program
is under way for the Puerto Rican Parrot (Amazona vittata) (Ehrlich
et al 1992). San Diego Zoo has successfully participated in reintroduction
programs for California Condors (Gymnogyps californianus) and San
Clemente Loggerhead Shrikes (Lanius ludovicianus mearnsi).
CONCLUSIONS
Status of the Montserrat Oriole
The Population Viability Analysis has revealed an unequivocal result:
the Montserrat Oriole is a species at great risk. Under realistic assumptions
about current conditions there is a substantial risk of extinction. There
is essentially little chance that the species can persist in the wild with
a viable population. Only if we assume that the ongoing eruptions cease,
or that they have no effect on Orioles, do the models predict any probability
of a population persisting at present levels.
Even these results are not the worst possible case. There is a finite
chance of a catastrophic eruption (report of Montserrat Volcano Observatory),
that will result in immediate extinction of the bird.
Conservation biologists have developed guidelines for determining the
degree of risk faced by a species. Mace and Lande (1991) proposed an evaluation
of a species status on the basis of extinction risks. Their categories have
been adopted by IUCN and other conservation organizations:
CRITICAL 50% probability of extinction within 2 generations
ENDANGERED 20% probability of extinction within 10 generations
VULNERABLE 10% probability of extinction within 100 years
Under these objective criteria, the Montserrat Oriole would be rated
as either ENDANGERED or CRITICALLY ENDANGERED.
One of the functions of PVA is to identify which factors have the greatest
impact on results. Attention may then be focused on these critical factors,
to determine whether they are accurately represented. In this PVA, the greatest
impact derives from three factors: continuing volcanic activity, and the
effects of this activity in reducing Oriole population growth rate, and
variation in this rate. The probable course of the ongoing activity at the
Chances Peak volcano is discussed in the appended document from MVO. The
authors conclude that continuing eruptions are to be expected for the foreseeable
future. The impact of ashing and other volcanic effects on Orioles is less
well documented.
Management
Several options are available for managing Oriole conservation. Nature
can, for instance, be left to 'take its course'. It could be argued that
extinction is a natural event, and that we should allow the species to take
it's chances. This would be tragic for the species, and for the island of
Montserrat, for whom the Oriole is an important national symbol in these
difficult times. Moreover this extinction would not be natural - the Oriole's
distribution has been restricted to the volcanic areas by agricultural development
of lowland habitat (Arendt and Arendt 1984). The volcano is the 'proximate
agent' of extinction (sensu Noon et al in press), but the real culprit is
ongoing human destruction of habitat.
A second option is to remove Orioles from the population. These could
be kept as a genetic reservoir, as an insurance against the fate of the
wild population. If the volcano stops activity the birds could be returned
to their native habitat when it has recovered sufficiently. These birds
could also be augmented by a captive breeding program, where birds raised
in captivity are also released, to augment the damaged population. Advantages
of this approach include eliminating the risks of complete extinction; disadvantages
include high costs, and the potential risks to birds during capture and
transport. If volcanic activity is prolonged, birds may be kept in captivity
for several years. It is essential therefore to have facilities which mimic
field conditions as far as possible, and keep the birds behaving as naturally
as possible.
A third option is to monitor the situation at Montserrat, and to remove
birds when the situation there deteriorates further, or when the population
shows signs of crashing. The advantages of this approach are that it costs
little initially, and keeps the birds in their native environment as long
as possible. The disadvantages are firstly, that the situation may deteriorate
rapidly, with abrupt escalation of volcanic activity - this appears to be
the pattern developing at Chances Peak volcano. Secondly, removing birds
from a population close to extinction is more damaging than removing them
from a more numerous population. Thirdly, there is an added risk to humans
capturing birds in a highly active volcanic area - the fewer the remaining
birds, the longer the people must remain exposed to danger in order to capture
them. Lastly, there is a small but real possibility that the volcano will
undergo a catastrophic eruption, eliminating the species before any rescue
attempt is possible.
Our models have addressed the potential negative effects of removing
birds from the population. We conclude that these risks are real, but small
in the current situation (either small or undetectable increases in risk).
They are certainly much smaller than the very large risks of extinction
if volcanic activity continues.
An important point here is the effect of being wrong in our assumptions.
Noon et al (in press) discuss the 'risks of being wrong'. If we have over-estimated
the risks facing the Oriole, and proceed with a captive breeding program,
the consequences will be that we will spend money unnecessarily, and we
will have a small negative effect on the population. If on the other hand,
we under-estimate the gravity of the situation, and do not act quickly enough,
we risk losing the entire species by our inaction.
We conclude that, if sufficient resources exist, and adequate holding
facilities are available, then there is no incentive to monitor the species.
We believe that an immediate effort to set up a captive population is the
only biologically and ethically defensible option.
Table 1
Effect of Volcano on Montserrat vegetation and on Oriole habitat
Percentage of Island occupied by Orioles (pre-eruption) 29.0%
Prime Habitat 5.0%
Sparse Habitat 24.0%
Distribution of Prime Habitat
Centre Hills 24.1%
Chances Peak 58.7%
South Soufriere 17.2%
Percentage of Habitat destroyed
Prime Habitat 76 %
Sparse Habitat 53 %
Percentage of Habitat lost, by area
Centre Hills 0 %
Chances Peak 94 %
South Soufriere 81 %
Percentage of Whole island
Destroyed by Pyroclastic Flows 17 %
Destroyed by Gas emissions, etc. 15 %
Total 32%
Table 2. Survival Probabilities for the Montserrat
Oriole
Scenario
Reproductive rate of population R
Probability of extinction
Probability of population crashing to 10 % of current numbers
Best possible
(Volcano ceases; recovery)
1.04
0
7%
Optimistic
(No further effect of volcano)
1.04
0
10%
Realistic
(Volcanic impacts on survival)
0.92
0.53
0.98%
Pessimistic
(Volcano affects survival and habitat) 0.92 0.54 100%
Table 3 . Probability of Extinction in unmanaged
and managed (captive population removed) scenarios
Scenario
Unmanaged
Managed
Best case
0 0
Optimistic
0
0
Realistic
53% 52%
Pessimistic
54%
53%
Acknowledgements
The authors thank Wayne Arendt, (US Forest Service, Puerto Rico), Bob
Anderson (Weyerhaeuser Company) Dr. Willy Aspinall (MVO, and Aspinall and
Associates), Dr. C Chrisellson (US EPA) Dr. Lev Ginzburg (State University
of New York), Dr. Richard Hoblitt (Cascades Volcano Observatory), Dr. Irby
Lovette (Smithsonian Institute), Dr. Barry Noon (Colorado State University),
Dr. C. J. Ralph (US Forest Service), John Seltz (Sedgwick County Zoo) and
The Montserrat Volcano Observatory for information and input that greatly
contributed to the preparation of this manuscript.
Appendix
1. The Effects of Volcanic Eruptions on Wildlife
Volcanic threats to wildlife (including the Montserrat Oriole) comes from
three main sources
1. Outright mortality from pyroclastic flows and surges
2. Indirect mortality through the destruction of essential habitat, nesting
and food resources.
3. Increased mortality from ash exposure.
The direct and indirect effects from flows and habitat destruction
have been studied in previous eruptions. The 1980 Mount St Helens eruption
provided extensive information on volcanic effects on a large landscape.
The 1978 St Vincent Soufriere eruption, provided additional information
on bird mortality. Exposure to toxic volcanic ash is a health concern, and
the nature of the impacts make clear cause and effect difficult to determine,
particularly in long-term animal models. Some results are available from
field observations from the Mt St Helens eruption. In the sections below
we review and evaluate the information on volcanic impacts on birds and
their associated community.
Effects of Pyroclastic Flows and Tephra Deposits (ash) on Bird Populations
Pyroclastic flows kill birds (Butcher 1981, Anderson and McMahon
1985, Manuwal et al 1987, Chrisellson unpubl. ms. and personal communication).
For instance, the Mount St Helens eruption of 18 May 1980, destroyed an
area of over 600km2, including half of the upper subalpine forests. This
eruption instantly killed the majority of birds present in these forests,
including permanent residents, winter residents, and early migrants (Manuwal
et al 1987). Toxic gas from the 1979 Soufriere eruption killed a significant
number of St Lucia Parrots (Lambert 1985).
While heat surges and tephra can kill birds outright, their
larger effect may be through destruction of habitat. At Pine Creek, Washington,
heat surge from the Mt St Helens eruption scorched the forest. This led
to drastic reductions in tree-foliage insectivorous birds (Manuwal 1987).
Two and five years after the eruption, the guild at Pine Creek was 50% smaller
than at less impacted sites, and tree-foraging insectivores birds had been
replaced by low understory and ground insectivores. Areas that received
only a light dusting of ash were probably temporarily abandoned and soon
recolonized by insectivorous birds (Manuwal 1987). Insectivorous birds,
(including the Northern Oriole) are most severely affected by volcanic ash
(Butcher 1981, Manuwal 1987).
Ashfall causes birds to abandon territories and nests. Ash deposits
from the 18th May 1980 Mount St Helens eruption, totaled about 5cm at the
Wildlife Recreation Area, Naches, Yakima County, Washington. This ashfall
caused all seven pairs of Northern Orioles (Icterus galbula bullockii) in
the study area to abandon territories, including three females that were
building nests. Subsequently six of the fourteen birds returned to the area
(Butcher 1981). The remaining Orioles either died or migrated elsewhere.
One Lazuli
bunting (Passerina amoena) was found dead the day after the eruption. More
than 200 individuals swallows of four species, that had begun nesting prior
to the ashfall, had interrupted their activities in response to the ash
(Butcher 1981). At Sprague Lake, Ring-billed (Larus delawarensis) and California
gulls (L. californicus) suffered reduced reproductive success as a result
of ash deposition (approximately 3.5cm ash) (Hayward et al 1982). No adult
birds suffered mortality, but birds left their nests during ashfall, with
the result that eggs became blanketed in ash. Species varied in their ability
to remove ash from the eggs (61.2% of ring-billed nests remained buried
5 days after the eruption). Rainfall cemented eggs into the substrate, and
prevented tending by adults. Overall, reproductive success was reduced and
was inversely correlated with ash depth. In the Columbia Basin, the Mt
St Helens eruption led to initially high failure rates of nests in pheasants,
ducks and songbirds. This was accompanied by a large reduction in insects,
and reduced abundance and biomass of fish in streams that received heavy
ashfall (Foster and Myers 1985). Subsequent nesting was successful for most
game species.
Following the Mount St. Helens eruption, bird species gradually
recolonized some of the impacted areas. Although species richness increased
rapidly, there was a significant change in the species composition of the
bird community. Ground nesting birds became dominant in the areas affected
by the volcano (Andersen and MacMahon 1986), and birds appeared to suffer
increased predation after the eruption. Insectivorous birds were dramatically
less abundant after the eruption.
Volcanic Impacts on Insect Populations
The Mt St Helens volcano caused widespread mortality among insects
( Akre et al 1981, Edwards and Schwartz 1981, Brown and Hussain 1981, Howell
1981, Johansen et al 1981, Klostermeyer et al 1981, Fye 1983, Edwards 1985,
Mason et al 1985). This contributed indirectly to reduced survival and reproduction
in some bird populations (Foster and Meyers 1985). Dry volcanic ash from
Mt St Helens eruption was lethal to many insects: Volcanic ash caused cuticular
abrasion, and this together with excessive salivation during grooming led
to dehydration and death of many insects (Edwards and Schwartz, 1981).
Obstruction of the spiracular valves and accumulation of ash boli in the
gut were contributory factors. During the eruption, flows and surges killed
some species instantly (those on the surface and in vegetation), and many
of these have not recolonized (Vane-Wright and Ackery 1984). Further reductions
in insect abundance resulted from habitat loss. By contrast, fossilivorous
insects (living deep in the soil) were buffered from ash and flows and survived.
The snowpack (still present in May) insulated many insects from heat and
desiccation. It also helped to moisten and compact the ash so that insects
emerging insects did not necessarily encounter a lethal habitat (Edwards
1985). Aerial dispersal greatly contributed to recovery of many insect populations
Mount St Helens was a strong, pulse disturbance that dramatically
affected flora and fauna over a large landscape. But the eruption was over
in a couple of months. The Soufriere Hills volcanic eruption has been escalating
since July 1995 (over two years) and is predicted to continue for many more
years (MVO report, August 1997). Pyroclastic flows are now common, and
they destroy increasingly large areas of habitat. Ash deposition is chronic.
Rates of deposition and aerial extent of coverage have increased. For example
in September and October 1997, most of the island received significant ashfall
(MVO personal communication, and Brosnan, personal observation). During
that time we observed damage to vegetation in the far north of the island
(Brosnan and Seltz personal observation). The studies cited above report
a significant and long-term effect from Mount St Helens eruption on the
fauna of the region. It is highly likely that Montserrats wildlife is experiencing
similar stresses and impacts.
2. Health Effects and Risks to Wildlife from
Ash Exposure
The escalation in volcanic activity, especially since June 1997,
has produced significantly higher levels of free ash in the air and on the
ground (MVO 1997). Recent medical and geological studies of the effects
of ash on health have indicated that people and wildlife are exposed to
levels of respireable toxic ash, which exceed conventional safety standards
in the UK (MVO 1997). But because of the multitude of factors that affect
respiratory health, especially in cases of chronic exposure, direct causality
is difficult to show.
Birds, because of their highly efficient respiratory system, ,small
size, rapid metabolic rate and low fat content are notoriously susceptible
to respiratory diseases (Dumonceaux and Harrison, 1994, 1997), and are
thus likely to be at risk from ash exposure. Their efficient system rapidly
extracts harmful gases and particles from inhaled air, thus increasing sensitivity
to toxins. Avian pulmonary silicosis, rhinitis and other respiratory diseases
are caused by inhalation of fine particulates and toxic gases (Tully and
Harrison 1994, Dumonceaux and Harrison 1997, Hillyer 1997), and airborne
particulates are listed among the toxic etiologic agents of respiratory
disease (Dumonceaux and Harrison 1997) A well-ventilated, dust and toxin
free environment is fundamental to avian respiratory health (Hillyer 1997).
While no studies, that we are aware of, directly address the effects
of volcanic ash on birds, some information is available from previous eruptions.
California gulls showed evidence of eye-irritation following the Mt St Helens
eruption, while ring-billed gulls did not (Miller and Hill 1985). The same
study found some evidence of ash-induced mortality in chicks: a California
gull chick was found with its beak and throat obstructed by a plug of hardened
ash attached to its tongue, and a dead ring-billed gull chick was found
outside its nest with one foot firmly cemented in the ash.
Most studies on ash impacts emphasize short-term exposure, often in
controlled settings . Prolonged exposure is likely to have long-term effects.
Kiessling et al (1981) reported that rats exposed to Mount St Helens volcanic
ash showed an acute pulmonary inflammatory response, followed by granulomatous
and fibrotic reaction which persisted until the end of the 6 month study.
The authors also examined the lungs of two heavily exposed humans and found
lesions consistent with a reaction to volcanic ash. Their results indicate
a moderate fibrogenic potential. In a separate study comparing the effects
of soil and Mount St Helens ash on Fishers rats, Sanders et al (1983) found
an enhanced histological degree of granulomatous reactivity, lipoproteinosis,
fibrosis and bronchiolar hyperplasia in ash-exposed rats. Mediastial lymph
nodes were 8-18 times larger in ash-exposed animals. (Rats received 22 or
77mg of ash in two or seven weekly intervals). In a two-year exposure study,
Wehner et al (1986) found similar exposure related lung responses in rats,
and females showed a slightly elevated carcinoma rate. McLemore et al s
(1984) study on cultured human alveolar macrophages found decreased cell
viability after 24 hours exposure to Mt St Helens ash in 300mug/ml cultures.
Buist et al (1983) found no significant acute effect of ash in 101 children
exposed to 1.2cm ashfall in June 1980. A four year study on effects of ash
exposure in 712 loggers in Washington State, showed that for the first year
loggers experienced reduced lung-function that was directly related to exposure
levels. Subsequently there was no significant difference among groups (Vollmer
et al 1985).
Studies on fish and plankton show that volcanic ash can affect
feeding and survival. Juvenile salmon were exposed to ash during the Mount
St Helens eruption. Bioassays on juvenile salmon confirmed that airborne
ash can causes mortality within hours of exposure. The gills of juvenile
salmon were uniformly coated with mucous and volcanic ash, and impaired
oxygen exchange was suggested as primary cause of death (Newcomb and Flagg
1983). Bolding (1981) found that ash abraded fish gills. Ash interfered
with carbon assimilation in some zooplankton (Gaddy and Parker 1986). Caddisfly
larvae density decreased in streams following Mt St Helens eruption but
the role of ash versus increased runoff is difficult to separate (Cushing
and Smith 1982).
The above studies indicate an effect of ash on wildlife and humans.
However, because there have been no long-term studies or monitoring we cannot
extrapolate the results into quantifiable reductions in reproductive output,
or survival rates. However, chronic exposure to ash, as experience by Montserrat's
bird populations is likely to have important consequences for the long-term
health.
CONCLUSIONS: Volcanoes destroy habitats, kill wildlife, and alter community
composition. Results from studies described above suggest that Montserrat's
wildlife will already have experienced some losses. The remaining wildlife,
particularly that confined to the southern two-thirds of the island (the
evacuated zone) is under severe threat. As individuals become more concentrated
into remaining habitats, their risk of mortality is likely to increase.
A single pyroclastic flow through a refuge zone can destroy all the animal
life in the area.
3. Volcanic Activity on Montserrat: Past
activity and Predictions
Since the eruption began, the overall tendency has been for activity
to increase in a series of successive escalations. Each new peak of activity
is higher than the previous one (MVO report August 1977). In a recent report,
scientists at MVO unanimously agreed that the eruption is likely to be long-lived
and continue for many more years. Even without explosive activity, the
trend for increasing dome size is likely to lead to even bigger pyroclastic
flows (MVO, 1997), and thereby to greater habitat destruction. The scientific
assessment of August 1997 concluded that: 1. Further escalation of activity
to significantly higher levels is thought to be increasingly likely. 2.
The chance of an eruption two to three times as intense as any explosive
eruptions experienced so far may be as high as 1 in 5. The probability of
a powerful, sustained vertical explosive eruption occurring in about the
next 6 months is estimated at between 1 in 10 and 1 in 100. The possibility
of an explosion generating a high velocity pyroclastic surge, that might
threaten the northern habitats is in the range of 1 in 500 to 1 in 10,000.
(The chances for a cataclysmic event are estimated at 1 in 10,000, with
significant variability among individuals in risk estimate).
The pattern of volcanic activity, and the unanimous conclusion of volcanologists
indicate that further habitat destruction is likely. Pyroclastic flows,
often traveling in excess of 100mph and at temperature of over 1000C, kill
plant and animal life in their pathway (see below). Their high velocities
preclude escape. Chronic exposure to toxic gases will further destroy forests
and other ecosystems. When habitat is lost, populations decrease: Habitat
loss is cited as the main factor responsible for species extinction globally
(Wilson 1992). In the case of island populations (including birds), there
is no alternative habitat or refuge and the prognosis for persistence is
poor. Our conclusion is that not only is the wildlife currently threatened,
but that the probability for greater threats and local extinctions is high
(see below).
4. Conclusions
In a recent scientific report volcanologists unanimously agreed that
Increased activity is likely to lead to more extensive flows and will lead
to greater habitat destruction. When habitat is lost, population densities
decline. Indeed habitat loss is cited as the single most important cause
of global species extinctions. As long as activity continues in Montserrat,
the wildlife is at risk.
Volcanoes kill wildlife. The results from Mt St. Helens and other eruptions
show this conclusively. Birds are highly susceptible. Insectivorous and
tree-nesting birds are most strongly affected. Orioles, being insectivorous,
vegetation-nesting, and a forest species are a high risk species. (Ash led
to the disappearance of Northern Orioles from the vicinity of Mt St. Helens).Volcanoes
cause mortality, breeding failures, reduced density through habitat and
food loss, and they lead to long-terms alterations in bird communities.
Humans and wildlife have been subject to increased ash levels in the
air and ground. These levels exceed conventional safety standards in the
UK. Studies on humans and other animals indicated that exposure to volcanic
ash leads to respiratory problems. Birds are notoriously susceptible to
respiratory ailments and are thus likely to be at risk from mortality in
Montserrat.
5. Maintenance of Icterids in captivity
Orioles and other Icterids are maintained in a number of North American
zoos. The species most commonly kept is the Troupial (Icterus icterus).
As of June 1997, the captive population consisted of 31 birds (J.Seltz pers.comm).
The majority of these birds have been born in captivity; 3 have been raised
in the past year. This species of Icterus at least can be maintained in
captivity, and has increased in numbers due to captive breeding efforts.
The North American Regional Collection Plan for the Passeriformes Taxon
Advisory Group stated, regarding Troupials: 'a popular and steady zoo exhibit
bird. Even with relatively few importations, populations have been able
to sustain themselves in captivity. Icterus icterus is popular as well in
the private sector, so additional genetic material should be available.
Troupials have historically been kept and bred in mixed species aviaries.
….Because of past breeding successes at a few zoos, a studbook should
be established to prevent inbreeding and to monitor growth of the population
(J.Seltz PACT TAG REP 1996)'.
J.Seltz has identified the following issues regarding maintenance of
Montserrat Orioles in captivity:
1. There is likely to be intra-species aggression, so that birds should
be housed in pairs. During transit, males and females should be housed separately.
2. It will be important to 'meat off' birds at an early point - that is
to switch them to captive diets. This is a technique that has been well-developed
for Troupials and other species.
3. Birds on Montserrat are likely to be have respiratory problems. Handling
stress may further compromise their health - it will be important to stress
these birds as little as possible, and to maintain vigilance for secondary
infections.
4.Birds with respiratory stress should be maintained in a facility that
is warm and humid, e.g. in a tropical environment, where they will have
a better chance of recovery. A humid outdoor environment, where there is
little dust, will minimize re-exposure to particulates.
Based on these observations, it appears that Icterids can be expected
to survive well in captivity, and even to breed. Any Orioles removed from
Montserrat however are likely to be stressed, and should receive minimal
additional stress, and careful attention.
6. Frequently Asked Questions regarding removal
of Montserrat Orioles to captivity.
1. Why act now? Why not wait, monitor the population, and take action
when birds are seen to be dying?
There are several important reasons why a 'wait and see' strategy is
not advisable. Firstly, the population is currently in danger from further
eruptions. The Montserrat Volcano Observatory scientists have evaluated
the risks of escalating activity at the Soufriere volcano. In their most
recent (August 1997) evaluation, MVO scientists unanimously predicted increasing
volcanic activity. There is a risk of a volcanic surge passing over the
Centre Hills - this would eliminate the remaining population centre for
the Oriole.
Secondly, if birds are to be removed from the wild, it is better to
do so when there is a strong chance they will survive (before they are in
respiratory distress), and when removal will not significantly reduce population
numbers.
Thirdly, to maintain an adequate captive population, it is important
to have a representative sample of the wild population - it is far better
to remove 30 birds now, than 10 when the reduced population is in severe
distress.
Fourthly, there is a risk to the humans carrying out this work. The
longer we wait, the greater the probability that a rescue effort would be
carried out with the threat of large eruptions.
Given the low risks to the birds of captivity, the current availability
of resources to maintain them, and the high risks of extinction, it is inadvisable
to adopt a monitoring strategy.
2. Why not relocate the Montserrat Oriole to another island?
There are again several reasons why this is not a viable option.
Firstly, the species is an endemic of Montserrat; introduction to another
island would alter the ecology of the host environment.
Secondly, introductions have limited success without a clear understanding
of the problems faced by immigrants (new predators etc.). There would also
be the problem of recapturing the birds for relocation back to Montserrat.
Thirdly, it is not clear that there is a suitable island in which to
introduce the birds. Several other islands with the necessary habitat (such
as St.Lucia, Martinique, Guadeloupe) already have endemic Orioles of their
own. It would be irresponsible to introduce a competing and possible inter-breeding
species. Other islands either lack suitable habitat (e.g. Antigua), or have
predators that would threaten the Oriole (e.g. St.Kitts).
Fourthly, it is unlikely that the political negotiations necessary to
achieve such a transfer could be carried out very quickly. Most of the nearby
islands are independent nations.
3. How much will a capture and reintroduction program cost?
There are several elements to such a program. The initial capture effort
will be just the first phase. The most expensive (and indeterminate) phase
will be maintenance of birds in captive facilities until such time as they
can be re-released to the wild on Montserrat. An essential third component
is training for Montserrat personnel and staff, to ensure that local expertise
is available for reintroduction and conservation. Commitments have been
made to fund all these phases.
4. Who will control the fate of captive birds?
The goal of a captive maintenance program must be to return the species
to the wild at the earliest safe opportunity. Day-to-day care of birds will
be decided by the facilities where the birds are maintained. However the
ultimate decisions on reintroduction and disposition of the birds must remain
with the Government of Montserrat.
5. Where should the Orioles be located?
The birds should be placed in facilities that will have sufficient resources
and capacity to care for them. As indicated by J.Seltz (see above) there
may be advantages to having Orioles maintained in warm, humid conditions,
preferably outdoors. The IPEAT facility in Costa Rica is one option for
such facilities. At the same time, it is generally a good strategy to maintain
a captive population in more than one facility, in order to spread costs,
and the risks of infections.
6. If we see Orioles in the forest now, what is the problem?
It has been suggested that, since there are birds to be seen, there
is no problem, and that it is alarmist to argue otherwise. We strongly disagree
with the notion that since birds are present, they must represent a healthy
population. There are several reasons to think otherwise.
Firstly, all the remaining Oriole habitat is subject to heavy ashing.
It is unlikely that Orioles will either breed or survive well in these conditions.
The birds may instead be under severe stress.
Secondly, the Orioles that remain are almost certainly packed into the
little remaining habitat. It is probable that the number of these animals
now exceeds the capacity of the area to support them.
Thirdly, and most importantly, this argument says that you only have
a problem when you don't see birds - by then it will be too late.