Proceedings 9th International Coral Reef Symposium, Bali, Indonesia 23-27 October 2000

The future of coral reefs: Integrating climate model projections and the
recent behaviour of corals and their dinoflagellates
O. Hoegh-Guldberg1
ABSTRACT
Most major climate models project that rapid increases to ocean temperature (1-3oC per century) will continue if
atmospheric greenhouse gases continue to accumulate. This scenario, together with the explicit link between coral
bleaching, mortality and sea temperature, has led to the bleak prediction of annual mass coral bleaching event by the
end of the current century. This paper explores this projection for 12 sites in the Pacific Ocean, and improves the
projections by including intensity as well as frequency information. The behaviour of coral reefs over the past twenty
years indicates a strong predictive association between the size and length of sea surface temperature anomalies and
the intensity of mass coral bleaching (Degree Heating Months, DHM). The DHMs associated with severe bleaching
events seen in 1998 in Palau, Okinawa, Seychelles and Scott Reef (Australia) were 3.9, 3.0, 3.1 and 2.6 respectively.
By contrast, DHMs for the less affected reefs of Moorea, Cook Islands, and the southern and central sectors of the
Great Barrier Reef were 0.9, 0.4, 1.7 and 1.4 respectively. Using the Global Circulation Model, ECHAM4/OPYC3
(forced by the moderate IS92a scenario) sea surface temperatures were projected for the next 100 years at the 12
Pacific sites. In all cases, thermal stress increases so rapidly such that DHM values of greater than 10 (triple current
extreme DHM values) are projected to be commonplace by 2080. Data showing that corals are rapidly adapting to
these changes remains extremely scant, as is evidence that corals can rapidly swap their symbiotic algal for more
thermally tolerant ones. The inescapable conclusion from this study is that the projected changes in thermal stress are
likely to result in major negative changes in the distribution and abundance of corals and related organisms as this
century progresses.

Keywords Corals, Bleaching, Thermal stress, Climate
Introduction
Changes to the temperature of the earth is one of the
principal changes being forecast as a result of the change
to the greenhouse gas content of the earth’s atmosphere
(IPCC 1992). The global average surface temperature has
increased by 0.6oC ±0.2°C since the late 1800s.
Significantly, the 1990s appear to have been the warmest
decade (and 1998 the warmest year) in the instrumental
record beginning in 1861 (IPCC-WG1 2001). These
changes are also being seen in the ocean. Available
evidence indicates that sea temperatures are also changing
rapidly (Bottomley et al. 1990, Cane et al. 1997, Brown
1997, Winter et al. 1998, Hoegh-Guldberg 1999, Mumby
et al. 2001, Edwards et al. 2001). Data from other sources
such as coral cores from the central Pacific have
confirmed the recent warming trend (e.g. Kuhnert et al.
1998) and have confirmed its anomalous nature relative to
the last thousand years at least.
Scenarios developed as part of the IPCC Second
Assessment Report estimated increases in global mean
surface temperature of 1-3oC are highly likely under a
doubling of carbon dioxide equivalents in the atmosphere
by 2100. More recently the IPCC has developed a new
set of scenarios that result in a projected temperature
increase of up to 6oC by 2100 (Fig. 22, IPCC-WG1 2001).
Earlier estimates of an increase in sea temperature of 1-
2oC expected by the end of the current century in response
to enhanced atmospheric greenhouse gas concentrations
(Bijlsma et al. 1995) may be too low given the revised
rates of warming of the IPCC Third Assessment Report
(IPCC-WG1 2001). Many authors have pointed to the
significance of rapidly warming oceans for corals and
coral reefs (e.g. Goreau 1990, Glynn 1991, Hoegh-
Guldberg and Salvat 1995, Brown 1997). Potential
details of how climate might change the frequency of
mass coral bleaching and mortality have come to light in
studies that have combined projected changes to sea
temperature from climate models with the known thermal
thresholds of coral bleaching (Hoegh-Guldberg 1999).
The implications of the changes in sea temperature
projected under even moderate changes in atmospheric
greenhouse gas concentrations are quite severe. Annual
bleaching events have been projected for all coral reef
regions by the middle of the current century (Figs. 9, 11
Hoegh-Guldberg 1999).
How changes to the frequency of bleaching events is
likely to affect the distribution of reef-building corals and
their symbionts depends also on how reefs respond to
bleaching events. Communities of corals may completely
recover from thermal stress events (e.g. Harriott 1985) or
may experience almost complete mortality (e.g. Glynn
1988, Brown and Suharsono 1990, Spencer et al. 2000,
Edwards et al. 2001). Clearly, the intensity of mass
bleaching events is an important feature that must be
resolved if future projections are to be useful in projecting
impacts expected in a rapidly warming world.
This paper represents an attempt to derive further
predictive power from the combination of information
from past bleaching events and the behaviour of global
circulation models. Strong et al. (2000) use the length
and size of thermal anomalies to detect mass bleaching
events from satellites. This study uses the method of
Strong et al. (2000) in association with projected changes
in sea surface temperature at 12 sites within the Pacific to
investigate how the intensity (in addition to frequency) of

1
Centre for Marine Studies, The University of Queensland, St Lucia, Queensland 4072, Australia
mass bleaching events is likely to change in a rapidly
warming world. The results expand on previous
projections (e.g. Hoegh-Guldberg 1999) by showing that
the intensity (as well as frequency) of mass bleaching
events is set to rapidly increase.
Methods
Strong et al. (2000) use satellite-derived Degree
Heating Weeks (DHW) to calculate the accumulated
thermal stress that coral reefs experience. One DHW is
equivalent to 1 week of sea surface temperature that is
1oC above the expected summertime maximum sea
(surface) temperature. Using this approach, mass
bleaching events can be detected very successfully.
Strong et al. (2000) and Toscano et al. (1999) have also
observed that DHWs of greater than 10 in the past have
resulted in severe bleaching and mortality among reef-
building corals.
For the current study, the DHW methodology was
modified as follows. Because monthly mean temperatures
were the only data series available that go back to 1900,
Degree Heating Months (DHMs) were used instead of
DHWs. A DHM represents 4.3 DHWs and hence the
critical DHM for severe bleaching and mortality
identified by Strong et al. (2000) is 2.3.
Bleaching thresholds
DHMs were calculated for sea surface temperatures at
12 Pacific sites from November 1981 to August 2000.
Sea surface temperature data were obtained from the
Lamont Doherty Earth Observatory (http://rainbow.
ldgo.columbia.edu/).
Strong et al. (1996) show that deviations of 0.8oC or
more above the mean summertime maxima will usually
result in a bleaching event. In keeping with this, thermal
triggers for the 12 Pacific locations were calculated from
IGOSS data and are reported in Table 1. Calculated DHM
values were compared to literature reports of mass
bleaching events. Not all incidents of mass coral
bleaching are reported or described. Consequently, the
relationship between DHM and mass bleaching incidence
was explored for three fairly well documented sites (Fiji,
French Polynesia and Palau).
Future sea temperatures in the Pacific
Projections of future sea temperature were derived
from the global coupled atmosphere-ocean-ice General
Circulation Model (ECHAM4/OPYC3, Roeckner et al.
1996) for the period 1850 - 2100. This GCM was
developed by the Max-Planck-Institut für Meteorologie
(Bundesstrasse, Germany) and is currently used by the
United Nations for climatology simulations. In order to
include the most conservative scenario, runs were
undertaken using IS92a scenarios with and without the
cooling effect of aerosols taken into account. Horizontal
resolution is roughly equivalent to 2.8o x 2.8o latitude-
longitude. In order to reduce the drift of the unforced-
coupled model, a yearly flux correction for heat and
freshwater flux was employed. Simulation of the El
Nino-Southern Oscillation is essential for approximating
tropical climate variability especially within Pacific sites.
This factor is handled well by the ECHAM4/OPYC3
model (Roeckner et al. 1996, Oberhuber et al. 1998).
Table 1. Trigger temperatures for coral bleaching,
calculated using the mean maximum (summer) sea
temperatures for each region. Values were dropped from
the analysis if anomalies of more than 1.0oC occurred
within a year (e.g. 1998 for Palau).
Location SST trigger (oC)
American Samoa (14oS, 169.5oW) 29.3
Cook Islands Nth (9.9oS, 161oW) 29.6
o o
Cook Islands Sth (17.5 S, 163.2 W) 28.8
o o
Fiji (17 S, 178.4 E) 28.9
o o
French Polynesia (17.3 S, 149.9 W) 28.8
o
Kiribati (2.9 S, 171.4 W) o
29.2
o o
Nauru (2.7 S, 164.5 E) 29.9
o
New Caledonia (20 S, 164 E) o
29.2
Palau (7.3oN, 134oE) 30.0
o o
Solomon Islands (9.2 S, 160.8 E) 28.2
o o
Tonga (19.2 S, 173.5 W) 30.0
o o
Tuvalu (7.7 S, 177.2 E) 29.2
The changes in greenhouse gas concentrations that were
used in the model were as follows: Observed concen-
trations of greenhouse gases were used up to 1990 and
thereafter according to changes outlined in the IPCC
scenario IS92a (IPCC 1992). Greenhouse gases are
prescribed as a function of time: CO2, CH4, N2O and also
a series of industrial gases including CFCs and HCFCs.
Changes in aerosols were as follows: Observed concen-
trations of sulfate aerosols were used up to 1990 and
thereafter changes according to the IPCC scenario IS92a.
The tropospheric sulfur cycle was also incorporated into
projections but only with the influence of anthropogenic
sources considered. Natural biogenic and volcanic sulfur
emissions were neglected, and the aerosol radiative
forcing generated through the anthropogenic part of the
sulfur cycle only.
Changes in thermal stress
Calculating DHM values for the 14 Pacific sites
generated measures of thermal stress over the next 100
years. This was done by combining information on the
changes in sea temperature generated by the GCM with
the thermal triggers listed in Table 1. In order to explore
how the intensity of mass bleaching and mortality events
might change, thermal events were sorted into two
categories: Mass mortality events were deemed to occur
when DHMs rose above 2.3oC months (i.e. equivalent to
the 10 DHW threshold of Strong et al. 2000). Less severe
events (recovery expected) were scored when DHMs rose
above 0.5 oC months. 44
A. Fiji
33
Results
22
Model accuracy *
11

Degree Heating Months (DHM)

A simple comparison of the rates of warming over the 00
last century from the models with actual rates of warming 44
B. French Polynesia
in the Pacific revealed that rates for the 33
ECHAM4/OPYC3 model run without an aerosol effect
(0.65 + 0.06 oC per century) did not differ significantly 22 * *
(p>0.05) from temperature changes over the last century 11 * * *
(COADS/IGOSS data (1998 excluded: 0.79 + 0.29 oC per 00
century). Rates for the ECHAM4/OPYC3 model run with

82

84

86

88

90

92

94

96

98

00
*

19

19

19

19

19

19

19

19

19

20
44
C. Palau
aerosols were significantly lower (p<0.05) than actual
33

rates of temperature change whether or not temperatures

from 1998 were included (0.33 + 0.05 oC per century). 22

11
Table 2. Comparison of rates of warming measured and
00
estimated from 14 sites within the Pacific.

1985

1990

2000
82

84

86

88

90

92

94

19956
98

00
9
19

19

19

19

19

19

19

19

19

Data source Average rate of warming
(14 locations), mean + 95%
Confidence Interval
Measured (COADS/IGOSS) (oC per 100 years)
20
Fig. 1. Degree Heating Month values for three Pacific
sites. Asterisks indicate years in which mass bleaching
events were reported (Salvat 1991, Hoegh-Guldberg and
Salvat 1995, Mumby et al. 2001, Wilkinson 2000).

1998 anomaly excluded 0.79 + 0.29 Table 3. Comparison of recent Degree Heating Months
and mass bleaching mortality estimates.
1998 anomaly included 0.73 + 0.27
Model (ECHAM4/OPYC3) Location DHM Mortality Source
No aerosol effect 0.79 + 0.29 Severe events
Palau 3.9 70-90% J. Bruno, manuscript
With Aerosol effect 0.33 + 0.05 in preparation
Seychelles 3.1 Up to 75% Spencer et al. (2000)
Hind-casting Pacific bleaching events
Okinawa 3.0 90-95% Loya et al. (2001)
Variation in the DHM values generated for past sea Scott Reef 2.6 90-95% L. Smith and A.
temperatures appear to track the incidence of mass Heyward pers comm
bleaching events over the past 20 years (Fig. 1). At three
sites, the incidence of mass bleaching correlates well with Mean + 95% CI 3.2 + 0.46
DHM values greater than 0.5. In Fiji, the recent event in
March 2000 generated the highest DHM value for the Mild events
preceding two decades. All events reported over the past Southern GBR 1.7 10-30% P.L. Harrison and S.
20 years in French Polynesia are also shown. The major (reef crest) Ward pers. comm.
event in Palau in 1998 is shown as a very high DHM Central GBR 1.4 1-16% Marshall and Baird
value (3.9). While qualitative data on the relative (outer reefs) (2000)
intensity of bleaching events in these regions is lacking, it
Moorea 0.9 0% mortality Personal observation
is notable that the Palau event has the highest value and
(10% bleached)
was reported as being probably the most intense.
Cook Is (South) 0.4 0% mortality Personal observation
In an attempt to get a broad measure of whether
(<5% corals
intensity is roughly encoded by the DHM variable, the
bleached)
DHM values for four sites at which severe mass bleaching
occurred were compared to four sites at which milder Mean + 95% CI 1.1 + 0.49
impacts occurred (Table 3).
 As argued by Strong et al. (2000), the DHM values
give information on the intensity of a mass-bleaching
event. Severe events had a statistically higher DHM
value (3.2 + 0.46 oC.months, mean + 95% confidence
intervals) than mild events (1.1 + 0.49 oC.months). As
concluded by Strong et al. (2000), DHM values greater
than 2.4 (i.e. DHW values > 10) caused major declines in
living coral cover.
Projected changes in the frequency and intensity
The information on past changes was used with
climate projections of sea temperature to project changes
in the intensity and frequency of mass bleaching events.
Fig. 2 outlines how the frequency of bleaching events
(any event with a DHM > 0.5) will change for the 14
Pacific sites. The change in the frequency of major
mortality events (DHM > 2.4) is also shown. As with a
previous study of six coral reefs from the Caribbean and
Pacific (Hoegh-Guldberg 1999), the frequency of
bleaching events generally increases steadily until a
maximum value (10 per decade or annual events) is
reached about the middle of the current century (Fig. 7 A-
N, dotted line in each b. panel). Of some importance is
the predicted frequency of severe bleaching events
(events where DHM > 2.3). As expected, this category of
event shows the same trajectory toward a maximum later
in time century.
Discussion
This paper demonstrates that consideration of the
length as well as the size of a thermal anomaly yields
important information about the intensity of mass-
bleaching events. While this relationship needs to be
explored further, the success of Strong et al. (2000) in
detecting the scale of mass bleaching and the logical links
to the underlying mechanism (Hoegh-Guldberg 1999)
underpin its usefulness. This study has shown that
thermal stress on reefs will increase rapidly over the next
100 years, with changes in both the frequency and
intensity of events occurring at the same time. The
projections show the astounding feature that the thermal
stress on reefs will increase to almost three times that
which has been shown to cause major mortality events on
reefs over the past 10 years.
The Importance of secondary factors
Several other factors have been identified as important
in mass coral bleaching. In addition to temperature,
Photosynthetically Active Radiation (PAR) and Ultra-
Violet Radiation (UVR) have been identified as being
important contributing factors to whether or not corals
bleach (Brown 1997, Jones et al. 1998). While a strong
case cannot be made for either UVR or PAR solely
causing mass coral bleaching events (see discussion by
Brown 1997), there is ample evidence that they both
affect the extent of damage to coral tissues during and
after a thermal stress event (Brown 1997, Hoegh-
Guldberg 1999, Mumby et al. 2001). Similar arguments
can be made for the flow conditions existing around coral
colonies (van Woesik 2000). The net effect of these
factors is that warm conditions on still cloudless days will
have a greater effect than on overcast turbulent days or
when the water column above corals is turbid (Mumby et
al. 2001). These secondary factors are, therefore,
responsible for a small amount of variability in the
thermal threshold over and above which mass bleaching
at a site will be triggered. This variability does not
negate the use of threshold values in climate projections
as the light quality varies randomly around a mean value
over long periods and will hence give an average response
of reefs to a set sea temperature.
Thermal stress and the incidence of mass coral bleaching
and mortality over the next 100 years
Degree heating months (DHMs) appears to allow the
prediction of when and where mass bleaching is likely to
occur on coral reefs (Strong et al. 2000). This is the basis
for the highly successful use of “hotspots” viewed from
orbiting satellites to detect mass bleaching events (Goreau
and Hayes 1994). This study shows success in hindcasting
mass bleaching events that have occurred over the past 20
years. Not only did DHM values increase in association
with reported bleaching events, but the DHM values
appeared to encode the intensity of these mass bleaching
events. As predicted by Strong et al. (2000), mass
bleaching events that are associated with high rates of
coral mortality are associated with DHM values of greater
than 2.4 (Fig. 2, Table 3).
DHM values over the next 100 years are set to rapidly
rise (Fig. 3). Based on the fact that bleaching events are
likely to occur if DHM values rise above 0.5, the
frequency of conditions that cause bleaching are set to
become annual by the middle of the current century. This
is similar to the conclusions of a broader study of six sites
worldwide (Hoegh-Guldberg 1999). Perhaps of greatest
concern is that DHM values rise to well over 10 by the
end of this century. Concomitant with these increases in
sea temperature are rapid rises in the probability of severe
bleaching events like those seen in Okinawa, Palau,
Seychelles, Maldives and NW Australia. As the DHM
values seen by 2100 are almost four times higher than
values that are known to have caused these severe
mortality events, the implications for reefs are hardly
optimistic. There has been some discussion about the
possible role of genetic adaptation and symbiont-
switching being reasons for comfort (e.g. Done 1999,
Baker 2001). Unfortunately, evidence of either
mechanism being potent enough to allow reef-buildings to
remain abundant in the face of projected climate change is
either lacking or unconvincing. Given the graphic
examples in 1998 of the scale of thermal impacts, rapid
and immediate attention to reducing greenhouse gas
emissions appears to be the only hope for the world’s
coral reefs that have already been heavily impacted by a
changing climate.
Fig. 3 Projected patterns of bleaching projected for reefs surrounding 12 Pacific nations (up to 2100). Sea temperature
data from the ECHAM4/OPYC3 (Roeckner et al. 1996) model assuming minimal effect of aerosols (see text for other
details). (a) Degree Heating Months (DHM) (b) Frequency of bleaching events over the next century. All events (DHM
> 0.5, dotted line) and changes in the incidence of severe events (DHM > 2.4) are shown. Years indicate the beginning
year of the decade in which these conditions (either 3 major mortality events or annual bleaching) are projected to
occur.
of climate change: scientific technical analyses: the
References second assessment report of the Intergovernmental
Panel on Climate Change : 6-12. (Cambridge
Baker AC (2001) “Reef corals bleach to survive change”. University Press: New York.)
Nature 411: 765-766. Bottomley M, Folland C K, Hsiung, J, Newell R E, Parker
Bijlsma L et al. (1995) Coastal zones and small islands. D E (1990) Global ocean surface temperature atlas
In: Watson RT, Zinyowera MC, Moss RH (eds) (GOSTA, http://ingrid.ldeo.columbia.edu/
Climate Change Impacts, adaptations and mitigations SOURCES/.IGOSTA/
Brown B E (1997) Coral bleaching: causes and
consequences. Coral Reefs 16: 129-138.
Brown B E, Suharsono (1990) Damage and recovery of
corals reefs affected by El Niño related seawater
warming in the Thousand Islands, Indonesia. Coral
Reefs 8: 163-170.
Cane M A, Clement A C, Kaplan A, Kushnir Y,
Pozdnyakov D, Seager R, Zebiak S E, Murtugudde R
(1997) Twentieth-century sea surface temperature
trends. Science 275: 957-960.
Done T J (1999) Coral community adaptability to
environmental change at the scales of regions, reefs,
and reef zones. American Zoologist 39: 66-79.
Edwards AJ et al. (2001) Coral Bleaching and Mortality
on Artificial and Natural Reefs in Maldives in 1998,
Sea Surface Temperature Anomalies and Initial
Recovery Marine Pollution Bulletin 42: 7-15.
Glynn P (1988) “El Nino-Southern oscillation 1982-1983:
nearshore population, community, and ecosystem
responses.” Annu.Rev.Ecol.Syst. 19: 309-345.
Glynn P W (1991) Coral reef bleaching in the 1980™s
and possible connections with global warming. Trends
in Ecology and Evolution 6: 175-179.
Glynn P W (1993) Coral reef bleaching ecological
perspectives. Coral Reefs 12: 1-17.
Goreau T J (1990) Coral bleaching in Jamaica. Nature
343: 417-419.
Goreau T J, Hayes R M (1994) Coral bleaching and ocean
‚hotspots. Ambio 23: 176-180.
Harriott V J (1985) Mortality rates of scleractinian corals
before and during a mass bleaching event. Marine
Ecology Progress Series 21: 81-88.
Hoegh-Guldberg O (1999) Coral bleaching, Climate
Change and the future of the world’s Coral Reefs.
Review, Marine and Freshwater Research Mar.
Freshwater Res 50: 839-866.
Hoegh-Guldberg O, Salvat B (1995) Periodic mass
bleaching of reef corals along the outer reef slope in
Moorea, French Polynesia. Marine Ecology Progress
Series 121: 181-190.
IPCC (1992) The Supplementary Report to the IPCC
Scientific Assessment. (Houghton JT, Callander BA,
Varney SKV eds) (Cambridge University Press:
Cambridge.): 200 pp
IPCC-WG1 (2001) Summary for Policymakers, A Report
of Working Group I of the Intergovernmental Panel on
Climate Change: 1-20.
Jones R, Hoegh-Guldberg O, Larkum AWL, Schreiber U
(1998) Temperature induced bleaching of corals
begins with impairment of dark metabolism in zooxan-
thellae. Plant Cell and Environment 21: 1219-1230.
Kuhnert et al. (1998) A 200-year coral stable oxygen
isotope record from a high-latitude reef off Western
Australia. Coral Reefs (1999) 18: 1-12.
Loya Y, Sakai K, Yamazato K, Nakano Y, Sambali H,
van Woesik R (2001) Coral Bleaching: the winners
and the losers. Ecology Letters 4: 122-131.
Marshall PA, Baird AH (2000) Bleaching of corals on
the Central Great Barrier Reef: differential
susceptibility among taxa. Coral Reefs 19:155-163.
Mumby PJ et al. (2001) Cloudy weather may have saved
Society Island reef corals during the 1998 ENSO
event. MEPS (in press)
Oberhuber JM, Roeckner E, Bacher A, Christoph M, Esch
M, Latif M (1998) Predicting the '97 El Nino event
with a global climate model, Geophysical Research
Letters 25: 2273-2276.
Roeckner E, Bengtsson L, Feichter J, Lelieveld J, Rodhe
H (1996) Transient climate change simulations with a
coupled atmosphere-ocean GCM including the
tropospheric sulfur cycle, Journal of the Climate.
Salvat B (1991) Blanchissement et mortalite des
scleractiniaries sur les recifs de Moorea (archipel de la
Societe) en CR Academy of Science Paris 314: 353-
391.
Spencer T, Teleki T A, Bradshaw C, Spalding M (2000)
Coral Bleaching in the Southern Seychelles During the
1997-1998 Indian Ocean Warm Event. Marine
Pollution Bulletin 40: 569-586.
Strong A E, Barrientos C S, Duda C, Sapper J (1996)
Improved satellite techniques for monitoring coral reef
bleaching. Proceedings of the Eigth International
Coral Reef Society Symposium (Panama).
Strong A E, Gjovig K K, Kearns E (2000) Sea Surface
Signals from Satellites - An Update, Geophys Res Lett
27(11): 1667-1670.
Toscano M A, Strong A E, Guch I C (1999) Ocean
HotSpots and Coral Reef Bleaching: 1999 Update and
New Data Products, Reef Encounters 26:31 pp
van Woesik R (in press) Coral Bleaching: Differential
survival explained by passive diffusion and water
flow. Proceedings of the Ninth International
Symposium for Coral Reef Studies (Bali)
Wilkinson CR (2000) Status of Coral Reefs of the World
2000. Australian Institute of Marine Science: 363 pp
Winter A, Appeldoorn R S, Bruckner A, Williams E H Jr,
Goenaga C (1998) Sea surface temperatures and coral
reef bleaching off La Parguera, Puerto Rico
(northeastern Caribbean Sea). Coral Reefs 17:377-382.