GCRA  OVERVIEW  FAQ  NEWS  ARTICLES  PHOTOS  REEF ISSUES  RESTORATION  PAPERS  LINKS 

Caribbean Sea Surface Temperatures and Coral Bleaching
1989—1996

Thomas J. Goreau, Raymond Hayes, and Alan Strong

Costa Rica AMLC Draft September 16 1997 

ABSTRACT 

Caribbean coral bleaching between 1989 and 1996 was most intense following the warmest periods at each site recorded by satellite sea surface temperature (SST) data. As between 1982 and 1991, the most intense bleaching took place within "hot spots", areas at least 1 degree celsius above long term averages in the warmest months. The most extensive bleaching was in 1995, but temporal histories differed between sites. Bleaching and hot spots were less frequent during the three years of global cooling following the June 1991 Mount Pinatubo eruption. SST records of reef sites across the Caribbean showed increasing temperature trends at almost all sites. Caribbean reefs had higher rates of temperature rise than most South Atlantic, Pacific, and Indian Ocean reefs. A strong latitudinal pattern was found, with warming at most northern hemisphere sites and cooling at most southern hemisphere sites. Variations in warming between regions in 1989-1996 may relate to reduced sea surface exchange in nearly enclosed seas with restricted circulation like the Caribbean, Red Sea, and Persian Gulf, and increased upwelling at open ocean and southern sites. 

INTRODUCTION 

The onset of mass coral bleaching in coral reef regions around the world during the 1980s (Glynn,; Goenaga and Williams, ) was at first mysterious, and many possible causes were suggested, including excessive temperatures, pollution, sediment and freshwater discharge, diseases, and natural cycles (Williams & Bunkley Williams, ). This wide range of possible causes appeared plausible because bleaching is a general response of corals to any stress that is sublethal, and is also often an intermediate stage in lethal stress response (Glynn, ). Corals can be readily made to bleach in the laboratory in response to temperatures, salinities, or light levels that are too high or too low, and to excessive levels of sediment. Coral bleaching is also often seen in corals which are being killed by algae or invertebrate overgrowth, or in response to lethal disease, bacterial infection, and toxins. 

Detailed field observations revealed that the only common feature in the mass bleaching episodes was that they followed unusually prolonged periods of very hot sea surface temperatures. These were usually, but not always, concurrent with very low cloud cover, low wind, low water circulation, high water transparency, and high sunlight. A handful of bleaching events were associated with low or normal temperatures, high cloudiness, high rainfall, and exceptional flooding, associated with high levels of freshwater and sediment discharge onto coral reefs. Those bleaching events were confined to limited areas of obvious freshwater stress (Goreau, 1963), in contrast to the mass bleaching events which occurred over areas up to thousands of kilometers across, with a wide variety of local stresses or none at all. Freshwater linked bleaching events are classified as local bleaching events, because of their highly restricted distribution. 

The first convincing data linking bleaching to high temperatures were found during the 1983 bleaching in the eastern Pacific. Water temperatures reached more than 4 degrees above average when deep ocean upwelling failed during the most severe El Nino event reported, and most of the corals died in the Galapagos islands and in coastal reefs in Costa Rica, Panama, Colombia, and Ecuador (Glynn, ). Mass bleaching was confined to the core El Nino area in the eastern Pacific and limited parts of the Caribbean coastline of Panama affected by weather patterns originating in the Pacific (Lasker & Coffroth, ). In subsequent years mass bleaching occurred in coral reefs worldwide, both during periods of high, low, and normal El Nino conditions, and in parts of the world with both positive, negative, and zero correlations with the El Nino Southern Oscillation Index (ENSO). This suggests that high temperatures per se, not El Nino, was the primary cause of bleaching. 

Detailed comparison of in-situ and satellite sea surface temperature data from 1982 to 1991 across the Caribbean showed that bleaching at each site could be represented by a threshold temperature above which bleaching always took place, and below which it did not (Goreau et al, ). This threshold temperature was strongly correlated with mean annual temperature, and was highest in Jamaica and lowest in Bermuda. At each site the threshold temperature was found to correspond closely with a monthly average temperature near 1.0 degrees Celsius above the long term average in the warmest month (Goreau et al.,; Strong, ). Satellite derived sea surface temperatures were found to provide a highly reliable correlation with in-situ values only when the highest spatial resolution data ( 4 kilometres) were used (Strong,; Goreau et al, ). These data were found to lie within 0.2 degrees celsius of the measured values, but tended to systematically underestimate sea surface temperatures at the highest values, presumably because small clouds formed by increased evaporation contaminated the data, being too small to be identified in the pixels. In contrast the low-resolution data normally published by NOAA had poor and variable correlation with in-situ temperature measurements and with high resolution data, apparently because the processing algorithm effectively averages the data over around a 400 kilometre distance. 

Despite the poor performance of the low resolution data in the Caribbean, this global data proved to be adequate to identify areas of high sea surface temperature associated with bleaching events in the Pacific, Indian, and Atlantic Oceans (Goreau & Hayes, ). This appears to be because these areas are much larger than the Caribbean, and lack much of the short range spatial complexity of the temperature field caused by flow around the closely spaced Caribbean islands. With the exception of the Caribbean, the low resolution data was sufficient to identify areas of ocean 1.0  degrees celsius or above the average in the warmest month associated with the onset of every mass bleaching event worldwide reported between 1983 and 1991 (Goreau et al, ). These areas were called "Coral bleaching hot spots" (Goreau and Hayes; Strong,). 

Continued mapping of coral bleaching hot spots has allowed the authors to identify major mass bleaching episodes since 1990, often in advance of field observations. Fieldwork conducted in the South Pacific in 1994 tracked coral bleaching in the wake of a hot spot which swept across that region. Measurements or estimates of 15 different environmental variables were made at 19 different reef sites across the area affected. The results showed that many variables, all highly positively correlated with human population density, were negatively impacting coral cover, but only high temperatures were positively correlated with coral bleaching (Goreau & Hayes, 1994). This indicated that coral bleaching was equally likely in human impacted as well as pristine reefs, as long as water temperatures were sufficiently warm. 

The last global analysis of coral bleaching and hot spots (Goreau & Hayes, 1994) covered the period of 1983 to 1991. In 1991 an important natural experiment, the eruption of Mount Pinatubo in the Philippines on June 15 1991 caused a global lowering of temperatures for a three-year period. In this paper we extend the analysis to cover the period of 1989 to 1996 on a global scale, and present detailed values of sea surface temperature trends throughout the Caribbean and Tropical Americas, including Florida, the Bahamas, Bermuda, the eastern Pacific, and Brazil. This period provides a definitive test of links between global sea surface temperature trends and coral bleaching because of the remarkable cooling which happened following the Pinatubo eruption. 

METHODS 

NOAA sea surface temperature records are derived from the Advanced Very High Resolution Radiometer (AVHRR) instrument carried on satellites which provide global coverage several times per day. Data from the near infrared channel which is sensitive to thermal radiation is processed using an algorithm which contains a calibration of the measurement to in-situ oceanographic data, and which uses observations in the visible wavelengths to eliminate pixels with visible cloud contamination. The data used here has been averaged on a monthly time scale. Nearly 80 sites were examined worldwide. These sites cover all major coral reef regions in all oceans, and span the full range of conditions under which coral reefs are found. Data from each location for each month was plotted covering all of 1989 through 1996. These data were fitted using a linear least squares regression to obtain mean temperature trends during this period. These trends were plotted against latitude. Extreme hot and cold months were identified, and histograms made of their frequency in each year. These data were plotted against mean annual temperature. Finally, in-situ observations of coral bleaching were compiled wherever available from local divers at each site. 

RESULTS 

The results of the global analysis are presented graphically. Figure 1 shows the location of all sites by latitude and longitude. Figure 2 shows the maximum and minimum monthly values at each site plotted against the mean value for the interval. Figure 3 shows the extreme values at each site histogrammed by year. Figure 4 shows the trends for the period plotted as a frequency histogram. Figure 5 shows these trends plotted against latitude. 

Detailed temperature and bleaching patterns varied considerably at different sites across the Tropical Americas (Figures 6 through 33). These are discussed in detail for each site below in alphabetical order: 

1. Abrolhos. Brazil 

The highest temperatures were in 1995, followed by 1996. Bleaching is reported in both of these years, and some bleaching took place in 1994 according to Clovis Castro. Abrolhos is remote and only rarely visited and few long term observations are available. 

2. Anguilla 

Although mean temperatures have increased sharply over the period, mainly due to an increase in winter temperatures, the warmest period in the record took place in 1990. We have not managed to obtain information on past bleaching events in Anguilla, where diving activity is limited. 

3. Antigua 

The warmest period in Antigua took place in 1990. A bleaching event was noted in that period by John Birk and Ashton Williams. A very mild threshold bleaching was noted in 1996 by T. J. Goreau & M. Goreau, implying that normal maximum temperatures are very close to the threshold value for some corals. 

4. Arraial do Cabo. Brazil 

Bleaching has been reported here from time to time, but little information is available by year. The highest temperatures were in 1993, with 1989, 1990, 1994 and 1995 running behind it. The overall trend in temperatures at this site is negative. This site is located near to a localized upwelling zone, and the data suggests that the rate of upwelling may be increasing. 

5. Aruba

The two years which were exceptionally warm, 1990 and 1995, were years of severe mass coral bleaching. 

6. Barbuda 

The warmest year was in 1990. Barbuda was apparently affected by the same event which affected nearby Antigua in that year. 

7. Belize North 

The warmest year of all, 1995, was the first year in which large scale mass bleaching was reported in Belize. 

8. Belize South 

This record is very similar to that of North Belize with one exception, a large but brief warming event in 1994 which was the hottest period on the record. Large scale bleaching was not reported, but it was the following year when temperatures reached a slightly lower value, but remained warm for twice as long. 

            9.         Bermuda 

The warmest years, 1991, 1994, and 1996 were bleaching years in Bermuda. 

10. Bocas del Toro. Panama 

The warmest year, 1995, had the worst bleaching seen in Bocas. 

11. Bonaire 

The two hottest years, 1990 and 1995, were the two years with severe mass bleaching in onaire. 

12. Cartacena' Colombia 

The hottest year, 1995, had the worst bleaching reported in Cartagena and Santa Marta. 

13. Cozumel. Mexico 

The hottest years were in 1989 and 1995. The 1995 bleaching event was severe and well documented, the 1989 event was confirmed by local sport divers, and appears to be related to the event which affected Jamaica and Cayman in that year. 

14. Fernando de Noronha. Brazil 

-The warmest year was 1989. There is no information currently available on coral bleaching in Fernando de Noronha. 

15. Florida Keys. USA 

The warmest year was in 1993. Strong bleaching took place in that year, but some bleaching has also been reported in other, slightly cooler, years as well. 

16. Galapagos. Ecuador 

The warmest year in this period was in 1992. Severe bleaching took place in the 1983 El Nino and probably again in 1997, when temperatures were abnormally elevated by up to 3 or 4 degrees. Throughout the 1989-1996 period, a non El Nino interval, the overall temperature decreased. Galapagos is known to be strongly affected by the Peru upwelling area, the world's largest. This data suggests that deep ocean upwelling in the southeast Pacific has been generally increasing, except when interrupted by El Nino events. 

17. Grand Cayman 

The hottest years were in 1995, 1990, and 1989. Bleaching tool place in all three years, with 1995 being most severe. and 1989 the least, according to Phil Bush of the Cayman Natural Resources Unit. 

18. Gonave. Haiti 

There has been a strong warming trend for western Haiti, implying that water is remaining longer in the Bay of Gonave and the region between Haiti, Cuba, and Jamaica. The hottest year was in 1995. There are no bleaching observations available from Haiti 

19. Jamaica North 

The hottest years were in 1990, 1995, and 1989. Major mass bleaching events on the north coast of Jamaica happened in those years. 

20. Jamaica West 

The western end of Jamaica had fairly similar temperatures as the north coast with the hottest years in 1990 and 1995, but differed in two significant periods: temperatures in 1989 were somewhat cooler, and in 1996 somewhat warmer than the north coast. Interestingly while the worst bleaching in western Jamaica was in 1990 and 1995, there was less bleaching than the north coast in 1989, and more in 1996. Distinctive short mild bleaching events took place during the warmest periods in 1992, 1993, 1994, and 1996, implying that the threshold temperatures for mass bleaching lie between 29.3 and 29.6 degrees, with progressively increasing bleaching the higher temperatures rise. Figure X shows an estimate of the proportion of reef corals bleaching from 1987 through 1996. These estimates, which have an estimated precision of plus or minus 10%, are based on sites on the north coast of Jamaica from 1987-1991 and on the west end of the island from 1991 -1996. 

21. Nassau. Bahamas 

The hottest years in the period were 1990 followed by 1989. 1993 and 1995 were next. The strongest bleaching was reported in 1990. Mild bleaching was also reported from the Bahamas in 1993 and 1995. 

22. Las Perlas, Panama 

The hottest year was in 1993. Bleaching was reported in this area only in the severe 1983 El Nino event. This region is located in a small upwelling zone in the northeastern tropical Pacific, which is influenced by El Nino events. The record suggests strongly that the deep upwelling which took place in 1989 did not happen in subsequent years. 

23. Portobelo. Panama 

The hottest year, 1995, had the worst bleaching seen. However mild bleaching was observed in 1996, when it got as warm, but for a much briefer period. 

24. Saint Kitts 

The warmest year was in 1990. There are no bleaching reports available from .St.Kitts

25. San Blas. Panama 

The warmest year, 1995, had the most severe bleaching seen in San Blas. 

26. Sao Sebastiao. Brazil 

The hottest year, 1996, was a bleaching year. Bleaching was also seen in 1995 (Alvaro Migotto), a year which shared second hottest year along with 1993 and 1994. Bleaching was not reported in 1993 or 1994, and although they reached the same maximum as 1995, warm conditions remained much longer in 1995. 

27. Tobago 

The hottest year on record was 1990, and bleaching was seen in that year by local divers. 

28. Turks and Caicos 

The hottest year was in 1990. There were few reports of bleaching in the Turks and Caicos until 1997, a period not covered in this paper. 

Overall mean temperature trends for the Tropical Americas during the 1989 to 1996 period are tabulated below. 25 off the 28 sites have increasing temperature trends over this period. The 3 sites with negative trends are all located in peripheral regions which are known upweliing areas (Galapagos and Southeast Brazil). Upwelling in these Southern Ocean sites appears to be increasing. All other sites show distinct warming. The mean rate of warming of all 28 sites is: .0051191degrees C per month and the standard deviation is: .0062838 C/month. When the three cooling sites are separated, their mean cooling is -.0085575 C/month and standard deviation is .0104553 C/month. For the 25 warming sites the mean warming rate is .0067627 C/month with a standard deviation of .0030081 C/month. This mean rate, equivalent to 0.811 C/decade, greatly exceeds the mean global warming rate, and implies that there is either reduced upwelling or the rate of circulation of surface water in most coral reef areas is decreasing, resulting in increased residence time of surface waters, which would allow greater heating. 

TABLE 1.          LOCATION        MEAN TEMPERATURE TREND (0.001 C/month). 

            1.         Abrolhos 1.8285

            2.         Anguilla 9.2811

            3.         Antigua 7.5746

            4.         Arraial do Cabo -3.870

5.                   Aruba 4.1386

6.                               6.         Barbuda 7.7876

            7.         Belize North

            8.         Belize South

            9.         Bermuda 

10. Bocas del Toro

11. Bonaire

12. Cartagena

13. Cozumel

14. Fernando de Noronha

15. Florida Keys

16. Galapagos

17. Grand Cayman

18. Haiti

19. Jamaica North

20. Jamaica West

21. Nassau

22. Perlas

23. Portobelo

24. Sai nt Kitts

25. San Blas

26. Sao Sebastiao

27. Tobago

28. Turks and Caicos

6.4216 7.4986

3.8565

7.8581

7.6167

11.550

5.2604

1.716

3.3546

-20.559

5.084

12.839

4.757

6.279

2.679

11.635

8.474

6.9642

10.848

- 1.3036

7.3603

6.4094

DISCUSSION

Overall sea surface temperatures in coral reef ecosystems worldwide rose during this period. Were it not for the eruption of Mount Pinatubo this rate would have been even higher. There is considerable variation between sites, which reveals a dramatic and unexpected latitudinal trend. Almost all northern hemisphere sites warmed during the 1989-1996 period, but the majority of southern sites cooled. 

The eruption of Mount Pinatubo, and the three-year global cooling which followed is shown by the greatly elevated frequency of extreme cold events in 1991, and the decrease in extreme warm events. Those extreme warm events occurring in 1991 all took place in the southern hemisphere where the seasonal maximum had already taken place prior to the eruption in the middle of the year. Had it not taken place 1991 would likely have set a record for extreme warm events. Only in 1994 did the surface ocean re-warm sufficiently for large numbers of mass bleaching events to take place. The frequency of extreme events between 1994 and 1996 indicate that the record global warmth which occurred in the years immediately before the eruption of Mount Pinatubo returned after the temporary cooling ended after settlement of the high altitude stratospheric sulfur aerosols which had reflected sunlight back into space during 1991-1993. This "natural temperature experiment" provides definitive proof that high temperatures, and not pollution or natural events, are the driving force behind mass coral bleaching. 

The areas which showed the highest rate of temperature rise were the Caribbean, the Red Sea, and the Persian Gulf, which warmed considerably faster than the global average. These are all nearly enclosed seas with restricted circulation. It appears that in these areas the vertical stratification and surface flushing is changing in such a way that heat is being retained in the system faster than it can be mixed out or down, and that these areas have oceanographic circulation feedbacks which are amplifying the rate of global warming, thereby placing their reefs at greatly elevated risk of bleaching damage. 

In contrast, an unexpected cooling of many southern hemisphere sites can only be explained if the rate of deep ocean upwelling is increasing in this part of the world. The spatial patterns found here strongly supports the results of recent data showing global warming in most parts of the globe during the course of this century, but a cooling in the eastern Pacific, causing an increase in the east to west gradient across the Pacific, due to an apparent increase in upwelling rates (Cane et al., 1997), and supports theoretical ocean circulation models which suggest that global warming rates may be partially masked and delayed by increased overturning of the deep ocean (Cane et al., 1997). 

The similarity in global warming patterns over the eight years between 1989 and 1996 found here agrees with that found over the much longer 91 years from 1900 to 1991 by Cane et al. This suggests that there has been a persistent pattern in regional geographic fluctuations from the mean global pattern. If this continues certain regions of the globe, including the Caribbean, would be expected to be much more strongly affected by global warming than others. While the longer record shows a clear long term global warming, the 1989-1996 period is one without major el Ninos and in the middle of what appeared to be a plateau in the rise of global temperatures during the 1980s and 1990s. Intensification of deep ocean circulation will not change the total rate of global warming, but by redistributing the heat into the deep sea, it will delay the full warming at the surface. At the same time, increased upwelling will bring increased nutrients to the surface, accelerating the rate at which eutrophication of reefs takes place from natural causes. Recent work however suggests that if the rate of warming is sufficiently high, the deep circulation should eventually fail in the Atlantic, resulting in a potential large decrease in the rate at which heat and carbon dioxide are removed to the deep ocean, and thus accelerating warming at the surface (Stocker & Scmittner, 1997). 

It is likely that global warming trends will continue, although punctuated by rare and temporary volcanic cooling events, and that this trend will result in elevated frequency and intensity of hot spots, and increased severity of coral bleaching worldwide, in the coming years. Action is urgently needed by global climate change negotiators to recognize that coral reef ecosystems are uniquely threatened by global climate change because they are the most sensitive ecosystem of all to high temperature damage, threatening the biodiversity, tourism, fisheries, shore protection, and ability to adapt to rising sea level of over 100 countries. It is urgent that Negotiators to the Framework Convention on Climate Change take the steps needed to recognize that coral reef ecosystems are threatened by any further climate change, put in place monitoring of these ecosystems for climate-induced damage, and take action to halt any further climate change. 

REFERENCES 

Cane, M. A., A. C. Clement, A. Kaplan, Y. Kushnir, D. Pozdnyakov, R. Seager, S. Zebiak, & R. Murtugudde, 1997, Twentieth-century sea surface temperature trends, Science, 275: 957-960 

Glynn, P. El Nino and E. Pacific

Glynn, P. Bleaching reviews

Goenaga, C., & Williams,

Goreau, T. F., 1964, Freshwater bleaching

Goreau, T. J. Hayes, Clarke, Basta, Robertson

Goreau, T. J., & Hayes, R. L., Ambio hot spot paper

Goreau, T. J., & Hayes; R. L. Remote sensing calibration paper

Goreau & Hayes, Pacific study

Lasker, H., & M. A. Coffroth, San Blas 1983 

Stocker, T. F., & A. Schmittner, 1997, Influence of C02 emission rates on the stability of the thermohaline circulation, Nature, 388: 862-865 

Strong, A. C., Panama conference climatology paper

Strong AC Calibration of satellite and in-situ data

Williams and Bunkley Williams