Pacific Science (1976), Vol. 30, No.2, p. 159-1 66 Printed in Great Britain

Thermal Tolerance in Tropical versus Subtropical Pacific Reef Corals I

STEPHEN L. COLES,2 PAUL L. J OKIEL,3 AND CLAR K R. LEWIS4

ABST RACT : Upper lethal temperature toleranc es of reef corals in H awaii and at Enewetak, Marshall Islands, were determined in the field and under controlled laboratory conditions. Enewetak corals survived in situ temp eratures of nearl y 34° C, whereas 32° C was leth al to Hawaiian corals for similar short-term exposures. Laboratory determinations indicate that the upper thermal limits ofHawaiian corals are approximately 2° C less than congeners from the tropical Pacific. Differences in coral thermal tolerances correspond to differences in the ambient temperature patterns between geographic areas.

REEF CORALS are generally considered to be The purpose of the present investigation was stenothermic (Mayer 1914, Vaughan and Wells twofold: to measure the intensity and duration 1943, Wells 1957), with relatively fixed upper of maximum natural temperature elevations and lower lethal temperature limits (Mayer among living corals on tropical and subtropical 1918). Yet some corals have been reported to reefs, and to compare upper thermal limits of survive temperature extremes in nature well be­ tropical and subtropical corals under identical yond the limits established by classical experi­ experim ental conditions. ments (Gardiner 1903, Wood-Jones 1910, Yonge and Nicholls 1931, Orr and Moorhouse 1933, Motoda 1940, Kinsman 1964, MacIntyre METHODS and Pilkey 1969). Sufficient data exist (Mayer 1918; Edmondson 1928; Jokiel et al., in press) Studies were conducted at Enewetak (E ni­ to suggest geographi c differences in coral ther­ wetok)Atoll, Marshall Islands , and Kaneohe mal tolerance, although preliminary studies, Bay, Oahu, Hawaii. Seven stations were estab­ both classical (Mayer 1918) and recent (Jones lished at Enewetak among living corals in the and Randall 1973), have not confirmed this shallow waters off Igurin (Glenn) Island. Four po ssibility. Unfortunately, comparison of ex­ of these stations were located on the leeward isting data is difficult because of incomplete ocean reef flat, and the remaining three on the information concerning th e temp erature en­ lagoon side of the island. Continu ously re­ vironments of corals under natural conditions cording thermographs and maximum-mini­ and because of differences in experimental mum thermometers recorded temperature techniques applied by different researchers. among the corals at each station between 29 August and 3 September 1974. Mortality and conditio n of corals at each station were ob­ I Hawaii Institute of Marine Biology contribution no. 483. This study was partially fun ded by U.S. served and compared with temperature cond i­ Environmental Protection Agency grant R800906, tions that had occurred during the observation Atomic Energy Commission contract AT(26-1)-628 to period. the Mid-PacificMarine Laboratory, and by the Hawaiian Upper lethal temperatures of E newetak and Electric Company, Inc . Manuscrip t received 25 April 1975. Hawaiian corals were experimentally deter­ 2 Hawaiian Electric Company, Inc.,Environmental mined in 16-liter plastic aquaria flushed with Department, Post Office Box 2750, Honolulu, Hawaii continuous flows of temperature-regulated sea­ 96803. water. Temperatures were maintained with 3 University of Hawaii, Hawaii Institute of Marine quartz-glass resistance heaters regulated by Biology, Post OfficeBox 1346, Kaneohe, Hawaii 96744. 4 University of Hawaii, Hawaii Institute of Marine proportional controllers and by adjustments of Biology, Post Office Box 1346, Kaneohe , Hawaii 96744. the flow rates. Temperature in each aquaria 159 160 PACIFIC SOENCE, Volume 30, April 1976 was monitored continuously with a scanning (Wells 1954). Acropora (A. delicatula and A. thermistor tele-thermometer and recorder. Res­ palmerae) was dominant, and Pocillopora, as idence time ofwater within the containers was well as microatolls of Porites, were common. very short (4-8 minutes), dissolved oxygen Lcptastrea, Millepora, and Heliopora were also was maintained at near saturation by constant present. Pocillopora, Porites, and Millepora were aeration, and a natural daylight regime was found at the three lagoon stations, but Acropora used. Therefore, detrimental effect of factors was conspicuously absent. other than heat stress were minimized. The At the most shoreward reef-flat station, which same apparatus and procedures were used in represented the boundary of coral growth, experiments conducted by the same investiga­ water circulation was cut off during midday tors within a few weeks of each other at the low tides. The temperature here exceeded 34° C two locations, thereby reducing any chance of for 1-2 hours at low tide, killing many corals. differences between Enewetak and Hawaiian Inside ofthis station, where no corals occurred, results due to experimental procedure. temperatures above 36° C were measured. At Two thermal stress experiments were con­ the other ocean reef-flat stations, the tempera­ ducted at each location. In the first, corals were ture generally held at 31°-32° C for a several collected and allowed to acclimate overnight at hour period at low tide, with occasional short­ ambient temperature in the aquaria. Tempera­ term increases to as much as 34°C. Minimum tures were then raised at a rate of 2° Cfhr until water temperature reached 27° C during peri­ desired test temperatures were reached. The ods of low tide at night. Minimum and maxi­ specified temperatures were then held to the mum temperature extremes measured during end of the experiment. In the second experi­ this study were as much as 2° C greater than ment, corals were collected and acclimated as those measured previously by Wells (1951) in before. The temperature was then raised to 34° equivalent zones during the month of June at C within 10-20 minutes, held for 3 hours, and Arno Atoll, Marshall Islands, and variation then lowered again to ambient. At Enewetak, was also greater. On one occasion a midday this stress cycle was repeated six times, with storm decreased the temperature to a low of 6-hour ambient holding periods intervening 25.5° C which, after the storm, rose to 32° between cycles. In Hawaii, extensive damage within 1 hour. to the corals occurred on the first cycle, so The death of coral along the inner margin of only two 34° C cycles were imposed, separated the coral zone on the reef flat suggests that we by 14 hours at ambient. observed near-lethal natural conditions of temperature on the reef. Because most of the corals that died were not exposed to air at low tide, we attributed their death to prolonged RESULTS AND DISCUSSION exposure at 34°C. Also, much of the coral on Monthly mean seawater temperatures at the ocean reef that was subjected to brief Enewetak are 2°_5° C higher than Hawaii exposures to 34° Clost zooxanthellar pigment, throughout the year (Figure 1). This study was indicating severe thermal stress (Yonge and conducted in late summer, when ambient Nicholls 1931, Jokiel and Coles 1974). Bran­ temperatures at both locations were maximal. ches of Acropora that extended above the mini­ Tidal range at Enewetak was high during the mum low tide level were probably damaged period of field measurements, with low tide more by dessication (Mayer 1918, Edmondson occurring near midday. These factors, together 1928) than by high temperature. The air tem­ with calm, sunny weather, produced extreme perature during midday low tide was substanti­ temperature elevations on the shallow reef ally lower (28.5° C) than the water temperature flats. Enewetak ambient open-ocean water (32° C). temperature at the time of the survey was ap­ Temperature variations on the lagoon reefs, proximately 29.5° C. which did not uncover at low tide, were more The reef flat coral fauna and zonation were moderate, but longer periods of temperature very similar to those reported for Bikini Atoll elevation occurred. At one lagoon station a 30 1951-1964 E~WETAK 29

28

27 u -0 ui a=:: 26 :::> «~ a:: w 25 0- ~ W ~ 24 1957-1964 HAWAII (Kaneohe Bay) 23

22

J FMAMJ JA SON D MONTH

FIG UR E 1. Monthly mean surface wate r temperatu res for Kaneohe Bay, Oahu, and E newetak Atoll, Marshall Island s(U.S. Coast and Geodetic Survey 1965). Upper and lower limits represent monthly mean maxima and minima. 162 PACIFIC SCIENCE, Volume 30, April 1976 temperature of32° C held for nearly 6 hours on showed an even greater difference in the ability 31 August. On other days, temperatures ranged of corals from the two areas to withstand ther­ to 30°_31° C during midday maxima for up to mal stress. At Enewetak, all species tested sur­ 8 hours. No damage to corals occurred at the vived six cycles of 34° C exposure. Pocillopora lagoon stations. elegans and Acroporaformosa showed slight tis­ Similar in situ thermograph studies on a sue damage by the end ofthe experiment, while Hawaiian reef subjected to thermal enrichment Porites lutea and Acropora hyacinthes were un­ from a power generating station (Jokiel and damaged. By contrast, one cycle to 34° C in Coles 1974, Coles 1975) have shown that pro­ Hawaii killed P. meandrina and damaged Pocillo­ longed exposure to temperatures above 30° C pora damicornis, Porites lobata, and Montipora is sublethal to common Hawaiian corals and uerrucosa. A second cycle killed one to two speci­ that temperatures above 32° C are lethal. In mens of each of these species. Fungia scutaria contrast, such extremes were tolerated by simi­ was moderately affected, with one specimen lar Enewetak species. In Hawaii, the highest losing pigmentation. natural water temperature that we have mea­ These results indicate that in both subtropi­ sured over several summers among living cor­ cal and tropical environments large populations als on the shallow and protected Coconut of corals are exposed to temperatures precari­ Island reef flat of Kaneohe Bay was approxi­ ously close (within 1° to 2° C) to their upper mately 30° C during clear, calm summer per­ lethal limit during the summer months. High iods of midday spring low tides. Several hours temperature alone can account for the exclu­ of exposure to this temperature did not kill sion ofcorals from some shallow inshore areas. corals. On two occasions, Maragos (1972) ob­ Mean summer ambient water temperature at served temperatures of 32° C on the shallow Enewetak is approximately 2° C higher tha n Kaneohe Bay barrier reef that appeared to be it is in Hawaii (Figure 1), and a corresponding lethal to Hawaiian corals (Fungia scutaria and difference of about 2° C was observed between Pocillopora meandrina). Exposures to naturally the two locations for upper lethal temperature, occurring temperatures of 32° C did not harm upper sublethal temperature, and maximum corals at Enewetak. . reef flat temperature among living corals. At Because other deterimental factors (low both locations, increases of +2° C above dissolved oxygen, altered salinity, etc.) often annual maxima appear to produce sublethal co-occur with high natural temperatures in the effects, while an increase of +4 to +5° C is field, the insitu observations are not conclusive, lethal to most coral species. this fact necessitating use of controlled labora­ The primary purpose of this research was to tory experiments. reexamine Mayer's (1918) conclusion that sub­ Results (Table 1) verify that Enewetak corals tropical species of corals do not differ from can withstand substantially higher absolute tropical species in upper thermal tolerance. temperatures than can their Hawaiian con­ Mayer did not base his conclusion on data geners. A mean temperature of 32.4° C killed from the same species. Therefore, for purposes most Hawaiian species tested, with 31.3° C of this study it was important to use common, being clearly detrimental, producing substantial shallow-water species which occurred where loss of zooxanthellae and some tissue damage insitu temperature data were taken at each loca­ and coral mortality. The same temperatures at tion, even though different species were present Enewetak for similar exposure periods pro­ at the two locations. It is possible, however, for duced little or no damage. Corals at 31.6° C one to evaluate the species effect using our remained pigmented and were often observed data along with data taken from the classical to have expanded polyps. Slight damage was literature. Upper lethal limits for the widely noted at 32.7° C, suggesting that this tempera­ distributed coral Pocillopora are available from ture approaches a critical value. Mortality was a number of geographic localities. The taxon­ nearly complete at 35° C, although one Porites omy of this genus is confused, and it has been lutea survived this treatment. suggested that Pocillopora damicornis, danae, Results from the thermal shock experiment oerrucosa, meandrina, elegans, breuicornis, lobilifera, Thermal Tolerance in ReefCorals-i-Cor.es, ]OKIEL, AND LEWIS 163

TABLE 1 SURVIVAL OF CORAL SPECIMENS TO TEMPERATURE ELEVATIONS AT ENEWETAK AND HAWAII

ENEWETAK Temperature (0C)* 29.1 31.6 32.7 35.6 Exposure Time (hrs) 96 93 60 '; 10 Condition] N I D N I D N I D N I D Pocillopora elegans 3 3 2 1 3 Acropora ~yacinthes 2 1 2 1 2 1 3 Acroporaformosa 3 1 2 1 1 Porites baea 1 1 3 3 1 1 Fungia scutaria 1 1 1 1 1 Totals 10 2 0 10 1 0 10 4 0 0 1 9

HAWAII Temperature (0C)* 27.1 31.3 32.4 Exposure Time (hrs) 96 95 50 Condition] N I D N I D N I D Pocillopora meandrina 3 1 2 3 Pocillopora damicornis 3 1 2 1 2 Montipora uerrteosa 3 3 3 Porites lobata 3 3 3 Fungia scataria 3 3 2 1 Totals 15 0 0 4 9 2 0 3 12

NOTE : Numbers in body of table represent individual colonies . * Standard errors of mean temperatures are less than 0.1° C based on hourly samplings . t N, normal pigmentation and good condition; I, intermediate condition with loss of pig- mentation and/or tissue; D, death .

and others probably are part of a continuous vations. The 2° difference in temperatu re toler­ series that might represent a single species ance between Hawaiian and tropical corals in­ (Vaughan 1907: 100; 1918: 78; Crossland 1952: dicated by the present study is substantiated 109). Figure 2 shows all available data on the throughout the temperature range for these upper temperature tolerance of three species combined data. of Pocillopora from Hawaii and from tropical This analysis indicates that the natural areas in the Pacific Ocean. Survival time for temperature environment at a geographic loca­ both Hawaiian and tropical Pocillopora shows a tion is far more imp ortant than taxonomic highly significant (P < 0.01) decreasing ex­ distinctions based on minor structural differ­ ponential relationship with temperature. An­ ences in determining coral temperature toler­ a lysis of covariance indicates no significant ance. Although Pocillopora damicornis appears difference between the slopes of the two re­ to be slightly more tolerant of elevated tem­ gression lines (P < 0.50; F = 0.039, df = 1, peratures than are P. meandrina in Hawaii or 12) but a hig hly significant (P < 0.01; F = P. elegans in the tropics, it does follow the 58.97, df = 1, 13) difference between their ele- same temperature-survival time relationship.

I I HPS 30 1 0 4 ~--r--:------,r------'------.------r---:I , -, -,, \ '., Tro p ic a l Pocillopora

Y =1O-O.603 x + 22 .2 78

r =0.90

2 • ", 10 -, ", o (/J 0:: o => 0 " , J: 10 -, ' 'Q Z , -,, w -,, • , tJ. ~ • -,, ,, ... 1.0 , • -, ..J • • -,, ~ o ,, > Hawa i ian eQ~jJJQP' o r a • ,, -> ,, 0:: Y= 1O-O.6 26x +21.636 " -, => 0.1 " (/J r =0. 9 9

• 011-._--JL...- --JL...- --.-I ---I. ~_.a..;~_~•• 30 32 34 36 38

TEMPERATURE tC)

t, • t FIGURE 2. Semilogarithmic plot of temperature versus survival time, in hours, for Hawaiian and tropical Pacific Pocillopora. The data have been taken from six sources , as follows . 1. Edmondson 1928, Hawaii : solid circle, Poeil­ lopora meandrina; solid hexagram, P. cespitosa (sny. P. damicornis). 2. Mayer 1918, Great Barrier Reef, Australia: open triangle, P . bulbosa (sny. P . damicornis). 3. Mayor 1924, American Samoa: open diamond, P. damicornis. 4. Jones and Randall 1973, Guam: open circle, P. damicornis, 5. JokieI et aI., in press, Hawaii: solid diamond, P. damicornis, 6. Present study: Enewetak, open square, P. elegans ; Hawaii, solid triangle, P . damicornis; Hawaii, solid square, P. meandrina, The regression calculated for Hawaiian Pocillopora excludes the Edmondson (1928) data at 32° C. Edmondson's experiments were conducted inclosed containers in which accumulated toxic metabolites probably biased experi­ ments lasting 1 or more hours. Such artifacts were eliminated by use of an open system in the present and other recent studies (JokieI et al., in press). Thermal Tolerance in Reef Corals-COLES, JOKIEL, AND LEWIS 165

Tropical P. damicornis is clearly more thermally tuations on reef corals at Kahe Point, Oahu. tolerant than is its subtropical Hawaiian Pac. Sci. 29(1): 15-18. counterpart. CROSSLAND, C. 1952. Madreporaria, H ydro­ These results contradict the classical concept corallinae, Heliopora, and Tubipora. Great (Mayer 1918) that a fixed physiological bound­ Barrier Reef Expedition 6: 86-257. British ary determines coral upper lethal temperature Museum (Natural History). limits, and that corals from different geographic EDMONDSON, C. H. 1928. The ecology of an locations subjected to different temperature re­ Hawaiian coral reef. Bull. Bernice Bishop gimes have the same upper thermal limit. Mu s. 45. 64 pp. Studies on the effect of temperature on calcifi­ EKMAN, S. 1953. Zoogeography of the sea. cation (Clausen 1972) and carbon fixation Sidgwick & Jackson, London. 417 pp. (Coles 1973) in the same species of Enewetak GARDINER, J. S. 1903. The fauna and geo­ and Hawaiian corals have shown physiological graphy of the Maldive and Laccadive Archi­ differences in corals from the two regions. pelagoes . Vol. 1. At the University Press, These studies provide insight into possible Cambridge. 471 pp. mechanisms responsible for the observed differ­ JOKIEL, P. L., S. L. COLES, E. B. GUINTHER, ences in lethal limit. G. S. KEy, S. V. SMITH, and S. J . TOWNSLEY. It may be assumed that the predecessors of In press. Effects of thermal loading on the Hawaiian corals, being derived from the Hawaiian nearshore marine biota. U.S. En­ tropical Indo-Pacific fauna (Ekman 1953), vironmental Protection Agency, final report were originally resistant to high temperature of project no. 18050 DDN. stress. However, water temperatures in Hawaii JOKIEL, P. L., and S. L. COLES. 1974. Effects of seldom naturally exceed 300, but do undergo heated effluent on hermatypic corals at larger annual fluctuations at a lower tempera­ Kahe Point, Oahu. Pac. Sci. 28(1) : 1-18. ture range than in the tropics. The process JONES, R. S., and R. H .RANDALL. 1973. A which has enabled establishment of reef corals study of biological impact caused by natural in Hawaiian waters has apparently reduced the and man-induced changes on a tropical reef. capability of many species to withstand tem­ Univ. Guam Mar. Lab., Tech Rep . 7. 184 pp. peratures above 30° C. It remains to be demon­ KINSMAN, D. J. J. 1964. Reef coral tolerance of strated whether the observed differences in high temperatures and salinities . Nature 202 : thermal tolerance at the two locations result 1280- 1282. from selective processes acting on many gen­ MAcINTYRE, 1. G., and O. H. PILKEY. 1969. erations, or whether temperature resistance in Tropical reef corals : tolerance of low tem­ corals can be changed by physiologica l accli­ peratures on the North Carolina continental matization to gradual increases in temperature shelf. Science 166: 374-375. over long time periods. MARAGOS, J . E. 1972. A study of the ecology of Hawaiian reef corals. Ph. D. Thesis. Univer­ sity of Hawaii, Honolulu. 290 pp . MAYER, A. G . 1914. The effects of temperature LITERATURE CITED on tropical marine animals . Carnegie Inst, CLAUSEN, C. 1972. Factors affecting calcifica-. Washing~on Publ. 183: 3-24. tion processes in the hermatypic corals --. 1918. Ecology of the Murray Island Pocillopora damicornis and Porites compressti. coral re~f. Carnegie Inst. Washington Publ, Ph.D. Thesis. Loma Linda University, Lorna 213 : 3-48. Linda, California. 95 pp. MAYOR, A! G. 1924.Structure and ecology of COLES, S. L. 1973. Some effects of temperature Samoan reefs. Carnegie Inst, Washington and related physical factors on Hawaiian Pub!. 340: 1-25. reef corals. Ph. D. Thesis. University of MOTODA., S. 1940. The environment and the Hawaii, Honolulu. 133 pp . life of massive coral; Goniastrea aspera Verrill ---. 1975. A comparison of effects of ele­ inhabiting the reef flat in Palao. Palao Trop. vate d temperature versus temperature flue- Biol,Sta: Stud. 2: 61-104.

II-2 ,166 PACI FIC SCIENCE, Volume 30, April 1976

O RR, A. P., and F. W. MOORHOUSE. 1933. WELLS, ]. W. 1951. The coral reefs of Arno Variatio ns in physical and chemical condi­ Atoll. Atoll Res. Bull. 9: 1-13. tions on and near Low Isles Reef. Sci. Rep. ---. 1954. Recent corals of the Marshall Great Barrier Reef Exped. 2(4): 87- 98. Islands. U.S. Geol. Surv., Prof. Pap . 260- 1: U.S. COAST ANDGEODETIC SURVEY. 1965. 385-486. Publ. no. 31-3 (revised). U.S. Government --. 1957. Coral reefs. Pages 609-632 in Printing Office, Washington, D.C. J. W.Hedgepeth, ed. Treatise on marine VAUGHAN, T. W. 1907. Recent Madreporaria ecology and paleoecology. Vol. 1. Geo l. of the Hawaiian Islands and Laysan. U.S. Soc. Am., Mem. 67. 1296 pp. Nat. Mus., Bull. 59. ix+427 pp. WOOD-JONES, F. 1910. Corals and atolls. --.-. 1918. Some shoal-water corals from Lovell Reeve & Co., London. 392 pp . Murray Island, Cocos-Keeling Islands and YONGE, C. M., and A. G. NICHOLLS. 1931. Fanning Island. Pages 51-234 in Carnegie Studies on the physiology of corals. IV. The Inst, Washington Publ. 213. structure, distribution and physiology of VAUGHAN, T. W., and J.W. WELLS. 1943. Re­ zooxanthellae. Sci. Rep. Great Barrier Reef vision of the suborders, families and geI1;era Exped . 1(6) : 135-1 76. of the Scleractinia. Geol. Soc. Am., Spec. Pap. 44. 363 pp.