The Impact of Cyclone Isaac on the Coast of Tonga 1

COLIN D. WOODROFFE2

ABSTRACT: Cyclone Isaac passed through the limestone and sand islands of the Ha 'apai and Tongatapu groups ofthe Kingdom ofTonga on 3 March 1982 and was probably the most severe storm experienced in southern Tonga in more than 100 years, causing extensive damage to buildings and crops. The limestone islands most affected by the storm are composed principally ofreefal limestones, and the coastal terraces are late Pleistocene in age, many having been dated to the last Interglacial. There is little evidence of past storms preserved on the coast, and storm blocks and extensive rubble deposits on reef flats are rare or absent. The worst hit islands were in the Ha'apai group. Here the raised reefal limestone cliffsofthe eastern shores ofislands on the barrier reef resisted the storm, except for blocks ofup to 2 m in diameter detached from the upper visor ofa wave-cut notch on Ha'ano. Coastal scrub, however, was stripped, .coconut palms were felled and Pandanus was broken for more than 50 m on Ha'ano, 30 m on Lifuka , and locally on Foa and 'Uiha. Some regeneration of the dominant shrubs , M esserschmidia argentea, Hibiscus ti/iaceus, and Bikkia tetrandra, was observed at the time of survey, 11-15 weeks after the storm. Both deposition and erosion were noted on pocket beaches between cliffs of reefal limestone , with at least 28 rrr' per m ofsand removed from the beach face on one beach on Lifuka. The island ofTatafa was severely devastated; the cyclone had passed directly over the southern end, causing the sandy shoreline to retreat by 200 m, scouring a channel through the island, leaving scour holes on the western side, and destro ying much of the coastal scrub. Tongatapu also sustained much damage to buildings and crops, and much of the north coast was flooded. The greatest change was observed to have occurred on sand cays off the northern coast. Four of these underwent change since a survey in 1969, and much of the change could have been attributed to Cyclone Isaac. The greatest change occurred on the smallest island, Manima, with coconut stumps left on the reef flat to the east indicating at least 12 m of shoreline retreat. On Oneata and Pangaimotu deep-rooted individ uals of Excoecaria agallocha were left exposed more than 8 m on the eastern shore by destruction ofthe coastal scrub . On Makaha'a retreat of the shoreline was localized, and along much of the coast a sand cliff occurred which had changed little since it was mapped in 1969. Undercutting of the shore and formation ofa sand cliffappear to be related to the presence ofbeachrock at the foot of the beach, and the storm did not cause much shoreline retreat. On the northern sand cays off Tongatapu patterns of change were variable. On Fafa recession occurred on the eastern and northeastern shore , but also occurred to the south , reflecting the configuration of the reef flat. Monuafe and Tufaka lost some scrub but the lack ofwoody vegetation meant that little evidence offormer island shape remained. Deposition occurred on Malinoa, the sand cay nearest to the path of the storm, which is situated on a narrow reef flat and is open

I This research was funded by the Auckland University Private Bag, Auckland, New Zealand. Present address: Research Fund. Manu script accepted 4 March 1983. Australian National University, Nor th Australia Research 2 University of Auckland ,Department of Geograph y, Unit, P.O. Box 41321, Casuarina, N.T. 5792, Australia. 181

,.-,,-.;.JU .. "",-_~__...... _~_ :lii J5 182 PACIFIC SCIENCE, Volume 37, July 1983

to a long fetch and deep water to the east. Few of the topographic changes brought about by Cyclone Isaac will remain for long before they are obscured, except the reshaping ofTatafa. Similarly, rapid regrowth of vegetation implies that coastal vegetation will soon be restored, but changes in composition may persist. Though Cyclone Isaac was a severe storm, a type which occurs in frequently, its effects on the coastal landforms and vegetation ofTonga appear to have been less than those ofsimilar storms on more isolated atolls elsewhere in the Pacific.

THE MAJORITY OF THE CORAL reefs and islands In contrast to the studies on isolated atolls ofthe world lie in areas which are susceptible or on the Great Barrier Reef, relatively little to tropical cyclones. Although the frequency is documented about the effects of intense of catastrophic storms varies, those of great storms on raised limestone islands and sand est magnitude can be important in shaping cays of island groups in the southwest Pacific. reef morphology, reef island topography, and There are reports of the impact of tropical coastal vegetation communities (Stoddart cyclones in the Lau Islands, Fiji (Phipps and 1971). In the last two and a half decades Preobrazhensky 1975, McLean 1977), and of knowledge of the effects of storms on such sedimentological effects of a cyclone in New environments has increased considerably. De Caledonia (Baltzer 1972). This study describes tailed studies have been undertaken, not only the impact of Cyclone Isaac on the coast of on the immediate impact of storms (Blumen both limestone and sand islands in the King stock 1958, 1961; McKee 1959; Craighead dom of Tonga. Cyclone Isaac, which hit the and Gilbert 1962; Stoddart 1962, 1963; Ball, Ha'apai group and passed close to the Tonga Shinn, and Stockman 1967; Maragos, Baines, tapu group on 3 March 1982, was probably and Beveridge 1973; Baines, Beveridge, and the most severe storm experienced in southern Maragos 1974), but also on the effects ofsub Tonga this century. sequent storms on the same coast (Sauer 1962, McIntire and Walker 1964, Perkins and Enos 1968) and on landform adjustment and vege GEOLOGY AND GEOLOGICAL HISTORY tation recovery (Blumenstock, Fosberg, and OF THE AREA Johnson 1961; Stoddart 1965, 1974; Baines and McLean 1976). The Kingdom of Tonga consists of more Studies of the effects of Typhoon Ophelia than 150 islands with a total surface area ofa (1958) on Jaluit Atoll in the Marshall Islands little over 700 km' . Most ofthe islands lie in a and Tropical Cyclone Bebe (1972) on Funa discontinuous chain, from Tafahi in the north futi Atoll in Tuvalu have shown the immense (16°10' S, 173°40' W) NNE-SSW to Ata in erosional and depositional impact that storms the south (22°08' S, 176°08' W). To the west of hurricane force can have on isolated reefs along the submerged Tofua Ridge there are a in the Pacific. On both of these atolls rubble series ofvolcanic islands , several ofwhich are ramparts were deposited on the eastern reef still active, composed predominantly ofbasal flats of the atoll, and these deposits subse tic andesite, andesite, and dacite rocks (Bauer quently migrated landward, adding a new 1970, Bryan, Stice, and Ewart 1972, Ewart, lithological unit to the reef islands (Blumen Bryan , and Gill 1973). To the east the Vava 'u, stock, Fosberg, and Johnson 1961, McLean Ha'apai, and Tongatapu groups ofislands are 1974, Baines and McLean 1976). Cyclones composed mainly ofreefal limestones ofPlio have also been observed to have had impor cene to Recent age, but with exposures of tant effects on the morphology of reef islands older rocks on some islands (Figure 1). These on the Great Barrier Reef (Moorhouse 1936, islands represent at least five distinct fault Gleghorn 1947, Stephenson, Endean, and blocks-the Vava'u, northern Ha'apai, south Bennett 1958, Hopley 1974, Flood and Jell ern Ha 'apai, Nomuka, and Tongatapu blocks 1977). -on the Tongan frontal arc which extends Impact ofCyclone Isaac on Coast ofTonga-c-Woonnosss 183

HA'A PAI GROUP J'\. VAVA'U '.....'1fJ GROUP Late Is. 0

HA'APAI GROUP Kao o Ha'ano \ d Foa () Lofanqa, LlUfuka Tofua Uol eva f!- Tatafa O'Uiha 20' .'

Nomuka to () l~· ··'" /f~'~ u Malinoa1:J ,, :;- D \ :.... ) )~\, () F ; ; t ~~ ~t) l; ~1~: :~ ~__';/:':~'~u~~~ ~ 'Euaik::q ~ta p u

• d 'Eua

FIGURE 1. The Kingdom of Tonga, showing the path of Cyclone Isaac and islands in the Ha'apai and Tongatapu groups (insets).

SSW to join the Kermadec arc (Taylor 1978). Tongatapu, and the sand cays offthe northern The frontal arc is separated from the Tofua coast of Tongatapu, and it was these islands Ridge to the west by the Tofua Trough, and that were visited in this study. The Ha'apai is bounded to the east by the Tonga Trench group consists of a series oflimestone islands, with water depths increasing rapidly to more the largest, Foa, being 7.3 km long and 2.3 km than 10,500m. The fault blocks have evidently wide, lying along a barrier reef on the margin had independent tectonic histories continu of the Tonga Trench. To the west and south ing at least into the Pleistocene (Katz 1976); there are numerous smaller islands, the largest, Eua has undergone pronounced uplift, while Lofanga , being 1.8km long. The southern Vava'u, with a deeply embayed coastline, Tongatapu group is dominated by two larger appears to be subsiding. islands, Tongatapu itself, 35 km long and Cyclone Isaac had its more devastating 25km wide, and Eua, 20 km long and 7 km effect on the islands of the Ha 'apai group, wide. 184 PACIFIC SCIENCE, Volume 37, July 1983

Coastal Geomorphology ofthe Ha 'apai and The larger islands in the Ha'apai group lie Tongatapu Groups on the barrier reef (Figure I) and are generally limestone-cored. Lifuka is typical ofthe islands The coastal geomorphology of the lime from Ha'ano to 'Uiha, On this island several stone and sand islands ofTonga is a function profiles were surveyed to illustrate both coastal of the geology and lithology of the islands, morphology and the effects of the storm tectonic movements of the fault block on (Figure 2). Smaller islands such as Uoleva and which the islands lie, oscillations of sea level Tatafa are composed of sand. To the east of during the Quaternary, and the supply ofsedi the barrier reef water depths increase rapidly ment and energy conditions, including storms, into the Tonga Trench; the nearly horizontal to which the coast is subjected. The geology of reef flat, 50m-I km wide, has a pronounced Eua is best known as this island, reaching a algal rim, often prominently terraced and in height of312 m, has undergone the most rapid several instances rising 50 em or more above uplift. The stratigraphy consists of Eocene high-tide level(profile la, Figure 2); elsewhere tuffs, gabbros, and basic to intermediate lavas isolated outcrops of raised reefal limestone into which dykes of basalt have intruded; occur on the reef flat (profile 2a, Figure 2). To Eocene foraminifera-bearing fine calcareous the west of the islands the reef flat is less tuffs; Eocene massive foraminifera-bearing continuous, with patches of reef more than limestones; and Eocene highly-fossiliferous I km from the islands. limestones, overlain unconformally by Mio The limestone-cored islands of the barrier cene volcanoclastic sediments, above which reef rise with a gradual gradient from the are Pliocene and Quaternary reefal limestones western shore, much ofwhich is a sandy beach, (Hoffmeister 1932, Stearns 1971, Tongilava to a maximum height ofmore than 15m toward and Kroenke 1975,Taylor 1978). Evidence of the eastern shore . The raised reefal limestones former still-stands of sea level is preserved in form 11 cliff on the eastern shore, with a num the form of terraces identified by Hoffmeister ber of pocket beaches between headlands. The (1932) at heights ofapproximately 30, 60, 105, cliff rises steeply 3 m from the reef flat and 120, 170, and 230 m, several being tilted to then less steeply to 6 m high (ie, profiles 1a, 1b, the northwest. Taylor and Bloom (1977) and 2a, 2b, 6, and 8, Figure 2). Within this lime Taylor (1978) identified a further terrace at stone numerous corals can be seen in their approximately 7 m, in which corals up to position of growth. The coastal terrace evi 5.5 m above high-tide level were dated using dently represents an emergent fossil coral reef, U-series radiometric dating to 133,000 ± equivalent to the makatea of other Pacific 12,000 years B.P. , and a lower emerged coral islands. Taylor and Bloom (1977) and Taylor reef, first described by Ladd and Hoffmeister (1978) reported U-series ages of 130,000 ± (1927), l.Om above high-tide level dated at 12,000 and 158,000 ± 14,000 years B.P. on 6120 ±110 years B.P. by 14C and 5700 ± 500 corals up to 4.2 m above high-tide level from years B.P. by U-series. Foa and imply that a raised reef of last Inter The other limestone islands of the Ha'apai glacial age surrounds older limestone cores on and Tongatapu groups do not reach as great these islands. There are a number oferosional an elevation as Eua and have not been uplifted sea-level features on these cliffs between high as rapidly; and the most extensive rocks are tide leveland 3 m above high-tide level(Figure the reefal limestones of Pliocene to Recent 2). These often combine to form a complex age. Although the underlying stratigraphy is composite notch up to 3 m high along the probably the same as on Eua , the Miocene eastern coast, though Taylor found an absence volcano clastics are exposed on only a few ofdatable material in similar features on Eua other islands in the Nomuka-Mango group and Tongatapu and was unable to relate (Lister 1891); they do not appear on the islands erosional features to constructional sea-level of the Ha'apai barrier reef and are not exposed features (Taylor 1978). on Tongatapu, where they have been encoun Taylor (1978)did not find a raised Holocene tered at 137m in bores and are more than reef in his brief survey of Ha 'apai. However, 1500m thick (Katz 1976). there are corals in position of growth and Impact of Cyclone Isaac on Coast of Tonga- WOODROFFE 185

~ge ta\ lo~ I ~pe d 1a 1b ~mm=/~ - - ~------" V77777777777/77J7>. Vegetation stripped

~~~~~ U~~ e st on e 2a 2b ---In;;;7;;;;;;;;;;;;-;07A--

%v;;j~~;;~;~~7~;~ ~,:: ~,~ :--, ~;;: ~,-:' -:;;,-;:' -::;;-;:';-;:,;~ Coastal scr ~ b 5 v~~7;;~~;;;;;;;;;;;;;7;;;~ --

6 . -~---- - W~~~~0~~~~~~~~~ ------Ve g e t a t i o n damaged .:';. -Sand sca rp ..;, 7 ~.. .. : ; . : ; :: .:.>; -;;:;;;;/-;;/T~~;;;~;;;;;;/~-;;;-;;;;;;w;;;;;;;;;r;w~ -

~ 8 . ---- ~~;;;;~;:: ~;;;~~~~;;;;;;;;/T~- -

9 -- . C~M/Wffm~~M~~~~~~m~~~;- 9 b

o me tres 100 10 I I

Raised reef limestone ES3 Sand with corals r

Reef fl at and Beach rock @ alga l rim L1FUKA o km 1 ~

FIGURE 2. Profiles of coastal morphology and storm damage, Lifuka (with respect to high-tide level). 186 PACIFIC SCIENCE, Volume 37, July 1983

TABLE I RADIOCARBON AGES OFRAISED CORALS FROM HA'APAI AND TONGATAPU

SAMPLE N.Z. 14CNO. LOCATION GRID REFERENCE MATERIAL 14CAGE'

TongajfWI 6133 Foa 853174 Raised in situ coral 3200 ± 40 TongajfW2 6134 Foa 841158 Raised microatoll 5810 ± 60 TongajfW4 6136 Lifuka 789122 Raised in situ coral 5600 ± 60 TongajfW5 6137 Ton gatapu 842635 Raised microatoll 4490 ± 60

'Radioc arbon ages determined at the Institute of Nuclear Sciences, New Zealand Radio carbon Dating Laboratory. Ages quoted on 4 the basis of 1 C half-life of 5568 ± 30 years, and reservoir corrected with respect to Fiji Shell Standard . above the present limit of coral growth at 200 m wide. In several places this reef flat has several places around Lifuka: at high-tide a well-developed algal rim, prominently ter level on profile 10,40 em below high-tide level raced, while elsewhere the reef flat is replaced on profile 9, and 1m below high-tide level on with a ledge or narrow platform at the base of profile4 (Figure 2), and there are raised corals the raised reefal limestone cliffs (Ladd and and degraded microatolls on the north and Hoffmeister 1927). west coasts of Foa. A 14C age of 5600 ± 60 Lister (1891) described large coral heads in years B.P. was determined for a coral from the the streets of Nuku 'alofa, and these occur up raised reef on profile 9, and ages of 5810 ± 60 to 1.0m above high-tide level. Taylor (1978) and 3200 ± 40 years B.P . were determined for has reported U'-series ages of 6200 ± 300, a microatoll on the west coast and for a coral 7600 ± 800, 6200 ± 300, 5900 ± 900, and a from a raised reef on the north coast of Foa 14C age of 6360 ± 110 years B.P. on these (Table 1). These are evidently remnants of a corals, showing them to be part of a Holocene Holocene raised reef and are of similar age to raised reef. In addition, there are numerous the raised reef on Eua, with the coral at the degraded emergent microato lls which occur north of Foa appearing rather younger. on the reef flat to the west of Nuku'alofa and Tongatapu is the largest island in Tonga can be seen amongst the mangroves on the and is roughly triangular. It reaches a maxi reef flat at least as far as Fatai and to the nort h mum elevation of 65 m to the southeast, where ofNukunukumotu. These occur up to at least there is a palaeo-island core of Pliocene 53 em above the present limit to coral growth Pleistocene limestones. Much of the island is on this reef flat; a 14C age of 4490 ± 60 years composed of a terrace at 20-35 m, which B.P. was determined on one north of Sopu represents a series of reefs fringing the palaeo (Table 1) implying that they are of an age island core, and adjacent patch reefs: Else similar to or slightly younger than that of the where there is a lower terrace at about 7 m, features described and dated by Taylor from within which Taylor and Bloom (1977) and Nuku'alofa. Taylor (1978)describe two barely distinguish The island of Tongatapu decreases in ele able coral reefformations; the older but lower vation toward the northern coast, which is reef, with corals reaching to 3.0 m above generally low lying and which is fringed with high-tide leveldated to 240,000 ± 40,000 and mangroves. The reef flat is most extensive on 205,000 ± 20,000years B.P. by U-series dating; the northern coast, reaching more than 7 km the younger, with corals reaching to 5Am to the north; on this reef flat extension and on above high-tide level, dated to 135,000 ± separate reef flats to the nort heast there are 15,000 years B.P . These late Pleistocene reefs, several small islands. The larger ones, such as though with older limestones exposed at some Atata, are cored by reefal limestone, while the locations , form cliffs around much of the smaller islands are sand cays. Stoddart (1975) southwest, south, and east coasts of Tonga has described and mapped four of the sand tapu , often with basal erosional features and cays and noted deposits of beachrock and fringed with a generally narrow reef flat up to undercutting along parts of their shore. Impact of Cyclone Isaac on Coast ofTonga-WooDROFFE 187

Taylor and Bloom (1977)and Taylor (1978) east and may affect part of the Tonga chain; propose, on the basis of the similar age and others, such as Cyclone Isaac, travel to the elevation of late Pleistocene raised reefs of southwest and can be felt along the entire Eua, Tongatapu, and Ha'apai, that there has chain. Visher (1925) recorded 58 storms been no net differential movement of the fault affecting Tonga in the 100 years before 1924, blocks on which these islands lie since the last of which 16were severe. That is an average of Interglacial. Their interpretation is compli more than one in two years, but in several cated by the presence of an emergent Holocene years more than one storm was recorded and reef on Eua and Tongatapu and its apparent in 1848 two severe storms occurred during absence from Ha'apai. Radiocarbon ages two weeks in April. Kerr (1976) noted 19 reported in Table 1 indicate that there are additional storms in southern Tonga up to remnants ofan emergent Holocene reef on the 1969, and there were storms in 1970 and reef flat of islands inthe Ha'apai group and 1973 and two in 1977(Revell 1981). Those in that microatolls on the northern coast of 1912, 1937, 1961, and 1963 were particularly Tongatapu are part ofa raised Holocene reef devastating. which is more extensive than previously re ported on that island. These corals are pres ently intertidal, and none reach the level of 1.0m above high-tide level recorded for the CYCLONE ISAAC Holocene reef on Eua or beneath Nuku'alofa, Cyclone Isaac was undoubtedly one of the Nevertheless, they reinforce the suggestion worst storms to hit Tonga this century. It that all three areas have undergone similar claimed six lives and caused enormous devas histories since the late Pleistocene. tation to buildings and crops; the damage Little direct evidence of past storms is pre to buildings on Tongatapu alone exceeded served on the coast of Tonga . There are rela Tonga $10 million, and to crops more than tively few storm blocks on the reef flats, and Tonga $8 million (Tonga Chronicle, 12March rubble ramparts or partially consolidated 1982). The cyclone developed about 160km rubble fields are generally absent. northeast of Western Samoa. It traveled south west at an average speed of 12 knots, hitting the Ha'apaiand Tongatapu groups on 3 March 1982. The systemdeepened as it moved toward CLIMATE AND PREVIOUS TROPICAL STORMS Tonga; it was a small diameter storm in which Tonga lies in the belt of the southeast hurricane-force winds were probably confined trades, and prevailing winds come from the within a 50-km-wide zone. It was felt in east and southeast and less commonly from Vava'u, though relatively little damage was the south and northeast. Mean annual rainfall done there. It passed through the Ha 'apai is 1801 mm in Ha'apai (range: 1093-2664mm, group of islands where it is believed that sus period:1947-1980), and 1878mm in Nuku'a tained winds of 80 knots with gusts of 120 lofa,Tongatapu (range:1254-2655mm, period: knots were experienced, though records were 1947-1980), the wettest months being January lost after a wind speed of 57 knots (N.Z. to March and the driest June. Mean daily Meteorological Service). Cyclone Isaac then temperature is 27.5°C and mean daily mini passed to "the northwest of Tongatapu, and mum is 21.9°C for Ha'apai; comparable fig in Nuku'alofa a peak gust of 92 knots was ures for Nuku'alofa are 26.9°C and 20.5°C, measured. Wind direction appears to have respectively. changed during the passage of the storm and Lying in a chain across 6° of latitude, the trees fell in all directions. The barometric Tonga islands are susceptible to tropical pressure dropped to 976.4mb at 1:45 pm cyclones formed to the north. The hurricane (local time). More than 120mm of rainfall fell season extends from November to April, with on Tongatapu on 3 March. most storms recorded in January and March. Tides in Tonga are semidiurnal; the tidal Many of these cyclones travel south or south- range is about 1.22m at springs and 0.98m at 188 PACIFIC SCIENCE, Volume 37, July 1983 neaps. Cyclone Isaac coincided with a high RAISED REEFAL LIMESTONE COASTS: The lime spring tide of 1.2 m at Nuku'alofa, predicted stone cliffsof the eastern shores, rising 2-3 m for 1:27 pm on 3 March. No tide gauge was in from the present reef flat, offered substantial operation at the time of the storm and no resistance to the high seas of the storm and record of the height of the storm surge is have undergone little morphological change. available, but the extensiveflooding of coastal An exception is along the northeastern coast areas on the north coast, and the fact that ofHa'ano where several fallen blocks of reefal almost the whole of Nuku'alofa was under limestone, up to 2 m in diameter and with water, testify that the water level was several white, recently exposed fractured surfaces, meters above the high-tide level. were observed detached from the upper visor of the wave-cut notch and dropped onto the reef flat, in some cases shifted a few meters to METHODS the east. The observations described in this paper The coastal scrub of the raised reefal lime were made during a visit to Tonga from 24 stones was devastated. The prestorm vegeta May to 21 June 1982, 11-15 weeks after the tion of the cliff tops evidently comprised a passage of Cyclone Isaac. Particular attention scrub dominated by Bikkia tetrandra and was paid to the impact of the storm on coastal Hibiscus tiliaceus, in which Messerschmidia landforms and coastal vegetation. The smaller argentea, Clerodendrum inerme, Excoecaria sand islands were mapped using a rapid pace agal/ocha, Terminalia litoralis, and Desmo and compass traverse in the same manner as dium umbel/atum wereimportant, with dwarfed the 1969 survey offour of the islands by David Pemphis acidula to seaward, and much Pan Stoddart (Stoddart 1975). These surveys are danus sp. forming a zone about 10m from the therefore comparable and are both drawn seaward edge. To landward of the Pandanus with respect to magnetic north, though the zone Cocos nucifera became increasingly absence of relocatable reference points and prominent. imprecisions of closure and scale determina In the worst-hit areas the vegetation was tion inherent in the technique prohibit direct severely damaged by the wind, and soil was overlaying of maps. Beach profiles were sur completely removed, with local deposition of veyed with a clinometer, and a series of more coral gravel suggesting flooding of the cliffed detailed profiles using a Sokkisha automatic coastline up to 6 m above high-tide level. The levelwere made and related to high-tide level. damage was worst on the northeastern coast of Ha 'ano, from which the vegetation was stripped for more than 50m inland, and in places up to about 80 m. Woody plants gen RESULTS erally lost all foliage, but the root systems The Ha'apai Group remained interwoven amongst the rock. Along this coast of Ha'ano the low Pemphis, less Cyclone Isaac passed through the Ha'apai than 30 em above the rock surface, remained , group of islands and caused widespread dam having been little damaged. Most evident in age to buildings, crops, and coconut wood the area are the broken stumps of Pandanus , land, felling about 15% of the coconut palms. almost all snapped off either just above the Extensive damage was obvious on much of prop root system or just below the first branch. the coast along the barrier reef. The islands Few retained any branches and even fewer along this barrier reef were studied in detail had any foliage left. Several of the shrubs also with visits to Ha 'ano, Foa, Lifuka, Uoleva, remained rooted on the rock, though their Tatafa, and 'Uiha, and with a further visit to canopy and most branches were gone. Many Lofanga (Figure 1). The storm also affected had commenced regrowth from the stump, the isolated outlying islands in the western including Messerschmidia, Terminalia , Hibis and southern Ha'apai group, but these were cus, Bikkia, Desmodium, and Excoecaria. Less not visited. frequent , but in some cases having survived Impact of Cyclone Isaac on Coast of Tonga- WOODROFFE 189

FIGURE3. Coastal scru b stri pped from the raised reefal limestone coast of northeastern Ha 'a no . the storm, were Scaevola taccada, Euphorbia On Lifuka Bikkia, Canavalia, Leucaena insu ramosissima, Leucaena insularum , Xy losma larum, Wedelia biflora, Euphorbia cyathophora, simulans, Acacia simplicifolia, and Canavalia Terminalia, Clerodendrum, and Portulaca sp. sericea. Hernandia nymphaefolia and Pisonia survived locally on this raised reef limestone grandis had not been common along this coast, shore. although individual trees were seen which had Lofanga has a similar raised reefal lime been left standing but entirel y stripped of fo stone coa st. In the center of the island man y liage and of the smaller branches. Coconut big trees were uprooted, but the coastal vege palms had also been snapped off in this belt tation did not suffer such severe damage. On of destruction. The coconut woodl and behind those coasts on which damage was sustained, the cleared area consists of Pandanus and and from which soil was removed, regenera coconut palms still standing but with damage tion will presumably be slow and restricted to their crowns (Figure 3). initially to regrowth from the stumps. Similar devastation occurred along the eastern coast of Lifuka and was observed BEACH MORPHOLOGY: The greatest changes locally at sites visited on the east coast of Foa to beaches in the Ha 'apai group occurred on and 'U iha, The width ofthe belt was less than the pocket beaches found along the eastern on Ha'ano, but exceeded 30 m on Lifuka (pro shore ofthe islands, between headlands of the file la,Figure 2). To the southeast of Lifuka raised reefal limestone. Erosion took place on stumps of Pandanus, Hibiscus, and Messer the shorter beaches and depo sition on many schmidia were left after the storm, with low of the longer beache s. No study of prestorm Pemphis on some of the seawa rd cliff tops. beach form s exists, but some record of the 190 PACIFIC SCIENCE , Volume 37, July 1983

prestorm profile was derived from outcrops of with a total beach width of 40-50 m. At limestone on which a line was detectable (like other sites along the eastern coast of Lifuka, that described by Ogg and Koslow, 1978,from the lower beach reaches 7-8°, while sand beaches after a storm on Guam) between white deposited on the upper beach is at 5° or less; limestone previously buried beneath sand and total beach widths are 18-30 m, again fre not exposed long enough to have an algal quently with buried stumps of Messerschmidia coating and that subaerially exposed lime in the mid-beach showing signs of regrowth. stone on which algal growth had discolored By contrast, little evidence of change was the rock to a dark gray. Several such locations found on the beaches of the western coast of were found on the east coast ofLifuka and on the islands on the barrier reef, or on any of the east coast of Ha 'ano, illustrating removal the shores of Uoleva, which was presumably of as much as I m depth of sand. The greatest sheltered behind Lifuka . On these beaches depth of removal was observed at the site anglesof the lower beach are 5-7°, and beaches surveyed on profile 2a (Figures 2 and 4), where are rarely more than 20 m wide. On the north more than 1.9m of sand was removed at the eastern shore of 'Uiha erosion was evident landward edge of the beach. The limestone along the western end of the beach and a sand outcrop does not extend to the foot of the cliffup to 1.3m high was scoured , but deposi beach and hence the change of volume near tion appears to have occurred on the upper the foot is unknown, but at least 28 m' per m part of the eastern end of the same beach. The of sand was removed from the beach face angles of the lower beach are 7-8° along the from the mid to upper beach. entire beach, but to the east the beach is con Elsewherealong the eastern shore of Lifuka siderably wider, up to 35m. there was deposition of sand , particularly on The beaches that have received sand in the thelonger beaches. Profile 7 (Figure 2) is one Ha 'apai group have angles of the lower beach site at which sand was deposited. The vegeta of 7-9°, while the unaffected beaches of the tion at the back of the beach, predominantly western shores and of Uoleva have angles Messerschmidia, Hibiscus, and Pandanus, was of the lower beach of 5-7°. This steeper beach broken and uprooted and is fronted by a sand may result from the deposition of sand by the cliffI m high. Stumps of M esserschm idia, with cyclone, or it may be a feature of the higher branches broken off but with some regrowth energy experienced along these coasts; the line of leaves visible, occur up to 9 m from the between the white and the algal-colored rock present vegetated sand cliff, on the beaches, on Ha'ano shows a prestorm beach angle of with sand deposited around them. These in 9°, while that on Lifuka (Figure 2, profile 2a) dicate the distance of vegetation retreat. The indicates an angle of the upper beach of 4-6°. beach has a prominent berm at about high Deposition appears to have occurred mainly tide level, and the lower beach is relatively above high-tide level.No obvious relationship steep, 8°, while the upper beach is at a lower was observed between poststorm beach mor angle and about 30 m wide. At the foot of the phology and the width of the reef flat. beach there isan outcrop ofbeachrock, which would seem to imply that there has been little TATAFA: The island ofTatafa was probably progradation or recession of the foot of the the most severely devastated in Tonga as a beach; though the beachrock need not have result of Cyclone Isaac. Tatafa is a sand island coincided with the foot of the prestorm beach lying to the north of 'Uiha and separated from (beachrock is presently found 45m to seaward Uoleva by a deep channel more than I km of the foot of the beach on profile 4, Figure 2). wide. The island is oriented NW-SE and is Beachrock was not observed at the foot of all composed of sand, with no exposure of reefal of the beaches on which deposition occurred; rock though with beachrock around much of on some it may have been buried. its perimeter.The cycloneseemsto have passed A similar depositional beach with a steep directly over the southern end of Tatafa re berm occurred on eastern Foa, where the lower sulting in total removal of much of that part beach reaches 9° and the upper beach 4-7°, of the island. The Directorate of Overseas Impact of Cyclone Isaac on Coast ofTonga-WOODROFFE 191

FIGURE 4. Removal of sand from the beach, profile 2a, Lifuka; the white limestone was previously buried by sand, while the darker limestone has been exposed and has an algal cover. TATAFA

~ Grass

~ Hibiscus

c:::::;] Hern andi a

~ M esserschmidia

5J Pandanus

~ M ixed coasta l scrub CillJ Sand RceJ1:~~~hs c~~ g e v a s t a t e d • -,: . o Intertid al sand ~ Coconut woodland ~ Beachrock

~ m etre s 300 Uproot ed plant debri s Undercut shore Iine ,

FIGURE 5. The island of Tatafa . -'-0 00 \;J Impact of Cyclone Isaac on Coast ofTonga-e-Woonaorrn 193

FIGURE 6 The channel scoured throu gh the southern end of Tatafa, seen from the west, showing the coconut and other uprooted plant debris scattered on the leeward sand flats.

Surveys (DOS) topographic map (l : 25,000), to the southeast of the island but may be based on aerial photography in 1968, shows buried beneath the sand flats. Some beachrock Tatafa to have been 1710m long, separated in this area was exposed as a result ofremoval from 'Uiha by a shallow channel about 240 m of sand by the storm. acro ss. In 1982, after Cyclone Isaac, the shape The eastern beach is generally wider and ofthe northern end of the island had changed steeper than the western beach and has had little since 1968(Figure 5), but the total length sand deposited on it like the eastern beaches is 1484m, and the distance from the south of of Lifuka and Foa . Vegetation has retreated Tatafa to 'Uiha has increased , with more than and been buried as shown by the presence of 450 m of intertidal sand flats to the south of partially defoliated stumps of Messerschmidia Tatafa. emergent from the upper part of the beach. In addition to removing much ofthe south Around much of the island the shoreline is ern end of the island, Cyclone Isaac also undercut; there is a sand cliffup to 80 em high scoured a channel through the middle of the on the northern side and found intermittently island (Figure 6), leaving two smaller islands, to the northwest, sometimes at the top of the larger one 1398m long and 348 m wide, beaches which have received sand. A sand cliff and the smaller 70 m long and 52m wide. The occurs to the southeast and is up to I m high channel scoured between these is 16m wide, through the scoured channel. Along the south sand floored, and flooded by the highest tides. ern shore it is 1.2-1.5 m high. In addition, Around most of the island there are inter along this part of the coast there are semi mittent outcrops of beachrock varying from circular scour holes formed on the leeward I-2m wide to 7-8m wide, and to the north side by water passing over the island, up to west there is a larger outcrop of beachrock, 25 m across and often with a narrower neck heavily cemented , which appears to rise above only 8 m acros s. These have plant debris in the highest tide level. The beachrock gener some of them, but others have been scoured ally occurs at the foot ofthe beach and implies clean, and to such a depth that there may be that the island to the northwest has undergone small pools of water retained in the deeper little change in configuration since the beach part. rock formation. Beachrock is less developed The vegetation of most of the island is 194 PACIFIC SCIENCE, Volume 37, July 1983

coconut woodland in which Scaevola and coast of the barrier reef islands ofthe Ha 'apai Phymatosorus grossus are found, and it is this group, except to the northeast of Tongatapu coconut woodland that overhangs the eroded in the vicinity ofNiutoua. Here the vegetation sand cliff. Many trees have been felled and was stripped; Pandanus was broken off, gener are lying in all directions around the island allyjust above the prop roots, and shrubs such (Figure 6), while the canopies of those still as Messerschmidia, Bikkia, and Hibiscus were standing have been badly damaged. defoliated and had most branches snapped Around much of the perimeter of Tatafa off; but these were already sprouting from the there is a belt ofcoastal scrub or woodland, up stumps at the time of survey. to 30m wide. In many places this is dominated by Hibiscus, elsewhere by M esserschmidia, THE NORTH COAST OF TONGATAPU: Cyclone and in other places by a mixed scrub of both Isaac left a trail of destruction in Nuku'alofa, Hibiscus and M esserschmidia, with Scaevola, both along the waterfront and inland as a Acacia, Suriana maritima, and occasional Pan result of widespread inundation. The seawall danus and Hernandia emergent. Along the was badly damaged and 10-20 em of fine northern coast there was a belt of Pandanus, grained sediment was deposited on Vuna 15-30 m wide before the storm; many ofthese Road. The mangroves- along the northern trees were snapped or felled during the storm. coast suffered surprisingly little damage as a Adjacent coastal scrub was also stripped; most result ofthe storm. Partial defoliation ofindi of the branches and canopy were lost from vidual mangroves occurred in Sopu and on scrub in a badly devastated area to the east, the reef flat to the west ofNuku'alofa. Clumps but rapid regrowth is occurring from the of Rhizophora samoensis on the reef flat were stumps. Hibiscus in particular had leaves particularly affected, as well as a few trees sprouting from the damaged trunks at the of Lumnitzera littorea in plantations inland. time of survey, and isolated Messerschmidia Along much of this northwest coast of Ton were also growing back. gatapu there is a seaward belt of Rhizophora, The greatest destruction occurred at the replaced to landward by a belt of Excoecaria eastern end ofTatafa. The leeward beach and agallocha. In the last week in May Excoecaria sand flats here were strewn with plant debris, was observed to be losing its leaves as a result much of which was coconut and the rest ofa heavy infestation ofthe caterpillar Achaea largely unidentifiable. Several large trees were janata which was grazing the young regrowth. toppled , including the only individual of The seawall enclosing an area of reclaimed Casuarina seen on the island, and several mud flat (Straatmans 1954)had been breached Hernandia. The changes in island configura by the storm, and renewed tidal inundation in tion that have taken place appear largely the area had meant loss of several hectares of irreversible. grassland. Much ofthe storm surge associated with Cyclone Isaac must have been buffered by the great width ofreef flat along this coast, Tongatapu and its impact was not enormous. Cyclone Isaac passed north of Tongatapu BetweenPolo'a and Nuku'alofa, about 50m and caused extensive damage to low-lying from the reef edge, there is a spread ofgravel settlements along the north coast, such as and rubble-sized material on the reef flat. This Sopu and Fatai which were flooded and Kolo forms a low ridge of debris in a gravelly matrix vai which was devastated by high wind speeds. ofcoral sticks, with larger coral fragments up The raised reefal limestone cliffs of the west, to blocks 30-40 em in diameter. The larger south, and east coasts suffered little damage. corals are generally flat or pedestal shaped. Many trees, in particular coconut palms, The material did not appear to be fresh, some breadfruit, and mango, were felled by the having a partial covering of the encrusting wind on the island, but the coastal scrub on algae, and much probably predates Cyclone the raised reefal limestone cliffs was not de Isaac. Some of the platy corals, however , are vastated in the same way as on the eastern still covered with delicate projections and have Impact ofCyclone Isaac on Coast of Tonga-c-Woonaosss 195 undergone minimal weathering; this material with a complex of buildings to the south of the may have been added during the storm. island and extensive gardens in the center. Coconut woodland and woodland of Hernan BEACH MORPHOLOGY: No systematic study dia nymphaefolia and Calophyllum inophyllum of the beaches around Tongatapu was under appear less extensive than in 1969, and to the taken; however, local accounts reported de north there is an area of grass, probably Lep position of sand on the eastern beaches, most turus repens, with Excoecaria agallocha, Vi/ex notably Oholei beach. Accordingly, profiles trifolia, and Hibiscus tiliaceus. Coast~1 woo~ were surveyed on Oholei and Laulea beach land forming a belt around the Island IS and on beaches to the northeast and north of dominated by Hibiscus with Scaevola taccada, the island. Deposition of sand had occurred Pandanus sp., and Clerodendrum, and with on the upper part of both Oholei and Laulea monospecific stands of Thespesia populnea beach, with a prominent berm at or just above and Cordia subcordata on the west coast, and high-tide level. On Oholei beach the lower appears more disturbed than in 1969. beach was observed to be steep, 8°, and 15 Cyclone Isaac had its greatest impact on the 18m wide, with an upper beach I-2°, and up eastern shore, which had evidently retreated. to 14m wide. Laulea beach is steeper; the Uprooted plant debris was particularly wide lower beach is 8-11°, and 13-18 m wide, while spread, composed especially of Hibiscus , but the upper beach varies considerably in angle including several large undercut and toppled and width. Calophyllum and Hernandia, as well as nu merous coconut palms. The retreat of the Sand Cays ofTongatapu vegetation has left deep-rooted individuals of Excoecaria isolated now on the foreshore; these The islands offthe northern coast ofTonga trees appear to have survived the storm, but tapu consist of limestone-cored islands and were undergoing defoliation in May through sand cays. Those formed of the raised reefal the grazing of the caterpillar, Achaea. limestone, such as Polo'a, Alakipeau, and The coconut palms on Oneata were not Velitoa, were little influenced by the storm, s~nd devastated as on other islands, though the apart from some scrub damage. The foliage is badly knocked and few nuts remain; islands, however, had been more susceptible most of those felledor broken wereon the east to the storm surge associated with Isaac. A coast. series of sand cays, in a north-south transect approaching the path of the cyclone, were MANIMA: Manima is a small sand island visited and surveyed. The four southernmost lying NW-SE situated about 200m to the islands-Oneata, Manima, Pangaimotu, and northwest of Oneata on the same reef flat. Makaha'a-had previously been surveyed by Neither reefal rock nor beachrock are exposed Stoddart in 1969 (Stoddart 1975), and the on the island, and Figure 8 shows that !he more northerly islands were mapped for the major change since 1969 has been aggradat~on first time in this study. The resurvey of the of the western end of the island and reshaping islands surveyed in 1969 allowed assessment of the intertidal sand spits. The vegetation of of the change that had occurred in 13 years; this western end of the island is grass with not all ofthis change resulted from the passage Ipomoea pes-caprae, Vigna marina , and Eu of Cyclone Isaac. phorbia chamissonis, with isolated .Col~brina ONEATA: Oneata isan elongated sand island asiatica, Scaevola, and Vitex. The mtenor of lying less than 500m to the north of Nukunu the island is composed of coconut woodland, kumotu on an extension of the reef flat of the overall area of which appears less exten Tongatapu. It has no exposure of reefal rock, sive than in 1969, and which was particularly nor of beachrock, and is similar in shape in badly ravaged by the storm. Many of those 1982to that mapped in 1969(Figure 7). Many palms left standing lost their crown and nuts, of the changes which have occurred between and there are low-lying areas in the center of surveys result from settlement on the island, the island stripped bare of undergrowth with 196 PACIFIC SCIENCE, Volume 37, July 1983

me t re s 100 ? -_....:.::.::.:.:..:.-_-- I o Cordia

~ Grass • T hespesia ~ Man go

~ Scaevola c=:=;] Excoecaria ~ Cocon ut Wood lan d ~ ~ . ~,;: I ~ Hi biscus Casua rina ~ Upr ooted Plant debris

rH""I Hern andia ~ T aro IT2J Sa nd L....!!J f Her nand ia ~ WOOd la nd~ lhO;:: Cassava~~~::~ld o Inter tidal Sand r;;----,c a c e i '0,"'''0 m I ne I saac.

F IGURELoJ7. The IS. Iand of Oneata, 1969 (after Stoddart 1975) and 1982, after Cyclo Impact of Cyclone Isaac on Coast of Tonga -WooDROFFE 197

MANIMA

- ...... -

1969

\ \

~ Pioneer grass & shrub

~ Hib iscus

~ Hern andia

~ Call oph yllum [SSj Coconut wood land ~ San d

~. Excoecaria ~ Coconut stump o Int ertid al sand • Th espesia ~' "'"\."~~, Upr oot ed debri s Undercut shoreline FIGURE 8. The island of Manima, 1969 (after Stoddart 1975)and 1982, after Cyclone Isaac. 198 PACIFIC SCIENCE, Volume 37, July 1983 dead and sometimes broken palms in them. where there was a discontinuous tangle of Elsewhere around the coast is a woodland up plant debris, especially of uprooted Hibiscus, to 20m wide but less continuous than in 1969, Thespesia, and Cerbera manghas, along with composed predominantly ofHibiscus in which several fallen coconut palms . Where retreat of Pandanus, Calophyllum, and Colubrina are the coastal woodland has occurred, outposts important. ofExcoecaria have been left on the foreshore, The main evidence of Cyclone Isaac's im often with individuals of Thespesia directly pact occurs along the northern coast. Despite landward. Comparing the 1982 and 1969 sur the existence of undercutting for more than veys, there appears to have been about lO 200m of this coast in 1969 (Stoddart 1975), II m of retreat along this coast in 13 years. there was no pronounced undercutting of the Obviously not all of this need have resulted shore in this survey (Figure 8), except for a from Cyclone Isaac: indeed, Stoddart recorded small sand cliffless than 40 em high fringing a (1975: 5) that erosion was taking place at grassed area to the east. There was, however, the eastern point ofthe island with Hernandia much tangled, uprooted plant debris, partic trees exposed on the coast. Hernandia pres ularly of Hibiscus and Thespesia, as well as ently seems less extensive than in 1969; the fallen and defoliated Calophyllum and numer retreat of coastal woodland around the Ex ous coconut palms. Further evidence ofreces coecaria trees left on the foreshore implies that sion of the coconut area was observed to the at least 8m of the erosion did occur during the east where several stumps or root remains of cyclone. palms were found on the reef flat, indicating MAKAHA'A: Makaha'a is a small sand island retreat of the shoreline by more than 12m. lying on an isolated reef flat, separated from PANGAIMOTU: Pangaimotu is the largest that on which Pangaimotu lies by a narrow island in the group immediately north of the channel about 10m deep. Much ofit is covered coast of Tongatapu. The island was mapped by coconut woodland which appears to have and described by Stoddart (1975) and has been little damaged by the storm and which changed little in its general configuration since has been described by Stoddart (1975). Seed 1969 (Figure 9). The interior of the island is lings ofCasuarina and Rhizophora recorded in composed of a coconut woodland, and the 1969seem to have taken hold and young trees coastal woodland around the outside is domi ofboth species were observed to the southeast nated by Hibiscus, with Colubrina and Clero ofthe island. Coastal scrub and woodland, up dendrum abundant and Cordia, Scaevola, to 20 m wide, is dominated by Hibiscus, with Pandanus, Calophyllum, Hernandia, Grewia several Hernandia to the west of the island crenata, and Diospyro s elliptica locally impor apparently having become prominent since tant. Isolated low individuals of Rhizophora 1969. samoensis grow at the northern end ofPangai Makaha'a shows several marked changes motu and appear to have changed little in ex from 1969to 1982.The patterns ofbeachrock tent since 1969.The most apparent differences mapped in the two surveys are broadly similar since 1969result from expansion ofsettlement and serve as a reference for the rest of the on the island, especially to the south and west, island. Changes in the beachrock itself may with extension ofa picnic area. reflect burial or exhumation ofbeachrock, or The only extensive outcrop ofbeachrock is erosion and degradation of beachrock between to the west ofthe island, extending 95 m along surveys.An area ofsettlement with three build the shore up to 6 m wide, with further degraded ings has grown up on sand which has aggraded outcrops to seaward. Also in this area the to the southwest ofthe island since the earlier shoreline has been undercut, and a sand cliff study. Also in this area of sand build up up to 80 em high occurs at the top of the is pioneer vegetation of grass, Vigna and beach. Messerschmidia. The impact of Cyclone Isaac was most An undercut shoreline was recorded along apparent along the northeast and east shore the eastern margin of the island in 1969

metres 100 !

,....., ~sa nd § Pioneer grass and vines l!!!!J Casuarin a ~ T he spesia L.!..J Cassava Inte rtidal sand rhl Coastal wood land. mostly D ~ ~ A rt ocarpus ~ Coconut woo dland L....!!.J Hib iscus Rhizophora ~ Beachrock ~ Herna ndia [:;] Excoecaria ~ Ma ngo ~ Uproot ed plant debris ~ Unde rcut shoreline

F IGURE 9. The island of Pangaimotu, 1969 (after Stoddart 1975) and 1982, after Cyclone Isaac. 200 PACIFIC SCIENCE, Volume 37, July 1983

MAKAHA'A

\ I

~ Pioneer strand vegetat ion ~ Leucaena ~ Up root ed pl ant debri s ~ ~ Scaevo la Casuarin a tiEill Sand

~ Hibi scus ~ Rhiz ophora ~ Beachrock

~ Hernandia ~ Coconut w oodl and "1'TTM"'fTT" Und ercut shore line

FIGURE 10. The island of Makaha'a, 1969 (after Stoddart 1975)and 1982, after Cyclone Isaac.

(Stoddart 1975) with sand cliffs ranging from showing little or no evidence of disturbance 1.5 to 2.0 m high to a maximum of 3 m high in resulting from Cyclone Isaac. It would appear the southeast. Stoddart (1975: 3) concluded that the present undercutting predates the that the island was migrating across the reef recent storm, and from its similar configura flat to the northwest. An undercut shoreline tion to the central eastern coast of the 1969 was also observed along the eastern shore in survey, and the similar distance from the 1982, rising from 104m at the south to 2m to beachrock outcrop, it seems to be a "fossil" the southeast, and being approximately 1m feature having changed littlefrom that mapped high on the eastern shore. However, along by Stoddart (Figure 10). much of the central part of this eastern shore To the southeast where the sand cliff is not the cliff is covered by growth of Hibiscus protected by growth of Hibiscus, some retreat presently extending 3 to 4 m to seaward and appears to have occurred; coconut woodland Impactof Cyclone Isaacon Coastof Tonga-WooDROFFE 201 is exposed on the coast and coconuts are was flattened by the cyclone has only the base undermined. In these areas retreat of 20m or remaining (Figure 12). Around much of the more has occurred since 1969, and there was northern and western shores coconut wood debris of Pandanus, Hibiscus , and Hernandia. land reaches to the coast, while to the east MONUAFE: Monuafe is a sand island, 176m there is a localized patch of grass with Aca cia long and 56m wide, lying to the northeast of a and a stand of Leucaena, 4-5 m tall, which has ?eep channel, called the Narrows, separating suffered some defoliation. It from Makaha'a, There are outcrops of Beachrock isdiscontinuous along the south beachrock to the north and east (Figure 11), ern shore and occurs to the northeast where it exceptionally reaches 9-10 m wide. It is also indicating the island to have been more exten sive,or to have migrated since the time of that along these shores, as well as much of the beachrock formation. An expanse of intertidal eastern shore, that undercutting was recorded. sand, exposed at low tide and up to 70m wide, The sand cliffis up to 80 em high to the north separates the vegetated island from a discon and west, 1m high in the southwest , and 1.2 tinuous line of beachrock to the east. 1.3m high along much of the south shore The vegetation of Monuafe is not well de reaching 2 m in the center of the latter. ' veloped and is composed largely of grass and The coastal woodland belt, up to 25m wide on the southern shore and composed of Hi scrub. Grass, probably Lepturus, and Ipomoea ?is~u~ , pes-caprae cover much of the west and south Pandanus , Ochrosia oppositifolia , and of the island, and here the only coconut palms individual Pisonia, with coconut palms rela occur, several well-spaced young palms less tively infrequent, was extensivelydamaged by than 7 m tall. The dominant plant in the scrub the cyclone; there was much uprooted plant debris. More coconut debris occurs to the on Monuafe is Leucaena leucocephala form ing a thicket up to 3 m tall, with occasional southwest where coconut woodland is found emergents to 6 m tall. On the eastern shore on the coast, and the outlying beachrock there is a small stand of Leucaena which has indicates that recession of the shoreline of more than 35m has occurred since the forma been partially defoliated . Little evidence of storm damage has been tion of that beachrock. Retreat ofa shoreline observed on this island, partly because the on which coconut woodland is found is also scrub has little woody vegetation which might indicated to the northeast (Figure 12). Not only has Fafa been hit on the eastern be preserved in contrast to the tangle of up rooted plants on the islands to the south . The and northeastern shores, as have the islands only undercutting of the shoreline has been to to the south, but it has been severelydamaged the north of the island where a sand cliffup to on the southern shore also. This reflects the 70em high can be traced for 30m along the configuration of the reef flat on which Fafa shore, at the point where the beach is nar lies, which is extensive to the east, and offers some protection to the island, but which is rowest, and beachrock is found at the base of much narrower, in some places only 50m wide, the beach. The large intertidal flat to the east on the southern shore and which was much less results from recession of the island but there is no indication whether this occurred during effectiveas a buffer against the storm surge. the cyclone or whether it predates the storm. TUFAKA: Tufaka isa sand island lyingabout FAFA: Fafa is a rectangular sand island, 8 km to the northwest of Nuku'alofa on the 422m long and 216 m wide, lying to the west reef flat which extends out to Atata. It lies on an elongated reef flat offshore of Tonga approximately 1km to the west of the reef tapu which extends northeast as far as Tau. edge and is 320 m long and 84 m wide. The Unlike most other islands in the chain, Fafa island was badly damaged by the cyclone, and has no exposure of reefal limestone. It is the coconut palms in particular were devas covered with coconut woodland with a broad tated; many palms have fallen and several area of grass and Ipomoea to the northwest, have been broken off2-3 m above the ground. where one building is standing and one that Of the 15 palms that were left standing, only 202 PACIFIC SCIENCE, Volume 37, July 1983

MONUAFE

': '

..... / /

r.:1;,'

······Z Gra ss & l ow co c onu t ~ Hibiscus Q Leu caena

r:;a Defoliated Leuca ena

~ Colubr ina

[!!J Panda nu s

r~':'~1.' ''t%:·"<~ Upr oot ed plant debri s

~ Sand

, Int er t idal sand EIJ

~ Beach ro ck

, Undercut shoreline

o metres 10 0 I I

F IGURE II . The island of Monuafe after Cyclone Isaac.

t_ ed' Impact of Cyclone Isaac on Coast of Tonga-e-Woonnorss 203

FAFA

o, m etres 100,

~ Grass

"' ... ~ Coconut wood land

~ U prooted plant debris

~ Hibiscus Q Leu caena ~ Sand

~ Hernand ia ~ Ochro sia ~ Beachro ck [AI . ~ Calophyllum ~ A cacia ~ Undercut shoreli ne

FIGURE12. The island of Fafa after Cyclone Isaac.

I I had crowns on them, and these severely 2 m tall, but with emergents to 4 m. In man y knocked (Figure 13). area s to the east L eucaena has been defoliated , To the north ofthe island there is a sand spit but the thicket is regenerating fast, with new 63 m long and to the south a spit 88 m long, shoots growing from ground level amid st the which is particularly broad to the east and divaricating leafless branches. may have received deposits of sand during the Around much ofthe island there are broad storm. The vegetation is dominated by a scrub deposits of well-consolid ated beachrock up to of L eucaena leucocephala, rarely more than I 1-12 m wide to the east and 32 m to the 204 PACIFIC SCIENCE, Volume 37, July 1983

TUFAKA

I • \.:'-: '.:.: :.::."

\,,', t,

I . \ .:..; . I . .. . , \ :. ,: ..' : ~'i \

\:'. \ . : \::,e.: .. ', ;'.:.. \ ... ", . . ~ Gr ass \ .:::."::. \:..:...:':.:.'. t=J Hibi scus I\:.:~·:::i:::~ :: :::. : . [S] Leucaen a \ ,:: . : :'>. : ~ : . tz*a Def ol iated Leucaen a

l~-'-%_'It.~'~ Upr ooted plant deb ri s

[TIm Sand B Beachrock

Und ercut shoreli ne

a m et re s 10 0 \ )'ij I I \

F IGURE 13. The island of Tufaka after Cyclone Isaac. Impact of Cyclone Isaac on Coast of Tonga-WooDROFFE 205 north. Outliers ofbeachrock to the northwest on reef flats of coral rubble and storm block s illustrate past changes in the size and config removed from the forereef and adjacent reef uration ofthe island. An undercut shoreline is flat. Although Ostergaard noted that in 1926 present along most of the eastern shore, gen reefs north of Nuku 'alofa, and particularly erally I m high and vegetated with Leucaena , those of Makaha'a, were "heavily littered reaching 1.2m high to the southeast. The with hurricane blocks" (Ostergaard 1935: 9), sand cliffis particularly distinct where there is Cyclone Isaac appears to have moved little beachrock at the foot of the beach. Plant material ofthis size; little rubble was observed debris consists of Acacia and Pandanus and on reef flats and storm blocks were rare. Iso is generally not widespread , as on Monuafe, lated boulders generally less than 1m in dia because little of the vegetation is woody. meter were recorded on the eastern reef flat of the Ha'apai barrier reef, but fewwere observed MALINOA: Malinoa is a small sand island, to have fresh faces on them and many were 285m long by 100m wide, which lies on an probably deposited before the passage of isolated reef flat more than 10km north of Cyclone Isaac. Rubble and boulders on the Nuku'alofa. The most apparent result of northern coast of Tongatapu also probably Cyclone Isaac has been the deposition ofsand predate the cyclone, though some ofthe more on the eastern and southern shores, where a delicate material may have been added re steep beach is found on which the sand has cently. To the north of 'Uiha and Uoleva in been undermined to leave a sand cliff 50 em the Ha'apai group it was noticed that several high. Sand, up to 1m thick, has buried a stand microatolls of coral had been detached from ofM esserschmidia leaving only the uppermost their growth position on the reef flat and branches exposed, and partly buried adjacent either tilted and strewn on the reef flat or in scrub of Suriana and of Scaevola and Acacia. many cases carried to the foot of the beach , This island shows a greater diversity of where the coral polyps had died. vegetation types than seen on other islands to Not all storms move rubble, and the absence the south (Figure 14). Malinoa is the only of rubble on the reef flats in Tonga might imply island visited on which Pisonia is an impor that Cyclone Isaac was not of sufficient mag tant element, forming several stands along nitude to move such large fragments. Alter the western shore. The interior of the island natively, absence of rubble may reflect the has a coconut woodland fringed with stands short supply of material of such size on the of Pandanus, rarely more than 5 m tall; both adjacent reef flats and offshore reefs but, as no show little evidence of storm damage. underwater observations were made on the Beachrock occurs around the western and nature of the offshore reefs and no recent northern shore of Malinoa, in a broad band accounts ofthese exist, this cannot be verified. generally 14m wide,but exceptionally reaching Conversely, rubble need not always be depo 37m wide. Undercutting has also occurred to sited on the reef flat. In Belize, Stoddart (1963) the north of the island, and a sand cliff up to recorded that in many cases Hurricane Hattie 1m high was observed . Little plant debris was had carried material entirely across the reef found on the coast ofMalinoa which, despite flat and deposited it in deeper water to lee being the closest ofthe islands to the center of ward, and in the Lau Islands Hurricane Val the storm, has experienced deposition rather deposited rubble in 20 m of water (Phipps and than erosion. Some trees of Pandanus have Preobrazhensky 1975). It may be that Cyclone been toppled along the eastern shore. Isaac has destroyed reefs around Tonga and deposited coral rubble on offshore terraces. It is interesting to note, however, that rubble ramparts are often found on islands in areas DISCUSSIO N which experience hurricanes and are evidence One of the major effects of catastrophic that storm events can move material of con storms on coral reefs and reef islands docu siderable size and deposit it above water level mented in studies elsewhere is the deposition (Hernandez-Avila, Roberts , and Rouse 1977). 206 PACIFIC SCIENCE, Volume 37, July 1983

M A LIN OA

~ : . : \ -.'.'\ '.:,:,:.>:\

' . . ,::\ . ' ..... \ ~~g~'~\~~:}i i ii)i\

~ Grass

~ Scaevol a & A cacia '~IIIII~;~_llj::f(, ~ ; )i; \

[SJ Hibiscus "'::,' .::::';}.: ·.: .· ·:: .: i.: ;' :~/j ~ Calophyllum bJ Pisonia . ~1111111~2· ··.•hJi;1 ~ Pandanus . . . . I .' ". >1 . ' : "1 Q Leucaena ~ ~ Sur iana

M esserschmidia Ed buri ed by sand ...,.. ,'. Coconut woodland ,. t§3 .,.. \. . I ' ~~.tj Up roo ted plont debr is :.:.' ~ ~: : . ': . ". " '.. ' '" ', -: '" -: " . [/.\ /1 Sand k:·\:. :~ ··: · ·\: ·· ~ : ·< :: 1:' . . , .. .' ...... ~ Beoch ro ck

~ Unde rcut shoreline

o m etre s 100

FIGURE 14. The island of Malinoa after Cyclone Isaac. Impact of Cyclone Isaac on Coast of Tonga-e-Wooonossa 207

Similar coastal deposits are found in the Lau reported from Jaluit and from the Belize cays group despite the observation that rubble (McKee 1959, Stoddart 1963). The changes deposition by Hurricane Val, occurred pre on the sand cays off the north coast of Tonga dominantly on offshore terraces. Rubble de tapu were less pronounced than on Tatafa. posits are not prominent landforms on the The four islands nearest to the shore, Oneata, coast of any of the islands visited in Tonga, Manima, Pangaimotu, and Makaha'a, had and there is as yet no indication that they are changed by different amounts since the survey extensiveoffshore. As Cyclone Isaac was prob in 1969 (Stoddart 1975), with greater change ably the most severe storm experienced in the on the smaller islands of Manima and Maka southern Tonga islands in 100 years or more, ha'a , which both showed aggradation to the it is implied that storms rarely deposit rubble west. Some retreat had occurred on the eastern sized material on the reef flat or on the coast in shore of each of these islands, though not Tonga and are not major factors determining all of this retreat necessarily resulted from island topography. Cyclone Isaac. There was direct evidence of The raised reefal limestone coasts under scour during Isaac, the retreat of the foreshore went little morphological change as a result having left individuals of Excoecaria and to a of the storm, except for the detachment of lesserextent Thespesia outstanding, indicating blocks from the upper part of the wave-cut as much as 8 m of recession on Oneata and notch on Ha 'ano. Both erosion and deposi Pangaimotu, or, where stumps of coconut tion occurred along the pocket beaches of palms remained on the reef flat, indicating up these raised reefal limestone coasts, with ero to 12m of retreat on Manima. sion prominent on the smaller beaches and Some undercutting of the shoreline had deposition on the upper part of the longer taken place on most cays, and sand cliffs beaches. Beach changes were comparable to reached up to 2 m high on Fafa and Makaha'a. those previously reported after storms from However, this undercutting was in most cases pocket beaches on Mauritius (McIntire and related to the occurrence of beachrock around Walker 1964) and on Guam (Emery 1962), the island, and whilethe cliffprovides evidence and the volume of 28m' per meter of beach of removal of sand, the presence of beachrock face for sand removed from the beach at the at the foot of a generally narrow beach implies site of profile 2a on Lifuka is within the range that little retreat of the shoreline occurred. 25-32 m' per m reported for other beaches It is true that the beachrock may have been under storm conditions (Morton 1976, Ogg formed several decades ago, and could sub and Koslow 1978, Simpson and Riehl 1981 ). sequently have been buried and exhumed by Deposition of sand had taken place on eastern the storm, as found to the south of Tatafa and beaches facing the open water, over reef flats reported from Belize by Stoddart (1963), but of 100-500 m width, and had resulted in steep this is unlikely to have happened in many lowerbeaches 7-8°, a prominent berm at high cases, and has not been a trend observed on tide level, and deposition on the upper beach those islands mapped in both 1969 and 1982. with variable angles of slope. Presumably the beachrock is resistant to The greatest coastal change resulting from scouring by the water (McIntire and Walker Cyclone Isaac was observed on the sand 1964), but it also appears to concentrate that islands. In the Ha'apai group Uoleva, shel scouring up the beach. The ineffectiveness of tered in the lee of Lifuka, underwent little shoreline undercutting as a means ofshoreline change, but Tatafa was perhaps the worst retreat is further indicated by the apparently affected island in Tonga. It had completely "fossil" sand cliffon Makaha'a which shows lost more than 200 m of sand shoreline from little change in plan with respect to the beach the southern end of the island, with a further rock from that mapped by Stoddart, and channel scoured through the southern part of which has Hibiscus growing to seaward of it. the island. There was extensive undercutting, Patterns of change on the northern sand and water overtopping the island had scoured cays off Tongatapu are more variable . Fafa holes in the leeward side like the scour holes shows recession of the shoreline ofthe eastern

-...... , 208 PACIFIC SCIENCE, Volume 37, July 1983 and northeastern shore, but in addition has Damage to coastal scrub on the raised reefal retreated on the southern shore, presumably a limestone shores directly exposed to Cyclone function of the reef flat configuration. The Isaac, on the northeastern shores ofthe islands degree of change on Monuafe and Tufaka is ofthe Ha'apai barrier reef and ofTongatapu, harder to reconstruct because the vegetation was largely by wind, though erosion of soil is not so woody and little evidence of the and deposition of coral gravel were evidence former shape of the islands exists. Malinoa, of high water levels on these coasts. Much of the island closest to the path ofCyclone Isaac, the vegetation was snapped off, and Pandanus has experienced deposition rather than erosion in particular was conspicuously broken gen with sand deposited on the upper beach to erally just above the prop roots. Regrowth the southeast. This pattern of accumulation of several shrubs, notably Messerschmidia, resembles that on the eastern beaches of the though also of Hibiscus and Bikkia, was islands of the Ha 'apai barrier reef and of occurring from the stumps. Tongatapu, and presumably reflects the rela tive isolation ofMalinoa, the narrow reef flat, and the open and deeper water to the east, from where the sand may have been derived. CONCLUSIONS While the effects of Cyclone Isaac on coastal landforms were not as far-reaching as Evidence of past storms is not preserved experienced on isolated atolls, coastal vegeta conspicuously on the coast ofTonga. Cyclone tion was in some cases as severely devastated. Isaac was one of the most severe storms to Coconut palms were badly damaged on many strike the southern islands of Tonga during islands, man y were felled, others snapped off, the time for which records are available and and ofthose remaining standing most had had caused extensive destruction to property and their crowns damaged and lost nuts and in to crops. It had relatively little impact on florescences.Other introduced economic trees, the morphology ofthe raised reefal limestone such as breadfruit and mango, also suffered, coasts and deposited little rubble-sized mate especially in the center of the larger islands , rial on the reef flats or beaches. Its effect on with man y big trees uprooted by the wind. sand cays was most devastating on those Casuarina was also noted to have been par directly in its path, such as Tatafa, while both ticularly susceptible to windfall, as noted on deposition and erosion occurred on those off Mauritius (Sauer 1962) and to a lesser extent the north coast ofTongatapu, depending upon on Jaluit (Fosberg 1961). the configuration of, and shelter provided by, Around the sand islands much coastal wood the reef flat. Deposition occurred both on sand land and scrub was removed by scouring, or cays and on the longer beaches exposed to the in other places buried by sand . Hibiscus was east or protected by only a narrow reef flat. severely hit, but on many shorelines, as on While many of the changes to coastal land Tatafa, it was showing signs of recovery and forms were relatively minor and will be rapidly regrowth from the broken stumps. Messer obscured, destruction ofthe coastal vegetation schmidia was in many cases buried on the was more extensive, particularly on the raised eastern beaches, as on Malinoa and eastern reefal limestone coasts. Regrowth, especially Lifuka, and it too was showing regrowth from from the stumps of the dominant coastal the stumps. Big trees such as Hernandia and shrubs, has been rapid, but changes in species Ca/ophyllum were undercut on the coast, and composition which have resulted from the where these had partly fallen they were storm will be longer lasting. Nevertheless, obviously going to continue to grow in this storms ofhurricane force would appear to be position. Scouring of coastal vegetation had a much lessimportant factor in shaping coastal left deep-rooted individuals of Excoecaria, landforms and coastal vegetation communities otherwise not a conspicuous element of the on raised limestone and sand islands in Tonga scrub, isolated on the foreshore . Leucaena was than on more isolated atolls elsewhere in the partially defoliated on the smaller islands. Pacific. Impact of Cyclone Isaac on Coast of Tonga- WOODROFFE 209

ACKNOWLEDGMENTS EMERY, K. O. 1962. Marine geology ofGuam. Considerable assistance was received from U.S. Geol. Surv., Prof. Pap. 403-B: BI -B76. the Ministry of Agriculture, Forestry, and EWART, A. , W. BRYAN, and J. GILL. 1973. Fisheries, Tonga, and from Seletute Falevai Mineralogy and geochemistry ofthe younger and Sela and Atolo Tuinukuafe. islands of Tonga, S. W. Pacific. J. Petrol. I thank David Stoddart, Roger McLean, 14:429-466. and Peter Hosking for comments on a draft FLOOD, P. G., and J. S. J ELL. 1977. The effect of this paper, Bill Sykes for identification of of Cyclone David (January, 1976) on the plant specimens, Jan Kelly for drawing the sediment distribution patterns on Heron diagrams, and Margaret Jacobson for typing Reef, Great Barrier Reef, Australia . Proc. the manuscript. 3rd . Int. Coral Reef Symp. 2 : 119-125. FOSBERG, F. R . 1961. Typhoon effects on in dividual species of plants. Pages 57-68 in LITERATURE CITED D. I. Blumenstock, ed. A report on typhoon effects upon Jaluit Atoll. Atoll Res. Bull. BAINES, G. B. K., P. J. BEVERIDGE, and J. E. 75. MARAGOS. 1974. Storms and island build GLEGHORN, R . J. 1947. Cyclone damage on ing at Funafuti Atoll, Ellice Islands, Proc. the Great Barrier Reef. Rep. Great Barrier 2nd. Int. Coral Reef Symp. 2: 485-496. Reef Com. 6 : 17-19. BAINES , G. B. K. , and R. F. M cLEAN. 1976. HERNANDEZ-AvILA, M . L., H. H. ROBERTS, Sequential studies of hurricane deposit and L. J. ROUSE. 1977. Hurricane-generated evolution at Funafuti Atoll. Mar. Geol. waves and coastal boulder rampart forma 21 :M I-M8. tion. Proc. 3rd. Int. Coral Reef Symp. 2: BALL, M . M ., E. A. SHINN, and K . W. 72-78. STOCKMAN. 1967. The geologic effects of HOFFM EISTER, J. E. 1932. Geology of Eua, Hurricane Donna in South Florida. J. Tonga. Bull. Bernice P. Bishop Mus. 96 : Geol. 75: 583-597. 1-93. BALTZER, F . 1972. Quelque effets sedimen HOPLEY, D . 1974. Coastal changes produced tologiques du Cyclone Brenda dans.la plaine by tropical cyclone Althea in Queensland, alluviale de la Dumbea. Revue Geornorph. December 1971. Aust. Geogr. 12 :445-456. Dyn. 21 :97-114. KATZ, H. R. 1976. Sediments and tectonic BAUER, G . R. 1970. The geology of Tofua history of the Tonga Ridge, and the prob Island, Tonga. Pac. Sci. 24 :333-350. lem of the Lau Basin. Pages 153-165 in BLUMENSTOCK, D. I. 1958. Typhoon effects at G. P. Glasby, and H . R. Katz, eds. CCOP/ Jaluit Atoll in the Marshall Islands. Nature SOPAC Tech. Bull. 2. (London) 182 : 1267-1269. KERR, I. S. 1976. Tropical storms and hurri ---, ed. 1961. A report on Typhoon effects canes in the southwest Pacific. N .Z. Met. upon Jaluit Atoll. Atoll Res. Bull. 75 : 1-105. Svc. Misc. Pub. 148: 1-114. BLUMENSTOCK, D . I., F. R. FOSBERG, and LADD, H. S., and J. E. HOFFMEISTER. 1927. C. G . JOHNSON. 1961. The resurvey of Recent negative shift in the strand line in typhoon effects on Jaluit Atoll in the Mar Fiji and Tonga. J. Geol. 35 : 542-556. shall Islands. Nature (London) 189: 618 LISTER, J. J. 1891. Notes on the geology ofthe 620. Tonga Islands. Q. J. Geol. Soc. London. BRYAN, W. B., G. D. STICE, and A. EWART. 47: 590-617. 1972. Geology, petrography, and geochem MciNTIR E, W. G. , and H. J. WALKER. 1964. istry of the volcanic islands of Tonga. J. Tropical cyclones and coastal morphology Geophys. Res. 77 : 1566-1585. in Mauritius. Ann. Ass. Am. Geogr. 54 : CRAIGHEAD, F . C., and V. C. GILBERT. 1962. 582-596. The effects of Hurricane Donna on the McLEAN, R. F. 1974. Morphology of hurri vegetation of southern Florida. Q. J. Fla. cane banks at Funafuti Atoll, Ellice Islands. Acad. Sci. 25: 1-28. Proc. 8th . N.Z. Geog. Conf., pp. 269-277. 210 PACIFI C SCIENCE, Volume 37, July 1983

1977. The hurricane hazard in the STEARNS, H. T. 1971. Geologic setting of an eastern islands of Fiji: An historical analy Eocene fossil deposit on Eua Island, Tonga. sis, 1875-1975. UNESCOjUNFPA Fiji Geol. Soc. Am. Bull. 82 :2541- 2552. Island Reports 1:9-63. STEPH ENSON, W., R. ENDEAN, and I. BENNETI. McKE E, E. D. 1959. Storm sediments on a 1958. An ecological survey of the marine Pacific Atoll. J. Sediment. Petrol. 29: 354 fauna of Low Isles, Queensland. Aust. J. 364. Mar. Freshwater Res . 9 :261-318. MARAGOS, J. E., G .•B. K . BAINES, and P. J. STODDART, D. R . 1962. Catastrophic storm BEVERIDGE. 1973. Tropical Cyclone Bebe effects on the British Honduras reefs and creates a new land formation on Funafuti cays. Nature (London) 196:512-515. Atoll. Science 181 :1161-1164. - - -. 1963. Effects of Hurricane Hattie MOORHOUSE, F . W. 1936. The cyclone of 1934 on the British Honduras reefs and cays, and its effect on Low Isles with special October 30-31, 1961. Atoll Res . Bull. 95 : observations on Porit es. Rep. GreatBarrier 1-142. Reef Com. 4 :37-44. ---. 1965. Re-survey of hurricane effects MORTON, R. A. 1976. Effects of Hurricane on the British Honduras reefs and cays. Eloise on beach and coastal structures, Nature (London) 207: 589-592. Florida Panhandle. Geology 4: 277-280. --- . 1971. Coral reefs and islands and OGG, J. G . and J. A. KOSLOW. 1978. The catastrophic storms. Pages 155-197 in J. A. impact of Typhoon Pamela (1976) on Steers, ed. Applied Coastal Geomorphology. Guam's coral reefs and beaches. Pac. Sci. Macmillan , London. 32: 105-11R. ---. 1974. Post-hurricane changes to the OSTERGAARD, J. M. 1935. Recent and fossil British Honduras reefs: Resurvey of 1972. marine mollusca of Tongatapu. Bull. Ber Proc. 2nd. Int. Coral Reef Symp. 2 :473 nice P. Bishop Mus . 131: 1-59. 483. PERKINS, R. D., and P. ENOS. 1968. Hurricane --. 1975. Sand Cays ofTongatapu. Atoll Betsy in the Florida-Bahama area-geo Res. Bull. 181 : 1- 8. logic effects and comparison with Hurricane STRAATMANS, W. 1954. Reclamation of tidal Donna. J. Geol. 76 :710-717. mud flats in Tonga. Tech. Pap. S. Pacif. PHIPPS, C. V. G. , and B. V. PREOBRAZH ENSKY. Comm. 53 : 1-18. 1975. Morphology, development and gen TAYLOR, F. W. 1978. Quaternary tectonic and eral coral distribution of some reefs of the sea-level history, Tonga and Fiji, South Lau Islands, Fiji. Proc. 2nd. In t. Symp. sur west Pacific. Ph.D. thesis, Cornell Univer les coraux et recifs coralliens fossiles, Paris, sity. 355 pp. pp. 440-455. TAYLOR, F. W., and A. L. BLOOM. 1977. Coral REVELL, C. G . 1981. Tropical cyclones in the reefs on tectonic blocks, Tonga Island ar c. southwest Pacific. N.Z. Met. Svc. Misc. Proc. 3rd. Int. Coral ReefSymp. 2:275-281. Pub. 170: 1-53. TONGILAVA , S. L., and L. KROENKE. 1975. Oil SAUER, J. D. 1962. Effects of recent tropical prospecting in Tonga, 1968-1974. S. Pac. cyclones on the coastal vegetation of Mar. Geol. Notes 1: 1- 8. Mauritius. J. Ecol. 50: 275-290. VISH ER, S. S. 1925. Tropical cyclones of the SIMPSON, R. H., and H. RIEHL. 1981. The Pacific. Bull. Bernice P. Bishop Mu s. 20: Hurricane and its impact. Louisiana State 1-163. University Press, Baton Rouge. 398 pp .

j .,