251

GEOTHERMAL ASSESSMENT OF THEFIJI ISLANDS : AN OVERVIEW

Malcolm E. Cox

Geothermal Institute, Geology Department, University of Auckl and

ABSTRACT L I This report attempts to summarise geothermal investigations i n and come to some conclusions PACIFIC i n regard to the utilisation potential of geothermal PLATE resources i n that country. The work carried out, institutions involved and survey results are rised. Some dffferences i n philosophy o f approach and i n methods o f interpretation are apparent, but a l l studies have introduced some aspects worth considering. Useof a silica mixing model geothermometer for dilute waters provided excessively high temperature estimates. The concl sions reached i n this report are that the area o f has the greatest potential for AUS TR ALIAN utilisation, with reservoir temperature o f PLATE which could produce by a binary system. ternatively, substantial direct heat isation potential exists such as for drying or freezing plants. The next greatest potential lies in small scale direct utilisation of a fairly - extensive reservoir i n the area. Geothermal systems on and other localities within the Fiji group are indicated to be of low through- put with subsurface temperatures o f 90-115°C. Utilisation of such systems i s limited to very small scale use i n this environment.

INTRODUCTION FIG. Location o f Fiji Islands i n S.W. Pacific. Tectonic features are shown, and zones o f The islands o f Fiji are located i n the S.W. shallow 70 km) seismicity (after Pacific between 16" and and consist of two Johnson and nar, 1972). major volcanic islands and numerous smaller islands and atolls. They a r e situated within the southern As a compari geothermal occurrences of the part of the belt of volcanism and seismicity which S.W. Pacific area can be found summarised in borders the western Pacific and which i s related to another report (Cox, major crustal structures and subduction. Fiji, however, i s within a tectonically complex, low RECORDED HISTORY AND UTILISATION seismicity section of this belt (Fig. and has a volcanic history from Eocene to Recent times, with Thefirst recorded description of thermal most activity during the Miocene-P1iocene period springs i n Fiji wasof those at Savusavu on the 1967). About 60 locations of thermal southern coast o f Vanua Levu i n about 1840 (Wilkes. springs are distributed throughout the two main 1845). During the next years most o f the islands and on five of the smaller islands of the other springs were located and described. Fiji group (Fig. 2). These springs are the only Traditionally, many of the springs had been used surface features of geothermal activity and most have for bathing, and the springs are still temperatures i n the range. Boiling springs used f o r the cooking of root crops. During the do occur, however, i n two areas on the island of 1940's small thermal baths were constructed a t Vanua Levu. This report attempts to summarise the Savusavu springs. geothermal investigations carried out in Fiji and come t o some conclusion regarding the potential I n 1951 the colonial government conducted f o r development o f the geothermal resources. 252

COX I pilot experiments into the feasibility of producing

salt from seawater by evaporation using geothermal Vonuo Leru heat (Munro, 1961; 1961 ) . These experiments were so a t Savusavu, and used a 1800 1 tre main supply tank and various combinations of settling pans (2.2 and 4.8 set into the ground. The . final pilot plant produced a crude evaporite o f Levu 95.8% NaCl a t about 13.6 per day. The experi- ments were disbanded largely due to problems with temperature variations. (A sample of the salt was, 0 however, used i n the gaol to bake a successful . batch o f bread). The local timber yard a t Savusavu 0 attempted some crude experiments of drying timber LOU i n an open-ended drum set into hot ground; other, Group similar experiments were made drying copra, e .. but none were further. 0

OUTLINE OF GEOTHERMAL INVESTIGATIONS I I I assessment o f Fiji eothermal FIG. 2: Locations o f thermal springs i n the Fiji potential began i n 1956 when J. Healy Geological ands. Survey) visited many o f t h e thermal springs on Viti Levu and Vanua Levu. Healy (1960) described and sampled the occurrences, measured flow rates and temperatures, related the systems to local geology 1974-75: Detailed descriptions of thermal features. and sumnarised the literature to that time. He Spring sampling and measurements. concluded that the geothermal systems i n the Geological r photo interpretation. Labasa and Savusavu areas o f Vanua Levu warranted Ground temperature surveys (1 m depth). further work and noted that the characteristics . Rechecking I.R. anomalies. of the Savusavu springs suggested an intrusive Shallow resistivity soundings and traver- neat source. ses (Savusavu only; i n conjunction with UNDP geophysicist, S. During the many o f t h e spring groups were further described by Fiji Mineral 1976 : Spring and groundwater 1 sampling. Resources Department (MRD) geologists i n geological Geophysical surveys a t Labasa-Tabia: mapping surveys. Before the Savusavu salt resistivity soundings and bipole- dipole experiments, Ibbotson (1960) carried out a shallow mapping (equipment on loan from Hawaii soil temperature survey around these springs to Institute of Geophysics and initially determine the distribution of near-surface heat. set up by HIG geophysicist, J. Kauahikaua); Some additional thermal spring sampling was done ground magnetics; S.P. traversing. during the early 1970's by H. Colley (MRD) with a 1977: Stable isotope study o f Vanua Levu and bias to mineral isation associated with hydrothermal Viti Levu springs (in conjunction with systems . J.R. Hulston, Institute of Nuclear Sciences, DSIR, N.Z.). I n 1972 black and white thermal infra- red Reinterpretation of existing airborne imagery surveys were flown over the main geothermal magnetic surveys. areas o f Vanua Levu (Savusavu and Labasa) and the Sampling o f springs i n other parts of small area of Wainunu i n S.W. Vanua Levu. The Fiji. surveys were flown by Canadian Aero Service under Preliminary resistivity survey o f Sabeto contract to MRD, and used a Bendix unit area, N.W. Viti Levu (incomplete due to (sensitivity 11.5 - 12.5 m and resolution of 2.5 equipment failure). mrad.) at flight elevations of 610 and 1220 m. The surveys were initially interpreted by the Subsequent t o the above studies, i n 1978-79, contractor (Moreton, 1973) with subsequent ground regional reconnaissance surveys were conducted by checking by MRD geologists (Colley, 1975). Results Pacific Energy and Minerals Ltd (PEM), as part of o f the surveys were later reinterpreted i n respect a work commitmentto the Fiji Governmentfor t o measurements o f surface temperatures (Cox , 1980b; options on a petroleum exploration 1icence o f f the The I.R. surveys provided 1ittle additional north coast o f Viti Levu. These included: information about the geothermal areas, though several minor, previously unrecorded seepages Interpretation of magnetic surveys f o r were located. depths t o Curie Point temperatures. Sampling o f many spring groups and From MRD personnel began detailed interpretation of chemistry. studies o f Fiji's geothermal resources, initially Reconnaissance soil mercury surveys on in the more active areas o f Savusavu and Labasa- Viti Levu and Vanua Levu. Tabia. .

253

cox In 1979, K.H. Williamson (IGS, UK) visited (Cox and Hulston, which show the thermal Fiji to carry out an appraisal of geothermal waters to be isotopically similar to local surface prospects and surveys done to that time. This was water and to have nil enrichment. at the request of H. (Director, to Overseas Development Administration, UK . Regional assessment techniques were employed Will iamson visited thermal spring localities by PEM (Anderson and Austin, 1979). Using in northern Viti Levu, Labasa and Savusavu on data (apparently largely offshore Vanua Levu and collected samples for chemical and northern Viti Levu with some onshore overlap) stable isotope their contractors(Eureka Resource Associates) deduced depths to the Curie Point and constructed Following by PEM, temperature heat flow maps. (The Curie Point is that where gradient studies were conducted during 1980 and the crust has lost its magnetisation due to 1981 in northern Viti Levu, and a t Savusavu. elevated temperatures). They deduced an average Measurements were made by N. Skinner (University of thermal gradient for N. Viti Levu of the South Pacific, Suva) in holes drilled by MRD and for the thermal springs From as well as existing holes. Thermal gradients had these results .they selected the Ba area of NW been previously measured in 1972 by J. Sass (U.S. Viti Levu a s a primary target for geothermal Geological Survey) in 4 shallow mineral exploration exploration. Some question, however, apparently holes in northern Viti Levu, 2 a t Emperor Gold exists in regard to the method of interpreting Mine and 2 MRD holes. the depth to Curie Point (Williamson, 1980). From reconnaissance soil Hg surveys over major RESULTS OF GEOTHERMAL SURVEYS roads (sample spacing 2-5 km with closer spacings over anomalous areas) Anderson and Viti Levu - Healy (1960) concluded that the Austin (1979) considered the Ba area as being geothermal systems on Viti Levu were of non-volcanic anomalous in respect to heat flow. They reported origin with low temperature surface discharges values of 35-1105 ppb with backgrounds of 20-70 and probably had 1imi ted subsurface ppb. I t does not appear, however, that regional temperatures. From bottom temperatures in drill- backgrounds due to 1ithological changes and holes (to 600 m) a t Emperor Gold Mine (EGM) he iferous mineralisation have been fully estimated geothermal gradients of 30" - considered (Williamson, 1980). Consideration but noted that the higher values probably resulted of such background conditions are extremely from circulating thermal water (which discharges a t important in Hg surveys for geothermal several levels in the mine workings). exploration usman and Landress, 1979; Cox, 1981b) in addition to which most surface geothermal Hg Temperature prof 1es measured in dri11hol e s anomalies are of limited extent several at EGM by Sass were - km). Calcite deposited from the Rabulu confirming Heal s resul ts, and were considered contained 580 ppb Hg (Cox, which is to indicate near normal crustal gradients, because believed to indicate a high background of Hg of thermal water circulation. Gradients in the in that area. two MRD holes (100 and 220 m) a t Rakiraki and Balenabelo in N and NW Viti Levu were 11.8 and Anderson and Austin (1979) also used respectively. Preliminary data from airphoto interpretation to locate curvil inear recent gradient measurements by Skinner and MRD they consider reflect the presence show values of for Ba and of batholiths resulting from mantle upwelling, for Rabulu (N. Skinner, A. Green, and which they consider are features of all Assessment of chemical analyses of thermal spring major geothermal areas. There is some question waters (Cox, shows them t o be essentially as to the existence of such structures in Fiji dilute alkali chloride water with a slightly and especially their relation to geothermal alkaline values are 45-80 The systems in that environment (Will iamson, 1980). dissolved solids content depends on location, In this author's experience virtually all Fiji being greater in coastal localities largely due geothermal occurrences are associated with to saline groundwater mixing. The springs have systems of linear, normal faults. PEM also low flow rates ( 2 and subsurface collected samples of most springs on Viti Levu, temperatures estimated from Si and with chemical analyses of all major constituents are usually in he range of (except Na, K, F) being made in the field. 11 The distribution of springs throughout Limited analytical results are reported but the island shows that many o f them are isolated (mixing model ) and kali systems, and probably derived their heat from calculations were made. Subsurface water fairly deep circulation 1 to 2 km). Chemical temperatures of 161 to 300°C were estimated deposition a t some of t h e springs reflects the by methods for geothermal systems at Ba, chemical character of the rocks through which the Tavua Rabulu in N. Viti Levu, with lower waters migrate and the low order temperatures: temperatures for the kai i geothermometers. is being deposited by the Waibasaqa springs A basic conclusion of this writer i s that the in central Viti Levu, and a t Rabulu on the central silica mixing model geothermometer used by PEM N. coast, deposition of calcite indicates produces increasingly unreasonable temperature subsurface water temperatures of 1 Further estimates with increasingly diluted waters. confirmation of the low magnitude of temperatures Temperature estimates from the alkali geother- of these systems comes from stable isotope studies mometers a r e more reasonabl e. 254

cox Williamson (1980) collected thermal water al springs and were estimated a t 14.3 for samples f o r chemical and isotopic analysis i n the the Labasa zone and f o r Tabia. (These UK. He considered the thermal waters of Viti values were incorrectly calculated i n Cox, 1980d; Levu show 1 ttle evidence-of high temperatures Geological mapping and airphoto inter- a t depth, being low i n (except where seawater pretation confirmed that all the spring localities i s mixing) and and showing no enrichment o f are associated with faults (with a dominant NE Li, Rbor F. kali geothermometers he trend) a point noted by Healy. Total discharge ated temperatures o f 105°C a t depth. from the Labasa zone was 60 and from Tabia He concluded from stable isotope data that the 12 subsurface temperatures may not be high enough to allow isopotic exchange t b occur within the period Site specific geophysical surveys were of circulation of the groundwater. conducted in the Labasa-Tabia zones (Cox, 1978; Electrical resistivity techniques involved 12 vertical soundings 915 m) and bipole- dipole mapping from 5 sources with 180 measurements of potential, The resistivity mapping was difficult to interpret but basically outlined a shallow, elongate layer low apparent resistivity 20 ohm-m) of 20 km and limited deep conductive zones, Soundings indicate the con- ductive zone is at 7 to58 and is 150 thick, with some continuation to depths o f 800 m within several zones some associated with faults. Ground magnetic surveys delineated zones of low magnetisation partly coincident with the low areas, which are interpreted as being due to hydrothermal teration. Reinterpretation of existing aero- magnetic data over the area suggests that the Labasa geothermal area could occur within a large collapsed caldera structure, within which faulting is well-developed. S.P. traversing within the Labasa zone showed dipolar anomalies, some wibh peak-to-peak amplitudes of up to 100 near hot springs and associated faults, but overall the results were non-conclusive (Lienert Cox, 1981) . The geophysical surveys are believed to have outlined the approximate extent of a shallow aquifer of thermal water and indicate FIG. 3: Thermal springs of the Labasa area. some vertical extent of thermal features (? hot Boiling springs occur at Waiqel e water upflow) largely associated with faults. (Wainggele); Tabia (Tambia) springs are On the basis of these data and geochemistry, 80°C. Approximate boundaries o f low 3 exploratory drill sites were located. resistivity zones are outlined. Geochemical studies of the Labasa area (Cox, show the thermal waters are near Labasa Area - Thermal springs i n this area occur neutral with Na and Ca as major cations and over a 19 km long NE-trending belt, 8 km south o f and as major anions, with total dissolved the town o f Labasa (population 3000). In constitutents of 1100 The boiling addition, there i s a smaller group of hot C) springs at the centre of the belt are depositing springs a t Tabia, 12 km to the west o f the town. (gypsum) and minor . Chemical geother- The springs i n the centre of the Labasa group indicates a maximum temperature are boiling (Fig. 3). in the system of 120" - The Tabia system is similar chemically but appears to be a small, Healy (1960) visited the main springs and unique system. Subsurface temperatures there estimated a heat loss from them o f 4.4 are estimated at 110°C. Stable isotopes of He considered they may be associated with a single oxygen and hydrogen indicate the spring waters large reservoir, which is of relatively low tempe - are of meteoric origin and have the characteristics rature. He concluded that the waters were probably of fairly low temperature, hot water systems, of local meteoric origin. Cox located with recharge occurring locally (Cox and Hulston, and described a l l the occurrences, carried out 1980). Waters from the different springs are preliminary sampling, heat loss calculations and indicated to be of a comnon origin, but to have soil temeprature surveys. The soi1 temperature experienced varying degrees of mixing with surveys (lm depth), i n which 30°C was shallow groundwater, and possibly different considered anomalous, showed that heat flow i s residence times in a shallow aquifer. Many of essentially convective and that shallow ground these conclusions were initially suggested by temperatures rapidly dropped to ambient away from Healy (1960). the discharges. Estimates of heat 1 were greater than those o f Healy, but included addition- 255

cox Anderson and Austin (1 structually inter- This area was considered by Healy (1960) to preted airphotos o f the area i n respect to spring have the greatest potential for geothermal locations associated with curvilinear features ation i n Fiji; he concluded that the characteristics and suggested doming caused by igneous intrusions. o f the system suggested a hot intrusive body below Igneous intrusions do-occur i n the area the peninsula. Calculations o f heat loss from the (Ibbotson, 1969) and are reflected i n aeromagnetic springs provide a figure of data, but a different structural surface expression in agreement with Healy's conclusions and recal was interpreted by Cox which i s believed ated heat loss measurements. Soil to indicate subsidence. From their water urements show an area o f 0.06 km o f C chemistry Anderson and Austin (1979) determined a around the main springs a t Nakama and on the adjac- possible reservoir temperature o f - ent beach, which i s attributed to lateral movement estimate o f 128°C) which they consider to o f shallow hot water. However, anomalous ground be related to a regional, high heat flow. eratures were limited to the immediate vicinity of Their soil Hg survey produced a "moderate mercury the other spring localities. Limited geophysical anomaly" which was excluded from exploration target surveys were made around the locations (Cox, selection on the basis o f the water chemistry Ground magnetics outlinned areas o f low suggesting subsurface temperatures too low for magnetisation, coincident with soil temperature electric-power generation. anomal ies, and interpreted as being due to shallow hydrothermal teration. S. P. traverses showed From h i s geochemical studies, Williamson (1980) higher potentials near some thermal discharges, but determined a total dissolved solids content o f low potentials over the main springs; the results of 1200 and from content calculated a the surveys are not, however, considered reliable. reservoir temperature of Results o f his Shallow vertical electrical soundings = 350 m) isotopic studies confirmed the results obtained defined the saline water table and suqqest shallow Cox and Hulston thin lenses o f thermal water near the-main and an indication of deeper zones o f thermal' \ some possibly within faults.

Chemistry o f thermal and other waters (Cox, indicate the deep thermal water to be a near neutral type with an estimated temperature o f 160 C. Mixing o f 15-25%seawater occurs i n many o f t h e springs. The thermal waters are saturated with quartz and boiling springs dep- osit a quartz sinter, an indication o f higher level temperatures. All springs discharge from a system of faults, which are indicated to be hydrologically connected, and t o control subsurface fluid migration. On the south coast o f the peninsula a coastal coral reef may tend to restrict thermal water discharge. isotope data (Cox and Hulston, 1980) show the thermal waters to be essentially o f meteoric origin, with some positive oxygen shift, the slope ...... o f which overall indicates seawater mixina. Some minor enrichment from thermal conditions may however occur. Three locations for exploratory drilling (one 500 m and two 300 m holes) have been suggested At least six volcanic centres occur along the peninsula and further inland indicat- ing an easterly trend, which coincides with an elongate E-W dipolar magnetic anomaly, suggesting the existence o f .an intrusive mass at depths o f around 200-400 m. Lavas o f one o f the volcanic centres (Suvasuva) 7 km east o f the springs, may be appreciably younger than the surrounding volcanic FIG. 4: Savusavu area: inferred from field rocks. mapping and airphoto interpretation. types, formations and volcanic centres (after Anderson and Austin (1979) interpreted the Woodrow, 1976). Savusavu hot springs to be on the western side of a prominent curvilinear feature, centred over this Savusavu Area - The thermal springs a t Savusavu (?) younger intrusion. They consider this as the occur a t 3 main localities around the coast o f a "apparent intrusive centre" and to be the centre block- faulted peninsula 7 x 3 km o f the area interest, showing radial faulting peninsula i s formed o f Late Miocene-P1iocene andes- representing shallow doming. Their chemical geo- i lavas and breccias The main provides a "best reservoir temperature" sprin s occur within the town of Savusavu (population o f 190 C. Soil Hg traverses i n this region had a background ue o f 20-70 ppb, with anomalous values 256

cox of 300 ppb a t the Savusavu springs, Sinters from 1ikelihood of gypsum deposition i n the pipes could, these springs contain 150 ppb Hg however, be troublesome. Wi11 (1980) showed from chemi analys s Savusavu, Vanua Levu - The Savusavu system has the that the Savusavu springs have a high content and greatest potential for exploitation of the Fiji the highest in indicating quartz equilib- geothermal areas. Subsurface temperatures of 1 rium o f 140 C. He concluded that the appear reasonable, within a fault controlled reserv- high surface heat loss and relatively high o i r below the peninsula. The subsurface permeability suggest the area has some promise o f a high o f the lavas, however, i s an important factor. An system. extension o f the reservoir inland i s a possibility but there is currently no direct evidence f o r this. Thermal gradient measurements i n the MRD hole Further confirmation o f the higher temperatures i n drilled in 1981 approximately 3.5 km SE o f the main this system are quartz deposition from the spring springs are not yet available. waters and the higher heat loss. Both the thermal waters and sinters are low i n trace metals. Other Geothermal Areas - Healy (1960) compiled data from previous studies o f other thermal areas on Some 1 are evident. Import- Vanua Levu and t h e smaller islands. Rodda (1979) ant for exploitation i s the need f o r adequate described and led the springs on Vanua Balavu, recharge; the hydrological conditions o f the penin- and Woodrow ( 1 described the springs on Rabi sula are not known, especially i n regard to whether Island. These and other previous reports show that recharge o f groundwater i s entirely local or some these occurrences have low flow rates and i s derived from further inland. This also introduces temperatures o f C. Heat loss calculations for a potential problem o f drawing i n a significant these groups o f springs range from 0.05 to 0.8 component o f sal ine groundwater after prolonged Most o f the thermal springs on the smaller islands production. This could be especially important if have appreciabl e seawater mixing, Assessment o f the Ghyben-Herzberg basal groundwater model appl ies avai1 e water chemistry indicated, to the peninsula. I n respect to the above, a sub- water temperatures i n the range o f 90 - stantial increase i n salinity could produce corros- 115 C. ion problems; silica deposition should also be Anderson and Austin (1979) extended their expected (maximum measured spring = 179 ppm). reconnaissance surveys t o include the springs o f Cakaudrove Peninsula (eastern Vanua Levu) and Rabi Some rough approximations can be made about Island. From airphotos they interpreted the exist- the geothermal potential o f system. Assuming ence o f a 30 km diameter NE-SW elliptical structure a reservoir temperature o f 160 C and a potential centred on Rabi Island, and which intersects the total discharge from the system o f 196 t i p o f Cakaudrove Peninsula. They report "best" (equivalent of the natural discharge o f 60 f o r subsurface thermal waters of gives an output o f 36.5 Accounting f o r and 145 C for the Rabi springs. They also report a permeability variations etc. i n wells, would give minor Hg anomaly f o r these areas. an optimum o f about 10%of this available for utilisation 3.65 from several ideally sited geothermal wells. As the conversion o f CONCLUSIONS FROM INVESTIGATIONS thermal t o electrical energy i s about 30%efficient, this gives a conservative approximate potential of Viti Levu - The thermal systems on Viti 1.1 10%of this to run plant and aux- some with fairly deep circulation. No significantly el evated geothermal gradients are indicated, and The indicated temperatures of this system the higher gradients measured do n o t produce any suggest that it would be marginal for exploitation major heat input to the geothermal systems. The by conventional geothermal power plants and that a systems are o f low throughput o f locally derived binary- type system may be more suitable, Such a meteoric water, with subsurface migration largely set-up may require a system with at least three control1ed by faults. water temperatures wells, and a skid-mounted (dimensions around are on the order of 90 -115 C. As a consequence, any 3 x 8 m). Another consideration is the reinjection possible isation current conditions of discharge water from the binary heat exchanger, in o f low investment, very small scale if recharge problems are anticipated. direct use. A possibly economic ternative for Labasa, Vanua Levu - The Labasa system is indicated Savusavu woul d be direct heat applications of the to be a relatively extensive fault controlled thermal waters. These would require shallower with subsurface temperatures on the order of 125 C. wells, and lower production. Some form of drying Several exploratory dri11 sites were recommended or refrigeration plant are worthy of consideration near the centre of the zone. Although there exists in this rurally based area, especially considering the potential o f significant recharge to this the potential for port facilities and road access system, any utilisation is limited by the low temp- to Labasa. eratures. On this basis the area i s unsuited for consideration for electric power generation, but Current Power Usage - some smaller scale direct heat applications may be feasible. The dissolved solids i n the waters are In terms o f usage of electrical energy, the 1200 ppm and the waters contain very low amounts main requirements in Fiji are on the island of of trace metals; their saturation with and the 257

cox Viti Levu, on which is the capital ,Suva, the inter- Healy, J. hot springs and geothermal national airport a t , and on which the majority resources of Fiji. Ind. Res., of the population of 500,000 live. N.Z. ,Bull. 136, The present instal 1ed capacity (diesel generators) Ibbotson, P. 1960. Geothermal investigations at of Viti Levu is (from Williamson, with Savusavu. Geol. Surv. Fiji., Rep. 62, another 20 generated by other facilities (eg. unpub. Emperor Gold Mine) Hydroelectric installations are Ibbotson, P. 1969. The geology of east-central being constructed, and a 40 MW plant was due for Vanua Levu. Bull . Geol . Surv. Fiji, 16. commission in late 1981. The installed capacity of Johnson, T. and P. 1972. Focal mechanisms hydroelectric stat ons should be increased to 60 MW and plate tectonics of the Southwest by 1984 and 80 by 1987. On Vanua Levu, the Pacific. J. Geophys. Res. 77, 26, installed capacity a t Labasa is 2.6 (maximum 5000-5029. demand 1.6 MW) and at Savusavu 0.6 (maximum Kennedy, E.N. 1961. Second report on the use of demand 0.2 MW). geothermal heat to produce salt at Savusavu, Vanua Levu. Mines Dep. Fiji, ACKNOWLEDGEMENTS Rep. unpub. Klusman, R.W. and Landress, R.A. 1979. Mercury in I thank James Kanyua f o r comments on utilisation soils of the Long Valley, California, aspects, and Peter Rodda for assistance during my geothermal system, J. Vol . Geotherm. three years with MRD and his subsequent editing of Res., 5, 49-65. the geothermal reports. I also thank for Lienert, B.R. and Cox, M.E. 1981. Variations in analytical work during these studies. The opinions self-potential close to hot springs in in this report do not necessarily reflect those of the Lambasa area, Fiji. Geotherm. the Fiji Government. Resour. Counc., Trans. 5, 91 -94. G.E. 1973. Thermal imagery of Vanua Levu and , Fiji. Can. Aero. Serv. Ltd., Rep. unpub. REFERENCES Munro, J.N. 1961. The use of geothermal heat to produce salt. Mines Dep. Fiji, Rep. Anderson, J.P. and Austin, W.H. 1979. Geothermal unpub. study of the Fiji Islands. Rep, t o Fiji Rodda, P. 1967. Outline of the geology of Viti Min. Resour. Div., unpub. Levu. N.Z. J. Geol. Geophys., 10, Colley, H. 1975. A report on the thermal imagery 1260 - 1273. of Vanua Levu and Taveuni. Miner. Resour. Rodda, P. 1979. Hot springs of Vanua Balavu. Div., Fiji, Note unpub. Rep. 10, Miner. Resour. Div., Fiji. Cox, M.E. 1978. The Lambasa a r e a geothermal in- Williamson, K.H. 1980. An appraisal of the geo- vestigation, Fiji. Geotherm. Resour. thermal prospects of Fiji. Inst. Geol . Counc., Trans., 2, 121-123. Rep. 50 pp, unpub. Cox, M.E. 1980, a. Geothermal occurrences in the Wilkes, Commodore. 1845. Narrative of the U.S. Southwest Pacific. UN ESCAP, Exploring Expedition, 1838 -1842. 3 Tech. Bull. 3, 197-219. 196-199. Cox, 1980, b. Geothermal investigations M.E. in Woodrow, P.J. 1976. Geology of southeastern the Savusavu area. Geotherm., Rep. 5, Vanua Levu. Bull Miner. Resour. Min. Resour. Dept., Fiji, Fiji, 4. Cox, M.E. 1980, c. Chemical description of ther- mal waters in the Fiji Islands. Geotherm Resour. Counc., Trans, 4, 153 - 156. Cox, M.E. 1980, d. Preliminary geothermal investi- gations in the Labasa area, Vanua Levu. Geotherm., Rep. 2, Min. Resour. Dept., Fiji, Cox, M.E. 1980, e. The Lambasa geothermal investi- gation, Part - geochemsitry. Geotherm. Rep. 4, Min. Resour. Dept., Fiji , Cox, M.E. 1980, f . Preliminary geothermal investi- gations in the Savusavu area, Vanua Levu. Geotherm. Rep. 1 , Min. Resour. Dept., Fiji, Cox, M.E. 1981, a. The Lambasa geothermal investi- gation. Part I geophysics. Geotherm. Rep. 3, Min. Resour. Dept., Fiji., Cox, M.E. 1981, b. An approach to problems of a geothermal mercury survey, Puna, Hawai . Geotherm. Resour. Counc., Trans. 5, 67 -70. Cox, M.E. and Hulston, J.R. 1980. Stable isotope study of thermal and other waters in Fiji. N.Z. 23, 237-249.