Wesnousky.1988 - LETTERSTO NATURE ' trim/-1. - . - - B....

Info iconThis preview shows pages 1–4. Sign up to view the full content.

View Full Document Right Arrow Icon
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 2
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 4
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: LETTERSTO NATURE ' trim/-1. - . - - B. 991. m-sl/ /) (Sample depthg [n I 68°N 66°N 64°N . Baltic Proper 209,000 km? 13,500 km3 - SADN r ,: .,_ 51.°N u._1__.L..-_.1__L_1.'_l L 1 1 1 1 1 1 .L ‘-__1 1 1 10°E 15°E 20°E 25°E Fig. 1 Technetium concentrations in Baltic seawater collected in October—November 1986. limit implies that 99Tc/ 137Cs was less than 5 X 10‘“, again con- sistent with the above. Salo et al.19 have calculated the global fallout inventories of 90Sr and 137Cs in the Baltic Sea. The calculated inventory in any year was obtained20 by adding the contributions from direct deposition on the sea surface (precipitation), run-off (with rivers) and inflow through the Danish Straits, and subtracting the outflow through the Straits and the radioactive decay. If the subtraction of decay is omitted, we obtain the inventory that would pertain if the radionuclides (9°Sr or 137Cs) had been stable and thus comparable with 99Tc, which in practice behaved as a stable isotope during the period of observation. As the sedimentation and runofl characteristics of 99Tc are closer to those of 90Sr than to 137Cs (refs 2, 21), we used the 90Sr data. In 1981 Salo et al. estimated a 9OSr inventory of 552 TBq; this corresponds to an undecayed amount of 811 TBq 90Sr. According to UNSCEARZZ, the 137Cs/90Sr ratio in global fallout is 1.6; therefore, the 99Tc/9OSr ratio is 1.6 times the 99Tc/mCs ratio in undecayed global fallout. Hence we estimate the inventory of global fallout-derived 99Tc in the Baltic as 0.2 TBq. The total inventory of 99To in the Baltic Sea was calculated as follows. In the Gulf of Bothnia (comprising the Bothnian Bay and the Bothnian Sea) we assumed no halocline. Thus we used the mean of all eleven measurements, 0.060 Bq 99Tc m’3, multiplied by the water volume, 6,700 km3 (ref. 19) giving an inventory of 0.40 TBq 99Tc. In the Baltic Proper (see Fig. 1) and the Gulf of Finland there is a halocline at ~50m depth. The water volumes above and below the halocline are 11,100 km3 and 3,500 km3 respectively. From Table 1 we calculated the mean of the eleven samples from the Baltic Proper with a sampling depth $50 In and of the eight samples collected below 50m. These means, 0.050 and 0.076 Bq 99Tc m‘3 respectively, were multiplied by their respective volumes, giving 0.55 TBq 99Tc above the halocline, 0.27 TBq below and a total inventory of 1.22 TBq 99Tc. The latter is similar to the mean of all 30 samples (excluding those from the Cattegat) multiplied by the total volume, which gives 1.28 TBq 99Tc. Our best estimate of the inventory is 1.3 TBq 99Tc, and thus our estimate of the contribu- tion from reprocessing is 1.05 TBq 99Tc. As it takes 5—6 yr for discharges from Sellafield to enter the Baltic Proper”. this NATURE VOL. 335 22 SEPTEMBER 1988 amount of 99Tc represents discharges up to 1980. Compared with the 0.6 PBq 99Tc discharged from Sellafield from 1970 to 1980333, the 99Te inventory in the Baltic Sea from reprocessing amounts to ~0.18% of the total release. As the Baltic was the sea where maximum water concentration of 99To was most likely to occur after the Chernobyl accident, and as we were unable to detect any additional 99Tc in the Baltic sea water after the accident, we conclude that the Chernobyl accident did not contribute significantly to 99Tc levels in the north-eastern Atlantic Ocean. Discharges from nuclear repro- cessing in western Europe remain the dominant source of this radionuclide. This project was supported by the CEC Radiation Protection Programme. Dr Chen had been awarded a fellowship under the International Atomic Energy Agency. We thank the F/ S Gauss and its crew for hosting us during the cruise to the Baltic Sea. Received 7 March; accepted 3 August 1988. 1. ICRP Publication 38 (Pergamon, Oxford—New York»Frankfurt, 1983). 2. Aarkrog, A. et al. Estuar. Coast. ShelfSei. 24, 637—647 (1987). 3. Jefferies, D. F., Steele, A. K. & Preston, A. Deep Sea Res. 29, 713—738 (1982). 4. Kautsky, H. Dt. hydrogr. Z. 40, 49—69 (1987). 5. Livingston, H. D., Bowen, V. T. & Kupferman, S. L. J. mar. Res. 40, 1227—1258 (1982). 6. Devell, L., Aarkrog, A., Blomqvist, L., Magnusson, S. & Tveten, U. Nuclear Europe V1, 16717 (1986). 7. Tech. Rep. Ser. 247: Sediment de and Concentration Factorsfor Radionuclides in the Marine Environment (Int. Atomic Energy Agency, Vienna, 1985). 8. Holm, E., Rioseco, J. & Mattsson, S. in Technetium in the Environment (eds Desmet, G. & Myttenaere, C.) 61768 (Elsevier, 1984). 9, Meereskudliche Beobachtungen und Ergebnisse Nr 62 (Deutsches Hydrographisches lnstitut, Hamburg, 1987). 10. Buesseler, K. O. et al. Nature 329, 825—828 (1987). 11. Holm, E., Rioseco, J. & Garcialeon, M. Nucl. Instrum. Meth. Phys. Res. 223, 2044207 (1984). 12. Chen, Q, Aarkrog, A., Dick, H. & Mandrup, K. Risa-M4671 (Risa National Laboratory, 1987). 13. Aarkrog, A. et at. Risa-R640 (Risa National Laboratory, 1987). 14. Harley, J. H. HASL-300 (Environmental Measurements Laboratory, New York, 1972). 15. Till, J. E. in Technetium in the Environment (eds Desmet, G. & Myttenaere, C.) 1~20 (Elsevier, 1984). 16. Summary Report on the Post-Accident Review Meeting on the Chernobyl Accident Safety Series No 75-1NSAG-1 (IAEA, Vienna, 1986). 17. ApSimon, H. M., Macdonald, H. F. & Wilson, J. .l. N. J. Soc. radial. Protfi, 109—119(1986). 18. Environmental Health Series No 24: Health Hazards from Radiocaesium Following the Chernobyl Nuclear Accident (World Health Organization, Copenhagen, 1987). 19. Study of Radioactive Materials in the Baltic Sea IAEA-TEC-DOC-362 (Int. Atomic Energy Agency, Vienna, 1986). 20. Sale, A. & Voipio, A. Mar. Pollution Bull. 12, 218-224 (1981). 21. Blaylock, B. 6., Frank, M. L, Hoffman, F. O. & De Angelis, D. L. in Technetium in the Environment (eds Desmet, G. & Myttenaere, C.) 79—89 (Elsevier, 1984). 22. UNSCEAK Ionizing Radiation: Sources and Biological Eflects (United Nations, New York, 1982). 23, BNFL Annual Reports on Radioactive Discharges and Monitoring ofthe Environment (British Nucl. Fuels Ltd, Risley, 1978785). Seismological and structural evolution of strike-slip faults Steven G. Wesnousky Center for Earthquake Research and Information, Memphis State University, Memphis, Tennessee 38152, USA The mapped traces of strike-slip faults are commonly characterized by discontinuities that appear as steps in map-view. Here I present observations to show that the number of steps per unit length along the trace of major strike-slip fault zones in California and Turkey is a smoothly decreasing function of cumulative geological offset. When coupled with a growing body of evidence that indicates that steps in fault traces work to impede or arrest the propagation of earthquake ruptures, the apparent smoothing of fault traces with displacement is interpreted to suggest that the spatial distribution of' strength properties on a fault plane is a function of cumulative geological offset. A consequence of this structural evolution is that faults may undergo a seismological evolution a well, whereby the size and frequency distribution of earthquakes is also a function of cumulative offset. An earthquake in California or other region of similar tectonic style does not generally rupture the whole length of the fault NATURE VOL. 335 22 SEPTEMBER 1988 SAN ANDREAS? OFFSET 95, 250 km; LENGTH=1000 . ' 4 Pt Arena to Belmont , l ” ‘ l n 4. I966 M6) Park um San Juan Boutista Cholame to Little Rock Reservoir D 10 20 30 40 50 GARLOCK FAULT? TOTAL OFFSET—:64 km; LENGTHZZQO km Step'~3km\ _U<——- \. _.. F / 'emom van" WHITTIER-ELSINORE FAULT ZONE: TOTAL OFFSETle-‘lokm; LENGTHflBOkm ' Step~3km 0 \0 20 30 40 50km L_A_A_I_J_l snep~3xm / .u. o 10 20 so 53'"§'o'nm I_1_1__;l_l CALAVERAS'GREEN VALLEYiTOTAL 0FFSET§Z4km;LENGTH=220km HAYWARD-ROGERS CRK -MAACAMA-TOTAL OFFSETz‘PLENGTl-hZOOkrn “\9 ~ 0'5.” o I) 20 30 no son.“ s " {1711/ ‘svenfizxm L—‘—“"‘ b“ l / “-— -/‘__,_ ‘smpznm Sign : 7km 5km "£U/avem5 t [alyykvhljjj \ w_ v 7 V // Step 73km? “no. v.5» any -, .M u a word 1* -" ; Manager—w"- Rodgers unfit-:- ‘ v. -"’" ’ 1‘". / \‘“ ‘ $2972er ‘ ‘ ' (M52 -ME 5) 0 1D 20 30 40 50km |__A_..|_A_.|_J 2.2;; l9l6l 6.3) y "\_a—--.» i “.5 __... Svep~4km - "m / Hemet Stnp~4km ‘N -lN LEWOODITOTAL OF FSETzl-lOkm; LENGTH=60km “man PACIFIC OCEAN an at . . . . ~ - - ~ * 1933(M6.3) Fig. 1 Maps of the San Andreas”‘”, Garlock”, Whittier- Elsinorez“, Calaveras-Green Valley”, San Jacinto“ and Newport- Inglewood13 fault zones of California showing the location of steps in fault traces which are characterized by stepover widths W52 1km. Segments of fault which have ruptured during historical earthquakes are marked by hachures, stippling or brackets, and the dates and magnitudes of the respective earthquakes. Epicentre of 1933 Long Beach earthquake is marked by star and half-sided arrows indicate sense of displacement along faults. on which it occurs but, rather, only a fraction of the entire fault length‘. Observations also show that the endpoints of strike-slip earthquake ruptures are commonly associated with step-like discontinuities along mapped fault traces“. When slip occurs along a fault during an earthquake, stress concentrations will occur at step-like discontinuities along the faultz. Such discon— tinuities are commonly referred to as releasing- or restraining- steps, depending on whether the stress concentration is extensional or compressional in nature“, respectively, and the severity of the stress concentration will generally be a function of the width WS of the step as measured perpendicular to fault strikez. The stress concentrations at steps may be sufficient to impede or stop the propagation of an earthquake rupture and, on that basis, a number of investigators have attributed a causal association to the observed coincidence between the endpoints of earthquake ruptures and the location of steps along mapped fault traces”. In turn, laboratory models and scattered field LETTERS TO NATURE 341 O ‘ ~ .ESESI C \1939— a ’ V ERZINCAN KURSUNLUO 50km Fig. 2 a, The North Anatolian fault shows right-lateral movement and strikes westward across Turkey from a point near Karliova to west of Bolu. Larger-scale and overlapping strip maps of the fault (b-e) are taken from Barka and Kadinsky-CaseS and show location of steps greater than about 1 km in size. Dates and arrows show the year and extent of rupture during large historical earthquakes, respectively. observations have been the basis to suggest that, during the initial stages of development, strike-slip faults are characterized by a zone of discrete fault segments, each of which is separated from the next segment by a step, and that, with continued displacement, the individual fault segments ultimately coalesce to accommodate displacements along a principal and through- going fault“. The few observations available thus point to a close connection between the seismological and the structural evolution of fault zones and provided the impetus for this work: to examine whether or not there exists a systematic relationship between frequency of occurrence of step-like discontinuities and the total geological ofiset registered along the trace of active strike-slip faults. Strike-slip fault zones and earthquakes in California and Turkey are the principal basis for my observations. It is for these regions that the most comprehensive maps and data concern- ing the total geological offset and the location of historical earthquakes along the trace of active faults are available. More specifically, I have relied on the published literature to obtain estimates of the cumulative geological offset across the San Andreass, Garlock9, Whittier-Elsinore”, Calaveras-Green Val- ley“, San Jacinto12 and Newport-Inglewood13 faults of Califor- nia and the North Anatolian fault zone14 of Turkey. Similarly, the magnitude and rupture length of historical ruptures along strike of each of the fault zones has been the source of numerous previous studies”. Thus, the location and magnitude of his- torical earthquake ruptures, the total fault length, the cumulative geological offset and the locations and sizes of step-like discon- tinuities along strike of each of the fault zones considered have been detailed in a separate report15 and are summarized in the accompanying maps (Figs 1 and 2). Because of the scale of 342 NATURE VOL. 335 22 SEPTEMBER 1988 LETI'ERSTO NATURE STEE_ END OF I :~ 0 1 Ir - _ — _ _ _ — 1 RELEAsINe Resraammc FAULT IE ' 1 I1) Newport—Inglewood I END or RUPTuaE ) L l O 5 . : 2) WhittierI‘Elsinore I WITHIN auvruas . I 2 g .3 I 3) San Jacmto : ° 5 4) Garlock = § 5 .7 : ,c , : 5% E 5 2 t 0.01 2 I5 aaveras I E m g g 3 g 16) San Andreas l a. a - a l s I o — m ._ . 5% g g ‘2} g 4 L7.) El'érlaloflal‘ _ _ _ _ _ _. an I®l '— I 1,, Q G! I '7 I ‘ 5% 000 6 I U) . I I @Q as Step width W5 (km) 41 "*"" ‘1942 ANATOLIAN ‘4 4 1 966 PARK FIELD 5 0 K} 1 857 CALIFORNIA 1943 ANATOLIAN (1. 100 200 500 hOO Earthquake rupture length (km) Fig. 3 Earthquake rupture length versus the size Ws of steps in mapped fault trace observed at the endpoints (open symbols) and within (solid symbols) the rupture zone of large Anatolian and California earthquakes. Half-filled symbol indicates that steps of similar size occurred both within and at endpoint of rupture. Steps in fault trace are also subdivided according to whether they are restraining (squares) or releasing (circles) in nature. available maps, I have limited attention to steps in fault trace characterized by stepover widths of a kilometre or greater. Figure 3 illustrates the relationship of the size and location of steps along the traces of strike-slip faults to the length and location of earthquake ruptures. The figure is a plot of the rupture lengths of large California and Anatolian earthquakes versus the stepover width Ws for each step in the fault trace observed at both the endpoints and within the respective earthquake rupture zones. For example, Fig. 3 shows that the 1939 Anatolian earthquake produced ~320 km of surface rup- ture and that the endpoints of the rupture are associated with steps of about 5 and 10 km respectively. The figure also shows that the same rupture propagated through steps of 1 and 4 km; that is to say, the steps located within the 1939 rupture zone are smaller than the steps associated with the termination of the rupture. A similar pattern is found for the other earthquakes in Fig. 3, that is, steps associated with the endpoints of ruptures are commonly the largest discontinuities registered along the length of the respective rupture zones. The trend, admittedly, is not without exception. The 1943 Anatolian earthquake is a case in point, where a step interior to the rupture zone is larger than a similar step observed at the endpoint of the rupture. It is also apparent that the endpoints of all ruptures are not marked by steps in mapped fault traces, as evidenced by the southern limit of the 1857 California earthquake (Fig.1). Nonetheless, the observations are certainly sufficient to argue that the associ- ation of steps with the endpoints of earthquakes is not merely coincidence. Rather, the stress concentrations resulting from the discontinuities appear to play a principal and causal role in limiting the extent of earthquake ruptures. A number of recent studies have attempted to explain charac- teristics of the spatial, temporal and size distributions of earthquakes along faults by assuming an arbitrary or fractal distribution of strength properties along fault planesz7'3‘. The observations I have summarized thus far suggest that the so 100 150 200 250 Total offset (km) Fig. 4 The number of steps ( Ws 2 1 km) per unit length of mapped fault trace versus cumulative geological ofiset along major strike- slips faults in California and Turkey. geometrical complexity of fault traces may be used to gauge, in a general sense, the heterogeneity of strength properties along fault planes. Toward that end, the number of steps (IVS? 1 km) per unit length of a fault zone provides a simple, quantitative measure of fault trace complexity. The complexity of fault trace measured in this manner is plotted as a function of cumulative geological offset in Fig. 4. Fault trace complexity varies by nearly two orders of magnitude within the data set and is a smoothly decreasing function of cumulative geological oflset. The observation supports the idea that fault zones are characterized by a discontinuous trace during the initial stages of development and that increasing displacement along such faults works to remove discontinuities along the fault trace, efiectively smooth- ing the fault plane. Moreover, when coupled with the growing body of evidence that indicates that steps in fault traces work to impede or arrest the propagation of earthquake ruptures, the apparent smoothing of the fault trace with displacement may be further interpreted to indicate that the distribution of strength properties on a fault plane is spatially heterogenous, and that the degree of heterogeneity is a function of cumulative oflset. A consequence of this structural evolution is that faults may undergo a seismological evolution as well, whereby the size and frequency distribution of earthquakes on a fault is also a function of cumulative geological offset. I thank A. Barka and A. Johnston for reviews of the manu- script, A. Barka, T. Rockwell and D. Herd for preprints of unpublished papers and maps, Tanya George for drafting the figures and the NSF for supporting the research. Received 26 May; accepted 12 August 1988. Wesnousky, S. G. J. geaphys. Res. 91, 12587-12631 (1986). Segall, P. & Pollard, D. D. J. geophys. Res. 85, 4337—4350 (1980). Sibson, R. H. Geaphys. Monogr. 37, 157-167 (Am. Geophys. Un., Washington D.C., 1986). Crowell, J. C. Soc. econ. Paleontologists and Mineralogists spec. Pop. 22, 190-204 (1974). Barka, A. A. & Kadinsky-Cade, C. Tectonics (in the press). Wilcox, R. E., Harding, T. P. & Seely, D. R. Bull. Am. Ass. Petrol. Geol. 57, 74-94 (1973). Withjack, M. 0. & Jamison, W. R. Tectonophysics 126, 99-124 (1986). Hill, M. L. Geol. Soc. Am. Bull. 92, 112—131 (1981). Smith, G. 1. Bull. Am. Assoc. Petrol. Geol. 46, 185—104 (1962). Lamar, D. L. 8: Rockwell, T. K. in Neotectortics and Faulting in Southern California, Guidebook and Volume 82nd Annual Meeting of the Cordilleran Section of the Geol. Soc. Am. 149—158 (Los Angeles, California, 1986). 11. Kintzer, F. C., Brooks, E. R. & Cummings, J. C. Geol. Soc. Am. Abstracts with Programs 9, 6S (1977). 12. Sharp, R. V. Geol. Soc. Am. Bull. 78, 705—730 (1967). 13. Barrows, A. G. Spec. Rep. Califl Div. Mines Geal. 114, 115 (1974). 14. Barka, A. A. & Gulen, L. Spec. Publs Middle East. tech. Univ. (in the press). 15. Wesnousky, S. G. Proc. USGS Workshop on Fault Segmentation and Controls of Rupture Initiation and Termination A USGS Open-File Report (in the press). 16. Clark, M. M. USGS Misc. Field Invest. Map. [-1483 (1984). 17. Hope, R. A. USGS Open-file Rep. 69—130 (1969). 18. Ross, D. C. USGS Misc. Field Invest. Map [-553, (1969). 19. Vedder, J. G. & Wallace, R. E. USGS Misc. geoI. Invest. Map I—574 (1970). 20. Brown, R. D. USGS Misc. geol. Invest. Map 1—575 (1970). 21. Brown, R. D. Jr & Wolfe, E. W. USGS Misc. geoL Invest. Map [-692 (1972). 22. Matti, J. C., Morton, D. M. at Cox, B. F. USGS Open-file Rep. 85-365, 23 (1985). 23. Clark, M. M. USGS Misc. geol. Invest. Map 74, (1973). 24. Anderson, J. G., Rockwell, T. & Agnew, D. Earthquake Spectra (in the press). 25. Herd, D. G. USGS Misc. Field Stud. 1 : 250,000 Map of Recently Active Faults of San Francisco Bay Region and Surrounding Northern California Coast Ranges (in the press). 26. Sharp, R. V. Bull. Calif Div. Mines Geol 196, 187—194 (1975). PPPN9P‘PPNI‘ _. NATURE VOL. 335 22 SEPTEMBER 1988 27. Lay, T. & Kanamori, H. in Earthquake Prediction: an International Review (eds Simpson, D. & Richards, P.) 579—592 (Maurice Ewing Series 4. Am. Geophys. Un., Washington, D,C.. 1981). 28. Andrews, D. .l. J. geophys. Res. 85, 3867—3877 (1980). 29. Aki, K. in Earthquake Prediction - An International Review (eds Simpson, D. & Richards, P.) 566—574 (Maurice Ewing Series 4, A. Geophys. Un., Washington, DC, 1981). 30. Scholz, C. H. & Aviles, C. A. in Earthquakes Source Mechanics Geophys. Monogr. 37, 147-156 (eds Das, S., Boatwright, J. & Scholz. C.) (Maurice Ewing Series 6, Am. Geophys. Un., Washington, DC, 1986). 31. Smalley, R., Turcotte, D. & Solla, S. J. geophys. Res. 90, 18944900 (1985). An experimental determination of primary carbonatite magma composition Margaret E. Wallace & David H. Green Geology Department, University of Tasmania, GPO Box 252C, Hobart, Australia 7001 Carbonatites are uncommon carbonate—rich rocks usually found in continental intra-plate regions and often associated with rifting. There has been much debate as to whether carbonatite magmas are primary melts derived from partial melting of mantle peri- dotite, or are formed by exsolution of an immiscible carbonate melt fraction from phonolitic or nephelinitic magmas. Our experi- ments on the phase relationships of carbonate and amphibole- bearing peridotite (containing 0.3% H20 and 05—25% C02) show that sodic dolomitie carbonatite magma coexists with an amphibole lherzolite assemblage in a field ranging from 21 to 30 kbar and 930 to 1,080 °C, spanning a pressure and temperature interval between the solidus and the amphibole breakdown and melting curve. Thus primary carbonatite melts may occur under suitable geothermal conditions. The nature of the peridotite solidus and of the melting reactions differ considerably from published modelsl‘3. The carbonatite melt composition, determined by a series of ‘sandwich’ experiments, was found to be rich in Na, Mg, Ca and Fe, with a small dissolved silicate content. This melt quenches to an assemblage of dolomite and Na—Mg carbonate minerals, producing textures similar to those preserved in samples from Oldoinyo Lengai“, Homa mountains, Tanzania‘ and Kaiser- stuhl, Germany‘s. The majority of carbonites are rich in calcium and are com- posed of calcite (the calcite carbonatites, sovite and alvikite) and dolomite (beforsites or dolomite carbonatites) with minor amounts of ferroan carbonate. Cation abundance is in the order Ca> Mg> Fe> (Na+ K) (ref. 7). These minerals characteristi- cally contain high contents of Sr, Ba, Nb and rare earth elements, and are associated with ijolites and other alkali—rich igneous rocks, and commonly are surrounded by large sodic or potassic fenitization halos. A particularly important type of carbonatite has been erupted by Oldoinyo Lengai, northern Tanzania“. It is highly alkalic (Table 1, column 12) and is composed dominantly of NaZO, K20, CaO and C02, which form the mixed carbonate minerals nyerereite (Na0_22K0_,8)2Ca(CO3)2 and gregoryite (Nam,K0,(,5)2Ca0,17C03 (ref. 8). Several other car- bonatite lavas, now composed largely of calcite, have been interpreted recently as originally having had similar mineralogy9, subsequently replaced by reaction with calcium-rich ground water4’5’9. It is thus argued that alkali—rich carbonatite magma occurred more frequently than is indicated by the present relative abundance of calcitic carbonatites and natro-carbonatites. Isotope studies demonstrate that carbonatitic magmas are derived from mantle source areaslo'”, but do not distinguish between origins by partial melting of the mantle or by exsolution from a phonolitic or nephelinitic magma7"3"8. Experimental investigations of carbonatite genesis have explored the Ca0- MgO-SiOz-COZ—HZO system, its subsystems and other systems involving alkali elements with C02 and H2014'19‘21. This exten- sive literature has been summarized by Wyllie22 and Eggler”. LETTERSTO NATURE 3” Pyrolite + 0.3%H20 + (0.5-2.5°/oC02) 30 Gt Lhz + 01 melilitite ’C O .n x a) 25 L 3 Amph Lhz + 8 Carbonatite E Sp Lhz o n- 232:?” 01 Nephelinite 20 Amph Lhz CD Sub—solidus + Vapor mAbove solidus 15 \\Amphibole Carbonate 900 1,000 1,100 1200 Temperature ( ° C) Fig. 1 Phase relationships of amphibole-rich Hawaiian Pyrolite (0.3% H20) plus 5% dolomite or 1.4% magnesite. Carbonatite melt field is bounded by the carbonate solidus at 930 °C (heavy line) and amphibole breakdown and silicate melt field above 1,080 °C (25 kbar). Size ofthe box symbols indicates the uncertainty in pressure and temperature. All runs were buffered with magnesite ([02 : EMOG). These earlier studies confirmed the existence of low-temperature alkali-rich carbonate melts, explored liquid immiscibility between silicate and carbonate-rich melts, and documented sta- bility fields for carbonated peridotite under upper-mantle condi- tions. Other studies have investigated the solubility of C02 as (CO?) in silicate melts at high pressure24'25 and have empha- sized continuity at high pressures between kimberlite, olivine melilitite and carbonatite magma compositions. Although there is thus a generally accepted framework for carbonatite genesis in the upper mantle, there is a continuing debate on the relative importance of crystal fractionation (for example, to yield Oldoinyo Lengai natrocarbonatite magma as a differentiate from parental calcitic or dolomitic carbonatite”) and liquid immisci- bility (for example, to yield Oldoinyo Lengai natrocarbonatite magma as an exsolution from parental carbonated nephelinite or phonolite magma”). The specific problem of accurate deter- mination of both the solidus temperature ofa carbonate-bearing lherzolite and the nature of the melt at the solidus at high pressures has frustrated experimental petrologists and generated much debatel‘3’22'27. We have clarified the position of the peri- dotite solidus and we demonstrate that, because of the import- ance of the minor components Na, K, P and Ti, primary mantle melts of sodic, dolomitie carbonatite character can be produced under oxidizing conditions from amphibole and carbonate- bearing lherzolite. Phase relationships for (amphibole pyrolite minus 40% olivine27 plus 1.4% MgCO, or 5% CaMgCO3) are presented in Fig. 1. The (amphibole pyrolite minus 40% olivine) starting-mix was prepared in 100-mg batches27 and all phases analysed by microprobe. The peridotite composition used represents a vapour-undersaturated ‘fertile’ mantle peridotite, in which par- ...
View Full Document

This note was uploaded on 07/23/2008 for the course GEOSC 508 taught by Professor Marone during the Spring '07 term at Pennsylvania State University, University Park.

Page1 / 4

Wesnousky.1988 - LETTERSTO NATURE ' trim/-1. - . - - B....

This preview shows document pages 1 - 4. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online