A chemical comparison of oceanic ridge, Hawaiian tholeiitic and Hawaiian alkalic basalts

A chemical comparison of oceanic ridge, Hawaiian tholeiitic and Hawaiian alkalic basalts


334KB Sizes 0 Downloads 24 Views


A CHEMICAL COMPARISON OF OCEANIC RIDGE, HAWAIIAN THOLEIITIC AND HAWAIIAN ALKALIC BASALTS * Norman J.HUBBARD L~mont Geological Ob~en,atory of Cblumbia Unit,er~ity. Palisades. New York, 109fJ4 Received I0 December 1968 Abundance da~a (AI203, TiP 2, St, Zr and REE) for two important groups of basalts (oceanic ridge and Haw~tlian) are compared. Tlw average eoncentrationu of AI203, TiP 2 and Zr are quite diff,:rent for the two groups but extensive ~verlap also occurs. TiP 2 and Zr are inverst,Py covariant with AI203 for oceani~ ridge and I-lax~aiianbas',dts. The AI203 and TiP2 data are consistq.-nt wl Lh the experimental data of Green and Ringwt.~,a [5 q and MaeGregojr[7'51 and s~agg~st that file eoncentr.atJons of M203, TiP 2 and Zr in a wide range of basalt types are dependent on the depth of magmasegregation. The two basaltic groaps ~'.e ~dmostcomplelely different in their Sr concentration and REE abundance patterns. The concentrations of Sr and tile REE abundance patterns are essentially independent of the concentrations of AI203, l'iO2 and Zr. The Sr and REI" t~ata can be qualitatively explained by assuming vat]zingdegrees of partial melting to form rile diJ['ferentliquids ob~elrved at the surthee. i. Introduction The chemical composition and petrography of Hawaiian lavas llJave played a significant role in the classification of volcanic rocks arLd theories of their origin [11-5]. Recent investigations of the ocean floor indicate th:at volcanic rock~, at.~ much more abundant in oceanic regions tha~ might have been inferred from the occurrence of oceanic islands such as Hawaii [ 6 - 8 ] . Engel etal. [6~ have suggested that the volcanic rocks associated with the midocean ridges replresent the paren l liquid for a variety of more diverse volcanic liquids foutid in the central vent volcanic ce~nters of oceanic regions (i.e., volcanic islands). In order to ascertain the petrogenetic relations of the mid-ocean ridge vol,=anics and the Hawaiian volc~lnic rocks in more detail, certain cl'lemical characteristics of a large number of H~lwaiian and mid-ocean ridge basalts will be compared i~a this paper. This comparison will require objective criteria that distinguish the rock groups that ate being compared. Such criteria are particularly important for the mid-ocean ridge ba~lts, since they come from widely separated regions ~.n wl'fich individual rock units cannot at present be Lamont Geological Observatory Contribution no. 1229.

related in stratigraphic and structural terms. Several authors have described volcanic rocks dredged from oceanic ridges in the Atlantic, Pacific and Indian Oceans [6, 9 - 1 2 , 1;4, 15]. These rocks have been variously designated as oceanic tholeiites [6]. abyssal basalts [ 16], subalkaline basalts [ 17] and high alurnina-olivine tholeiit,~s [5]. Chemically they are characterized by olivine tholeiite normative compositions, and by low K and trace ele:ment eo:aeentrations. The similarities between these rocks and some of the tholeiitic rocks of Hawaii have beert noted by other authors [5]. The nomencIature of submarine basalts is at present obscured lay attempts ~o include both essentially geographic and chemical characterisl:ics in a single descriptive term. In the di:geussion that follows, the existence of a dis~finct group of basaltic rocks broadly similar to the rocks defined by Engel etal. [6] will be accepted as a working hypothesis. However, in order to avoid biasil~g the comparison which fial~ows only geographical and geologic;d criteria will be used to delimit the rocks being compared. The recks of the Hawaiian volcanic province are well defined according to these criteria, since they are associated

COMPARISONOF OCF,A~'iICR!DGE AND HAWAIIANBASALTS with a particular set of volcanic islands (only basatts, both tholciitic and alkalic,, as defined by Maodon,;dd and Katsura [ 18] are includl,~d in this discussion). In addition, the great majoJdty of Hawaiian rocks studied by petrologists are extruded from the central v,,,'ntatrd associated rift systems of ]large shield volcanoe'~. By comparison, rocks ~'r,~mthe o,cean floor are a less well deft]ned group. The samples discussed below are associated with oceanic ridges and were all obtained by dredging. Most of these were clearly extruded underwater, i.e. are described as pillow lavas, commonly with glassy edges. In most cases, these ro,,'ks are not associated with large readily distinguish~ble volcanoes, but with widespread fissure ert~ptions. '~n the discussion that follow~ they will be dl,~signatedas oceanic rMge basalts. The choice of geugrapbic anei geologic crileria to distinguish different volcanic rock types is not meant to imply that chemical criteria are invalid or without value. In fact it will i~e shown in this paper that oceanic ridge basalt:~ have readily definable and perhaps unique cbemic:d characteristics.

2. Presentation of data The elements included in the present study are A1, Ti, Zr, St, and the rare earth elen~ents (REE). Cornparison of AI203, TiO2, Zr ~nd ti;~"concentrations will be made for a number of individual basalts. Comparison of the relative and absolu;'~-• abundances of the REE will also be made. The dmta for this study come from several sotatces, lVlacdcnald and Katsura [ 18] and Macdonald {19] have p~blished a large number of major element compositions of Hawaiian wdcanic rocks; tile AI~O3 and TiGi:~ concentralions u~,,ed tbr Hawaiian rocks are t:om t~t~se studies. Using interealibrated emissinn spectrographic and X-ray fluorescence techniques the author ~20] has de,totrained the Zr and Sr concentrations for approxirr~ rely three-quJters of the sample", described by M,aodonald and Katsura [ ~8] and Macdonald [I 9]. Engcl and Engel [9], Engcl etal. 16[', Nicholls [101, Nicholls etal. [ I I ], Muir and Tilley [ 121, Aunrento [ I ;3], Wiseman [ 14 ! and F'oldervaarl and Green [I15 ] have reported chemical data for a lar~!,enumber of basalts from oceanic ridges of the Pacific, Atlantic and Indian Oceans. All published emissior spectrographic Sr and Zr analyses were irejected unle,,is direct inte;-lab


comparisons bad been made. Samples repo~tcd ~n the literature were rejected unless fresh or nea'dy fresh, i.e., having generally low water contents and glassy selvages. Samples with ~dkaline affinities were alsc~ rejected. Both m,~jor and trace element contents lor a number of additional samples from the East Pacific. Juan de Fuca and the Gorda Rises have been made by the author as part of a general study of rocks t'rom oceanic ridges. Zr was determined by X-ray I'luo~escenee. Sr was determined both by X-ray fluorescence and mass spectronretry. Major elements, includi.nl~. AI203 arid TzO2, were determined by X-ray fluorescence using the method of Rose [211. Additional element abundances and analytical details will be published in a forthcoming paper. The new samples used in this st,ady are fresh basalts with glassy selvages. Both the new observatinns and data taken from the lit,e~rature are shown in figs. 1-4. Petrographic descriptions are availabl,: for almost all the oceanic ridge basalts included in figs. I 'to 4, Aphyric rocks ale distingl:fished from rocks containing macroscopic plagioclasc phi:notrysts in f~g.2. Zr and TiO2 concentrations of oceanic ridge arid Hawaiian ha:~a;ts are compared in fig. I. The da,a shown here demonstrate: I ) that there is a good tierrelation between Zr and TiO2 in these basalts, 7~1that the oceanic ridge basalts arc generally lower in Zr and TiO2 than the Hawaiian tholeiltic basalts, which are generally lower in Zr and TiO2 than Hawaiian a'ik~iic basalts, 3) that all three group:' of basalts have an approximately constant Zr/TiO2 ratio, 4) that zhere is considerable overlap between adjacent groups of basalts. A1203 and TiO2 concentrations of tire bas:~lts compared in fig. I are compared in fig. 2. The large number of major element analyses available for llawaiian rocks would dominate this figure if all individual analyses were shown, thevefure only the average and the ~verage deviatiolls of analyses from individual volcanoes are given, except for the Pololu s~- , ~'ies lavas of Kohala. The correlat ion between AI203 ;;md TiO2 concentrations is remarkably good. It is also seen that the high AI203 and low TiO2 ro,:ks t'~om oceanic ridges are commonly apiwtic. The AI203 and "rio2 co;icentratio:~s overhtp (or the tv,~ groups of tholeiitic I:~asalls. However. as ~ group, the oceanic ridge basalts are higher in AI203 and town' in TiO2 than Hawaiian tholeiitic basalts. Tile overlap






.L: >~. ,s~s s= ~

~ ,r- •



Oceop¢ Ridqe I~solfs

Ho~Kllian th~ellt,C l ~ l l l ~ii~n

o~olc bosol~s

// l


@ ~ - - I

,O l l

1I ~ ' l

Iz Q'

l l 12:~ l l l l s:0l

l l l~ 5

l l l,~ I I l 4 5

wt % T~Ol

Fig. 1. 'fie2 versus gr for oceanic ridge, H~waiian tholciitic and Itawaiinn alt:aii¢ basalt:;, Data are from most of/he sources listed in the introduction.



/;VERAGE TItOI.EIITIC BASALT$ OF T H ( HA~Y~AIIAN ISLANDS (~g. dev. ',~ho~n ~ Wa;mloe A, We~ ;#~ui I




~s~ s



~r :':;::: -- o

!,; s





'd~ ....

,b' "






'JI % TiO=

Fig. 2. Ai203 versus Tie 2 for oc~zmiicridge, Hawaiian tholeiitic and Hawaiian alkalie ha,salts, Da~a arc from the source.; listed in the intJrocluction.


IOOC 90C t






• ~,°° +°/


Ig" ° ~





i s



°o . . . . . .

Oceaf~lc I!~Klqe I:+=~llts

ss s s~s s s s s s



.'lO0 i

Opm Zr


~.'ig. 3. St versusZ'r ibr the same ba"~ts as in f't~. 1.!


with Hawaiian tholeiitic basalts and the wide rang,~ of AI~'O3 con~:ents of the oceanic ridge basalts show~ that the oceanic ridge basalts include a wider specl:rum of compositions than high [email protected] olivine tholeiitesl. Clearly, high A1203 basalts similar to the Warner Basalt [21 occur on oceanic ridges,, however, this ia not a genenll characterization of subnlarine voleailJic rocks from oceanic ridges. There is nc reason from the present data to ascribe 16.0.-19.0 wt% AI203 in oceanic ridge ba,';alts to the accumulation of phgioclase feldspar, since such rocks may or may not have plagioelase phenoerysts. Sr and Zr contents of Hawaiian and oceanic ridge basalts are compared in fig. 3. It is seen here that there is very litde correlation between these two elemer~ts within the two rock types being eumpared, particularly within the oceanic ridge basalts. ~oreover, the Sr contents of the ocean!c ridge basalts and Hawaiian basatts do not overlap. The oceanic ridge basalts have Sr cot~centrations of less than 200 ppm ',and Hawaiian basalts h~ve Sr concentrations greater than about 250 ppm, There ar~ a few basalts frem oceanic ridges which h~tveSr concentrations falling within the range of Hawaiian basalis, and vice versa. These basalts (not plotted on fig. 3) are only I% to 2% of the total sample po,pulation and those from Hawaii are the result of extreme olivine accumulation. The oceanic ridge basalts wi~h


i!o " 6"7 ] ATOMICNUMBER Fig. 4. Ra.le earth abundance patiarn~ "or ocean ( r dg.c!)asalts,alkalie habits and Ilawaiiantlloleiitiebasaits, Data for atkalic busalts (includingIlawaiian)are from Hubbard and ,'3ast [;!'31, for oceanic ridgebasalts from Kay (pr~v~ e co nmunie~ o ) nd Hubbard and Gast {23l. and for II~+waiiantholefitie basalts from Schilling ]24J and :rey el a. [22L



atypical Sr eonten(s are distinguished from "'normal" oceanic ridge hasalts by other Llnustla] chemical char* acteristics. They will be diset:ssed elsewhere. Frey et al, !22] have shown that the oceenic ridge basatts have almost unique relative rare earth abundances. Kay (personal communication. 1968) and Gust and Hubbard (unpt~blished results) have analyzed more than 20 additional samples of oceanic ridge basalts. A summary of these new data is given in fig. 4. Data for atkalic basalts from oceanic islands have been taken from Frey et at,, [22], from an un,, published study [23] and are also given in t~g. 4. The ranges of rare earth abundar~ces for these two rock types are compared with data given by Schilling [24], Frey et at. [22] and Hubbard (unpublished results) for ilawaiia," volcanic rocks. The Hawaiian tholeiitic basalts appear to be systematically different from oceanic ridge basalts even when they have similar AI203 and Tie2 concentrations, including the Kohala basalts. Fig. 4 illustrates that the REE distributions of the Hawaiian tholeiitie basalts fall in a position that is intelmediate between alkalic basa]tS, including Hawaiian, and oceanic ridge basalts. TILeabundance patterns of Hawaiian tholeiitic basalts haue features in eoml'non with both the oceanic ridge basalts and the alkalic basa]ts. They have the light REE enriched patterns and low Yb corrceutrations of the alkalic bas:,lta and the slight relative depletion of La and Ce that is common for the oceanic ridge basalts.

3. Discussion The observations made here Show that the use of combined minor and major etement abundance pa~terns is a fruitful, if not nece:isari~,,approach in the chemical characterization of basaltic volcanic rocks. The observed variations are in agreement with several genetic interpretations bused on recent experimental studies. M~gGregor [25] has demonstrated that the "]['iO2concen~.ration of a euteetic melt in the syslem MgO-l"iO2-SiO2 increases markedly with pressure. MacGregor suggests that the Tie2 content o. ,,. basaltic liquid may be correlated with the depth of melting. Green and Ringwood [5 ] have further shown that high A1203 liquids may be the result of equilibration of silicate liquid and solid, with moderate degrees cfpartial melting, at rather modest depths

and that the A1203 concentration can be controlled by the depth where this equilibration took place and by tlhe degree of// partial melting. The observed cartelotion of AI2 C'3 and Tie2 concentrattions is thus in accord with this experimel~tal work if one postulates ithat the range of compositions found represents liquids formed, over a wide range of pressures, Seismic observations [26] suggest that magmas of the Kilauea w:tlcan6 separate and accumulate at 60-70 krn depths. "/"he experimeutal work of Green and Ringwood [5j indicates that tile range of depths involved for the production of tholeiitie liquids may extend from > 60 to < 15 lcm in the oceanic mantle, if the seismic observations made on Kilanean eruptions can be used as a depth calibration, we can infer from the chemical observations made here that tholeiitic liquids can be segregated over the entire depth range from < 15 to "460 km.~lt can also be postulated that tboleiitic bas',dts are s'~ldom the products of silicate liquids that segregated at depths much in excess of 70 km because Kilauean basallts are among the most Tie2 rich tholeiitic basalts kuown. Green and Ringwood [5] have proposed that the Hawaiian alkali basalts may be produced from an olivine tho~eiite liquid by fractional crystallization. They may also be produced from the same parent rock by a lesser degree of partial melting. It\ however, the degree of silica saturation in a volcanic liquid is controlled by the pxessure (depth) at which silicate liquid and surrounding silicate solids last equilibrate [27]. the Hawaiian alkali basalts carl equally well be produced by partial meltin.~ probably at slightly greater depths than the Kilaue;,o b:~salts. Whatever the raechanisn'r tbr generating Elawaiian alkalic basalts, the AI203 and Tie2 data indicate that th,e pressure ,(depth) at which these zlkalic basalts were generated wa!; quite similar to tin.~ pressure (~epth) associated with the ;ener~tion of Kilauean tholeiitic basalts. Gasl 116] has shown that the extent of ~artial mel;.inlg can greatly influence the concentration and relative abL.,ndance of mar~y :race elemen, lts, The extent e,f this partial melting would have to be some. what less (8-15%) than that inferred for the Hawaiian tholeiific liquids, in order to explain the higher Sr content and mort: extensive REE fractionution observed in the alkali basalts. Following an original suggestion by Kenne,dy. Gust [1(~] s~geesred that the degree of partial met~Iingmay be limited by pressure. This suggestion is not supported by the data and interpretations presented in this paper.

COB.IPARISI,)NOF OCEANIC RIDGE AND, tIAWAffAN IIASALTS Ttte observed range o f Al and Ti ubu, ndance~ indi::ates that the oceanic ridge basalts are formed over a i~,tirly w,d~ range o f pressures. The REE patterns and Sr contents,;~re rather constant for all oceanic ridge basalts, Since these trace element,~ are expected to be marke:lly afIected by differen'c degrees of partial melting, i~ appears that the degree o f partial, melling for the oceanic ridge basalts is I:~rgely independe;rt o f the depth where melting takes place, However, t!otb the differ~mt REE patterns and Sr eoncentratioris o f the oct:ante ridge basalts versus Hawaiian bas:ill~s can be explained in term, s o f the degree o f patti:ill melting, i.e., > 2 5 % for the oceanic ridge basalt!~!, 15-20% for the Hawaiian ~holeiitic basalts and 8 - ] 5% for the H~waiizn alkali basalts. However, the d::ta e f Green and Ringwood indicate fl~at high AI20 3 olivine tholeiitic (oceanic ridge) basalts are genenated by lesser degrees o f par:.ial melting than Hawaiian tholeiites, just tt~c opposite of the situation gi'aen in Gast's parlial melting model. The close correlation of TiO 2 and Zr and lack o f corrc!atiot~ o f Sr and Zr suggests that pressure rather than the degree o f partial melting is the dtm~nant factor controlling the Zr content o f basaltic hqnids.

4. Conclusion The oceanic ridge basaits cover a wide range n f AI20 3 and TiO 2 contents ranging from 14% to over 19% A I 2 0 3 :md 0,.7% to 2.8% TiO 2. The AI20 3 and TiO 2 contents o f both Hawaiian and oceanic ridge basalts are correlated. This correlaUon and the magnitude o f the AI20 3 and TiO 2 contents are ascribed to the depth (pressure) at whiclt different tboleiili: liquids :are segregated. The Sr and REE content,,; distingu~,sh the oceanic ridge tholeiites and Hawaiian tholeiiies from each other.

Acknowledgements The author is indeblLed to Gordon A.Macdom,ld for his services as thesis adviso~ and teacher o f Hawaiian geology. Discussions with Paul W.Gast have provided much o f the background o f ideas and questions from which this paper has grown, particularly

35 t

concerning o,:eanic ridge basaits. Research completed at the University o f Hawaii was supported by NSF Grant GA-1958, Research perform,;d at Lamont Geological Observalory was supported by NSF Grant GA-1188.

References I 1] C.E.Tilley. Some aspects of mudmatic err, lotion, Quat. 1. C,eol, Soc. London 106 (1950l 37. 121 EI.S.Yodcr Jr. and C.E.TilIey. Origin of b~lsallmagmas: An experimental study of natural and syTitheticrock , sysl:ems, J. Petrol 3 ( 19621 342. 13i (.;.L.M:~cdonald ~nd "r.Katsura. RelaUon~hipof petrographic suites in Hawaii, in Ctttst of the ?acifi¢ Ocean Basin. Amer. Ge~:,ph.Un, Monog. I ~19f.2~ 187. {~.] H.A,Powers, Composition and origin of b~tsalticmagnl:~ in the Hawaiian Nands, Gcochim. Co~nlochim. Acta 7 (19.~5) 77. [51 D.ll.Creen and A,E.Ringwood, The genesis of basaltic maL~mas.Control. Mineral. and Petrol, 15 ( 1967l 103. {61 A.E.J.Engel. C.G.Engel and R G.Ilavclts, Cilemical chax~cteristics of oceanic bas~dts and the upper mantle. Geol. Soc. Amer. Bull. 76 (1965) 729. [71 tl.W,Menard, Marine geology of tile Pacific (McGrawHill, 19645 271 pp. 181 T.r/. Van Andel and C.O,l'lo~in, Mid-Atlantic Ridge ben,veen 22 and 23° north !atitude and the tectol.lcs of mid-ocean rises, ,I. Geol. Res. 73 ll9681 [279. 191 A.E.J.Engc: ;rod C.G.Enge]. Igneou~ rocks o~"the l,a~t Pacific Rise, Scier~,:e 146 ( 19641 477. [ 111] G.D.Nicholls, Basafis from the deep ocean floor. Mineral, Ma~z. 34 (19651 373. I 11 ] G.D.Nieht~ll~..A.J.Nalwalk an~t E,E.llays. Tile nature and composition of rock samples dredged from tllc Mid-Atlantic Ridge between 22°N and 52°N..~.larine Gec,l. 1 (196'~.) 333. [ 121 I.b.~.luiJ and C,E.'lilley, Basalt~ from the northern p~rt of'the: rift zone of I:heMid-Atlantic Ridge, J. Pelroi, 5 (I~64) a09. [ 13] l:.Aunlt,~ato, Tile Mid-Atlantic Ridge near 45°N. II. Ila~'~tl~sI'J'ptn t;]lear,.'a of ('on federation Peak, t, anad. J. Earth Sci. 5 (19681 1. tl41 J.D.H.Wiseman, The Johtl M~arty I,'.xpediUon,1933 34. put,. Brit. ~.us. GeoL Mineral. Invest, 3 (19401 I. [ 15 ] A.l:'oldervaart and D.II,Grecn, Chemical an:lyses of submarine basalts, Am. Miner. 50 ( 19651 1723. t 161 P,W.Gast, Trace element l'ractionation and the tlrigin of tholeiitic and alkalic nl~rna types, Geochim.Cosn~ochim. Acta 32119681 1(157. [ 17 [ F.Chayes, Alkaline zmd subalkaline basalts, Amcr. J, Sci. 264 (19661 128 [ ! 8~ G.A.Macdonald and 'l.Katsura, Chemical colnposition of I lawaiian lavas. J. Petrol. 5 (19641 82.



[ 19} G.A.M;tcdo,~aM, Composi1ion and originof Hawaiian ]avas, GanL Soc. An1. Mcm. 116 (in presS). [20[ N.J.tlubhard, Some trace elem~mts in Hawaiian [avas, Ph. D. "Pi~esis,Univ. of I lab,vail (1967). [2 l [ It.G.Kose and F.J.l:lanagan, U.S.Geol. Surv. Prof. Paper 4.S0-B. 80 (1962L [22] F.A.[:rcy, M.A.Haskin, J.A.Po[~z and L.A.Haskin, Rare earth abundances in some basic rocks (submitted for publication), [231 N.J.Hubbard and P.W.Gas~ (unpubl!:hed studies).

[24] J.G.Schilling, Rare earth fractionation in Hawaiian volcanil: rocks, Ph.D. Thesis, M.LT. (1966), ; J.G.Schilling and J.A.Winchester, Rare eattll!; in tlawaiian basaits, Science t53 (1966) 867. [ 25 ] I.D.MacGregor, The effect of pressure on the minimum melting composition in the system MgO-SiOz-TiO2, Ann. Rcp. Dir. Geophys. Lab., Carnegie Inst. Yea~ Book (19641965) ]:35. [26] J.P.F, at,an and K.J.Murata, How volcanoes grnw, Science 132 (1960) 925. [ 27 t !.Kushiro, Compositions of magmas lbrmed by partial zone melting of the earth's upper mantle, J, Geophys. Res. 73 (1968) 619.