JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 100, NO. B12, PAGES 24,593-24,609, DECEMBER 10, 1995
Correlation and emplacement of a large, zoned, discontinuously exposed ash flow sheet: The 40ArP9Archronology, paleomagnetism, and petrology of the Pahranagat Formation, Nevada
Myron G. Best,' Eric H. Christiansen,' Alan L. Deino,' C. Sherman Gr0mm6,~ and David G. Tingeyl
Abstract. Many single-cryslal 4oAr/3gArages and thermoremanent magnetization directions have resolved the problematic stratigraphic correlation of the laterally and vertically zoned rhyolite ash flow sheet of the Pahranagat Formation in the southern Great Basin. This outflow sheet was previously designated by four different stratigraphic names in different locations over its highly discontinuous exposure area of 33,000 kmz.We show that it is a single cooling unit emplaced at 22.639+0.009 Ma around its source, the Kawich caldera. The volume of the outflow sheet was about 1600 h3after compensation for 50% post volcanic east-west extension. A comparablc volume of tuff likely accumulated inside the Kawich caldera. Modal and chemical compositions of bulk tuff and cognate pumice fragments, together with compositions of phenocrysts, show the preemption magma body was zoned from high-silica rhyolite (two feldspars, quartz, biotite, and titanomagnetite) to underlying, silica-poor, more mafic rhyolite and trachydacite (plagioclase, minor biotite, titanomagnetite, amphibole, and clinopyroxene). Initial cvacuation of the uppermost evolved zone produced proximal outflow hundreds of meters thick of relatively densely welded, pumice-poor, high-silica rhyolite tuff. As eruption pro- gressed, tens of meters of more mafic ejecta were deposited in distal areas and locally near the caldera and consist of less welded, pumice-rich ash flow tuff derived by physical mixing of pyroclasts from all zones of the magma chamber. This mixing during eruption invalidates direct comuariso~lof the eom~ositionof tuff and a particular part of the magma chambcr. Thc ~ahrana~atash flow sh'cet provides a rigorous test case for application of high-precision correlation tools because of the zonal emplacement of ejecta from the compositionally stratified magma chamber together with the subsequent tectonic dismemberment and erosion of the sheet that created widely scattered exposures.
Introduction is llall~perednot only by their sheer size, but also by widespread dismemberment due to subsequent faulting, erosion, and conceal- Unambiguous correlation of an ash flow deposit can be a clial- ment beneath extensive valley fill. Not only is it impossible to lenging exercise in geologic field work and laboratory analysis. physically correlate an individl~alcontinuous sheet, but added 'This is particularly the case where the ejecta were broadcast in a impediments to correlation are posed by lateral and vertical complex manner around the vent(s) from a compositionally zoned composirional variations within some sheets and the multitude of magma chamber, where exposures of the deposit are poor or compositionally similar cooling units in any one sequence. The widely scattcrcd, and whcrc othcr similar dcpasits occur in thc number of cooling units in a single section can he as many as 20; volcanic succession [Hildreth and Mahood, 19851. Many such 10 is common. deposits typify the middle-Tertiary Great Basin ash flow province We have made a multidisciplinary study of the ash flow tuff in thc northcrn Basin and Range province of the wcstcrn Unitcd sheet of the Miocene Pahranagat Formation that is exposed over an States. This vast ash flow province contains well over 1M) ash area of 33,000 km2 in the southern Great Basin of Nevada (Figure flow tuff cooling units related to about 70 sources, mostly malked by calderas [Stewart and Carlson, 1976; Best et al., 1989b; Best 1). Because it is one of the younger deposits found in the fault- bounded, uplifted, and variably eroded mountain ranges of this part and Christiansen, 19911. Some outflow sheets span as many as ten ,mountain ranges covering an area as great as 40.W km': pat.ts of of the Basin and Range Province, erosional remnants are highly discontinuous and widely scattered. Substantial lateral compos- a single caldera !nay be exposed in four ranges [Best et al., 1989al. Smdy of outflow sheets and associated calderas in the Great Basil' itional variations west to east in the sheet compound the problem of correlation by conventional stratigraphic. . methods. It is our intent to (1) confirm the correlation of thc outflow shcct, which
1 ~~~~~t~~~tof ~~i~h~~younc~~i~~~~i~~. prove. ~~~1,.was initially based on stratigraphic position and composition, by ~~~l~~~.-. . , . ~ ' Bcrkcl:) tic~~.'I~rdn~~l.q!r
Figure 1. Kawich caldera and outflow tuff sheet of tlte Paliranagat Formation. Thicknesses and contours in Imeters. Zero indicates sites where the outflow sheet is absent between older and yuunger deposits. Co~rtours drawn by Techbase computer program which grids sheet into cells 5 x 5 km and determines best fit contours by iterative extrapolation between data points. Sites of samples dated by the *uAr/i"At. method shown by triangle, paleomagnetic santple sites by circle. Unlabeled txaleumaenetic sa11111lesite suutheast of the Cactus .. as foliows: CR, Cactus; FICR, 1,Iut Creek; KR, Kawlch; MR. oill lor; PAR, Pancake; I'R, Pahranag;; RR, Reveille. Inset [nap shuws Grcat Basin (dotted line) in northern part uf Basin and Range \)rovi~~ce encompassing most of Nevada and western Utah. Rectangular area is main diagran~. CNCC, central Nevada caldera complex: KC. Kawich caldera: R, Reno; LV, Las Vegas: SIC, Salt I.;lke City.
The Pahranagat Formation thus serves as an example of how caldera sources were stacked cullformably olie upon anotlier [Besr correlation of widely scattered rrosional remnants of an outflow a,& Cltri,~tia~~,,lcetr.1991J. After about 22 Ma, ash flow volcanism sheet may necessitate upward adjustments in eruptive volu~lteand slowly declined over the next several milli~inyears, extrusiolts of reduction in number of eruptive events, thus affecting models of canpositionally diverse lava flows became Itlure prominent, and continental magma genesis and eruption processes. (2) Document widesprrad crustal exlensiorr created tlw bxin and range struc1ul.r. the chemical and mineralogical compositions of tuff and cognate A major part of tlle as11 tlou, volunie ill the province was derived pumice fragments to understand how the magma chamher was from two especially large, long-sustained lnaglna centers: the zoned and how it was systematically evacuated lo produce a Indian Peak [Best el nl., 1989aI and the central Nevada [Brsr et vertically and laterally zoned outflow deposit around the Kawich 01.. 19931. Tlte latter is marked by a cluster of as rllally as 12 caldera. nested calderas and two indefinite source areas which altogether generally are younger t~rthe southwest. Tl~evolumc uf as11 flow Geologic Setting of the Pahranagat Formation lull associated with this central Nevada caldera colnplcx, including iiitracaldcra deposils and surrounding outflow shects, is at lcast The ignimbrite flareup, which created the Great Basin ash flow 20,000 knl'. Large rhyolite outflow tuff sl~eetswere en~placed pruviu~e,wah acconrpa~tied, at least in its early stages, by hetwecn 35.3 and 18.3 Ma and witllirl this time interval, rliorc vol~tminauseruptions at 31.3 Ma yip1dr.d zo~~edrl~yolite to dacite " and at 27.3 Ma dacite. The rhyolilic I';~l~ranagatFormation was from about 34 to 22 Ma, the crust was tectonically (nore or less emplaced a1 22.6 Ma. Formally defined by ScoN d 01. 119951, this quiescent so that ash flow sheets outside their complexly evolving formation cansists of two facies: intracaldesa and outflow. BEST ET AL.: THE ZONED PAHRANAGAT ASH FLOW SHEET 24,595
Althougb intracaldera deposits within tbe Kawich caldera source tion. If indeed Pahranagat Formation, we are uncertain whetller are briefly considered below, tlie main focus of this report is the the considerable thickness (500i- '! m) ol the tutt in tlie Cactus correlative outflow facies slleet (Figure 1). Range reflects ponding of outflow facies in an uldcr caldera or We now describe the general geology of the Kawich caldera and con~titutesa remnant of tlie intracaldera facies. then the surrounding outtlow sheet of the Paliranagat Formation. More field work invulvir~gdrtailed sampling and mapping will 1:ollowing sections present pertinent 'oAr/39Ar chronologic and be necessary to fully chara~.terizctllc Kawich caldera with regerd paleotnagnetic data essential to the correlation of the intracaldera to its history of collapse, resurgence [Ekren er al., 1971, p. 341, and outflow tacies. Sample sltes, denoted by capital letters, such and the magmatic evululion of the intracaldera rocks in relation to as ST1 and MU, are shown in Figure 1. tlie outtlow facies rocks. Even though accurate knowledge rrf the struclurr and cvululion of thc caldera is not available. its mareill is fairly well cnnstraiztcd, allowing us to ehtimate its area at about Kawich Caldera 1550 ktn2. The original area of tlie caldera prior to east-west The Kawich caldera [Stewart and Carlsnn. 1976; Sargenf and crustal extension would be 1040 km2 (For this estimate we used Roggemack, 1984; Besr er al, 19931 that marks the source of the a uniform extension of 50%; the amount of strain differs lo Pahranagat Formati011 lies near the southern margin of the central different parts of the Great Basin but the 50% value is compatible Nevada caldera complex. Through our reconnaissance and the with independent seisntic and stmctnml data and aspect ratios of s~nall-scaleinallping of Gnrdttrr et ul. [1980], tlie perimeter of the Great Basin tuff sheets (K. R. Suliivan and M. G. Brrf, unpub- Kawich caldera has been found in three places (Figure 1): (I) On lished data, 1993). As the intracaldera luff was likely at least 1 km the west side of the Kawich Range, apparently older andesitic thick, its original vnlume was at least 1000 km" and probably caldera wall locks are in contact with heterogcneous intl-acaldera more. tuff of the Pahranagat Formation (samples ST1 and STO). Intercalated breccias are locally n~onolithologicand contaiu clasts Outflow Facies of the Pahranagat Formation of tuff as much as a few meter3 in diameter of tuff that are similar to the Pahranagat Formation. (2) At the north end of the Kawich 'The outtlow sheet of the Pahranagat For~nationis exposed over Range, just south of Highway 6, banked against older wall rock is a present area of 33,000 km' in the southern Great Basin (Figure 1). It is a compound cooling unit in thick proximal parts just east, what we interpret to be tlle intracaldcra facies of the Pahranagat Formation (called by Gnrdner et al. (19801 the intl.acauldron tuffs and possibly west, of the Kawich caldera, where it apparently of Kawich Peak, tuff of Clifford Spring, and tuff of Kawich ponded in older calderas. Elsewhere, the outtlow sheet was enlplaced as a simple cooling unit. The restored volume of the caldera undivided). These deposits extend southward into the ranpe, where they are locally more than 1 k111 in tl~icknessand outflow sheet, compensating for 50% postdeposition extension is about 1600 km3, n~akir~glhc total restored volu~~leof tlie etltilr constitutc a compound tuff cooling unit containing multiple Pahranagat For~natianat lcast 2600 km'. On a dense-rock vitropliyres and local accu~nulationsof blocks of sedimentary and igneous ruck, as much as 20 m in diameter, intercalated with equivalent basis this is at least 2000 km'. lcnscs of sedimentary deposits. Just north of Cedar Pass (site and 'The appearance of the outflurr, facies of the Pahranagat Forma- tion in the field varies both vertically and, especially, laterally, but ralnple CD),thick intracaldera tuff has the Pahranagat Formation must uutcrups can he classified into one of two petrographic types paleomagnetic direction. (3) In the ce~itralReveille Range, Fang bascd on dcgrcc of welding and abundance of pumice fragments Ridge (sites FR and RENW) is a dike-sill complex of dacitic rocks and matic phenocrysts. These contrasts may have played a role in Illat are compositionally unrelated to and younger by at least one the multiplicity of stratigraphic names applied to tlie formation hy million years (samples FR and RENW) than the Pahranagat previous geologists wrlrking in differerlt ranges (see Scoff el ul. Fortoation. However, magma creating this complex intruded a Sequence of epiclastic rucks which could be local caldera-fill [I9951 for co~npletelisting of references). The Saulshury pctro- sediment deposited near the topographic margin of the Kawich graphic type was designated 2s the tuff of Saulsbury Wash [c.g., caldera because rocks predating the Pahranagat Furn~atiu~tarc Whitebred. 19891 in the Monitor Range (Figure 1) and the tuff ol widespread to the north [Ekreir el al., 19731. To thc south, the White Blotch Spring southeast and southwest ofthe Kawich caldera Reveille Range is a thick pile of massive, more or less altered (see previous section). This type characterizes proximal parts of illtracaldcra tuff of the Pahranagat For~llatiurl(salrlplc RE), wlrich the outtlow sheet around the caldera. It is partly lo densely welded has local strong compaction fi~liationand a stretching lineation, (densities range from 2200 to 2M)O kgim3; Table 1 and the designated as the tuff of White Blotch Spring by Ekren et al. electronic supplement') and contair~ssparse pumice lumps and few r>"7,, matic phenocrvsts, mailllv hiotite and rare amnliihole: xe~iolithsare LL7r.J. Stratigraphically problematic rocks occur in the Cactus Rerlge practically absent. A speckled gray-hlack vitlriphyre ah~~uta meter where Ekren et al. [1971, p. 351 found what was designated as thc thick lies near the habe of the MU sectio~~(Figure 1); elsewllere tlie upper cooling unit of the tuff of White Blotch Spring. Tlle~erock? Saulsbury type is entirely devitrificd. Our fcw samplcs of thc were "indistinguishable from the rocks in tlie Kawich Range that are poor in lithic fragments." Because the Cactus Range, on the #Anelectronic supplement of thlr material ,nay he obclcnrd no a dlske~te Tonopah Test Range, is inaccessible to 11s. we cannot verify these or Anonymous FTP from KOSMOS.AGU.ORG. (LOGIN to AGU's FTP rocks as belurlgillg to the Pahranagat Forn~atiun. Howrvrr, a account using ANONYMOUS as the usrrname and GUEST as the password. Go lo the right dciectory by lyplng CD APEND. Type LS a, palromagnetic corrclativc of the formation has heen found jgst scc what filcs arc availahlc. Type GET &od the namc of the f~leu, gcl it. southeast of the Cactus Range (see Figure 1; Mark R. Hudson, Finally. type EXIT to leave the Fyrrern ) (Paper 95JR01690, Correlation written communicatiun, August 1994). Moreover, at the White and emplacement of a large, zoned, discontiiiuuusly erposcd ash flow Blotch Spring section lileasured by Ekren el ul. 119711 (site WHB), sheer: The b''Ar/'VAr chmi~ology,paleo~~~agoetisn~, and petrology of the Pallfan'aeat Formalinn, Nevaiia. bv M.G. Besr el all. Diskctre mav be we found twu coulillg urlits, tlie upper oi wliicll has the age and ordered from American ~rophysicaiUnion, 2000 ~lohdaAvmur. N:W, paleo~nagneticdirection (sa~nplcWHB) of the Pahranagat Forma- Washington. DC 20009: $15.00. Pdyr~lrlltmust ~CCOIIIP~UI~UT~CT. 24,596 BEST ET AL.: THE ZONED PAHRANAGAT ASH FLOW SHEET
Table 1. Descriptions and Physical Properties of Paleomagnetic Sites in the Pahranagat Formation
Locality Name Density, Surceptihility, Curie Temneraturec. 'C Site Structure. deg Distance, 1OOO kglm' 0.OWl emu Low Middle High Name Type Strike Dip km * 1 S. D. f 1 S. D.
Squaw Flat. LT A 270.0 6.0 47.0 1.83+0.06 4.03i0.04 390 545-575 --- Little Fish Lake Valley Black Beauty Mesa, CH A 0.0 0.0 25.0 1.7510.08 3.52+0.03 --- 536 618 Pancake Range Forest Home, PO A 321.0 17.0 70.0 2.02k0.08 4.40k0.04 --- 541 578+ East Grant Range Shingle Spring, SH A 25.0 39.0 119.0 1.84+0.05 2.8li0.03 --- 556 --- Egan Range White River Narrows, WHRN A 156.0 8.0 83.5 1.73f 0.04 2.92f 0.03 --- 540 625 Seaman Range Pahrac Sullmit, PAHS A 210.1 6.5 'M.0 1.88i0.09 2.7710.03 --- 576 --- North Pahroc Range Badgcr Valley, AL A 7.1 23.8 88.0 1.73i0.06 3.71f 0.04 --- 570 --- ~;st~ahranagat Range Southern Delamar PAHR A 326.0 13.0 122.0 1.56+0.06 1.64i0.02 --- 551 605 Mountains Mud Sorinc. Saulsbuw MIJ S 228.3 19.9 27.0 2.36f0.03 5.6310.06 --- 558 --- . u. Wash, Monitor Range Red Bluff Spring, ED S 80.8 26.8 22.0 2.43f 0.N 3,791004 --- .. Quinn canyon Range Queen Cily Summit, BL S 354.7 20.7 5.5 2.40i0.05 1.30f0.01 358 550 604 Quinn Canyon Range White Blotch Sorine.. -. WHB Chalk Mountain Coyote Summit, TE S 78.7 20.5 37.0 2.161.09 2.8410.03 --- 563 -- ~im~ahuteRange Longstreets Ranch. ST1 I 167.4 22.7 -0.3 2.40i.08 7.66+0.08 --- ... 617-628 North Kawich Range Cedar Pass. CD I 354.0 11.0 4.7 2.47' n.d. n.d. South ~akchRange Reveille Peak, RE I 105.4 22.4 4.9 2.53f .0l 6.55*0.07 343 --- 626 South Reveille Range - - Type: A. Alamo; S, Saulsbury; I, intracaldera. Structure is attitude with strike 90" to left of direction of dip. Distance is from nearest caldera margin. Curie temperamres are degrees mrasured in argon gas and a direct magnetic field of several kilogauss. 'Estimated value.
intracaldera facies are densely welded like the Saulsbury type, but, as much as 5 mm across. Subordinate (< 8%) proportions of as will be shown below, the two are compositionally distinct. The smaller mafic phenocrysts, chiefly biotite, and titanomagnetite, Alamo petrographic type was previously designated as the granite- occur in most samples, but ilmenite, amphibole, and clinopyroxe~le weathering tuff [e.g., Ekren el al., 19731 in the Hot Creek, are only locally present; microscopic apatite, zircon, and chev- Pancake, and Reveille Ranges (Figure 1) and the Pahranagat Lakes kinite are evident in many samples. Quartz phenocrysts are usually Tuff [Willim, 1967; Best et al., 1989h. p. 1201 in die southcast- deeply embayed and many plagioclases are pervasively corroded. ern part of the outflow sheet. It mostly forms distal parts of the Sanidines are typically intact subhedral grains. Their concentra- uutfluw shmt, espially tu the east. In the MU section (Figure 2) lions in the Alamo- and Saulsbury-type tuffs are not significantly just northwest of the Kawich caldera, Alamo-type tuff overlies the different; however, the quartzlplagioclase ratio tends to be lower Saulsbury and therefore erupted after it. Compared to the in the Alamo and mafic phenocrysts are more abundant. Amphi- Saulsbury, the Alamo-type tuff is less densely welded (densities bole is very sparse in the Saulsbury type and clinopyroxene is 1500 to 2000 kgim3; Table 1 and Appendix). It contains sparse absent. small lapilli (generally < 1 cm) of dark volcanic xenoliths and In the initial stages of the project, conventional methods of white to light gray pumice lapilli and local blocks, constituting at correlating the outflow sheet of the Pahranagat Formation in the least 10% of the tuff, which lie in a pink to pale orange or purple- field between the eastern, distal exposures and the western mostly gray devitrified matrix. About half the thickness of the Alamo- proximal exposures were uncertain. Petrographic differences type sheet at the WHRN, LT, and CH sites consists of a weakly between the Alamo and Saulsbury types of tuff made correlation welded, white ash flow tuff that contains blocks and lapilli of suspect, even though the two types of tuff were somewhat similar pumice and grades upward into thc more compacted but still only compositionally (Figure 6) and both lay in similar stratigraphic partly welded tuff which forms the rest of the deposit; a meter of position near or at the top of most volcanic sections. Nine K-Ar bedded tuff lies near the bases of these sections. ages on biotites and sanidines from three differently named parts Tuff in the Pahranagat Formation contains subequal amounts of of the Pahranagat Formation ranged too widely (20.0 to 23.8 Ma quartz, sanidinc, and plagioclase phcnocrysts (Figures 3-5) that are [Scon el al., 1995, Table I])to confirm correlation. To resolve BEST ET AL.: THE ZONED PAHRANAGAT ASH FLOW SHEET
Firmre- 2. Vertical zonation in a cotnolete section of the outflow sheet of the Pahrananat Formation at the MU itu I,. SIIrrr h I IIIIUIII rei IUII I Iul~atII. NLI~C [lie ..,II~IJ\I~111 l.,lal III~I'I.I~IIC.IIU.~) >Is 111d'1 10: lhdl LII~IIII~UI~~Ihc 1\1.411111 :trhI Srllll\llur! pclr.,gr:iplli; I!pr\ 131 tutt these uncertainties in correlation, two precise quantitative analyti- method have been presented previously by Deino el al. 119901 and cal tcchniques were employed: 4oAr13gArlaser-fusion dating and Deino and Pons [1990]. determinations of paleomagnetic direction. The aAr/39Ar analyses corroborate the correlation of the outflow sheet in its differently named parts and additionally shows the intracaldera tuff within the Kawich caldera to be time equivalent Argon401Argon39 Laser-Fusion Geochrouology with the outilow facies. The data set demonstrates the tight Thirteen samples of thc Pallrandgat Forn~ationwere dated by the chronologic control that can be obtained on a single tuff cooling "Ar/39Ar laser-fusion technique, in wllich a ligl~tly focused, unit and constrains the eruptive timing of a large-volume pyro- continuous laser beam fused individual sanidine phenocrysts. clastic event. The range of 'UArl'YAr ages is only 1.6% of the Table 2 contains full 'OAr/'%r analytical data for one representa- range in K-Ar ages cited above and the mean age of 22.639 * tive sanidine and one plagioclase sample; Table 3 sut~~marizes 0.009 Ma is somewhat older than the averagc K-Ar age of 22.3 analytical data and statistics for all samples. Further details of the Ma. In all, 108 sanidine grains from the Pahranagat Formation were analyzed. Not all analyses were of equal quality, however, reflecting variable analytical conditions (i.e.. counting noise as a function of grain size or incomplete fusion) or geologic reasons (i.e., tluid or glass inclusions, trapped air pockets, or alte~.ation). We have found that the most sensitive parameter for ranking analytical quality within a group of similar sanidine analyses is the percentage of radiogenic aAr (%"Ar*) relative to total radiogenic
tl Qtz.poor Rhyolite . Saulsbury Low&lOl Rhyolite + Intracaidsra
0 10 20 30 40 50 60 Quartz (~01%total phenocrysts) Fieure 3. Modal ~rooortions.. . corrected to dense-rock ~qui~~al~~tt~.dt'+i:%~t~:tttd tut.11 t~t.tit; po?nt>:r!,t, rel:at~\,ct~, tut.tl nhc~~~,.r\stsin alta~le r.12k. c,f the I'.tl~r~wintI'LIIIII~I~~. open boxesBre tuff samples. Solid boxes are pumice (see Table 5 and followine figures for additional characterization of compositional groups). Shaded error hoxes show lnaximrtm ranee in modal ~ro~ortionscounted in du~licatethin sections from one sample at high and low ends'uf range in quartz concentration. Least squares best fit line through pumice Figure 4. Ratio of quartz to plagioclase phenocrysts versus samples has a correlation cuefficient 01 0.96. TiO, cuncer~lratiurairt [he PahrauagaL Formation. 24,598 BEST ET AL.: THE ZONED PAHRANAGAT ASH FLOW SHEET
Quartz Pumlcc- -- Tuff ' 72 Oatz~rlchRhyote o Alarno BOtr~poorRhyoi#te - Sauisbury 1 , Trachvdaclte ~Low-S~OZRhvol8te + InIracaldera I Qe-r~chRhyollte 0 Alarno Qk-poor Rhyollte Saulsbury LOW-SIOZRhyollte
Low St02 Rhyolite\ Trachydacite si02 (&%) 17'3-24'3 2i'7 - A Figure 6. Com~ositionof rocks from the Pahrana~atForma- / ~~~~~~~~ tion in the ~nternationalUnion of Geological Sciences- chemical Plagioclase classification scheme. Figure 5. Modal proportions of felsic phenocrysts in foul. groups of pumice fragments in the Pahranagat Formation 5). ' ~ro~ortionsof mafic to total phenocrysts are able that enable most of the analyses to be retained, yet elitninates a indicated helow pumice group lahel. questionable minority of runs. Most of the lower radiogenic analyses were in fact incomplete fusions, reflecting the occasional and atmospheric 'OAr. Over 90% of the a~ialysesyielded greater difficulty in coupling visible light Ar laser with 11anipart.nt than 97% *Are, while another 10 analyses formed part of a tailing feldspars. An additional four grains that fell more than 2 S.D. distribution ofmAr*, to a minimum of 87.7%. Tu refine the data (standard deviations) beyond the weighted mean age, but were set, only the higher radiogenic anal-yses were retained in subse- otherwise analytically acceptable, were also culled from the data quent data analyses. Although the 97% cutoff is somewhat set. These ages may reflect alteration or the presence of inclu- arbitrary, it was selected at a perceived break in thc distribution sions, or simply analytical variability. Age-probability distribu-
Table 2. Representative "ArI3'Ar Analytical Data
Lab ID CaiK 36Ari3'Ar %('6Ar/3qAr),. "Ar*i3'Ar %l"Ar* Age (Ma) +lo
Sample ST0 I!~Y-OZ 0.016 20 1.088 1929-03 0.015 19 1.090 1929-04 0.017 17 1.090 1929-05 0.015 19 1.088 1929-06 0017 13 1.087 1829-07 0.016 6 1.091 1929-09 0.018 11 1,090 1929-10 0.017 12 1.087 Weighted average
Sonidine :Sample ST0 20 Beyond Mean 0.00003 10 1.084 99.3 Pln~ioclase: Sam~leRENW 5629-02 8.829 5629-05 8.683 5629-07 8.502 5629.06 9.079 5629-03 8.498 562941 8.424 5629-04 8.905 Weighted Average
Errors in age quoted for individual runs rre lo analytical uncerrainty. Weighted averages are calculated using the inverse variance as the weighting factor [Toylor, 19821. Errors in the weighted averages are lo standard error of the mean [Snmson and Alexander, 19871, CaIK is calculated from "Ari3'Ar using a multiplier of 1.96. %eAr* is the percentage of radiogenic argon of the total radlogeluc and ahnospherlc '"At. Variable A = 5.543 x 10"' yr'. Isotopic interference corrections: For both sampler, (36Ari"Ar),, = 2.58 x lo4 ? 6 x 10*, ('gAr/"Ar),, = 6.7 x lod f 3 x 10.'. and (wArl'YAr), = 8 x 10' k 7 x lo4. For STO,J = 0.01160 i 0.MM03; for RENW, J = 0.002097 i 0.m1. BEST ET AL.: THE ZONED PAHRANAGAT ASH FLOW SHEET 24,599
Table 3. aAr139Ar Analytical Data on the Pahranagat Formation and Associated Rocks
Sample Lab lD N J iJ CaIK iCalK 40t/39 i40*/39 Age, S.D. SEM Ma Without il With iJ
Colibrnriun 1fInrurrm1 Ynn~/t~r,/,WHRN-IJ
WHRN-ll* 474 WHRN-Il+ 7226 WHRN-11 I 7222 WHRN-llt 7219 WHRN-1J+ 7216 Weighted mean age Analyses of Olher Pahranagat TGSamples U~ingWHRN-lJ as the Standard 2.908E-03 6E-06 0.0081 00055 4.3384 OMIS6 22.62 7.289E43 1E-05 0.0169 0.0029 1.7330 0.0012 22.65
ST0 1929 ST0 3624 WHB 48 1 WHB 6054 Weighted mean age
Analyses of Post-Pahranagat TylfLavar and Dikes BELL 4086 1.488E-02 3E-05 0.0132 0.0016 0.7716 0.0011 20.58 0.L7 RENW 5629 2.096E-03 1E-05 8.7028 0.2445 5.7440 0.0110 21.57 0.11 FR 5631 2.0%E43 IE-05 0.7657 0.8245 5.6250 0.0187 21.13 0.19
Read 2890E-03 as 2.890 x 10'
tions [fluford er a/., 1984; Deino and Pons. 19921 for sanidines To avoid these problems. we recalibrated the age of the WHRN- of the Pahranagat Formation both before and after elimination of 11 sanidine against a commonly employed sanidine monitor from these outliers are shown in Figure 7. In only one case (1929) was another middle Tertiary ignimbrite, the Fish Canyon Tuff of the the mode affected, and in this instance by only 0.01 1n.y. San Juan Mountains, Colorado. (The age assigned to the Fish In all '0Ari39Ar dating of middle Tertiary ash flow tuffs of the Canyon sanidine was 27.84 Ma, based on the age reported by Great Basin, we have employed an internal standard as the monitor Cebula et al. [1986], adjusted for the revised age of the prlrnary mineral for determination of the reactor neutron flux parameter, I. monitor mineral, MMhb-1 [Samson and Alexander, 19871). We This standard is sanidine in the 25-30 mesh range extracted from achieved unusually high precision in the calibration experiment by the tuff of the Pahranagat Formation at the White River Narrows intermixing 25-30 mesh grains of Fish Canyon and Pahranagat (sample WHRN-IJ). An early calibration of the WHRN-11 sanidines in the same irradiation position (a pit approximately 2 sanidine against the international 4Ar/3qAr hnrnhlende standard, mm diameter by 2 mm deep), and then selecting grains at random MMhb-1, with an age of 520.4 Ma [Samson and Alexander, 19871, from a given position for analysis. The advantage of this approach yielded the relatively imprecise age of 22.60f 0.08 Ma. Much of is that no uncertainty is introduced by the usual problem of relating the uncertainty was attributable to the large inherent scatter in the flux from a monitor position to that of an unknown. The final measuring 1-3 grains of MMhb-l (in this experiment, overall uncertainty in the age of the unknown is then simply the combina- standard error of the mean was 0.3% with n=18). In addition, if tion of the uncertainties in the measurement of the two sanidines. monitor and unknown do not occupy the same position in an 'Sable 3 summarizes statistics for this calibration, in which irradiation package, as is usual and applies in this case, significant sanidines in four positions on the same disk were analyzed. The error arises from the spatial flux gradient. We position our spread of the age determinations of the WHRN-I1 sanidines is monitors and unknowns on the same level of an irradiation acceptably small, from 22.62 to 22.64 Ma. The overall weighted package, in small wells drilled in a 2.5- to 3-mm-thick Al disk. mean age of the four aliquots is 22.636 Ma with a S.D. of 0.009 However, experience with several reactors has shown that lateral m.y. and S.E.M. (standard error of the mean) of 0.007 m.y. flux gradients can be as much as 1 or 2% across a 14-mm-diameter [Samson and Alexander, 19871. This is within error of the result disk. Thus, unless the gradients are unusually tightly monitored, measured against MMhb-1. the error in determination of an appropriate J for an unknown is no This value represents our best independent age determination of better than about 0.2-0.3%. the Pahranagat Formation. All other sanidine IOAr/"Ar ages of the BEST ET AL.: TIIE ZONED PAHRANAGAT ASH FLOW SHEET
attitude, measured dcnsitics, and scalar magnctic prnpcrtics are shown in Table 1. The magnetic susceptibilities are typical values for felsic welded tuft's, with the exception of sites WHB and CD which are an order of magnitude lower; phenocrysts in the tuff at site WHB include a very low proportion of matic minerals, about 1%. The distribution of Curie temperatures in the sites indicates the presence of up tn three kinds of magnetic minerals. Values in the range 398°C to 343°C represent Litano~liagiietitewith about 30 to 40 ~iiol% of ulvospinel, considerably higher than measured by microprobe (see mineral compositions below), but such low Curie temperatures were observed at only three sites. Values in the range 521°C to 575°C represent low-Ti titanomagnetite (12 sites): the measured ulvospinel contents correspond to a narrow range of Curie temperatures from 4WDCto 5M)"C. That thcsc prcdominant Curic temperatures are higher than those implied by the microprobe measurements may be ascribed to the well-known breakdown of titanomagnetite into submicroscopic intergrowths of ilmenite and very low-Ti titanomagnetite that results from high-temperature oxidation. The natural remanent magnetization in these minerals is without doubt thermoremanent magnetization carried by oxide Age (Ma) grains (microphenoc~ysts)that passed through the eruption process. Figure 7. Age-probability spectra for samples ot Pahranagat Nine of the 16 sites also contain notable amounts of a magnetic Formation. The vertical axis of the plot is a relative probabil- mineral having Curie temperatures up to 628°C. significantly ity measure of obtaining a given age for a particular sample, higher than that of pure magnetite. This mineral is now known to based on the sum of the assumed Gaussian errors of the be maghemire that has formed in the tuff during postcompaclion individual single-crystal analyses. The shaded areas shown for cooling [Schlinger er al., 1988; Rosenbaum, 19931, and the natural several samples represent that part of the probability distribu- remnant magnetism that it carries must be a high-temperature tion trimmed by omitting outliers 20 beyond the weighted chemical remanent magnetization. Note that maghemite is the only mean ages. Removal of outliers has virtually no effect on the magnetic mineral present in the intracaldera site ST1 (Table 1). mode of the distributions. The paleomagnetic data are given in Table 4, and the directions are illustrated in Figure 8 after correction for the structural attitudes listed in Tat& 1. The angular standard deviation of the Pahranagat Formation are referenced against the WHRN-11 pales to these attitudes is 19", sufficient to allow a paleo~nagnctic standard, and so are not strictly independent measurements. fold test, and indeed for the 16 sites Fisher's precision parameter Nevertheless, comparison of ages between sanidines of the for the mean of site-mean directions increases from 17 to 50 after Pahranagat Formation remains valid. Seven other samples were structural correction, a positive result by any statistical test. All analyzed, some in replicate, from intracaldera and outflow sites have reversed polarity, and the mean direction is only 10' exposures, and from different directions in the outflow sheet from the predicted Oligocene-early Miocene (38-22 Ma) reversed relative to the Kawich caldera (Figure 7 and Table 3). All apes geomagnetic axial dipole field direction (inclination -56". declina- fall within a narrow span of 0.06 my., yielding an overall tion 172") at the average location of these sites, calculated from wcightcd mcan of 22.63910.009 Ma (lo S.E.M.; S.D. 0.02 Diehl et al. [1983]. Hence, except for the reversed polarity, there my.). This age is very close to that of WHRN-11, as expected if is nothing distinctive about these paleomagnetic directions. WHRN-11 is truly representative of the tuff in terms of age. The precision parameter 50 represents an angular standard Further, all outflow and intracaldera samples are analyt-ically deviation of 15" for all site means (Table 4), which is greater than indistinguishable in age; hence neither a geographic variation in might be hoped for. The three most divergent directions (Figure the outflow sheet nor an intracalderaloutflow difference in the age 8) are from sites PAHR, MU, and BL. Thc sitc PAHR is the most of thc Pahranagat Formation is dcmanstrable by thcsc data. In distant from the caldera source, and the tuff there is only weakly addition to the analyses of the Pahranagat Formation, we also dated welded, so it may be a less than ideal paleomagnetic recorder as sanidine from one sample (BELL) of postcaldera lava capping the suggested hy its rather high value of within-site angular standard formation in the Kawich caldera and plagioclase and anorthoclase deviation (Table 4). The other two sites are in the more densely from two dike samples (RENW and FR) in the Fang Ridge dike- welded Saulsbury type, and for these the divergence iliust be sill complex (Table 3). ascribed to difficulty in measuring the structural attitudes in these Paleornagnetism unusually thick sections of tuff which lack a recognizable compac- tion foliation because of scarcity of biotite and flattened pumice Sixteen sites in the Pahranaeat Formation were collected for lumns. Nn correlation exists between deviation of site-mean measurement of paleomagnetic and other physical properties; they directions from the overall mean and either distance from the were distributed as widely as possible over the outcrop area source caldera or structural dip. In particular, we sought evidence (Figure 1). Eight sites were collected in the Alamo petrographic of correlation between azimuth of deviation of paleomagnetic site- type, five in the Saulsbury type, and three in intracaldera tuff mean direction from the overall mean and azimuth from caldera to within the Kawich caldera. Methods of field collection and site. Such a correlation could be interpreted as evidence of doming laboratory analysis are described in the Appendix. Structural centered on the caldera; none was found. LIES1 ET AL.: THE ZONED PAHRANAGAT ASH FLOW SHEET 24,601
Table 4. Paleomagnetic Directions in the Pahranagat Formation
Site Treatment Site 1, D, VGP Name Type Latitude Longitude deg deg N R X a,, Del LatiNde LongiNde
LT LlNS 38.579 243.507 -70.9 200.3 8 7.9398 116 5.2 7.5 -68.5 96.1 CH LINS 38.272 243.893 -66.6 199.7 6 5.9872 389 3.4 4.1 -72.2 110.0 FO LlNS 38.318 244.715 -56.7 192.1 8 7.9M7 199 3.9 5.6 80.4 157.3 SH H400 38.591 245.140 -67.6 192.3 6 5.9363 79 7.6 9.2 -75.3 97.2 WHRN LlNS 37.838 244.9M) -63.6 175.7 6 5.9725 182 5.0 6.0 -81.9 43.0 PAHS, LlNS 37.622 245.CQ2 -58.5 177.9 8 7.8356 43 8.6 12.4 87.8 20.1 AL LlNS 37.255 244.777 -58.1 182.3 8 7.9890 636 2.2 3.2 -87.6 114.8 PAHR LlNS 37.095 245.107 -63.1 226.0 8 7.9268 96 5.7 8.3 -54.9 128.1 MU LlNS 38.159 243.200 60.6 153.8 8 7.9864 516 2.4 3.6 69.7 351.2 ED LINS 38.061 244.119 -54.5 182.7 8 7.9894 658 2.2 3.2 -86.3 207.5 BL LINS 37.771 244.059 -39.0 195.4 8 7.9959 1688 1.3 2.0 -69.4 199.5 WHB LlNS 37.538 244.076 -54.1 197.5 6 5.9818 274 4.1 4.9 15.6 160.6 TE LINS 37.557 244.328 -50.2 202.1 4 3.9958 718 3.4 3.0 -70.7 167.7 ST1 LlNS 38.995 243.472 -57.2 171.7 7 6.9622 159 4.8 6.4 -83.5 334.6 CD USP 37.974 243.889 -48.8 193.1 7 6.9707 205 4.2 5.7 48.8 193.1 RE LINS 37.853 243.850 51.3 200.1 3 2.9985 1341 3.4 2.2 72.5 167.5 Alamo type, mean direction 37.91 244.63 -63.9 192.3 8 7.9168 84 6.1 8.9 Saulsbury type, mean direction 37.87 243.91 -52.8 188.4 5 4.8871 35 13.0 13.7 Intraealdera sites, mean direction 37.87 243.63 -53.0 189.2 3 2.9713 70 14.9 9.7 Saulsbury type and intracaldera sites combined, mean direction 37.87 243.81 -52.9 188.7 8 7.8584 49 8.0 11.6 Overall mean direction 37.89 244.25 -58.4 190.2 6 15.6998 50 5.3 11.5 Overall mean virtual pole 16 15.5019 30 6.8 14.9 81.9 316.4
Tnamcnt typc: LlNS is principal component analysis of progressive AF demagnetization; H400 is blanket AF demagnetization at 400 peak Oersteds, USP is unspecified AF demagnetization. Site latitude, longitude, nonh latitude and east longihlde of sampling site. I, inclination (negative upward) and D, (positive eastward) of mean primary magnetization direction. Nis number of specimens (or sltes): K 1s vector resultant of Nutut vectors: K is Fisher precision parameter; %, is semiangle of 95% confidence cone centered on mean; Del is angular standard deviation of data around mean. VGP Latihlde, Longitude, are latitude (negative in southern hemisphere) and east longitude of virtual geomagnetic pole (polariry of mean pole is inverted to northern hcrnisphcrc). All anglcs arc measured in degrees.
Density decreases and paleomagnetic dispersion increases with similarity has little significance because of the small number of increasing distance from the nearest caldera margin (Figure 9), as sites and the compositional similarity of the intracaldera tuff to the would be expected, but the (negative) inclination also seems to Alamo type. The between-site mean direction for the Alamo type increase with distance. Because this might be the result of (Table 4) differs from that of the Saulsbury by 11.3', mostly in shallowing of inclination due to compaction of tuff during welding inclination. This difference is statistically insignificant at 95% of the more proximal sites [Rosenbawn, 19861, we show in Figure probability according to the tcst of McFadden andLower [19811. 10 the inclination plotted against density. As might be predicted, although such a test is not strictly applicable because the popula- inclination becomes more shallow with increasing density, but tions of directions being compared do not have circularly syrnmet- within either the Alamo type or the Saulsbury type (and intra- rical distributions about their means, that is, they are not Fisherian, caldera sites), no clear trends exist. Moreover, the difference is but are elongate (Figure 8). In summary, when considered in the accentuated by one divergent direction, from site BL (Table 4) at light of the degree of compositional evolution of the three main which the structural attitude was exceptionally difficult to measure. types of tuff of the Pahranagat Formation, thc corresponding scts Instead, the systematic difference might result from differing times of paleomagnetic directions illustrated in Figure 8 are part of a of eruption coupled with geomagnetic secular variation. To single population. illustrate this possibility, subsets of the paleomagnetic directions are shown separately in Figure 8. The between-site mean direc- Petrology of the Pahranagat Formation tions for the Saulsbury type and the intracaldera tuff are virtually Having now established that the severally named stratigraphic identical, but, in the context of a possible time difference, this tuff units are but one single cooling unit, we next document 24,602 BEST ET AL.: THE ZONED PAHRANAGAT ASH FLOW SHEET
A = ALANC I= SAULStlllPY = CALCE9A Figure 8. Equal-area projections on part of upper hemisphere of mean paleomagnetic directions for Pahranagat Formation, Figure 10. Paleomagnetic site mean inclinations plotted corrected for structural dip. Open circles are site mean against average densities of paleomagnetlc samples of Pah- directions, and open diamonds are means of site means. raliagat Formation. Error limits on i~iclinatio~ipuil~ts are Limits of 95% confidence are shown around each site mean angular standard deviations for sites. A11 data are from Tables direction as light circles and around each mean of site means 1 and 4. as heavy circles. All data are from Table 1.
petrologic aspccts that bear on thc cruption and cmplacement of the zoned outtlow sheet from a compositionally stratified magma chamber.
Pumice Fragments in the Outflow Facies Because at least some physical mixing of pyroclasts from diffcrcnt parts of erupting zoned magma chambers is inevitable, fragments of cognate pumice are more reliable samples of the magma body than their host tuff. Such pumice samples provide valuable insight into the compositional nature of the preeruptian magma chamber, its petrologic evolution, and how it might have been evacuated during eruption. To this end, 17 blocks of glassy cogllate pumice to as much as 30 cm in diameter from six localities in the outtlow sheet of thc Pahranagat Formation were studied (Table 5). All but three have Alamo-type tuff as a host. Felsic phenocrysts are as much as 8 mm in diameter. Where present, quartz phenocrysts are invariably 2500 m, :., .,&, , , * , , 1 ; ; embayed, but sanidines are intact and appear to have beer, rrla- Density tively more stable just before eruption. The pumiccs can be divided into four groups based on their elemental and modal * 1 A compositions nable 5 and Figures 3-6, 11) and the nature of their 1500 plagioclase phenocrysts. Three of the groups are rhyolite and one 0 50 100 is trachydacite. The quara-rich and quartz-poor high-silica rhyolite groups are co~~~positiunallygradatio~~al inlo one a~other. Distance from caldera rim, km. The two high-silica rhyolite groups have relatively homogeneous, A - ALAMO I= SAULSnUnY = CALDC7A euhedral to subhedral, more intact, plagioclase phenocrysts; these groups have low TiO,, low to modest amounts of biotite, Fe-Ti Figure 9. Paleomagnetic site mean inclinations, angular oxides, and most contain amphibole. together with large propor- (circular) standard deviations for sites, and density of Pahran- agat Formation plotted against distance from nearest point on tions of felsic phenocrysts. One quartz-rich pumice (AL-43PC) margin of Kawich caldera. Error li~nitson inclination points has very rare clinopyn~xene. Low-silica rhyolite pumice contains are f1 angular standard deviation. Intracaldera sites are more biotite and Fe-Ti oxides than the high-silica rhyolite pumices, shown with negative distances. All data are from Tables 1 and as well as more amphibole and clinopyroxene. Quartz and sanidine 4. phenocrysts are sparse or absent, but where present are as large as BEST ET AL.: THE ZONED PAHRANAGAT ASH FLOW SHEET 24,603
Table 5. Pumice Groups in Pahranagat Formation
Group TiO,, wt% Phenoclyst Proponions Sample Height, m Density, IOOO kg/m3
I. Quaw-rich high- 0.10-0.14 Q-S>P> >B-O+A aAL-43PC* 43 0.96 silica rhyolite aAL2PD 2 l.W 1.07 aAL-2PET 2 - aCH-40P 40 0.72
2. Quartr-poor 0.13-0.21 P-S>Q> >B>O-A aWHRN-IJBPC 2 0.79 high-silica 0.81 rhyolite aLT-EPA 2 0.92 aLT-EPC 2 0.88 0.95 aLT-EPFt 2 0.81 sMU4ZPt 42 1.99t:
4. Trachydacite 0.48 P>B>O>C>A aMU-12x5 1.02 1.07
Q, quartz; S, sanidine; P, plagioclase; B, biotite; A, amphibole; 0, Fe-Ti oxides; C, clinapyroxene. The a designates host tuff of Alamo type; s is host hlff of Saulsbury type. Height of sample above base of Pahranagat outflow sheet. Duplicate values by different methods (see electronic supplement). * Contains two small clinopyronene grains in one Ulin section. t Not chemically analyzed or listed in Table 6. i Partially compacted.
3 mm. Plagioclase is the dominant phenocryst and ranges widely rhyolite (Figure b and the electronic supplement). We do not in size fro1110.1 to 3 mm: most phenocrysts are intcnscly norlnally believe posteruption mobility of alkalies was sufficiien to change zoned, some are oscillatory zoned, and the largest are also the classification of the four low-silica rhyolite pumice clasts and intensely cor~uded. T~achydacite pumice has lcss silica, is one trachydacite pumice (66% SOl; MU-125P). The rhyolites are texturally and nodally similar to the low-silica rhyolite pumice, metaluminous to peraluminous, the latter probably because of but has no sanidine or quartz; plagioclase phenocrysts do not posteruption mobility of alkalies because samples lack peralum- range to the smaller sizes. inous minerals and high-Al biotites typical of peraluminous magmas. The rhyolites have low FeIMg ratios (less than 0.84 in Compositions of Tuff and Pumice Clasts fresh rocks) typical of calc-alkaline rhyolites in general. More- Despite some post-eruption mobility of a few elements (Na, K, over, in the trace element classification of Pearce el al. [1984], our and perhaps Rh and Y), it is clear that all of the tuff and two samplcs of thc Pahranagat Formation lie in the volcanic arc field, groups of pumice clasts of the Pahranagat Portnation are high-silica typical of rhyolite magmas erupted above subduction zones. We
QUARTZ- RICH + . . . . 1'1111736
2 POOR
r Low- /::.;., 1 807 SILICA , .;, I I\ 1AIi 11 TRACHY .. . ., - I . ...::. .. U DACITE I::l I I I I PLAG SANl QTZ BOT AMP OX AP ZlR CH T[OC) CPX Figure 11. Mineralogical and modal co~lipositionof pv~liicefragments in the Pahranagat Formation. Relative width of bands is indicative of proportions of phenocrysts. Temperatures based on feldspar georherrnomerry. BEST ET AL.: THE ZONED PAHRANAGAT ASH FLOW SHEET
302 (Wh) 302 (Web) 30, (WL) Figure 12. TiO, variation diagrams for the Pahranagat Formation. Solid squares, pumice; open squares, tuff.
have used TiO, concentrations to show chemical variations in the the early Saulsbury and late Alamo petrographic types are gener- Pahranagat Formation (Figures 4 and 12), because TiO, is well ally distinctive, but the Saulsbury is more evolved of the two. analyzed, shows a wide range in concentration, and is relatively However, a few samples with Alamo appearance have evolved immobile during hydration ar devitrification of glassy material. compositions and mineral proponions, suggesting that factors Samples with TiO, less than 0.18% to as low as 0.10% are besides magma composition controlled the degree of welding and ennsiderrd to he chemically evolved: more mafic samples with abundance of lithic and pumice fragments. TiO, greater than 0.18% to as high as 0.48% are less evolved. In general, low TiO, concentrations correlate well with mineralogical Compositions of Phenocrysts indicators of magma evolution, including low contents of mafic minerals and high quartzlplagioclase ratios (Figures 3 and 4). Plagioclase grains in the trachydacite and low-silica rhyolite TiO, concentrations are negatively correlated with incompatible pumice clasts are strongly zoned from An,, to An,,, whereas more elements (Si, As, Rb, Y, Nb, Sb, Cs, Tb. Yb, Lu, Ta, Pb, Th, silicic samples have more uniform plagioclase compositioi~s and U) but positively with colnpatible elements Wg, A!, P, K, Ca, (Figure 13). Plagioclase rim compositions become progressively Sc, V, Cr, Fe, Zn, Ga, Sr, Zr, Ba, La, Ce, Nd, Eu, and Hf). inore sodic in more silicic pumice fragments and bulk tuff, ranging Largely because of post-magmatic mobility or low concentrations, from An,, to An ,, , consistent with a systematic decrease in we found no systematic evolutionary trends for Na, K, F, CI, S, temperature. Alkali feldspar grains are essentially unzoned and, Mn, and Ni relative to TiO,. Pumice compositions generally within individual pumices, have relatively uniform compnsitions overlap those of tuffs (Figure 12). Compositional variations within ranging from Or,, in the trachydacite pumice to Or, in the most
Ab Fip~re13. Feldspar compositions in pumice (P in sample label) and tuff frurn the Pahranagat Forination. Tie lines connect compositions used in geothermometry [l'uhrman and Lindsley, 19881. BEST ET AL.: THE ZONED PAHRANAGAT ASH FLOW SHEET 24,605 evolved pumice. Calculated two-feldspar temperatures range flow deposits [e.g., Hildrelh, 19811, this sheet is vertically zoncd systematically from 800' to 735'C in trachydacite to high-silica from a more evolved basc to a lcss evolvcd upper part (Figures 2 rhyolite, respectively, using the thermometer of Fuhrman and and 14). Mineral proportions and assemblages change in comple- Lir~dsley[I9881 and a pressure of 5 kbar (see below). Tuff sarnple nientary fashion so that, for example, the quartzlplagioclase ratio FO of thc Alamo typc contains a mixed population of feldspars that declines with stratigraphic height as mafic mineral abundance appears to have been derived from separate parts of the magma increases. TiO, increases upward. As for the pumice fragments, chamber, including a few grains of sodic plagioclase and sanidine the early erupted Saulsbuly-type tuff hosts only high-silica rhyolite from the high-silica part of the chamber, hut containing chiefly clasts, whereas the later erupted Alamo-type luff contains all more calcic plagioclase and more potassic sanidine from the more pumice groups (Table 5): the one trachydacite fragment is hosted mafic part. by Alamo-type tuff. Thcsc relationships are compatible with a Unzoned biotites generally have relatively uniform compositions compositionally zoned magma body in which an upper, more in each sample. As a whole, Fe/(Fe+Mg) ratios are ahout 0.4 and evolved, liigh-silica rhyolitic magma was underlain by a compos- are slightly lower in the more evolved samples, probably because itionally distinct low-silica rhyolitic magma that was itself under- of higher f 0,. More evolved biotites also have lower Al,,,,,, Ti, lain by trachydacitic magma (Figure 11). The rhyolitic zones had and TiIFe, probably because of lower temperature, but greater Mn co~npositionalgradients within them. The obvious compositional and F than less evolved samples. Unzoned clinopyroxenes zonation of the erupted part of the Pahranagat mdglna body generally form homogeneous populations of relatively (for rhyolite) (Figures 3-6 and 11-12) cannot have been produced by the Mg-rich augite (average Ca,, Mg, Fe ,,). Amphiboles range to accumulation of separate magma batches produced by different relatively high Al,,,,, (2.0 to 2.1 per formula unit or pfu), indicat- degrees of partial melting of a common source. The large ing a crystallization pressure of 5.0 to 5.5 kbar, according to the depletions of compatible elements in pumice fragments precludes geobarometer of Johnson andRulheifurd [1989]. Surne amphibule this process. Magma mixing acting alone cannot have played an phenocrysts are zoned from high-Al cores to lower Al rims. The important role in producing the chemical variation in the magma a~~~pl~ibolesare quite magnesian, consistent with relatively because trends on compatible-incompatible element variation oxidizing conditions prevailing in the Pahranagat lnaglila chamber, diagrams of pumice clasts are arcuate, ratl~erlllarl linear (Figure and appear to have been in equilibrium with biotites. Titano- 15). On the other hand, elrrncnlal tre~~dsarc gerlerally cvr~sistc~~l magnetite is present in mast tuff and pumicc samples of the with fractionation of the obscrvcd phcnocrysts from a rhyolitic Pahranagat Fromation, progressively increasing to 3.5% of the melt [Christianren and Best, 19931 and are similar to those fuund total phenocrysts in the mast matic pumices. The mole fraction of in other silicic rocks. As predicted by this model, elemental trends ulvospinel ranges from 0.139 to 0.147. increasing will! evolutior~ for pumice fragmrnls in which TiO, is plotted against compatible of the host rock. MnO also systematically increases, whcrcas Al, elements generally are straight lines: also, incompatible elements V, and Cr contents decrease with whole rock evolution, apparently plotted against TiO, yield strongly curved variation trends which correlating with declining temperature and declining concentrations show somewhat Inore scatter (Figure 12). The continuity and of these elements in the magma. Because it occurs in such small linearity of the variation trends, the smooth variation of mineral quantities, ilmenite was only analyzed in one sample, quartz-poor composition with rock composition, and the complementary pumice LT-EPA. MgiMn ratios of ilmenite and tilanomagnetite in variations in the co~npositionsof the pumice fragnlents and their this sample indicate that they were in equilibrium [Bacon and minerals argue that all of the erupted magma was comagmatic and Hirschmann, 19881: these yield a temperature of 735°C [Ghiorso existed in a single zoned chamber. Nonetheless, our conclusion and Sack, 19911, similar to the two feldspar temperature for this that the Pahranagat magma body was mned by fractio~lalcrystal- sample. lization is based only upon permissive evidence and application of Occam's ruler, as is common in petrologic investigations. Other more complicated hypotheses for the origin of zonation cannot be Compositional Zonation in the Pahranagat Magma Chamber ruled out; these include (1) tapping two or more separate lnaglna Compositional zonation in the outflow sheet of the Pahranagat bodies each derived by partial melting of a different source and (2) Formation and the pumice fragments hosted in it indicate the mixing of already fractionated and zoned chambers which coinci- erupted part of the magma chamber was zoned. Like inany ash dentally lie along the same chemical trend.
Pahranagat Outflow Sheet
Kawich Caldera Pre-caldera Rocks
Figure 14. Idealized composite radial section through the outflow and intracaldera facies of the Pahranagat Formation showing TiO, concentrations in the tuff and relative stratigraphic positions of samples of the four groups of pumice fragments. The most evolved tuff (TiO, (0.13%) is found only in the lower part of the proximal sheet; less evolved tuff is found high in some proximal sections, but in general it occurs in distal parts of sheet. Note that the most mafic tuff (TiO,, 0.18%) hosts all pumice types, whereas more evolved tuff hosts only high-silica rhyolite pumice. The four intracaldera samples are of less evolved tuff in which TiO, contents are greater than or equal to 018%. BEST ET AL.: THE ZONED PAHRANAGAT ASH FLOW SHEET
5 10 15 20 25 30 Nb (PPm) Figure 15. Nb-Sr variation diagram for the Pahranagat Formation. Heavy solid curve labeled "Magmatic Variation Trend" is best fit to the salnples of pumice fr:igments (rZ = U.W2). Most samples of tuff fall to the right and above the ~~~ag~~raticvariation trend. suggesting that they formed by mixing of pyroclasts from different parts of the zoned magma chamber during eruption. To illustrate the chemical effects of this nlixing process, a series of straight mixing lines connect melt compositions on the magmatic variatinn curve with the most evolved melt magma (low Sr, high Nb). Light-dashed curves show the fraction of most evolved magma mixed with unevolved magma (high Sr, low Nb). For this nonunique scenario, mixing of as much as 35% pyroclasts from deep in the chamber with 65% pyroclasts from high in the chamber would form a mixture like tuff sample FO. Many other mixing combinations could create the observed tuff compositions.
Physical Mixing of Pyroclasts During Eruption mixtures represented in the tuff samples should lie above the magma trend but within the envelope defined by the pumice Regardless of the origi~lof the zonation in the magma chamber, fragment samples. The fact that pumice samples lack compos- the compositional variation of the tuff appears to retlect mixing of itional evidence of mirir~gargues that physical mixing of the pyroclasts from different parts of the chamber during explosive pyroclasts that make up tuff happened during explosive venting of eruption. Such eruptive mixing could produce the mixed mineral fragmented magma derived from different compositional zones in assemblages in samples of the outflow tacies of the Pahranagat the chamber (Figure 16). Formation. For example, plagioclase in the Alamo-type tuff In contrast to the Alamo-type tuff, tl~cSaulsbury closely sample FO has two distinctive compositions (Figure 13), morphol- resembles the quertz-rich rhyolite pumicc with regard to plagio- ogies (intact and corroded), and patterns of compositional zonation clase textures, proportions of phenocrysts, and bulk chemical (zoned and unzoned) which could have been derived by mixing of composition. This suggests that this type of tuff was derived by pyroclasts from magmas like evolved pumice AL-43PC and mafic explosive fragmentation of only thc magma in the upper part of the pumice WHRN-IJBPF. In addition, sample FO has clinopyroxene chamber represented in this pumicc. that coexists with abundant quartz and sanidine, a combination nut found in any pumice fragment. Entrain~nentot phenocrysts from Lateral Compositional Variatinns in the Outflow Sheet the upper, chemically evolved, and cooler pan of the chamber with trachydacitic lnagrlia CI~UIII deeper in the chamber provides a Stratigraphically complete sections of the outtlow tuff are quite reasonable cxplanation for these mincralagical features. More- different from place to placc within the sheel (Figurc 17). (This over, many samples of tuff, including FO, tllose frun~the MU has likely been a factor in the multiplicity of stratigraphic names section, and others, fall on trace element mixing trends between applied in different areas, as has been documented above.) The magmas, represented by pumice sao~ples, that had different [!lost evolved ejecta was drawn from the top of the magma compositions (Figure 15). For example, sample FO call be chamber and is restricted to the base ul proximal extracaldera modeled as a 3555 mixture of unevolved magma from deepcr in deposits where, for thc most part, the tuff is densely welded the chamber and evolved magma from the roof zone. The mixing Saulsbury type. More mafic ejecta, mostly of the Alamo type, is lines are not unique: many other mixing combinations involving locally preserved at the taps of proximal sections (MU) but is magmas in different parts of the cliat~~berare possible. The critical widespread in distal parts of the sheet north and especially east of observation here is that, givcn thc curvature of the incompatible the Kawich caldera. This ejecta represents mixing of pyroclasts versus compatible element variation trend for the magma, all from compositionally different zones of the cha~nbcr(Figurc 16) BEST ET AL.: THE ZONED PAHRANAGAT ASH FLOW SIIEET
Magma Chamber for Pahranagat Formation
1 Eruption lsochrons I I
Figure 16. Schematic successive eruption isochrons [Srhuruyfz ef 01.. 19891 suggest how the Pahranagat magma chamber. which was compositionally zoncd in thc manncr of Figure 11, might have been evacuated. during eruption. The Pahranagat eruption !nay have started as a volume of the erupted magma, represented in the outflow and more or less symmetric plume from which Saulshury-type ejecta intracaldera facies and taking into account 50% postemplacement spread rather equidistantly around the Kawich source. Subsequent east-west crustal extension, was at least 2000 km3. High-silica development of an asymmetric plume [Carey et ul., 19881 may rhyolite containing feldspar, quartz, biotite, and titanomagnetite have directed Alamo-type ejecta chiefly to the east. phenocrysts was erupted first and formed proximal parts of the outflow sheet that is hundreds of meters thick where it apparently Conclusions filled older calderas. Pyroclasts from deeper, low-silica rhyolite and trachydacite magmas that contained phenoc~ystsof plagioclase, High-precision, single-crystal 4oArI'9Ar ages and thermo- biotite, titanomagnetite, amphibole, and clinopyroxene were mixed remanent magnetization directions confirm that several lower with pyroclasts from the overlying more evolved rhyolite magma Miocene rhyolite tuffs in southern Nevada referred to by different during continued eruption. 'The resulting eruptively mixed ash stratigraphic names by previous workers are, in fact, one cooling flow deposit forms distal parts of the sheet and locally the upper unit whose present area of exposure covers 33,000 km'. This part of the proximal deposit. vertically and laterally zoned ash flow sheet of the Pahranagat Our multidisciplinary investigation of the outtlow sheet of the Formation was emplaced at 22.639k0.009 Ma while the Kawich Pahranagat Furmatiun lras rectified u~lcerlai~lliesin its correlatio~~ caldera was forming above an evacuating body of chemically and that occurred at first when unly traditional pctrugraphic and mincralugically ZUIIL.~I!lagnla. The total dense-rock equivalent stratigraphic tools had been employed. The substantial lateral, as
Figure 17. Lateral mmpositional variation in outtlow sheet of the Pahranagat Formation (see Figure 1). Boxes closed at top are complete sections ot the sheet; open boxes denote sections in which top has been eroded off. (a) 'TiO, cu~ee~~lrations.(h) Quartzlplagioclase ratio. 24,608 BEST ET AL.: THE ZONED PAHRANAGAT ASH FLOW SHEET well as the more commonly recognized vertical, com~ositional the northern Nellis Air Force Base Bombing and Gunnery Range, Nye zot~ing io lhe sheet, coll~poul~dedby its subsequent teclonlc County, Nevada, U.S. Geol. So.Pr@ Pap., 651. 91 pp., 1971. dismcmbcrmcnt and erosion into widcly scattered rcmnants ovcr a Ekren, E.B., C.L. Rogers, and G.L. Dixan, Geologic and Bouguer gravity very large area in the southern Great Basin has served as a map of the Reveille quadrangle, Nye County, Nevada, U.S. Geol. Sum. rigorous test case for correlation tools. Such cases necessitate Mop. 1-806, 1973. reduction in number of eruptive events, upward adjustments in Fuhrman. M.L.. and D.H. Lindslry, Ternary-feldspar modeling and volume of ejecta, and possible revision of models of ash flow thermometry, Am. Mineral., 73, 201-215, 1988. eruptions and of continental magma genesis. Gardner, J.N., A.C. Eddy, F.E. Golf, and K.S. Grafft, Reconnaissance geologic map of the northern Kawich and southern Reveille Ranges, Acknowledgments. Our rhanks to Donald M. Hudson for informarlon Nye County, Nevada, Lor Alornos Sci. Lab. Rtp. LA-83.90-MAP, UC- on the geology of the Kawich Range, to Steven I. Weiss and Jonathan T. 51, 1980. Hagstrum far considerable help in the field, to Mark R. Hudson for paleomagnetic data from the Nellis Air Force Bombing and Gunnery Ghiursu, MS., and R.O. Sack, Fe-Ti oxide geothernlollli.lry: Tl~crrnudy- Range, to William P. Wash for assistance with the electron microprobe namic formulation and the estimation of mntensivr variables in silicic analysis, and 10 Kim R. Sullivan fior volume calculations. Helpful reviews magmas, Co~~trib.Mirrrrul. Pelml., 108, 485-510, 1991. were provided hy Rnhen I.. Chrisriansen, Mark R. Hodqnn. Thnmaq P. Hildreth, W., Gradients in silicic magma chambers: Implications for Miller. and Thomas A. Vorel. This research was suooorted bv Brieham lithaspheric magmatism, J. Geophys. Res., 116, 10,153-10,192, 1981. Hildretll, W., and G. Mahood, Currelation of nsh-flaw ruffs, Geol. Soc. Am. BUN., 96, 968-974, 1985. References Hurford, A.J., F.J. Fitch, and A. Clarkc, Resolution of the age structure of the detrital zircon populations nf two 1.ower Cretaceous sandstones Bacon, C.R., and M.M. H~rschmann,MglMn partitioning as a test for fram the Weald of England by fisrinn track dating, Geol. Mag., 121, equilibrium between coexisting Fe-Ti oxides. Am. Mineral., 73, p. 57- 269-277, 1984. 61, 1988. Johnson, M.C., and M.J. Rutl~erfonl,Experimental calibration of the Bcst, M.G., and E.H. Christiansen, Limited extension during peak aluminum-in-hornblende gevbarometer with application to Long Valley Tenialy volcanism, Great Rasin nf Nevada and llmh, .I. Geophyr. RP.Y,, caldera (California) volcanic rocks. Geology, 17, 837-841, 1989. 96. 13.509-13.528. 1991. McFadden, P.L., and F.J. Lowes, The discrimination of mean directions Best, M.G., E.H. Christiansen, and H.R. Blank Jr., Oligocene calc- drawn from Fisher distributions, Geophys. J. R. Asfron. Soc., 67, 19- alkaline mcks of the Indian Peak volcanic field, Nevada and Utah, Geol. 33, 1981. Soc. Am. Bull., 101, 1076-1090, 1989a. Pearce, J.A., N.B.W. Harris, and A.G. Tindle, Trace element discrimina- Best, M.G., E.H. Chnstlansen, A.L. Deino, C.S. Fromme, F.H. McKee, tion diagrams for the tectonic insrpretation of granitic mcks, J. Petrol., and D.C. Noble. Eocene through Miocene volcanism in the Great Basin 25. 956-983, 1984, of the western United States, Mem. N M. Bur. Mines Mifler. Resour., Rosenbaum, J.G., Paleomagnrtic directional disgersion produced by plastic 47, 91-133, 1989b. deformation in a thick Miocene welded tuff. southern Nevada: Implica- Best, M.G., et al., Oligocene-Miocene caldera complexes, ash-fluw tions for welding temperaures, J. Geophys. Res., 91, 12,817-12,834. shseu, and user~iarnin lltr crnlral anal suuthcaste!n G~catBasin, in 1986. Cnunrl Evr,lurim r$ltlr Cirrul Burirr and ilrr Sierra Newdu, Field liip Rosenbaum, J.G., Magnetic grain-size variations through an ash-flow Guidebook for the 1993 Joint Meering of ihe Cordilleran/Roc!+ sheet: Influence on magnetic properties and irnpllcat~onsfor cooling Mormin Sections of the Geolugical Society of America, edited by M.M. history, J. Gropl~yr.Ref., 98. 11.715-1 1.727, 1993. Lahren, J.H. Trenler, Jr., and C. Spinosa, pp. 285-311, Mackay Samsun, S.D., and E.C. Alexander Ir., Calibration of the interlaboratory School of Mines, Reno, Nev., 1993. -ArP9Ar dating standard, MMhb-1, Chem. Geol. 6, 27-34, 1987. Carey, S.N., H. Sigurdmon, and R.S.J. Sparks, Experimental studies of Sargent, K.A., and K. Roggensack, Map showing oulcrops of prr- particle-laden plumes, J. Geophys. Res., 93, 15314-15328, 1988. Quaternary ash-flow Nffs and volcaniclastic rocks, Basln and Rangc Cebula, G.T., M.J. Kunk, H.H. Mchnm. C.W. Naeser. J.D. Obradovich. province, Nevada, U.S. Geoi. Sum. WaferRerour. Invest. Kep., 113- and J.F. Sutler, The Fish Canyon Tuff, a potential standard for the 4119-E, 1984. '!ArPYAr and fission-track methods (abstract), Two Cgnim, h 139- Schlinger, C.M., D. Gnscom, G.C. Papaefthymiau, and D.R. Veblen, 140, 1986. Christiansen, E.H., and M.G. Best, Lack of evidence for winnowing of The naNre of magnetic single domains in volcanic glasses of the KBS Tuff, .I.Geopltys. Res., 93, 9137.9156, 1988. glass during eluption and dispersal of a large volume rhyolite ash tlow, Geol. Soc. Am. Ahsrr. Progmms, 25 (6), 43-44, 1993. Schuraya, B.C., T.A. Vogel, and L.W. Younker, Evidence far dynamic Deino, A,, and R. Porn, Single-crystal u'Arl'VAr dating of the Olorgesailie withdrawal from a layered magma body: The Tapopah Spring Tuff, Formation. Southern Kenya Rift. J. Geophys. Res.. 95, 8453-8470, southwestern Nevada, I. Geophys. Res.. 94, 5925-5942, 1989. 19%. Scott, R.B., C.S. Gromme, M.G. Best, J.G. Rosenbaum, and M.R. Demo, A,, and K. Pons, Age-probability spectra for examination of singlc- Hudson, Srratigraphic relationships of Teniary volcanic rocks in central crystal "'ArPYAr dating results: Examples fram Oiorgesailie, Southern Lincoln County, southeastern Nevada, in Geologic Studies in thc Basin Kenya Rift, Quaf. Inrcrn., 718, 81-89, 1992. and Range to Colorado Plateau Transitiotl of the Nevada-Utah-Arizona Deino, A,, L. Taunr, M. MunagIran, and R. Drake, Single crystal Area, edited by R.B. Scott, U.S. Geol. Sum. Dull., 2056-A, in press, "'Ar1"Ar ages and the lltho and paleornagnetlc stratlgraphlrs of the 1995. Ngorora Formation. Kenya, I. Geol., 98, 567-587, 1990. Severinghaus, J., and T. Atwater, Celoruis pculrslry and tllcrrnal slalc ul Dield, J.F., M.E. Beck Jr., S. Brske-Diehl, D. Jacobson, and B.C. Hearn thc suMucting slabs bcncath wcstcrn North America, Mem. Geol. Soc. Jr., Palealnagnetism of Ulc Latc Crctaccous-Early Tzrtiary north-ccntral Am., 176, 1~22,1990. Montana alkalic province, J. Geophys. Rrs., 88, 10593-l0W9, 1983. Stewart, JH., and I.E. Carlsnn, Cenozoic rnckr of Nevada, Mop 52, Nev. Ekren, FH., K.E. Anderson, (:.L Rogers, and D.C. Noble, Geology of Bur. Mines Geol.. Reno. 1976. BEST ET AL.: THE ZONED PAHKANAGAT ASH FLOW SHEET 24,609
Taylor, J.R.. An Introduclion lo Error Analysis, UNV. Sci. Books, Mill Geology, Brigham Young University, Provo, UT 84602. (e-mail:
~~~ ~ [email protected]; eric [email protected]; dg@geology. Vallev.,. Calif.. 1982. .. . - - - .. Whitebread, D.H., Geologic map of the Tanopah 1- X 2. quadrangle, byu.edu) A. L. Drino, Berkeicy Genchrnnnlogy Center, Berkeley, CA 94709. (e- central Nevada, U.S. Geol. Sum. Misc. Field Stud. Map. MF-I877A, mail: adeinomhec.ore, - 7~ L. 1vb9. (:. S. Gro~~ull.'.I' S Cir.dI~gt;:~lSur,.:).. 31q hl~JJlcri