A clinopyroxene–basalt geothermobarometry perspective of Columbia Plateau (NW-USA) Miocene magmatism Graziella Caprarelli 1 and Stephen P. Reidel 2 1 Department of Environmental Sciences, University of Technology, Sydney, PO Box 123, Broadway, NSW 2007, Australia; 2 Pacific Northwest National Laboratory, MS K6-81, PO Box 999, Richland, WA 99352, USA Introduction After more than three decades of studies of the tholeiitic c. 17–6 Ma Miocene Columbia River Basalt Group (CRBG) in NW-USA, no unequivocal conclusion has been reached with regard to the origin of these magmas. Sources postulated include: depleted, primitive and me- tasomatized asthenosphere, subconti- nental lithosphere, and combinations of these, with various degrees of mixing and contamination with cru- stal materials (e.g. Carlson, 1984; Hart and Carlson, 1987; Brandon and Goles, 1988; Hooper and Hawkes- worth, 1993; Brandon and Goles, 1995). This uncertainty has hampered interpretation of the contemporary tectonic conditions, although a gen- eral extensional environment of erup- tion is undisputed. Interpretations of tectonism and geodynamic evolution are in turn connected to the cause of CRBG magmatism: the Yellowstone mantle plume (Brandon and Goles, 1988; Hooper and Hawkesworth, 1993; Brandon and Goles, 1995; Camp and Ross, 2004); decompres- sion melting because of back-arc spreading (Hart and Carlson, 1987; Catchings and Mooney, 1988; Smith, 1992); lithospheric pull-apart (Ander- son, 1994); rifting over an anomal- ously hot mantle (White and McKenzie, 1989). Insights into the tectonic evolution and magma sources of the Columbia Plateau (CP) can be gained by deter- mining the pressures and temperatures of the rising magmas, because these reflect residence times and lithospheric evolution. Our in-depth clinopyroxene – basalt geothermobarometric study of the 16.5–15.6 Ma Grande Ronde Basalt (GRB; Caprarelli and Reidel, 2004) indicated that a normal asthen- ospheric mantle source and fraction- ation of magmas during ascent, are compatible with the major element chemistry of these rocks. Further studies of the physical evolution of magmas during their ascent through the lithosphere should provide valu- able information with regard to plumbing, ultimately related to the stress state of the lithosphere. This approach could prove useful in even- tually breaking the cycle of inter- dependent geodynamic and source hypotheses for the CRBG. Acme of CRBG tholeiitic volcan- ism occurred between 17 and 14.5 Ma. Volcanism waned with the eruption of members of the Saddle Mountains Basalt, which lasted until 6 Ma. In this paper, we expand on our geothermobarometric GRB investigation (Caprarelli and Reidel, 2004) by adding results from other formations, and applying [A] and [B] models of Putirka et al. (2003). These algorithms are specifically calibrated to recover values of pressure and temperature from low Fe 3+ clinopy- roxene and whole-rock compositions, provided the rocks are nearly aphyric. Analytical methods and results Samples spanning the CRBG strati- graphy were collected in the State of Washington. Outcrop sampling sites and a simplified stratigraphic se- quence are shown in Fig. 1, where the positions of borehole BN1–9 and GRB sampling area (grey shaded rectangles), investigated by Caprarelli and Reidel (2004), are also indicated. Descriptions of GRB samples are in Caprarelli and Reidel (2004) and are not repeated here. We selected the most aphyric samples available. Table 1 and Fig. 2 provide a descrip- tive synthesis of the main petrographic characteristics of the samples. No glass was found in the groundmass of any sample. Bulk rock major and trace element compositions (Table 2) were meas- ured on fused discs and pressed pellets, respectively, using a Philips PW 1480 X-ray spectrometer at the Department of Geology and Geo- physics of the University of Adelaide (Australia). Microanalyses of clino- pyroxenes were carried out at GE- MOC, Macquarie University (Australia), using a Cameca SX50 electron microprobe, and at the As- tromaterials Research and Explora- ABSTRACT The origin of NW-USA Columbia River Basalt Group Miocene magmatism and its relation to tectonism has been widely debated and is still open to study. We investigated the pre- eruptive evolution of the magmas, to constrain pressures and temperatures of the ascending magmas, and plumbing condi- tions. We determined major element concentrations of 17–6 Ma tholeiites, and applied clinopyroxene – liquid geothermobaro- metry to calculate pre-eruptive pressures and temperatures. These ranged from 0 to 0.66 GPa and 1120 to 1222 °C, respectively, defining two age-related parallel trends in a P–T diagram. This indicates a consistent crustal evolution of the magmas, and records at least two distinct initial temperatures. Using clinopyroxene interdiffusion coefficients we estimated magma ascent speeds ‡ 0.6 km yr )1 . Possible geological explanations for the calculated parameters are: lower-crust magma chamber processes; magmatism and tectonism feed- back consistent with an extensional environment. Terra Nova, 17, 265–277, 2005 Correspondence: Graziella Caprarelli, Department of Environmental Sciences, University of Technology, Sydney, PO Box 123, Broadway, NSW 2007, Australia. Tel.: 00 61 2 9514 1776; fax: 00 61 2 9514 1755; e-mail: [email protected]Ó 2005 Blackwell Publishing Ltd 265 doi: 10.1111/j.1365-3121.2005.00611.x
Transcript
A clinopyroxene–basalt geothermobarometry perspective ofColumbia Plateau (NW-USA) Miocene magmatism
Graziella Caprarelli1 and Stephen P. Reidel21Department of Environmental Sciences, University of Technology, Sydney, PO Box 123, Broadway, NSW 2007, Australia; 2Pacific
Northwest National Laboratory, MS K6-81, PO Box 999, Richland, WA 99352, USA
Introduction
After more than three decades ofstudies of the tholeiitic c. 17–6 MaMiocene Columbia River BasaltGroup (CRBG) in NW-USA, nounequivocal conclusion has beenreached with regard to the origin ofthese magmas. Sources postulatedinclude: depleted, primitive and me-tasomatized asthenosphere, subconti-nental lithosphere, and combinationsof these, with various degrees ofmixing and contamination with cru-stal materials (e.g. Carlson, 1984; Hartand Carlson, 1987; Brandon andGoles, 1988; Hooper and Hawkes-worth, 1993; Brandon and Goles,1995). This uncertainty has hamperedinterpretation of the contemporarytectonic conditions, although a gen-eral extensional environment of erup-tion is undisputed. Interpretations oftectonism and geodynamic evolutionare in turn connected to the cause ofCRBG magmatism: the Yellowstonemantle plume (Brandon and Goles,1988; Hooper and Hawkesworth,1993; Brandon and Goles, 1995;Camp and Ross, 2004); decompres-sion melting because of back-arcspreading (Hart and Carlson, 1987;Catchings and Mooney, 1988; Smith,
1992); lithospheric pull-apart (Ander-son, 1994); rifting over an anomal-ously hot mantle (White andMcKenzie, 1989).Insights into the tectonic evolution
and magma sources of the ColumbiaPlateau (CP) can be gained by deter-mining the pressures and temperaturesof the rising magmas, because thesereflect residence times and lithosphericevolution. Our in-depth clinopyroxene– basalt geothermobarometric studyof the 16.5–15.6 Ma Grande RondeBasalt (GRB; Caprarelli and Reidel,2004) indicated that a normal asthen-ospheric mantle source and fraction-ation of magmas during ascent, arecompatible with the major elementchemistry of these rocks. Furtherstudies of the physical evolution ofmagmas during their ascent throughthe lithosphere should provide valu-able information with regard toplumbing, ultimately related to thestress state of the lithosphere. Thisapproach could prove useful in even-tually breaking the cycle of inter-dependent geodynamic and sourcehypotheses for the CRBG.Acme of CRBG tholeiitic volcan-
ism occurred between 17 and14.5 Ma. Volcanism waned with theeruption of members of the SaddleMountains Basalt, which lasted until6 Ma. In this paper, we expand onour geothermobarometric GRBinvestigation (Caprarelli and Reidel,2004) by adding results from otherformations, and applying [A] and [B]models of Putirka et al. (2003). These
algorithms are specifically calibratedto recover values of pressure andtemperature from low Fe3+ clinopy-roxene and whole-rock compositions,provided the rocks are nearlyaphyric.
Analytical methods and results
Samples spanning the CRBG strati-graphy were collected in the State ofWashington. Outcrop sampling sitesand a simplified stratigraphic se-quence are shown in Fig. 1, wherethe positions of borehole BN1–9 andGRB sampling area (grey shadedrectangles), investigated by Caprarelliand Reidel (2004), are also indicated.Descriptions of GRB samples are inCaprarelli and Reidel (2004) and arenot repeated here. We selected themost aphyric samples available.Table 1 and Fig. 2 provide a descrip-tive synthesis of the main petrographiccharacteristics of the samples. Noglass was found in the groundmassof any sample.Bulk rock major and trace element
compositions (Table 2) were meas-ured on fused discs and pressedpellets, respectively, using a PhilipsPW 1480 X-ray spectrometer at theDepartment of Geology and Geo-physics of the University of Adelaide(Australia). Microanalyses of clino-pyroxenes were carried out at GE-MOC, Macquarie University(Australia), using a Cameca SX50electron microprobe, and at the As-tromaterials Research and Explora-
ABSTRACT
The origin of NW-USA Columbia River Basalt Group Miocenemagmatism and its relation to tectonism has been widelydebated and is still open to study. We investigated the pre-eruptive evolution of the magmas, to constrain pressures andtemperatures of the ascending magmas, and plumbing condi-tions. We determined major element concentrations of 17–6 Matholeiites, and applied clinopyroxene – liquid geothermobaro-metry to calculate pre-eruptive pressures and temperatures.These ranged from 0 to 0.66 GPa and 1120 to 1222 �C,respectively, defining two age-related parallel trends in a P–T
diagram. This indicates a consistent crustal evolution of themagmas, and records at least two distinct initial temperatures.Using clinopyroxene interdiffusion coefficients we estimatedmagma ascent speeds ‡ 0.6 km yr)1. Possible geologicalexplanations for the calculated parameters are: lower-crustmagma chamber processes; magmatism and tectonism feed-back consistent with an extensional environment.
tion Science Office of NASA (Hous-ton, TX, USA), by a Cameca SX100electron microprobe. Operating con-ditions were 15 kV and 20 nA for allelements.
Whole rock silica contents rangefrom 49.56 wt% (LM0110-a) to 51.29wt% (Po0111-b). Abundances of totaliron, measured as Fe2O3, range from11.77 wt% (Po0111-b) to 15.71 wt%
(G013). Magnesium oxide contentsare 4.09 wt% (G013) to 6.71 wt%(Po0111-b). Magnesium numbers[Mg/(Mg + Fe2+)] range from 0.36(G013) to 0.56 (Po0111-b). Concen-trations of Cr range from 13 ppm(G013) to 103 ppm (Po0111-b), attest-ing to the fractionated nature of themagmas.All analysed pyroxenes were aug-
ites. Where the minerals were suffi-ciently large (e.g. ‡ 100 lm), wecarried out traverse point analyses toinvestigate any compositional differ-ences between centres and rims of thecrystals. Figure 3 illustrates the com-positional overlap between centre andrim data points.
Model pressures and temperatures
As no glass is present in the rocks thatcould be analysed by EPMA, nodirectly measured liquid compositionswere available. Given the aphyricnature of the rocks we used wholerock compositions as proxies for themelts. After calculating all pyroxeneanalyses per 6 oxygen basis, theamount of Fe3+ in all analysed clino-pyroxenes was calculated by massbalance (Lindsley, 1983), and foundto be minimal (Table 3). Use of geo-thermobarometric models of Putirkaet al. (2003) required the followingpreliminary steps: (i) calculation ofcation fractions from whole rockmajor element analyses; (ii) calcula-tion of clinopyroxene componentsDiHd (diopside + hedenbergite),EnFs (enstatite + ferrosilite), andSum of the total components (defini-tions in Putirka et al., 2003) for theanalysed clinopyroxenes; (iii) calcula-tion of the DiHd, EnFs and Sumparameters for model clinopyroxenesthat would be at equilibrium with aliquid having the same chemical com-position of the samples; (iv) compar-ison of these values with the DiHd,EnFs and Sum components of themeasured pyroxenes. Where the modelDiHd and EnFs parameters were thesame as those calculated from themeasured clinopyroxenes within 2rerror, and the sum within ±0.1 ofunity (Fig. 4), the analysed clino-pyroxene compositions were taken tobe at equilibrium with the whole rockanalyses. After thus filtering the data,we were left with 118 analyses. Ta-ble 3 contains the selected data and
Columbia River
124 122
Extent of CRBG
120 118 116
48Idaho
Washington
RF0115
C4184
1E11-34.4;
1m5-15.0DC8/1075.1
BN1-9
ES0113
PR0112;
Po0111LM0110
Snake River
a
b
N
G013
AncientSalmon /ClearwaterRivers
46
44Oregon
GR019
DY018GR014IB017;Dy016
IB015;
DY01DY018888888888888
B0BBBBBBBBBBBIIIGR
50 km
Fig. 1 Geographical extent of main Columbia River Basalt Group Miocene lavas,and locations, ages and simplified stratigraphic positions of samples discussed in thispaper. In (a) the thick grey line represents the areal distribution of the lavas, and thebold black dots depict the locations from where the samples (labelled) were collected.The rectangular dark shaded areas correspond to the geographical locations of theGrande Ronde Basalt (GRB) samples (not all reported in this figure) investigated byCaprarelli and Reidel (2004), including the position of borehole BN1–9. Only thestratigraphic positions of the samples discussed in this paper are indicated in (b).These include GRB samples already described by Caprarelli and Reidel (2004) andnot shown in (a). Small case letters were added to some of the sample runningnumbers (e.g. IB017-b) when there was more than one representative fragmentcollected from an outcrop, and indicate which fragment of rock was analysed anddiscussed in the paper. Ages are after Tolan et al. (1989). Bold horizontal linesindicate erosional unconformities.
Columbia Plateau Miocene magmatism • G. Caprarelli and S.P. Reidel Terra Nova, Vol 17, No. 3, 265–277
synthetic information on the size ofthe crystals and location of analysedpoints of the new samples. The GRBanalyses already presented in Capra-relli and Reidel (2004) are not repea-ted in the table. The fact that nospatial compositional variations occurin the pyroxene crystals and that coreand rim compositions equally fit thepredicted equilibria within ±2r, con-firmed that clinopyroxenes crystallizedin equilibrium with their melts and didnot re-equilibrate after crystallization.After calculating pressures and tem-peratures (Table 3) and finding that
Table 1 Synthesis of petrographic characteristics of samples
Sample Texture
Microadglomerates
(cpx + plag)
Microphenocrysts
(£0.5 mm)
Microphenocrysts
(£0.2 mm)
Matrix and
groundmass
LM01110 Aphyric cpx, plag, rare ol cpx, plag, mt
Po0111 Aphyric cpx, plag cpx, plag, mt, ilm
1m5–15.0 Aphyric cpx, plag cpx, plag, mt, ilm
1E11–34.4 Aphyric Yes cpx, plag, rare ol cpx, plag, mt, ilm
C4184 Aphyric Yes cpx, plag, rare ol cpx, plag, mt, ilm
DC8/1075.1 Aphyric cpx, plag cpx, plag, mt, ilm
RF0115 Aphyric plag cpx, plag,
rare ol (corr)
cpx, plag, mt, ilm
G013 Aphyric plag cpx, plag, rare ol cpx, plag, rare ilm
IB017 Aphyric Yes cpx, plag, rare ol cpx, plag, mt, ilm
Fig. 2 Examples of back-scattered electron (BSE) images of Grande Ronde Basalt (a–b; samples BLS30 and DOG28, respectively),Pomona (c; sample 1E11–34.4) and Lower Monumental (d; sample LM0110-a) members, obtained by the Cameca SX 50 electronprobe. Rock textures are aphyric, with mineral grain sizes all <0.5 mm. Plagioclase (dark) and pyroxene (grey) are ubiquitous asmicrophenocrysts and in the rocks matrix and groundmass. Low modal percentages of olivine (light grey) are present in somesamples (e.g. LM0110-a). Bright minerals are titanomagnetite and ilmenite. These are often skeletal, and almost exclusively presentin the matrix and in the groundmass.
Terra Nova, Vol 17, No. 3, 265–277 G. Caprarelli and S.P. Reidel • Columbia Plateau Miocene magmatism
the calculated temperatures (Putirkaet al., 2003) matched clinopyroxenesaturation temperatures (Putirka,1999) within 1r error (Fig. 5), wewere satisfied that the clinopyroxenecomponents–melt relationships arenot coincidental and, consequently,that the P–T estimates are valid. Nocorrelation between crystal size andP–T values is evident (Fig. 6).Averaging pressures and tempera-
tures calculated from clinopyroxeneanalyses provides a more precise esti-mate of intensive variables comparedwith individual analyses (Putirkaet al., 2003). Therefore, we plottedsingle clinopyroxene averaged valuesin a P–T diagram (Fig. 7). The dataare distributed along two parallellinear trends: a general CRBG trend,coincident with Caprarelli and Rei-del’s (2004) GRB trend, and a highertemperature Pomona Member trend.
The highest pressure calculated is0.66 GPa.
Interpretation
Thermodynamic intensive variables(P, T) and chemical potential arerelated. The Pomona Member rockshave the highest Mg number amongthe analysed samples and reflect high-er temperatures of crystallization.Therefore, we are confident that thevalues we obtained are internally con-sistent, and that the two trends iden-tified in Fig. 7 are significant withregard to temperature. The modelpressure error (±0.17 GPa) is equiv-alent to the spread of pressure valuescalculated for each individual sample(Fig. 7). This suggests that the modelhas at least as much tolerance withrespect to variations of major elementcontents both in the clinopyroxenes
and in the whole rocks. Thus, theequilibrium clinopyroxene–melt pairsrepresent a �frozen� record of thedepth conditions of clinopyroxenecrystallization compatible with thepreserved bulk rock chemical compo-sitions.Considering model clinopyroxene
grains of radius ranging from 0.01 to0.1 mm (consistent with the dimen-sions our crystals), and Brady andMcCallister’s (1983) 0.25 GPa clino-pyroxene interdiffusion coefficients,ranging from 2.0 · 10)16 cm2 s)1
(1150 �C) to 8.0 · 10)16 cm2 s)1
(1200 �C), CRBG clinopyroxenesshould take from a maximum of c.16 kyr to a minimum of c. 40 yr to re-equilibrate. Given the lack of a posit-ive correlation between crystal sizeand P–T values (Fig. 6), the smallercrystal estimate represents the upperlimit of diffusion: the clinopyroxenesmust have ascended all the way to thesurface in £ 40 years to retain therecord of their deeper conditions ofcrystallization. Taking the maximumrecorded depth (c. 25 km) the averagespeed of ascent was ‡ 0.6 km yr)1.This is a conservative estimate,because Brady and McCallisters’(1983) diffusion coefficients experi-ments were conducted for Fe-poorcompositions and at constant pres-sure. The effect of pressure increase(and corresponding temperatureincrease) and of presence of Fe in themedium is that of enhancing thediffusion coefficient values, thus lead-ing to even faster rates of ascent. Thelinear P–T trends defined by the rocksindicate that the magmas did notstagnate in the upper crust for anylength of time sufficient for the clino-pyroxene–melt systems to evolve andchange the compositions of the higherpressure clinopyroxenes, otherwise allthe high pressure clinopyroxene–liquid pairs would have reset theirgeobarometers to lower pressure val-ues. Melting and assimilation of uppercrustal material could not haveoccurred, because this process wouldhave caused a modification of theliquids, and the clinopyroxenes pres-ently yielding higher values of pres-sure would not have been atequilibrium with the whole rocks.Also this process would have resultedin lower values of maximum estimatedpressures. It is therefore legitimate tointerpret the linear trends in Fig. 7 as
Table 2 Bulk chemical compositions
Sample IB017-b G013 RF0115-a Po0111-b LM0110-a
SiO2 50.06 50.48 50.43 51.29 49.56
TiO2 2.81 2.98 3.2 1.70 2.90
Al2O3 13.79 13.23 13.66 14.61 13.69
Fe2O3 14.915 15.71 14.55 11.77 15.49
MnO 0.2 0.19 0.22 0.18 0.22
MgO 4.69 4.09 4.11 6.71 5.01
CaO 8.94 8.03 8.81 10.72 8.67
Na2O 3.13 2.91 2.8 2.32 2.75
K2O 0.995 1.273 1.396 0.636 1.458
P2O5 0.4 0.60 0.68 0.24 0.64
Total 99.93 99.50 99.86 100.18 100.39
Mg# 0.41 0.36 0.38 0.56 0.42
Cr 84 13 41 103 24
Fe2O3 is total iron. Mg# were calculated assuming a Fe3+/(total iron) ratio ¼ 0.1.
Fig. 3 Pyroxene quadrilateral with data points representing the compositions of coresand centres (white circles) and rims (grey diamonds) of Columbia River Basalt Groupaugites. Correction methods for Na and Al and 1 bar isotherms are after Lindsley andAndersen (1983). Isotherms are plotted for illustrative purposes only. Seventy-threecore/centre and 39 rim analyses from all samples (including Caprarelli and Reidel’s,2004 Grande Ronde Basalt data) are represented in the diagram. Centre and rim datapoints overlap, with only three rim data points of a total of 112 analyses (or <2.7%)stranding from the rest of the data. This indicates no substantial difference betweenpyroxene core/centre and rim compositions.
Columbia Plateau Miocene magmatism • G. Caprarelli and S.P. Reidel Terra Nova, Vol 17, No. 3, 265–277
a clinopyroxene–melt closed systemrecord of magma crustal ascent paths.
Geological explanations
The parallel trends in Fig. 7 indi-cate that CRBG magmas hadsimilar crustal histories, but at leasttwo different initial temperatures. Thishas several possible explanations:1. Higher temperature magmas re-
flect higher mantle potential tempera-tures. The hotter magmas are those of12 Ma Pomona Member, whicherupted during the waning stages ofCRBGactivity, that is, after the bulk ofthe volume of CRBG had already beenemplaced. We expect higher potentialtemperatures to translate into large
volumes of erupted magmas, howeverfractionated, so this hypothesis contra-dicts the geological record.2. Temperature differences reflect
magma chamber heterogeneities (a).Magma fractionation involves growthand erosion of solidification fronts(Marsh, 1995), the thickness of whichdepends on temperature, chemistry,shape and development of magmachambers and fracture zones. Frontscomprise rigid crust, highly viscouscrystallinity region, mush zone, bor-der capture zone and suspensionzone. Physical conditions vary acrossthe solidification front, with highertemperatures and lower viscosities atthe interface between suspension zoneand crystal free magma. The high
viscosities of the external zones of thefront prevent them from rising to thesurface, but movement of the mushand suspension zones across themcan cause their rupture and theirmaterial can be entrained in ascend-ing magmas. Various combinationsof these simple factors can produce alimitless range of compositions andtextures. It is possible to surmise thata fortuitous set of �just right� condi-tions involving suitable variations insolidification front thickness, compo-sitions, temperatures and viscositiesthroughout the 17–6 Ma history ofthe CRBG, produced the highly uni-form rock textures and compositionsof our samples. This would be anextraordinary (and, in our opinion,
Fig. 4 Test for clinopyroxene–liquid equilibrium, carried out according to Putirka’s (1999) model, based on clinopyroxene EPMAanalyses and assuming whole rocks compositions to represent the true liquid. Parameters DiHd (diopside + hedenbergite), EnFs(enstatite + ferrosilite), and total component sums for both measured clinopyroxene and ideal compositions are compared.Parameters calculated from the analytical values are read on the horizontal axis (measured), predicted values are read on thevertical axis. Diagonal lines are equilines (1 : 1 lines) and indicate total agreement between measured and predicted compositions.Vertical bars are statistical 2r errors. Only analyses which are within 2r from the equiline are considered to represent conditions ofequilibrium between mineral and melt. Equilibrium tests for samples of Caprarelli and Reidel (2004) Grande Ronde Basalt are notrepeated here.
Columbia Plateau Miocene magmatism • G. Caprarelli and S.P. Reidel Terra Nova, Vol 17, No. 3, 265–277
unlikely) case of convergence, partic-ularly when ascent of magmasthrough the filter of about 35 km ofcrust is considered.
3. Temperature differences reflectmagma chamber heterogeneities (b).In a simpler magma chamber model,the low temperature trend could be
generated by eruption of cooler andmore fractionated magmas concentra-ted along the magma chamber upperwall, more readily expelled to thesurface than the high temperaturemagmas, likely concentrated in thecore of the chamber (Marsh, 1995).Random and localized conditions inspace and time can thus account forthe 12 Ma high temperature trend.Magma chamber heterogeneities
cannot be unequivocally discountedon the grounds of our geothermobaro-metric evidence and are thereforeplausible geological explanations forthe observed trends. However, at pre-sent there is no geophysical evidencefor lower-crust magma chambers inthe CP. Therefore, we propose analternative scenario as a possibilitydeserving further investigation.4. The parallel trends reflect a funda-
mental temperature difference, arisingprior to magma chamber evolution. Inthe context of the dynamics of magmaemplacement during extension of con-tinental lithosphere (Bonini et al.,2001), magma bodies at depth weakenthe lithosphere, localize strain andthus increase lithospheric deformation(White and McKenzie, 1989). There-fore, the 12 Ma Pomona Membertemperatures could indicate a shorterresidence time of these magmas in alithosphere thinned as a consequenceof voluminous earlier magmatic activ-ity. By the time of eruption of the lowvolume Lower Monumental Basalt,6 Myr subsequent to Pomona volca-nicity, the basalt ascent path hadreturned to that normal for CRBG,suggesting a counterflow of ductilematerial which re-thickened the litho-sphere. Upon thickening of the mantlelithosphere, termination of extensionoccurred, possibly because of thestronger rheology of the mantle asopposed to that of the crust (House-man and England, 1986).This scenario reconciles extensional
tectonic regime and flood magmatismand is worth pursuing, because it mightlead to the definition of an interpretat-ive framework in which to test CPmagmatism and tectonism models.
Conclusions
Two principal conclusions can bedrawn from our geothermobarometricstudy: (i) The upper-to middle-crustresidence times of the CRBG magmas
Fig. 5 Clinopyroxene saturation temperatures (Putirka, 1999), determined using thecalculated pressures and whole rock compositions, are plotted against the modelcrystallization temperatures (Putirka et al., 2003). The diagonal line is the equiline.Almost all the saturation and model temperatures match within 1r error.
Fig. 6 Two-class frequency histograms of pyroxene–liquid equilibrium calculatedtemperatures and pressures (a, b), and temperatures and pressure ranges (c, d), forsmaller (<100 lm; white areas and symbols) and larger (>100 lm; grey areas andsymbols) pyroxene crystals. In (a) and (b) the frequencies are: 46% for smaller grainsand 54% for larger grains in the 1120–1170 �C class, and 54% and 46% in the 1171–1220 �C class, for smaller and larger grains, respectively; 42% (smaller crystals) and50% (larger crystals) in the 0–0.33 GPa class, and 58% (smaller) and 50% (larger) inthe 0.331–0.66 GPa class. The spreads of temperatures and pressures in (c) and (d) aresimilar for smaller and larger crystals, although the temperature and pressure highestvalues were obtained from larger grains. This, however, is compensated by the factthat all frequencies (a and b) are substantially close to 50%. The combined evidenceprovided by the frequency and distribution of temperature and pressure valuesindicates lack of a correlation between calculated temperatures and pressures.
Terra Nova, Vol 17, No. 3, 265–277 G. Caprarelli and S.P. Reidel • Columbia Plateau Miocene magmatism
were short and equal throughout theentire volcanological history of theCP; (ii) Prior to middle- and upper-crustal ascent, the residence time forthe 12 Ma magmas may have beenshorter than for the rest of the studiedmagmas. This possibility deserves fur-ther investigation.We are therefore conducting addi-
tional geothermobarometric studiesaimed at identifying any other tem-perature regime changes in the 17–6 Ma history of this magmatic region.
Our geothermobarometric approachis promising and might provide in-sights into the geodynamic evolutionand tectonic setting of this region.
Acknowledgements
GC thanks Norm Pearson and CraigSchwandt for access to the Cameca SX50and SX100, John Stanley for performingthe XRF analyses, Evan Leitch for criti-cism of earlier versions of this paper.Thoughtful reviews by Vincent Famin
and Angelo Peccerillo focused and im-proved the paper. Carlo Doglioni handledscientific editing of the paper. Anne Borc-herds promptly helped with all mattersrelated to uploading and on-line editing.
References
Anderson, D.L., 1994. The sublithosphericmantle as the source of continentalflood basalts; the case against thecontinental lithosphere and plume headreservoirs. Earth Planet. Sci. Lett., 123,269–280.
Bonini, M., Sokoutis, D., Mulugeta, G.,Boccaletti, M., Corti, G., Innocenti, F.,Manetti, P. and Mazzarini, F., 2001.Dynamics of magma emplacement incentrifuge models of continental exten-sion with implications for flank mag-matism. Tectonics, 20, 1053–1065.
Brady, J.B. and McCallister, R.H., 1983.Diffusion data for clinopyroxenes fromhomogenization and self-diffusionexperiments. Am. Mineral., 68, 95–105.
Brandon, A.D. and Goles, G.G., 1988. AMiocene sub-continental plume in thePacific Northwest: geochemical evidence.Earth Planet. Sci. Lett., 88, 273–283.
Brandon, A.D. and Goles, G.G., 1995.Assessing subcontinental lithosphericmantle sources for basalts. Neogenevolcanism in the Pacific Northwest,USA, as a test case. Contrib. Mineral.Petrol., 121, 364–379.
Camp, V.E. and Ross, M.E., 2004. Mantledynamics and genesis of mafic magma-tism in the intermontane Pacific North-west. J. Geophys. Res., 109, B08204, doi:10.1029/2003JB002838.
Caprarelli, G. and Reidel, S.P., 2004.Physical evolution of Grande RondeBasalt magmas, Columbia River BasaltGroup, north-western USA. Mineral.Petrol., 80, 1–25.
Carlson, R.W., 1984. Isotopic constraintsin Columbia River flood basalt genesisand the nature of the subcontinentalmantle. Geochim. Cosmochim. Acta, 48,2357–2372.
Catchings, R.D. and Mooney, W.D., 1988.Crustal structure of the Columbia Plat-eau – evidence for continental rifting.J. Geophys. Res., 93, 459–474.
Hart, W.K. and Carlson, R.W., 1987.Tectonic controls on magma genesis andevolution in the northwestern UnitedStates. J. Volcanol. Geotherm. Res., 32,119–135.
Hirschmann, M.M., 2000. Mantle solidus:experimental constraints and the effect ofperidotite composition. Geochem. Geo-phys. Geosystems, 1, paper no.2000GC000070.
Hooper, P.R. and Hawkesworth, C.J.,1993. Isotopic and geochemicalconstraints in the origin and evolution of
Fig. 7 Pressure–temperature diagram illustrating the crystal-averaged pressures andtemperatures of the magmas calculated from clinopyroxene-liquid equilibria ofColumbia River Basalt Group samples (data points; symbols in legend reported invertical order, starting from the youngest member samples on the top). The depths ofthe lower crust (LC, grey area) and of the Moho (thick dark grey horizontal line at35.5 km depth) were calculated by Caprarelli and Reidel (2004). Dry peridotitesolidus (after Hirschmann, 2000), and position of the adiabat calculated considering apotential temperature of 1280 �C and a gradient of 20 �C/GPa, are given as reference.The largest pressure and temperature errors inherent in the model (arrowed errorbars) are ±0.17 GPa and ±33 K, respectively (Putirka et al., 2003). Two trends areclearly discriminated beyond temperature model error: a lower temperature trend(R2 ¼ 0.8510) and a higher temperature trend (R2 ¼ 0.8263). The higher temperaturetrend is that of Pomona magmas. The lower temperature trend comprises all othersamples, and includes data from Caprarelli and Reidel (2004) Grande Ronde Basaltdata set (GRB, black diamonds) recalculated using equations of Putirka et al. (2003)and crystal averaged. The thin line along the lower temperature trend represents thevalues of pressure and temperature of clinopyroxene crystallization for GRBcalculated using models of Putirka et al. (1996) (Caprarelli and Reidel, 2004). Bothmethods give highly consistent and reproducible results. The statistically significantP–T correlations in both trends, i.e. the close alignment of the points along the trendswithout substantial horizontal displacement, suggest that the clinopyroxenes do notrecord any intratelluric ponding event during ascent of the magmas through themiddle and upper crust. The dashed lines with question marks represent eitherpossible P–T ranges of evolving magmas in lower crust magma chambers, or lowercrust paths of magmas ascending from a deeper reservoir, possibly located below theMoho, where the dashed lines converge. The geothermobarometric method we useddoes not recover a record of magmatic evolution at these depths.
Columbia Plateau Miocene magmatism • G. Caprarelli and S.P. Reidel Terra Nova, Vol 17, No. 3, 265–277
Lindsley, D.H. and Andersen, D.J., 1983.A two-pyroxene thermometer. J. Geo-phys. Res., 88 (Suppl.), A887–A906.
Marsh, B.D., 1995. Solidification frontsand magmatic evolution. Min. Mag., 60,5–40.
Putirka, K., 1999. Clinopyroxene + liquidequilibria to 100 kbar and 2450 K.Contrib. Mineral. Petrol., 135, 151–163.
Putirka, K., Johnson, M., Kinzler, R.,Longhi, J. and Walker, D., 1996. Ther-mobarometry of mafic igneous rocksbased on clinopyroxene-liquid equilibria,0–30 kbar. Contrib. Mineral. Petrol.,123, 92–108.
Putirka, K.D., Mikaelian, H., Ryerson,F. and Shaw, H., 2003. New clino-pyroxene-liquid thermobarometers formafic, evolved, and volatile-bearinglava compositions, with applications tolavas from Tibet and the Snake RiverPlain, Idaho. Am. Mineral., 88, 1542–1554.
Smith, A.D., 1992. Back-arc convectionmodel for Columbia River basalt gen-esis. Tectonophysics, 207, 269–285.
Tolan, T.L., Reidel, S.P., Beeson, M.H.,Anderson, J.L., Fecht, K.R. and Swan-son, D.A., 1989. Revisions to the esti-mates of the areal extent and volume ofthe Columbia River Basalt Group. In:Volcanism and Tectonism in the ColumbiaRiver Flood-Basalt Province (S.P. Reideland P.R. Hooper, eds), Geol. Soc. Am.Spec. Pap., 239, 1–20.
White, R.S. and McKenzie, D., 1989.Magmatism at rift zones: the generationof volcanic continental margins andflood basalts. J. Geophys. Res., 94, 7685–7729.
Received 16 September 2004; revised versionaccepted 3 January 2005
Terra Nova, Vol 17, No. 3, 265–277 G. Caprarelli and S.P. Reidel • Columbia Plateau Miocene magmatism
