Open System Evolution of Peralkaline Trachyte and Phonolite from the Suswa Volcano, Kenya Rift

Home , Menengai, Mount Sidley, Mount Suswa, Phonolite, Suswa, Trachyandesite

Eastern Kentucky University Encompass

EKU Faculty and Staff choS larship

11-2012 Open system evolution of peralkaline trachyte and phonolite from the Suswa volcano, Kenya Rift. John C. White Eastern Kentucky University, [email protected]

Vanessa V. Espejel-Garcia Universidad Autonoma de Chihuahua

Elizabeth Y. Anthony University of Texas at El Paso

Peter Omenda Geothermal Research Corporation, Kenya

Follow this and additional works at: http://encompass.eku.edu/fs_research Part of the Geology Commons, and the Volcanology Commons

Recommended Citation White, J.C., Espejel-García, V.V., Anthony, E.Y., and Omenda, P., 2012, Open system evolution of peralkaline trachyte and phonolite from the Suswa volcano, Kenya Rift. nI : Peralkaline Rocks and Carbonatites with Special Reference to the East African Rift G( . Marks, M.A.W. Markl, and A. Zaitsev, Eds.) Lithos, v. 152, p. 84-104. (doi: 10.1016/j.lithos.2012.01.023)

This Article is brought to you for free and open access by Encompass. It has been accepted for inclusion in EKU Faculty and Staff choS larship by an authorized administrator of Encompass. For more information, please contact [email protected]. This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright Author's personal copy

Lithos 152 (2012) 84–104

Contents lists available at SciVerse ScienceDirect

Lithos

journal homepage: www.elsevier.com/locate/lithos

Open System evolution of peralkaline trachyte and phonolite from the Suswa volcano, Kenya rift

John Charles White a,⁎, Vanessa V. Espejel-García b,1, Elizabeth Y. Anthony b, Peter Omenda c a Department of Geography & Geology, Eastern Kentucky University, 521 Lancaster Ave., Roark 103, Richmond, KY 40475, USA b Department of Geological Sciences, University of Texas at El Paso, 500 W. University Ave., El Paso, TX 79968, USA c Geothermal Development Company, Taj Tower, Upper Hill, P.O. Box 10076–00101, Nairobi, Kenya article info abstract

Article history: Suswa is the southernmost volcanic center in the Central Kenya Peralkaline Province (CKPP) and represents Received 14 October 2011 the only salic center to have erupted signiﬁcant volumes of peralkaline silica-undersaturated lavas and tuffs Accepted 25 January 2012 (trachyte, nepheline trachyte and phonolite). The eruptive products of Suswa can be clearly divided into two Available online 3 February 2012 series, which correspond closely to the volcano's eruptive history. The earlier series (C1) includes lavas and tuffs that built the initial shield volcano (pre-caldera, unit S1) and erupted during the ﬁrst caldera collapse Keywords: (syn-caldera, units S2–S5); these rocks are dominated by peralkaline, silica-saturated to mildly under- Trachyte Phonolite saturated trachyte. The later series (C2) includes lavas and tuffs that erupted within the caldera structure Central Kenya Peralkaline Province following the initial collapse (post-caldera, units S6–S7) and during the creation of a second smaller, nested caldera and central “island block” (ring trench group, RTG, unit S8); these rocks are dominated by peralkaline phonolite. In this study, we combine mineralogical evidence with the results of major-element, trace- element, and thermodynamic modelling to propose a complex model for the origin of the Suswa volcano. From these results we conclude that C1 is the result of protracted fractional crystallization of a fairly “dry”

alkali basalt (b1 wt.% H2O) under relatively high pressure (400 MPa) and low oxygen fugacity (FMQ to FMQ-1). Although C1 appears to be primarily the result of closed system processes, a variety of open system processes are responsible for C2. We propose that crystallization of C1 trachyte resulted in the formation of a syenitic residue, which was assimilated (Ma/Mc=0.1) during a later stage of recharge and differentiation of alkali basalt to produce post-caldera ne-trachyte. Post-caldera (S6-7) phonolites were in turn the result of fractional crystallization of this ne-trachyte. RTG phonolites, however, are the result of feldspar resorption prompted perhaps by magma recharge as evidenced by reverse zoning in alkali feldspar and linear compatible trace element patterns. © 2012 Elsevier B.V. All rights reserved.

1. Introduction (~1 km) amounts of uplift. The dome is apparently in isostatic equilibrium, being supported by the loading of anomalous mantle within the Mount Suswa is a large (>700 km2) trachytic shield volcano with underlying lithosphere (Smith, 1994). Salic volcanism in the CKPP has two nested summit calderas that erupted peralkaline trachyte and pho- been focused on five centers. These volcanic centers include: (a) two nolite lavas and tuffs from approximately 240 to b100 ka (Baker et al., dome fields: the Eburru Volcanic Complex composed of trachyte and 1988; Johnson, 1969; Nash et al., 1969; Skilling, 1988, 1993). Suswa pantellerite (Omenda, 1997; Ren et al., 2006; Velador et al., 2003), is the southernmost volcano within the Central Kenya Peralkaline and the Greater Olkaria Volcanic Complex (GOVC), dominated by comen- Province (CKPP), an area within the Kenyan branch of the east African dites with minor outcrops of trachyte (Clarke et al., 1990; Macdonald et rift system (EARS) between approximately 0° and 1.5°S latitude al., 1987, 2008a; Marshall et al., 2009; Omenda, 1997); and (b) three (Fig. 1a; Macdonald and Baginski, 2009; Macdonald and Scaillet, trachytic caldera volcanoes: Menengai with a comenditic to pantelleritic 2006). The location of the CKPP coincides with the apical region of the trachyte composition (Leat et al., 1984; Macdonald et al., 2011a), Kenya dome, an area of crustal upwarping associated with minor Longonot composed of pantelleritic trachyte and mixed (trachyte/ hawaiite) lavas (Clarke et al., 1990; Heumann and Davies, 2002; Rogers et al., 2004; Scott and Bailey, 1984), and Suswa, with trachyte and phono- ⁎ Corresponding author. Tel.: +1 859 622 1276; fax: +1 859 622 3375. lite lavas and tuffs (Johnson, 1969; Nash et al., 1969; Scott and Skilling, E-mail addresses: [email protected] (J.C. White), [email protected] 1999; Skilling, 1988, 1993). The maficlavafields Elmenteita, Ndabibi, (V.V. Espejel-García), [email protected] (E.Y. Anthony), [email protected] and Tandamara, lie in the rift floor adjacent to Eburru, Olkaria and (P. Omenda). 1 Present address: Facultad de Ingeniería, Universidad Autónoma de Chihuahua, Suswa, respectively, and are composed of alkali and transitional basalt, Circuito No. 1, Campus Universitario 2, C.P. 31125, Chihuahua, Chih., México. basaltic trachyandesite and trachyandesite.

J.C. White et al. / Lithos 152 (2012) 84–104 85

Fig. 1. (a) Map of the Central Kenya Peralkaline Province (CKPP; Macdonald and Scaillet, 2006). The subject of this study, Suswa, is the southernmost volcano in the CKPP. (b) Ternary diagram indicating silica saturation for CKPP volcanic centers. GOVC, Greater Olkaria Volcanic Center; S1–S5 Suswa pre-and syn-caldera units (also referred to as C1 in this paper); S6-S8 Suswa C2 units. See Table 1 and text for further information.

The Suswa volcanic system includes silica-saturated (hypersthene- the result of protracted fractional crystallization of Ndabibi-like alkali normative [hy-]trachyte) and silica-undersaturated (nepheline- basalt and the silica-undersaturated suite (C2) primarily the result of normative [ne-]trachyte and phonolite) lavas and tuffs, and represents coupled assimilation-fractional crystallization from a similar parent the only salic volcanic system in the CKPP not dominated by silica- (with the syenitic residue from C1 petrogenesis serving as the assimi- oversaturated materials (quartz-normative [q-]trachyte, comendite, lant) plus feldspar resoprtion triggered by magma recharge. This com- and pantellerite) (Fig. 1b). Coexisting oversaturated and undersaturat- plexity reflects the open-system behavior of EARS volcanoes, where ed alkaline igneous suites have also been described from several intra- mafic dikes periodically intrude compositionally evolved shallow plate magmatic provinces, including the Massif Central, France magma chambers. Examples of this phenomenon occurred in 2005 in (Wilson et al., 1995); the Cameroon line, west Africa (Fitton, 1987); the Afar region of Ethiopia, where an episode of recharge/diking opened the Marie Byrd Land province, west Antarctica rift system (LeMasurier a 60 km rupture in the crust (Sigmundsson, 2006; Wright et al., 2006; et al., 2003, 2011; Panter et al., 1997); the Canary Islands (Carracedo Yirgu et al., 2006), and in 2007, when a magma-driven earthquake et al., 2002); and the Trans-Pecos Magmatic Province, Texas USA swarm struck northern Tanzania for a period of two months and trig- (Potter, 1996; White and Urbanczyk, 2001). In each of these provinces, gered an eruption at Oldoinyo Lengai volcano (Baer et al., 2008). Similar the origin of coexisting silica-oversaturated and –undersaturated suites recharge is thought to have happened at Suswa and a number of the ad- is attributed to either fractional crystallization from a similar alkali jacent volcanoes based on geodetic (InSAR) data (Biggs et al., 2009) basaltic parent with variable degrees of assimilation leading to the for- Modelling also suggests that the silica-undersaturated nature of mation of two different suites (e.g., Foland et al., 1993), variable degrees Suswa may be due to fractional crystallization of alkali basalt at relative- of partial melting from a similar asthenospheric source followed by ly high pressures (~400 MPa) compared to the silica-oversaturated fractional crystallization (e.g., Wilson et al., 1995), or fractional crystal- volcanic systems in the CKPP. lization at differing depths (LeMasurier et al., 2003, 2011). The goal of this study is to model the magma processes that generated the 2. Geologic setting trachyte-phonolite association at Suswa. We propose that fractional crystallization, assimilation, magma mixing, and feldspar resorption The growth of Suswa has been divided into three major stages played significant roles, with the silica-saturated suite (C1) primarily (Table 1, Fig. 2; Johnson, 1969; Nash et al., 1969; Omenda, 1997; Author's personal copy

86 J.C. White et al. / Lithos 152 (2012) 84–104

Table 1 trachyte tuffs; (b) the Ring Feeder Group (RFG, S3), including Suswa volcanic stages (following Johnson, 1969, and Skilling, 1993). trachytic lava flows; (c) the Western Pumice Group (WPG, S4) con- Stage Unit Group Group name sisting of pumice-lapilli tuffs; and (d) the Enkorika Fissure Group (EFG, S5), with trachytic lavas and agglutinate flows (Skilling, Pre‐caldera S1 PLG Pre‐caldera Lava Group Syn‐caldera S2 SPG Syn‐caldera Phreatomagmatic Group 1993). The basaltic trachyandesite ash and spatter in the SPG (S2), S3 RFG Ring Feeder Group which resemble geochemically the mafic flows exposed adjacent to S4 WPG Western Pumice Group the central volcanoes, are a key aspect of the evidence for periodic S5 EFG Enkorika Fissure Group magma intrusion and recharge at Suswa. Their presence caused Post‐caldera S6 EPLG Early Post‐caldera Lava Group S7 ODNG Ol‐Doinyo Nyoke Group Skilling (1988, 1993) to hypothesize that caldera collapse occurred S8 RTG Ring Trench Group either by magma-water interaction or introduction of denser trachybasaltic magma into the magmatic chamber. The post-caldera stage (S6-7), erupted from the Ol-Doinyo Nyoke vent on the caldera floor and is divided into two units (Nash et al., Randel and Johnson, 1991; Skilling, 1988, 1993). The pre-caldera 1969; Omenda, 1997; Skilling, 1993): the Early Post-caldera Lava stage (S1) contains the Pre-caldera Lava Group (PLG), represented Group (EPLG, S6), which contains aphyric to sparsely phyric trachyte by the eruption of dominantly trachytic lavas. The age of group is to phonolite lava flows and ash fall deposits; and the Ol-Doinyo presently best estimated by a K–Ar date of 240 ka on “early shield- Nyoke Group (ODNG, S7), which consists of strongly porphyritic pho- building lavas” (Baker et al., 1988; Rogers et al., 2004); a preliminary nolite lava flows. Baker et al. (1988) reported K-Ar ages of 100± 40Ar/39Ar age of ~120 ka has been obtained for lavas exposed in the 10 ka for EPLG (S6) lavas. The eruption of Ol-Doinyo Nyoke was fol- main caldera wall (A. Deino, personal communication). The syn- lowed by a second caldera collapse and resurgent uplift of the central caldera stage (S2–5) saw incremental collapse of the caldera roof “island block.” Formation of the second caldera and the subsequent and eruption of carbonate-rich trachytes through ring fractures to uplift led to collapse of a large part of the Ol-Doinyo Nyoke cone, form ignimbrite sheets and trachyte flows, with carbonatite immisci- which produced a debris flow deposit (Skilling, 1993). The youngest bility in some units (Macdonald et al., 1993). Syn-caldera trachytes episode of phonolitic magmatism, the Ring Trench Group (RTG, S8), contain syenite xenoliths (wall rock). The individual units within is associated with the resurgence of the island block in the center of the syn-caldera stage include: (a) the Syn-caldera Phreatomagmatic the second caldera and a vent on the southern flank of the volcano Group (SPG, S2), composed of trachyte pumice lapilli tuffs, basaltic (Skilling, 1993). RTG (S8) phonolites are strongly porphyritic, similar trachyandesitic ash and spatter deposits, and mixed carbonate- to the ODNG (S7) rocks (Omenda, 1997).

Fig. 2. Simpliﬁed geologic map of Mount Suswa following Skilling (1988, 1993). Author's personal copy

J.C. White et al. / Lithos 152 (2012) 84–104 87

Mafic lava fields formed on the rift floor in areas between volcanic b0.3 mm and subhedral with compositions about Fa76 (Table 5). complexes. These mafic lavas extruded from fissures and cinder Magnetite (Usp70) forms small (0.2–0.5 mm), euhedral crystals; in cones, and have basaltic to trachyandesite compositions (Clarke et one sample (IS139) a magnetite grain surrounds a few crystals of al., 1990; Omenda, 1997). Elmenteita lava field, located next to rutile. Apatite occurs as euhedral to subhedral crystals (up to Eburru volcanic complex, has porphyritic basalt and basaltic tra- 80 μm), either as glomerophenocrysts with clinopyroxene, olivine chyandesite, with plagioclase as the major phenocryst (Clarke et al., and magnetites or as isolated crystals in groundmass. Alkali amphi- 1990; Omenda, 1997). Ndabibi lava field, situated beside the GOVC, bole also occurs as a rare, interstitial groundmass phase (Skilling, has weakly porphyritic basalts with augite and plagioclase as 1988). Fluorite occurs as small crystals and globules both as a vapor major phenocrysts (Davies and Macdonald, 1987; Omenda, 1997). phase filling vesicles and as a primary groundmass phase. Tandamara mafic field, adjacent to Suswa in the northern flank, contains basaltic trachyandesite rocks with plagioclase and clinopyrox- 4.1.3. Ring Feeder (S3), Western Pumice (S4), and Enkorika Fissure ene as major phenocrysts (Clarke et al., 1990; Omenda, 1997). Groups (S5) The Ring Feeder Group (RFG) and Enkorika Fissure Group (EFG) 3. Methodology are similar unites comprised of trachyte lavas; the Western Pumice Group (WPG) consists of aphyric trachyte pumice lapilli tuffs Whole rock analyses of samples from Suswa, Elmenteita, and (Skilling, 1993). The phyric assemblage in the RFG and EFG lavas are Tandamara are presented in Table 2. Analyses were by ICP-OES at similar to the PLG lavas, although the alkali feldspar is more Or-rich Texas Tech University, INAA at the University of Texas at El Paso (An0-3Ab57-60Or40). Olivine, clinopyroxene, and magnetite form sub- (Omenda, 1997) and by ICP-OES and ICP-MS at Activation Laborato- hedral to euhedral crystals and occur together as glomeropheno- ries, Ontario. Mineral chemistry was obtained by electron probe mi- crysts. Clinopyroxene compositions are similar to the PLG (En21- croanalysis at UTEP on a Cameca SX50 instrument equipped with 29Fs27-32Wo45-46), but olivine is more fayalitic (Fa80). In one sample four wavelength dispersive detectors and a Rontec solid state energy of the RFG, magnetite composition is Usp68. Rare accessory phases in- dispersive detector (see Espejel-García, 2006, for additional details). clude pyrrhotite in an RFG glomerophenocryst and relatively low- A standard block from the Smithsonian Institution was used for cali- fluorine (~2 wt.%) apatite in EFG. bration. Additional Smithsonian and Astimex standards were analyzed each day to monitor accuracy and precision. Representative electron probe analyses are presented in Table 3 (feldspar), Table 4 4.2. Silica-undersaturated suite (C2) (clinopyroxene), Table 5 (olivine), Table 6 (Fe-Ti oxides), and Table 7 (apatite). 4.2.1. Early Post-Caldera Lava (S6) and Ol-Doinyo Nyoke (S7) Groups The post-caldera sequence includes the Early Post-caldera Lava 4. Mineralogy and petrography Group (EPLG) and Ol-Doinyo Nyoke Group (ODNG). Both groups consist of geochemically similar phonolites, but differ in terms of their 4.1. Silica-saturated suite (C1) mineralogy: the EPLG is largely aphyric to sparsely phyric, while the ODNG consists of strongly porphyritic phonolite lavas with alkali feld- 4.1.1. Pre-caldera lava group (S1) spar, clinopyroxene, olivine, magnetite, and apatite comprising the Pre-caldera Lava Group (PLG) rocks represent the earliest Suswa phyric assemblage. ODNG alkali feldspars consist of zoned anortho- lavas. These rocks are porphyritic, with dominant alkali feldspar clase crystals ranging in composition from An20Or23 to An5Or35 with (~25 vol.%) and subordinate clinopyroxene and magnetite as the rims of sanidine (Or40-45). Clinopyroxene, olivine, magnetite, and ap- major phyric phases. Accessory minerals include apatite and small ol- atite occur together in these rocks as glomerophenocrysts. Clinopyr- ivine grains. The groundmass consists of alkali feldspar, devitrified oxene in these rocks is more magnessian (Wo45–48Fs20–24En30–35) volcanic glass, rare alkali amphibole, and small vesicles. Alkali than in the pre- and syn-caldera samples. Olivine (Fa56-75) is consid- feldspar phenocrysts are tabular (2–4 mm), euhedral, and normally erably more abundant in the post-caldera samples compared to the zoned from Or to Or with low An (b8 mol%). Clinopyroxene crystals syn-caldera samples, comprising up to 5 vol.% of the mode. Fluorian 26 36 – are smaller (b0.5 mm) and subhedral with compositions that straddle (5 7 wt.%F) apatite frequently occurs in melt inclusions in clinopyroxene and olivine from the post-caldera groups, and has a composi- the join between diopside and hedenbergite (En26-31Fs26–29Wo43–45). Magnetite forms small (0.2–0.5 mm), euhedral crystals with a variable tion similar to groundmass apatite. composition (Usp58-72). Fluorian (~5 wt.%F) apatite occurs as small (30–60 μm), euhedral crystals frequently found in clusters with ferro- 4.2.2. Ring Trench Group (S8) magnesian minerals. Similar apatite has also been documented in The history of the RTG alkali feldspars is substantially more com- quartz trachytes and rhyolites at the adjacent GOVC (Macdonald et al., plex than for the previous units. Large feldspar phenocrysts (3– 2008b). 5 mm) are zoned with euhedral overgrowths surrounding resorbed cores. Feldspar cores have two distinct compositions: (a) high-K

4.1.2. Syn-caldera phreatomagmatic group (S2) cores (An0–5Ab50–60Or30–40) similar to pre-caldera feldspars, and (b) The Syn-caldera Phreatomagmatic Group (SPG) rocks consist of low-K cores (An3–15Ab65–70Or15–20). Zoning rims have compositional several trachyte tuffs and minor trachybasalt tephras (Skilling, variations between the anorthite (Ca) and orthoclase (K) contents, al- 1993). Trachytes are porphyritic with alkali feldspar, clinopyroxene, bite (Na) compositions are constant in all zoned overgrowths. Line olivine and Fe-Ti oxides as phenocrysts with accessory apatite and scans throughout zoned feldspars conﬁrm the oscillatory composition ﬂuorite. Trachybasalt tephras consist of clasts of pumice, alkali feld- between cores and overgrowths, and the distinct core composition spar, resorbed fragments of broken plagioclase xenocrysts, and sye- between different feldspars (Fig. 3). Zoned overgrowth compositions nite xenoliths. Alkali feldspar phenocrysts comprise ~25 vol.% of the overlap those from the cores, although the highest anorthite (An20-25) trachyte tuffs and are smaller than (1–2 mm), but compositionally contents correspond to the overgrowths. The thin rims (~0.01 μm) at similar to, crystals in the PLG lavas. One sample of trachyte (KS/94/ the edge of the crystal contain similar composition (An11Ab71Or18)to 25) contains alkali feldspars with anorthite contents up to 20 mol% groundmass feldspars (Espejel-García, 2006). A sample (KS/94/01) and resorbed xenocrysts of plagioclase (An45-75Ab25-55). Clinopyrox- from the Ring Trench Group contains a melt inclusion inside an alkali ene grains are small (~0.5 mm) and euhedral to subhedral, with com- feldspar, which also precipitated a small (50 μm), euhedral Fe–Ti positions similar to PLG lavas (En27Fs29Wo44). Olivine grains are oxide. Author's personal copy

88 J.C. White et al. / Lithos 152 (2012) 84–104

Table 2 Whole-rock analyses.

Stage: Elementeita Tandamara

Sample ES11 ES12 ES10 ES8 ES9 KS/94/08 KS/94/10 KS/94/13 KS/94/09 KS/94/11

Class BTA B TA B TA BTA BTA BTA BTA TA

SiO2, wt.% 50.07 48.34 54.95 48.41 55.46 55.29 54.46 54.77 55.22 57.02

TiO2 1.64 2.21 1.19 2.55 1.16 1.47 1.52 1.68 1.30 1.43

Al2O3 15.41 15.36 16.24 14.71 16.26 15.71 15.96 15.11 15.89 16.54 ⁎ FeO 9.76 11.99 8.40 12.78 8.12 9.80 9.24 10.58 8.13 8.28 MnO 0.25 0.21 0.21 0.21 0.20 0.24 0.18 0.25 0.17 0.19 MgO 3.90 6.50 4.05 5.75 3.74 3.34 4.05 3.55 3.49 2.27 CaO 8.72 10.52 6.66 9.70 6.30 6.92 7.41 7.29 6.75 5.29

Na2O 3.90 3.18 4.27 3.26 4.31 3.86 3.84 3.59 3.85 4.64

K2O 2.22 0.81 3.00 1.07 3.20 2.78 2.43 2.46 2.84 3.52

P2O5 0.40 0.50 0.33 0.62 0.29 0.41 0.35 0.50 0.31 0.42 Total 96.27 99.62 99.30 99.06 99.04 99.82 99.44 99.78 97.95 99.60 Mg# 0.48 0.56 0.53 0.51 0.52 0.44 0.51 0.44 0.50 0.39 PI 0.57 0.40 0.63 0.44 0.65 0.60 0.56 0.57 0.59 0.69 Fe/Al 0.90 1.11 0.73 1.23 0.71 0.89 0.82 0.99 0.73 0.71 Q, mol% 0.00 0.00 0.00 0.00 0.00 0.54 0.00 1.64 1.13 0.12 or 13.65 4.83 17.76 6.46 18.99 16.56 14.47 14.73 17.14 20.87 ab 31.42 28.38 38.42 29.90 38.88 34.95 34.75 32.67 35.33 41.82 an 18.72 25.50 16.32 22.83 15.65 17.48 19.29 18.09 18.08 13.96 ne 3.02 0.28 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 di 18.87 19.14 11.71 17.76 11.09 11.55 12.37 12.20 11.26 7.70 hy 0.00 0.00 4.55 2.67 6.07 14.85 14.12 15.98 13.58 11.68 ol 9.89 16.30 7.92 13.92 6.14 0.00 1.03 0.00 0.00 0.00 ac 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 mt 1.18 1.41 0.98 1.52 0.95 1.15 1.08 1.25 0.97 0.97 il 2.38 3.11 1.66 3.63 1.62 2.06 2.13 2.37 1.85 2.00 ap 0.87 1.06 0.69 1.32 0.61 0.86 0.74 1.06 0.66 0.88 NMS 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Be, ppm 1.3 1.4 1.9 Sc 27.6 38.3 24.1 37 22.4 25.4 27.7 28 24.6 18.9 V 182 329 134 346 128 123 162 138 121 84 Cr 74 95 74 46 64 30 37 27 64 17 Co 32.1 28 49 23.8 27.7 17.2 Ni 24 55 44 31 28 23 48 16 21 13 Cu 38 53 18 28 21 30 38 28 50 27 Zn 76 93 83 108 86 93 92 94 86 93 Ga Ge Rb 27 16 33 22 36 53 65 50 69 77 Sr 319 493 296 499 282 264 358 287 278 264 Y 31313335354147424647 Zr 108 128 143 171 175 208 233 260 334 459 Nb 31 28 39 35 43 48 88 48 76 85 Mo Ag Sn Sb Cs 0.587 Ba 624 406 774 562 790 1301 901 1153 981 1259 La 22.8 31.8 35.7 45.5 47 84 Ce 41.1 57.7 62 83.5 87 141 Pr Nd 22.4 28.9 34.7 41.1 42 62 Sm 5.4 6.6 7.6 8.11 8.52 10.8 Eu 2.45 2.62 2.45 2.99 2.76 2.98 Gd Tb 0.84 1.01 1.07 1.14 1.23 1.43 Dy Ho Er Tm Yb 2.7 3.05 3.09 3.65 3.86 4.36 Lu 0.485 0.541 0.527 0.547 0.536 0.63 Hf 2.69 3.76 4.11 5.04 5.12 9.25 Ta 1.53 1.98 1.95 2.52 2.54 4.35 W Tl Pb Bi Th 1.96 2.94 3.37 6.23 6.1 11.6 U 1.42

Key to analyses: Normal font, ICP-OES; Italics , INAA; Bold, ICP-MS; bblank>, not analyzed. Author's personal copy

J.C. White et al. / Lithos 152 (2012) 84–104 89

Suswa, Pre-caldera stage (S1) Suswa, Syn-caldera stage (S2-5)

KS/94/33 KS/94/26 KS/94/24 KS/94/19 KS/94/21 KS/94/38 KS/94/41 KS/94/21A KS/94/43 KS/94/05 KS/94/29 KS/94/29 KS/94/06 KS/94/06 ne-TR ne-TR q-TR hy-TR hy-TR hy-TR hy-TR hy-TR ne-TR ne-TR ne-TR ne-TR Ph Ph

60.80 59.93 60.75 59.28 61.14 60.97 61.07 61.01 59.96 60.11 59.86 60.43 58.97 58.55 0.88 0.90 0.79 0.68 0.85 0.79 0.79 0.80 0.60 0.56 0.57 0.63 0.57 0.60 16.35 16.58 14.20 13.38 14.93 14.91 14.88 15.12 15.41 15.59 14.93 15.04 14.99 14.93 6.41 6.85 8.45 8.67 7.48 7.07 7.08 7.05 7.87 7.55 7.31 7.57 7.69 7.84 0.27 0.29 0.34 0.39 0.24 0.33 0.34 0.34 0.37 0.38 0.35 0.24 0.39 0.23 0.55 0.70 0.42 0.65 0.47 0.48 0.51 0.50 0.34 0.37 0.43 0.41 0.34 0.32 1.49 1.87 1.69 2.65 1.24 1.12 1.10 1.23 1.09 1.12 1.25 1.30 1.20 1.27 6.52 6.41 6.43 6.83 6.83 6.54 6.96 6.88 7.18 8.02 7.79 7.95 7.78 8.21 5.10 4.73 4.71 4.29 5.08 5.11 5.20 5.27 5.02 4.82 4.96 4.97 4.75 4.83 0.14 0.17 0.08 0.04 0.09 0.12 0.11 0.11 0.09 0.11 0.10 0.07 0.08 0.05 98.51 98.43 97.86 96.86 98.35 97.45 98.04 98.30 97.93 98.63 97.55 98.61 96.76 96.83 0.17 0.19 0.10 0.15 0.13 0.14 0.14 0.14 0.09 0.10 0.12 0.11 0.09 0.09 0.99 0.94 1.10 1.19 1.12 1.09 1.15 1.13 1.12 1.18 1.22 1.23 1.20 1.25 0.56 0.59 0.85 0.92 0.71 0.67 0.68 0.66 0.73 0.69 0.70 0.71 0.73 0.75 0.00 0.00 0.05 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 30.12 27.99 28.23 25.89 30.10 30.59 30.83 31.17 29.77 28.18 29.35 29.09 28.36 28.71 56.12 55.58 50.40 48.71 51.62 51.87 50.68 51.46 50.87 50.14 50.26 49.97 49.84 46.79 0.28 2.50 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.45 1.25 0.00 0.00 0.00 0.00 0.00 0.00 2.28 3.54 1.21 1.36 2.69 3.90 4.95 4.54 6.38 10.53 4.46 3.87 3.81 4.31 3.87 3.83 4.45 4.75 4.39 4.81 0.00 0.00 8.70 4.60 4.79 6.93 5.76 2.38 0.00 0.00 0.00 0.00 0.00 0.00 4.45 5.33 0.00 2.47 2.51 1.00 1.90 4.16 6.87 6.56 6.27 6.09 6.61 6.29 0.00 0.00 3.98 4.12 3.49 3.33 3.30 3.28 3.67 3.47 3.40 3.48 3.61 3.66 1.12 1.20 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.23 1.26 1.12 0.97 1.19 1.12 1.10 1.11 0.84 0.77 0.80 0.87 0.80 0.84 0.29 0.36 0.17 0.09 0.19 0.25 0.23 0.23 0.19 0.23 0.21 0.15 0.17 0.11 0.00 0.00 0.96 2.64 1.66 1.04 2.37 1.89 1.64 3.27 4.06 4.24 3.53 4.89 4.6 5.9 8.3 12.5 8108 10.2 10.7 10 16 3.47 3.5 3.6 2.6 2.6 434 2.2 2 2 1.7 2 1.7 222226 b5 b5 12b5 2 b5 2 11 4 11 6 b20 b20 b20 7 b20 2 b20 1.09 2.26 1.34 0.85 23 17 b1 1.227 0.91 b12 5 1 16 1 b1 b1 b1 15 b1 0 b1 28677 18 7 7 6812 14 7 11 173 193 245 289 257 191 233 200 286 271 310 284 310 371 39 38 36 47 49 2.1 2.1 1.8 2.4 2 115 148 164 143 176 170 192 162 208 208 234 130 216 221 30 56 26 29 13 12 9 12 8812 13 32 31 83 93 129 156 132 124 108 109 116 137 147 151 203 206 558 772 1006 1203 983 1028 1031 1031 1104 1145 1190 1214 1755 1778 250 337 421 532 463 312 305 303 528 549 431 559 431 586 334 4 5 2.1 2.4 2.3 3 3.6 11 8 7 12 16 >200 4.4 b0.2 4 b0.2 0.952 1.54 1.44 0.426 0.8 1.8 1.4 2.37 2.55 2.8 3.6 781 341 86 32 28 97 50 28 40 54 54 53 167 162 110 144 184 224 181 162 174 223 216 188 220 182 247 318 373 300 301 310 361 359 338 401 33.4 31.4 32.3 36.9 44.7 70 89 121 137 106 95.9 100 125 122 112 141 12.9 16.4 22.2 25.1 20.3 19.2 20.2 23.1 23 21.5 27.7 3.15 2.89 3.38 3.16 3.45 3.21 3.34 3.05 3 2.94 4.64 20.3 16.2 17.2 20.4 26.1 2.04 2.48 3.52 3.96 3.54 3.02 3.3 3.62 3.5 3.6 4.89 21.4 18.7 20.4 22.4 29.6 4.55 4.09 4.38 4.87 6.33 14.2 12.5 13.2 14.9 19.4 2.16 1.93 2.09 2.26 2.84 8.1 10 13.7 16.7 14 13 13.6 14.4 14.8 14.4 16.8 1.09 1.38 1.87 2.29 2.09 1.94 1.98 1.99 2.01 2.23 2.48 12.7 18.1 24.4 29 23.2 22.3 24 27 27.4 27.3 36.5 11.3 16.4 20 23.7 23.3 23 24 24.4 25.2 27.2 27.2 181 153 4.5 2 1.9 0.13 0.15 0.14 0.2 0.17 24 25 17 27 26 0.2 0.4 0.1 0.1 0.3 17.9 26.1 32.4 39.5 31 29.5 31.6 40.5 41.2 37.8 45.6 2.8 5.2 5.7 6.7 5.56 4.67 5.67 8 7.7 7.11 6.18 Author's personal copy

90 J.C. White et al. / Lithos 152 (2012) 84–104

Table 2 (continued)

Stage: Suswa, Post-caldera stage (S6-7) Suswa, Ring Trench Group (S8)

Sample KS/94/03 KS/94/23 KS/94/39 KS/94/23 KS/94/07 KS/94/31 KS/94/22 KS/94/02 KS/94/44 KS/94/01 KS/94/01 KS/94/49

Class ne-TR Ph Ph Ph Ph Ph Ph Ph Ph Ph Ph Ph

SiO2, wt.% 57.84 56.67 57.53 56.31 56.39 57.23 56.70 57.51 57.19 56.80 57.82 57.24

TiO2 1.24 1.15 0.74 1.10 1.09 1.06 0.94 0.74 0.76 0.79 0.78 0.71

Al2O3 16.96 16.30 17.57 16.53 16.43 16.31 16.24 17.75 18.52 17.50 18.15 18.00 ⁎ FeO 6.94 8.35 8.44 8.16 8.34 8.67 8.62 6.79 6.89 6.74 6.69 7.30 MnO 0.26 0.24 0.34 0.34 0.34 0.36 0.38 0.24 0.29 0.28 0.29 0.32 MgO 1.03 0.75 0.57 0.82 0.82 0.68 0.69 0.68 0.97 0.88 0.75 0.66 CaO 2.92 2.31 1.94 2.27 2.31 2.10 2.00 2.20 2.65 2.53 2.47 2.02

Na2O 7.04 7.01 8.38 7.11 7.30 7.50 7.47 8.16 7.75 8.03 8.51 8.62

K2O 4.35 4.79 4.98 5.04 5.03 4.85 4.93 4.66 4.34 4.53 4.54 4.73

P2O5 0.38 0.20 0.15 0.24 0.26 0.25 0.27 0.19 0.22 0.25 0.25 0.16 Total 98.96 97.77 100.64 97.92 98.31 99.01 98.24 98.92 99.58 98.32 100.25 99.76 Mg# 0.26 0.17 0.14 0.19 0.19 0.15 0.16 0.19 0.25 0.23 0.21 0.17 PI 0.96 1.03 1.09 1.04 1.06 1.08 1.09 1.04 0.94 1.04 1.04 1.07 Fe/Al 0.58 0.73 0.68 0.70 0.72 0.75 0.75 0.54 0.53 0.55 0.52 0.58 Q, mol% 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 or 25.46 28.45 28.41 29.82 29.62 28.38 29.05 26.98 25.02 26.40 25.88 27.08 ab 47.71 41.69 34.64 37.56 35.87 39.13 38.14 38.88 42.15 38.35 37.83 35.81 an 1.81 0.00 0.00 0.00 0.00 0.00 0.00 0.00 2.86 0.00 0.00 0.00 ne 8.95 11.58 17.74 13.79 14.33 12.40 12.73 17.45 15.46 17.68 19.12 19.40 di 8.07 8.17 6.68 7.77 7.78 6.96 6.51 7.58 6.86 8.62 8.20 6.96 hy 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 ol 4.30 5.48 6.33 5.91 6.26 6.58 6.94 4.63 5.00 4.56 4.34 5.35 ac 0.00 1.83 3.79 2.72 3.86 3.99 4.00 3.07 0.00 2.64 3.00 3.29 mt 1.20 0.78 0.00 0.40 0.00 0.00 0.00 0.01 1.17 0.17 0.00 0.00 il 1.71 1.61 1.00 1.53 1.51 1.46 1.31 1.01 1.03 1.08 1.05 0.96 ap 0.79 0.42 0.30 0.50 0.54 0.52 0.56 0.39 0.45 0.52 0.50 0.32 NMS 0.00 0.00 1.12 0.00 0.22 0.58 0.76 0.00 0.00 0.00 0.08 0.83 Be, ppm 5 8.1 889.8 9.6 4.7 6 5.7 6.5 Sc 5 3.5 2.3 453.7 3.6 2.2 3.4 4 3.1 2.2 V 10 52b5 b5 3351617 13 6 Cr b20 212b20 b20 64 16b20 64 Co 41 1.88 2213.02 3.33 4.86 45 4.22 2.86 Ni b1 1491 b1 452 3 6 Cu 7 68676156 11 9 10 7 Zn 190 226 232 170 190 235 245 176 162 140 172 199 Ga 33 33 32 32 Ge 1.8 1.6 1.4 1.3 Rb 116 169 174 159 161 182 190 135 123 132 132 154 Sr 264 77 39 81 63 62 37 192 194 189 209 96 Y 70.1 107 100 90.6 89.8 118 119 89 74 73.7 71 92 Zr 491 841 870 883 927 1002 1008 591 605 622 648 745 Nb 200 411 421 288 278 471 481 330 301 240 308 374 Mo 4466 Ag 1.1 2.2 2 1.2 Sn 4664 Sb 2.9 0.3 b0.2 b0.2 Cs 0.8 1.86 1.2 1.8 2.15 2.21 1.18 1.4 1.31 1.57 Ba 1060 392 301 415 355 307 150 1075 1016 1005 1028 566 La 95.6 180 188 165 194 196 129 120 145 150 Ce 176 310 304 291 319 330 224 216 231 258 Pr 19.2 32.5 30.4 22.8 Nd 61 110 102 95.9 112 117 80 71.3 80 91 Sm 11.4 19.2 19.5 18.8 20.8 21.4 13.2 13.3 15.6 15.2 Eu 3.52 3.36 3.7 3.41 3.34 3.08 3.5 3.56 3.07 2.78 Gd 10.2 14.6 15.5 12.6 Tb 1.75 2.86 2.86 2.98 3.19 3.23 1.98 2.09 2.08 2.29 Dy 10.1 18.4 17.5 12.3 Ho 2.16 3.71 3.79 2.59 Er 6.44 11.5 11.5 8.08 Tm 0.934 1.78 1.83 1.18 Yb 6.06 10.9 11.5 11.7 12.5 12.7 7.5 7.59 7.9 9 Lu 0.963 1.5 1.69 1.69 1.74 1.74 1.03 1.17 1.11 1.23 Hf 10.9 20 22.8 21.5 23 24 13 14.1 14 17 Ta 12.3 21 22.5 23.4 22 23 14 17 15 17 W 269 4.4 194 356 Tl b0.05 0.08 0.1 b0.05 Pb 10 19 14 9 Bi b0.1 b0.1 b0.1 b0.1 Th 15.3 33 32.7 31.4 38 38 22 21.9 24 28 U 2.49 6 4.75 6.72 98 44.78 55 Author's personal copy

J.C. White et al. / Lithos 152 (2012) 84–104 91 Tandamara ES-12 ES-12 KS94/08 MAFIC LAVA FIELDS 7fspar 36 phenoc xenoc 10 14 mtx plag 9 phenoc 77 phenoc 5 phenoc 22 0.0229.32 0.14 26.20 0.05 23.31 20.57 0.14 30.83 0.00 32.84 0.02 27.84 0.00 31.26 0.01 ) SUSWA 55phenoc phenoc 19 phenoc phenoc 21 phenoc xenoc 9 xenoc 7 phenoc phenoc phenoc core 17 rim 5 rim phenoc 38 core 45 rim rim 3 rim core 6 rim 4 mtx 8 10 6 30 25 45 20 6 5 0.0819.81 0.13 21.45 0.03 19.45 19.33 0.12 22.63 0.12 28.20 0.13 32.74 0.11 19.55 19.87 0.12 20.74 0.03 18.79 0.05 20.07 0.01 21.66 20.78 0.24 21.20 0.08 22.72 0.00 21.96 24.06 0.05 21.79 20.29 0.01 21.83 0.03 0.10 0.01 0.04 0.03 continued ( wt.% 65.40 64.24 66.41 65.65 62.63 54.71 48.87 65.72 66.29 65.48 64.63 65.22 63.41 65.29 64.39 62.79 63.48 60.46 63.76 65.05 64.37 wt.% 52.85 57.18 63.26 64.43 51.34 47.74 55.82 50.38 3 3 O 6.93 7.39 7.11 7.40 7.27 4.93 2.67 6.51 6.63 7.59 4.40 7.18 7.67 7.46 7.86 8.12 8.14 7.82 8.26 7.36 7.66 O 4.47 5.98 5.98 6.16 3.83 2.42 5.48 3.75 2 2 2 2 O O 2 2 O 6.25 4.50 6.74 6.61 3.08 0.59 0.18 7.09 6.52 4.84 10.67 5.31 3.46 3.75 4.27 2.39 2.97 1.70 3.25 5.56 4.12 O 0.54 1.33 4.22 7.25 0.18 0.09 0.77 0.26 2 2 2 2 K TotalAn, mol%Ab 3.8 99.46Or 99.54 60.4 8.1 35.8 99.95 65.6 0.6 99.66 26.3 99.29 61.2 0.8 38.2 99.62 62.5 16.7 36.7 64.3 100.25 19.0 53.4 99.64 43.0 99.59 3.6 75.7 99.95 23.2 99.08 2.2 1.1 57.0 99.34 0.7 40.8 60.3 99.38 100.05 5.3 39.0 66.7 100.09 28.0 61.4 0.1 99.91 99.71 38.5 99.29 64.0 4.9 99.41 67.0 31.1 13.2 99.15 66.7 19.8 100.31 11.2 22.1 66.6 9.5 23.9 69.3 17.3 69.4 13.4 13.9 67.2 16.7 23.2 70.7 9.6 11.0 18.3 64.1 4.1 67.0 31.8 9.3 23.7 TiO Stage:Group: Pre-caldera stageSample: PLG KS/94/33n: KS/94/33 PLG IS118SiO IS108 Syn-caldera stage PLG KS/94/25 KS/94/25 SPG KS/94/25 IS326 SPG IS196 IS SPG 147 KS/94/03 SPG RFG EFG IS342 Post-caldera stage EPLG KS/94/01 ODNG ODNG ODNG RTG RTG KS/94/44 RTG RTG RTG KS/94/49 RTG RTG RTG Al FeOMnOMgO 0.19CaO 0.00Na 0.00 0.18 0.80 0.00 0.01 0.09 1.65 0.00 0.00 0.35 0.01 0.13 0.03 0.34 0.00 0.16 0.00 0.63 3.23 0.00 0.02 0.75 10.42 0.00 0.03 14.90 0.21 0.00 0.00 0.12 0.44 0.00 0.00 0.16 0.13 0.02 0.00 0.56 1.08 0.00 0.00 0.02 0.33 0.00 0.00 0.31 0.99 0.05 0.01 0.52 2.73 0.00 0.00 0.30 2.26 0.00 0.00 2.03 0.20 0.00 0.20 0.00 0.00 3.68 0.24 0.00 0.01 2.94 0.00 0.00 0.00 4.90 0.00 2.34 0.00 0.00 0.38 0.00 0.00 0.85 0.00 1.92 Stage:Group: Sample:n: SiO Elmenteita ES-9 ES-9 K TotalAn, mol%AbOr 57.1 99.63 39.8 3.1 100.05 39.9 52.4 7.7 100.09 15.8 57.5 26.7 99.75 4.9 53.6 41.5 100.02 65.0 34.0 1.0 99.77 78.1 21.4 99.79 0.5 45.9 49.5 99.38 4.6 65.3 33.2 1.5 TiO Al FeOMnOMgOCaONa 0.82 0.01 0.00 11.61 0.91 0.06 0.00 8.24 0.26 0.03 0.01 2.97 0.18 0.00 0.00 1.02 0.46 0.00 0.13 13.25 0.62 0.04 0.07 15.94 0.53 0.00 9.19 0.15 0.37 0.00 13.37 0.00 Table 3 Representative analyses of feldspar in Suswa, Elmenteita and Tandamara rocks. Table 3 Author's personal copy

92 J.C. White et al. / Lithos 152 (2012) 84–104

Table 4 Representative analyses of clinopyroxene in Suswa, Elmenteita and Tandamara rocks.

SUSWA MAFIC LAVA FIELDS

Stage: Pre-caldera Syn-caldera Post-caldera Elmenteita Tandamara

Group: PLG PLG SPG RFG RFG EFG EPLG ODNG ODNG RTG

Sample: IS118 KS/94/33 IS108 870156 IS326 IS196 IS 147 870037 KS/94/03 KS/94/01 ES-9 ES-12 KS/94/08

n: 10 23 7 6 6 5 3 5 16 15 5 4 10

SiO2 wt.% 50.40 50.63 50.30 49.81 50.85 49.33 50.65 49.47 51.21 50.95 50.30 45.82 51.57

TiO2 0.58 0.85 0.64 0.64 0.63 0.43 0.95 1.01 0.94 0.69 0.76 2.97 0.61

Al2O3 0.56 0.65 0.40 0.74 0.49 0.57 1.25 1.86 1.44 1.43 2.62 4.75 1.36 FeO 16.81 15.52 16.96 19.61 16.21 18.95 15.66 15.53 12.05 13.75 9.98 14.88 11.68 MnO 1.38 1.05 1.31 1.33 1.23 1.34 0.95 0.90 0.69 0.64 0.46 0.25 0.31 MgO 8.49 10.09 8.69 7.14 9.09 7.09 9.50 8.82 11.61 9.71 15.10 9.50 15.73 CaO 20.09 19.79 20.16 18.58 21.17 21.26 20.88 20.74 21.11 21.57 19.54 19.67 17.69

Na2O 0.88 0.56 0.92 0.82 0.58 0.78 0.62 0.87 0.64 0.72 0.36 0.68 0.20

K2O 0.00 0.01 0.01 0.00 0.01 0.01 0.01 0.01 0.00 0.00 0.02 0.15 0.00 Total 99.18 99.15 99.39 98.67 100.26 99.75 100.47 99.20 99.70 99.46 99.41 99.12 99.15 Wo, mol% 44.6 43.0 44.3 42.4 45.6 46.3 45.1 45.9 45.2 47.1 40.4 44.2 36.4 En 26.2 30.6 26.6 22.7 27.2 21.5 28.5 27.2 34.6 29.5 43.5 29.7 44.9 Fs 29.2 26.4 29.1 34.9 27.2 32.2 26.4 26.9 20.2 23.4 16.1 26.1 18.7

Table 5 Representative analyses of olivine in Suswa, Elmenteita and Tandamara rocks.

SUSWA MAFIC LAVA FIELDS

Stage: Pre-caldera Syn-caldera Post-caldera Elmenteita Tandamara

Group: PLG SPG RFG EPLG ODNG ODNG RTG RTG

Sample: IS118 IS322 870156 IS 218 IS33 KS/94/03 IS342 KS/94/49 ES-12 ES-12 KS/94/08

n: 1 162614111421820

SiO2 wt.% 31.32 31.50 30.66 31.99 31.13 33.30 31.52 31.42 31.66 37.08 35.27

TiO2 0.25 0.16 0.05 0.20 0.12 0.09 0.10 0.03 0.00 0.00 0.00

Al2O3 0.00 0.02 0.03 0.00 0.06 0.01 0.00 0.01 0.00 0.00 0.00 FeO 54.98 53.18 56.99 49.83 53.61 44.07 51.42 52.01 53.94 26.70 35.35 MnO 4.02 3.95 4.32 3.16 3.68 2.16 3.04 2.99 1.17 0.43 0.71 MgO 9.43 9.57 6.89 14.13 9.53 19.69 12.47 11.55 11.83 35.53 28.42 CaO 0.58 0.52 0.54 0.75 0.87 0.54 0.78 0.82 0.77 0.43 0.30

Na2O 0.02 0.06 0.00 0.00 0.05 0.03 0.02 0.05 0.06 0.02 0.00

K2O 0.08 0.10 0.00 0.06 0.03 0.00 0.04 0.01 0.01 0.00 0.00 Total 100.67 99.04 99.47 100.11 99.08 99.88 99.39 98.88 99.44 100.19 100.05 Fa, mol% 76.6 75.6 82.3 66.4 75.9 55.7 69.8 71.6 71.9 29.7 41.1

Table 6 Representative analyses of Fe-Ti oxides in Suswa and Elmenteita rocks.

SUSWA MAFIC LAVA FIELD

Stage: Pre-caldera Syn-caldera Post-caldera Elmenteita

Group: PLG PLG SPG RFG EPLG ODNG RTG

Sample: KS/94/33 IS118 IS108 870156 IS218 KS/94/03 KS/94/44 ES-12 ES-12

n: 22 6 7 15 3 15 12 2 5

SiO2 wt.% 0.00 0.00 0.08 0.00 0.00 0.00 0.00 2.63 0.80

TiO2 24.93 20.83 24.79 24.69 22.79 25.26 20.17 22.92 46.65

Al2O3 0.98 0.06 0.43 0.46 1.01 1.83 1.85 1.31 0.17 FeO 66.36 72.85 69.18 69.43 68.76 68.44 71.29 67.39 48.38 MnO 2.00 2.18 1.96 2.37 1.78 1.67 1.66 0.69 0.83 MgO 1.40 0.37 0.78 1.07 1.38 0.66 1.20 0.55 0.79 CaO 0.06 0.13 0.08 0.10 0.09 0.05 0.14 0.28 0.35

Na2O 0.00 0.00 0.02 0.03 0.00 0.04 0.01 0.43 0.02

K2O 0.05 0.11 0.02 0.07 0.07 0.03 0.06 0.24 0.08

ZrO2 0.44 0.47 0.10 0.46 0.06 0.01 Sum 96.22 96.99 97.44 98.21 95.88 97.98 96.83 96.49 98.08

Fe2O3 18.85 28.42 20.43 21.59 23.28 18.27 27.93 15.56 9.10 FeO 49.40 47.27 50.80 50.01 47.81 52.00 46.15 53.38 40.20 Total 98.12 99.83 99.49 100.38 98.21 99.81 99.63 98.04 99.00 Xusp 0.72 0.58 0.70 0.68 0.70 0.75 0.59 0.76 Xilm 0.91 Author's personal copy

J.C. White et al. / Lithos 152 (2012) 84–104 93

Table 7 Representative analyses of F-apatite and ﬂuorite in Suswa rocks.

SUSWA (F-apatite) SUSWA (Fluorite)

Stage: Pre-caldera Syn-caldera Post-caldera Syn-caldera

Group: PLG SPG EPLG ODNG ODNG RTG SPG

Sample: IS118 IS214 IS 147 KS/94/39 KS/94/03 KS/94/44 IS322 ⁎ n: 3 2 6 1 (d. m.) 5105

SiO2 wt.% 0.51 0.37 0.41 0.89 0.26 0.53 0.09 FeO 0.34 0.16 0.96 0.70 0.41 0.37 0.22 MnO 0.17 0.13 0.21 0.12 0.19 0.15 0.01 CaO 50.46 52.06 52.92 53.58 52.53 52.87 72.86

Na2O 0.29 0.21 0.10 0.14 0.08 0.14 0.39

P2O5 38.17 38.90 38.88 36.33 40.09 38.64 0.00

La2O3 0.57 0.51 0.53 0.55 0.22 0.39 0.00

Ce2O3 1.67 1.36 0.81 1.43 0.48 0.81 0.00 F 5.18 4.74 5.73 5.27 5.30 5.39 46.34 Cl 0.04 0.04 0.07 0.05 0.07 0.09 0.00 Total 97.39 98.49 100.62 99.06 99.63 99.38 99.12 ⁎ (d. m.) =daughter mineral.

5. Whole-rock geochemistry quartz lies below the 95% conﬁdence interval, which strongly suggests that their normative assemblage may be an artifact of signiﬁcant 5.1. Assessment of post-crystallization element mobility Na loss and that the original magmatic composition was less silica- oversaturated, if not silica-saturated or -undersaturated like all

Signiﬁcant post-crystallization losses of Na2O and other compo- other samples. In addition to Na2O, Cs, Cl, and F may also suffer signif- nents in glassy peralkaline rocks have been noted in previous studies icant post-crystallization losses, high-ﬁeld strength elements (e.g., Zr, (e.g., Bailey and Macdonald, 1975; Baker and Henage, 1977; Noble, Nb), Rb, and Ba are generally not affected, and Sr concenrations may

1967). Na2O is particularly mobile, and its loss will result in a lower increase slightly (Baker and Henage, 1977; Weaver et al., 1972). peralkalinity index (P.I.=mol (Na+K)/Al). Likewise, the normative mineral assemblage will appear both less peralkaline and less silica- 5.2. Classification undersaturated, and if losses are large may even appear metalumi- nous and silica-oversaturated. To assess post-crystallization Na2O Whole-rock samples from Suswa and from maficvolcanicfields in mobility in peralkaline rocks, White et al. (2003) devised an index the rift valley, including Elmenteita (this study), Ndabibi (Macdonald for peralkalinity, FK/A (= mol (Fe+K)/ Al, with all Fe calculated as et al., 2001, 2008a), and Lolonito/Akira/Tandamara (this study; Fe2+), which utilizes components that are not highly mobile. They Macdonald et al., 2001, 2008a) are classified first with the total alkali- demonstrated that for non-hydrated peralkaline rocks, FK/A and P.I. silica (TAS) diagram (Fig. 5a; Le Bas et al., 1986). Silicic samples are fur- correlate strongly, and that FK/A may be useful to assess Na2O loss ther classified as peralkaline (alumina-undersaturated) if their P.I. is ≥1 for a wide variety of peralkaline suites (e.g., Jutzeler et al., 2010; or if FK/A≥0.60; using these criteria all Suswa trachyte and phonolite Macdonald et al., 2011b; Marshall et al., 2009; Ren et al., 2006; samples are peralkaline. The silica-saturation state can only be deter- Thorarinsson et al., 2011). A plot of FK/A versus P.I. is presented in mined after calculation of a normative mineral assemblage (Fig. 5b): Fig. 4. Of the major-element data presented in this report, half (13 samples with normative quartz (q) and hypersthene (hy) are oversatu- of 26) demonstrate loss of Na2O. Losses are especially high for the rated (e.g., quartz- or q-trachyte), samples with normative hy and oliv- pre-caldera (S1) and post-caldera (S6-7) suites. Notably, all samples ine (ol) are saturated (e.g., trachyte), and samples with normative ol from Suswa have FK/A values greater than 0.6, which implies peralk- and nepheline (ne) are undersaturated (e.g., nepheline- or ne- alinity (i.e. estimated P.I.≥1), and the single sample with normative trachyte). We use this terminology whether or not quartz or nepheline are present in the modal assemblage. Molecular norms were calculated

Fig. 3. Electron microprobe zoning proﬁle for anorthite (An) and orthoclase (Or) in a single alkali feldspar from the Ring-trench Group (RTG). X-ray images and additional Fig. 4. FKA versus P.I. (White et al., 2003). Many of the pre- and post-caldera samples analytical data can be found in Espejel-García (2006). have Na2O loss, which affects calculation of traditional P.I. Author's personal copy

94 J.C. White et al. / Lithos 152 (2012) 84–104

accumulation/resorption, volatile enrichment, or both. The behavior

of K2O is strongly decoupled from Na2O, increasing only slightly with Zr before maintaining a nearly constant (~5 wt.%) concentration through both suites.

5.4. Trace-element geochemistry

Variation diagrams that use Zr as a differentiation index plotted against three incompatible trace elements (ITE: Rb, Y, and Nb) and

three rare earth element ratios (LaN/SmN,CeN/YbN, and EuN/YbN; REE normalized following Boynton, 1984) are presented in Fig. 7. All three ITE demonstrate a positive linear correlation with Zr, with variable slopes. Zr/Rb ranges from 2.1 to 8.4 within the mafic suites, with most values around 5.6 (±1.8). There is no correlation between Zr and Zr/Rb. Average Zr/Rb within the Suswa suites is similar (5.7± 1.3); however, unlike the mafic suite there is a clear positive correlation between Zr and Zr/Rb, which increases from 4.2 to 6.4 as Zr increases from 490 to 1200; four samples with >1200 ppm Zr have higher Zr/Rb values (8.0–9.3). Zr/Y values range from 3.4 to 9.8 within the mafic suites. For samples with Zr between 79 and 237 ppm, Zr/Y increases slightly from 3.4 to 5.1; basaltic trachyandesite to trachyandesite samples from Tandamara with higher Zr (>260 ppm) have notably higher Zr/Y (6.2–9.7). Zr/Y values within both Suswa suites are fairly constant (8.4±0.9) and similar to the more evolved samples from Tandamara. There is also a weak positive correlation between Zr and Zr/Nb in the mafic suites, which rises from 3.0 to 5.4 (average fi Fig. 5. (a) TAS diagram for Suswa and adjacent volcanic elds (basalt through tra- – chyandesite). See Fig. 1 for location of adjacent fields. C1, pre- and syn-caldera rocks; 4.0±0.7). Four samples do not follow this trend and have lower (2.6 C2, post-caldera and Ring Trench Group. (b) Silica saturation diagram for C1 and C2. 2.7) values, including three samples from Ndabibi and a single sample of Tandamara basaltic trachyandesite. Zr/Nb ratios from Suswa are fairly constant, and considerably lower than those from the mafic with a FeO/FeO* ratio of 0.90 for basalt-trachyandesite lavas and 0.85 for suites (2.5±0.6). The median Zr/Nb ratio in the Suswa suite is 2.2; trachyte-phonolite lavas (cf. White et al., 2005).Thesevalueswerede- the average is skewed higher by four samples with Zr/Nb up to 4.0. termined following the method of Sack et al. (1980) for tempera- Scott and Skilling (1999) suggest that these higher-Zr/Nb trachytes tures between 1200 and 800°C and oxygen fugacity along the may represent batches of Longonot trachyte (Zr/Nb average 4.9) fayalite–magnetite–quartz (FMQ) buffer (Frostetal.,1988). Some that erupted from the Suswa volcanic center. mafic to intermediate lavas are generally hy-normative or weakly ne- or Through all suites, EuN/YbN decreases with increasing Zr before q-normative, and may thus be considered “transitional” rather than “alka- becoming approximately constant at >1100 ppm Zr, possibly due to line” or “subalkaline” (cf. Barberi et al., 1975). Within the mafictointer- the increasingly incompatible behavior of Eu with respect to alkali mediate suites the samples tend to become silica-oversaturated with feldspar as the melt becomes increasingly peralkaline (cf. Mahood increasing differentiation. Pre- and syn-caldera (S1–5) samples classify and Stimac, 1990; White, 2003; White et al., 2003). EuN/YbN corre- 2 primarily as peralkaline trachyte and peralkaline ne-trachyte to peralka- lates strongly (R =0.93) with Eu/Eu* (=EuN /√[SmN∙GdN]or=EuN / line phonolite; post-caldera (S6-7) and Ring Trench Group (S8) samples [TbN −2∙(TbN −SmN)/3]) throughout these suites (not shown). A classify primarily as peralkaline phonolite. negative Eu anomaly (Eu/Eu*b1.00) is indicated by EuN/YbN b1.65 (Eu/Eu*=0.5354∙[EuN/YbN]+0.1094.) Within the mafic suites, LaN/ 5.3. Major-element geochemistry SmN increases with Zr from 2.7 to 4.9, but remains fairly constant within the Suswa C1 and C2 suites (5.7±0.3). CeN/YbN also demonstrates a SiO2 is a poor differentiation index for the Suswa rocks: in addition positive correlation with Zr through the mafic suites, increasing from to varying only slightly (~2 wt.%) within each major suite, it tends to 3.9 to 8.4. CeN/YbN is systematically higher in the C2 suite (6.4–7.6) decrease with increasing differentiation as measured by indices such than the C1 suite (5.3–6.2). Other trace element variation diagrams are as Zr; therefore, we present major element oxide variation diagrams not presented; generally, for similar values of Zr and EuN/YbN,compared that use Zr as a differentiation index (Fig. 6). Within increasing Zr, to the C1 suite the C2 suite has lower Ce/Y and lower concentrations of Y the mafic-intermediate lavas increase in SiO2,Na2O, and K2O, while and HREE, and higher concentrations of K, Ti, P, Rb, Sr, Nb, Ba, and LREE. decreasing in FeO*, MgO, and CaO. Contrasting trends are seen with

Al2O3 and P2O5 concentrations: both decrease with increasing Zr 6. Petrogenesis within the Ndabibi suite, while increasing within the Lolonito– Akira–Tandamara suite. Both Suswa suites demonstrate considerable 6.1. Results of major-element modelling overlap with respect to their Zr concentrations, although most C1 samples have >1000 pm Zr and most C2 samples have b1000 pm Major-element modelling is herein accomplished with both mass Zr. Within both Suswa suites, silica remains fairly constant with in- balance (e.g., Bryan et al., 1969) and thermodynamic (MELTS; creasing Zr, with the C1 suite showing a slight decrease. In both Ghiorso and Sack, 1995; Asimov and Ghiorso, 1998) techniques. suites, there is a clear negative correlation with respect to Al2O3 and Mass balance modelling tests the feasibility of fractional crystalliza- a clear positive correlation with respect to FeO*. Also, with increasing tion hypotheses and constrains the relative proportion of the crystal-

Zr, both suites show decreases in MgO, CaO, and P2O5, with an in- lizing phases. MELTS modelling accomplishes the same results, but crease in Na2O. As with Al2O3, the Ring Trench Group (S8) shows an also can provide constraints on the order of crystallization as well as especially high concentration of Na2O, especially considering its rela- the conditions under which crystallization occurred (e.g., tempera- tively lower concentration of Zr, which may suggest a role for feldspar ture, pressure, and oxygen fugacity.) Author's personal copy

J.C. White et al. / Lithos 152 (2012) 84–104 95

Fig. 6. Zr as a differentiation index vs. major and minor elements. Symbols as in Fig. 4.

Maﬁc rocks with low Zr/Nb ratios (b3.1) in or near the CKPP in- suite, KS/94/03, was selected as the model daughter for Stage 3 petro- clude the most primitive (Mg#>0.50) alkali and transitional basalts genesis (alkali basalt to phonolite). of the Quaternary Ndabibi Formation (Olkaria), lower Pleistocene alkali basalts of the Thiba Formation (east of Aberdare), Miocene neph- 6.1.1. Mass balance modelling elinites of the Ngong Formation, and the Miocene Theloi alkali basalts All calculations were managed with Microsoft Ofﬁce Excel 2007. (Macdonald et al., 2001). We consider the Ndabibi basalts more likely Representative major-element mass balance models that test frac- than the others to be parental to the Suswa magmas, since they are tional crystallization hypotheses for the origin of C1 ne-trachyte the nearest in terms of both space and time. A sample of Ndabibi alka- from alkali basalt or gabbro (Stage 1), C2 phonolite from alkali basalt li basalt, ND64 (Davies and Macdonald, 1987; Macdonald et al., 2001, (Stage 3), and C1 phonolite from ne-trachyte (Stage 2) are presented 2008a), was selected as the model parent because of its low Zr in Table 8. Mineral analyses used for modelling Stages 1 and 3, except (76.4 ppm) and Th (1.74 ppm), and relatively high Mg# (mol Mg/ for magnetite and apatite, are from a sample of Ndabibi alkali basalt (Mg+Fe2+)=0.56, all iron calculated as Fe2+), Ni (42 ppm) and Cr (ND154a; Macdonald et al., 2008b). The model magnetite composi-

(72 ppm). Mass balance models using similar parental magmas, tion is stoichiometric Usp55Mgt45; the model apatite composition is such as ND154a (Mg#=0.57, 77 ppm Zr) and ND30 (Mg#=0.58, stoichiometric Ca5(PO4)3(OH)2. Mineral analyses used for modelling 79 ppm Zr), produce results nearly identical to models that use compositional variation in Stage 2 are from samples KS/94/33 (alkali ND64. A sample of pre-caldera ne-trachyte with the lowest Zr feldspar, clinopyroxene, and magnetite) and IS118 (olivine, apatite). (558 ppm) in the C1 suite, KS/94/33, was selected as the model All rock and mineral data used for modelling are normalized anhy- daughter for Stage 1 fractionation (alkali basalt to ne-trachyte); this drous. These major-element models suggest that the origin of and sample also serves as the model parent for the second stage of compositional variation within the C1 suite may be the result of magma petrogenesis (compositional variation within the C1 suite), ﬁrst ~93% fractional crystallization (F=0.07) of an assemblage of for which the selected model daughter is KS/94/29. A sample of 52% plagioclase, 31% clinopyroxene, 9% magnetite, 7% olivine, and post-caldera phonolite with the lowest Zr (491 ppm) in the C2 1% apatite from alkali basalt followed by an additional ~61% fractional Author's personal copy

96 J.C. White et al. / Lithos 152 (2012) 84–104

Fig. 7. Zr as a differentiation index vs. key trace elements and trace-elemental ratios. Symbols as in Fig. 4. crystallization (F=0.39) of an assemblage dominated by alkali feld- Sample ND64 was selected to represent the composition of the parent spar with subordinate clinopyroxene, magnetite, olivine, and apatite, or source rock as discussed above. Liquids with a composition similar for a total of ~96% crystallization. The sum of the squares of the resid- to KS/94/33 (ne-trachyte) were produced by high-pressure uals (Σr2) for each model are excellent (0.018 and 0.174, respective- (400 MPa) fractional crystallization models that started with low- ly). A major-element model for the origin of C2 phonolite from alkali H2O (0.5–1.0 wt.%) alkali basalt at low oxygen fugacity (≤FMQ). basalt was also calculated. This model suggests that the C2 suite could These models suggest that under these conditions, ne-trachyte may have a similar origin as the C1 suite (i.e., lays on the same liquid line be the result of 82–88% crystallization (F=0.12–0.18, approximately of descent), but is the result of a slightly smaller degree of fractional twice the value predicted by the mass-balance model) of a fractionat- crystallization (~91%, or F=0.09). These hypotheses will be further ing assemblage of 35–45% plagioclase, 40–54% clinopyroxene, 5–10% tested with thermodynamic and trace element modelling. magnetite, 0.6–13% olivine, and 0.4–0.6% apatite from alkali basalt. Higher proportions of plagioclase and lower proportions of clinopyr- 6.1.2. MELTS modelling oxene—similar to the results of mass-balance modelling—are associ- Magmatic differentiation processes involving solid–liquid equilib- ated with crystallization at both lower oxygen fugacities and initial ria can be thermodynamically modeled through use of the MELTS al- water contents (i.e., 0.5 wt.% H2O and FMQ-1). As one would expect, gorithm (Asimov and Ghiorso, 1998; Ghiorso and Sack, 1995) via the the relative proportions of magnetite and olivine is solely a function Adiabat_1ph (version 2.0) software package (Smith and Asimow, of oxygen fugacity. Generally speaking, lower pressures, higher initial 2005). For isobaric processes, MELTS computes phase equilibrium re- water contents, and higher oxygen fugacities result in less peralka- lations by minimization of Gibbs energy given a bulk composition, line, silica-oversaturated melts (q-trachyte) that are the result of a temperature, pressure, and oxygen fugacity. To model fractional crys- smaller degree of fractional crystallization (F=0.18–0.29). tallization, MELTS first determines the liquidus temperature and liquidus phase for the bulk composition and specified parameters. 6.1.3. Constraints on pre-eruptive conditions Mineral phases (except for a very small fraction, 0.01%) are then sub- The pre-eruptive conditions of the Suswa C1 magma chamber may tracted, a new liquid composition is calculated, and the temperature be constrained by comparing the observed mineral assemblage and is lowered by 5 °C. MELTS then calculates a new equilibrium phase as- whole-rock geochemistry with the results from MELTS, QUIlF, and ex- semblage, and the process repeats until the solidus temperature, or a periments on phonolitic liquids from literature. Values for intensive temperature specified by the user, has been reached. thermodynamic parameters for samples of peralkaline trachyte (S1, The results of a total of twenty four models of fractional crystalli- IS118; S3, 870156) were determined using QUILF95 (Table 10; zation are presented in Fig. 8 and summarized in Table 9. Models vari- Andersen et al., 1993). Calculated clinopyroxene-olivine tempera- ables include two oxygen buffers (FMQ and FMQ-1), four starting tures range from 858 to 885 °C at a pressure of 100 MPa and from water contents (0.1, 0.5, 1.0, and 1.5 wt.%), and three pressures of 885 to 919 °C at a pressure of 400 MPa; magnetite compositions in crystallization (100, 400, and 500 MPa). For all models, temperature these samples suggest oxygen fugacities between FMQ-0.8 and decreased incrementally from a superliquidus guess of 1400 °C. FMQ-1.1. MELTS results for crystallization at 400 MPa and oxygen Author's personal copy

J.C. White et al. / Lithos 152 (2012) 84–104 97

Table 8 Major-element mass balance models.

Stage 1. Fractional Crystallization of Alkali Basalt to ne-Trachyte (C1)

ND64 KS/94/33 PL-core PL-rim CPX MGT OL AP Calc. Residuals FC model

SiO2 wt.% 47.60 61.72 52.01 56.54 52.72 0.00 38.07 0.00 47.60 -0.01 KS/94/33 0.067=F

TiO2 1.86 0.89 0.00 0.00 0.67 19.35 0.00 0.00 1.94 -0.08 PL-core 0.355 38.1%

Al2O3 15.94 16.60 29.78 26.66 2.64 0.00 0.02 0.00 15.95 0.00 PL-rim 0.131 14.1% ⁎ FeO 11.30 6.50 0.61 0.86 6.76 80.65 23.18 0.00 11.28 0.02 CPX 0.284 30.5% MnO 0.20 0.27 0.00 0.00 0.21 0.00 0.38 0.00 0.11 0.09 MGT 0.087 9.3% MgO 7.37 0.56 0.00 0.00 16.77 0.00 38.05 0.00 7.38 -0.01 OL 0.068 7.3% CaO 12.27 1.51 13.53 9.72 19.97 0.00 0.31 55.82 12.25 0.02 AP 0.007 0.7%

Na2O 2.68 6.62 3.89 5.66 0.25 0.00 0.00 0.00 2.64 0.04 0.999

K2O 0.50 5.18 0.19 0.56 0.00 0.00 0.00 0.00 0.49 0.01

P2O5 0.28 0.14 0.00 0.00 0.00 0.00 0.00 42.39 0.30 -0.02 99.93 0.018=Σr2

Stage 3. Fractional Crystallization of Alkali Basalt to Phonolite (C2)

ND64 KS/94/03 PL-core PL-rim CPX MGT OL AP Calc. Residuals FC Model

SiO2 wt.% 47.60 58.45 52.01 56.54 52.72 0.00 38.07 0.00 47.59 0.00 KS/94/03 0.088=F

TiO2 1.86 1.25 0.00 0.00 0.67 19.35 0.00 0.00 1.95 -0.09 PL-core 0.370 40.6%

Al2O3 15.94 17.14 29.78 26.66 2.64 0.00 0.02 0.00 15.95 -0.01 PL-rim 0.100 11.0% ⁎ FeO 11.30 7.01 0.61 0.86 6.76 80.65 23.18 0.00 11.28 0.02 CPX 0.285 31.2% MnO 0.20 0.27 0.00 0.00 0.21 0.00 0.38 0.00 0.11 0.09 MGT 0.085 9.4% MgO 7.37 1.04 0.00 0.00 16.77 0.00 38.05 0.00 7.39 -0.02 OL 0.066 7.3% CaO 12.27 2.95 13.53 9.72 19.97 0.00 0.31 55.82 12.27 0.00 AP 0.006 0.6%

Na2O 2.68 7.11 3.89 5.66 0.25 0.00 0.00 0.00 2.70 -0.02 1.000

K2O 0.50 4.40 0.19 0.56 0.00 0.00 0.00 0.00 0.51 -0.01

P2O5 0.28 0.38 0.00 0.00 0.00 0.00 0.00 42.39 0.28 0.00 100.04 0.018=Σr2

Stage 2. Fractional Crystallization of ne-Trachyte to Phonolite (C1)

KS/94/33 KS/94/29 KFS CPX MGT OL AP Calc. Residuals FC Model

SiO2 61.72 60.17 65.76 51.07 0.00 31.14 0.40 61.65 0.08 KS/94/29 0.392=F

TiO2 0.89 0.57 0.08 0.86 26.04 0.25 0.00 0.87 0.02 KFS 0.541 87.0%

Al2O3 16.60 15.01 19.92 0.66 1.02 0.00 0.00 16.71 −0.11 CPX 0.031 5.0% FeOT 6.50 7.34 0.19 15.65 69.32 54.66 0.17 6.50 0.00 MGT 0.022 3.5% MnO 0.27 0.35 0.00 1.06 2.09 4.00 0.14 0.33 −0.05 OL 0.028 4.5% MgO 0.56 0.43 0.00 10.18 1.47 9.38 0.00 0.78 −0.22 AP 0.000 0.0% CaO 1.51 1.26 0.80 19.96 0.06 0.58 56.70 1.58 −0.07 1.014

Na2O 6.62 7.83 6.97 0.56 0.00 0.00 0.22 6.86 −0.24

K2O 5.18 4.99 6.28 0.00 0.00 0.00 0.00 5.36 −0.18

P2O5 0.14 0.10 0.00 0.00 0.00 0.00 42.37 0.05 0.09 100.68 0.174=Σr2 fugacity between FMQ and FMQ-1 are consistent with QUIlF geother- magnetite found in Suswa lavas, coupled with the lack of phases mobarometic results for S1 and S2 magmas. Additionally, evidence such as nepheline, leucite, biotite, and titanite, may further indicate from experimental petrology suggests that the relatively high TiO2 crystallization at relatively high T (>800 °C), P (>300 MPa), and (>0.60 wt.%) in these samples indicate temperatures >890 °C H2O (>4 wt.%) (Freise et al., 2003; Scaillet et al., 2008). Such pre- (Berndt et al., 2001). Both MELTS and trace element models (which eruptive conditions differ markedly from Vesuvius where, for exam- assume DH2O ≈0) suggest that 82–88% fractional crystallization of al- ple, experimental studies have shown that the phonolitic magma kali basalt with an initial concentration of H2O between 0.5 and chamber has become increasingly shallower with time (Scaillet et 1.0 wt.% will result in a slightly water-undersaturated melt (4.1– al., 2008).

5.5 wt.% H2O). These high predicted water values are consistent with the presence of alkali amphibole, which occurs as an interstitial 6.1.4. Pearce element ratio diagrams groundmass phase in most Suswa lava units and as a common (0.5 to Pearce (1968) element ratio (PER) diagrams graph ratios of major 6.0 vol.%) phase in syenite nodules recovered from unit S2 (Skilling, elements or combinations of major elements with an incompatible el- 1988) and the Kilombe volcano (Ridolﬁ et al., 2006). ement (e.g., Si/Zr, Si+Al/Zr). PER diagrams are based on the stoichi- Pre-eruptive conditions have been experimentally constrained for ometry of rock-forming minerals, and slopes of data distributions several trachytic-phonolitic systems, including the Laacher See volca- are equal to major element ratios of minerals lost or gained during no (Berndt et al., 2001; Harms et al., 2004), the Kerguelen Archipela- differentiation of a cogenetic suite of rocks (Russell and Nicholls, go (Freise et al., 2003, 2009), the Canary Islands (Andújar et al., 2008, 1988; Stanley and Russell, 1990). For example, data plotted with Si/ 2010), and Mount Vesuvius (Scaillet et al., 2008). These studies ﬁrst Zr on the ordinate and Fe/Zr on the abscissa will form a linear trend indicate that our modelled estimates for pre-eruptive water concen- with a slope that varies depending on the fractionating assemblage trations are similar to these other systems, which range from ~3 to from vertical for a phase that lacks stoichiometric Si such as magne-

6 wt.% H2O. Under these relatively hydrous conditions, amphibole co- tite (Fe2O3; Fe:Si=2/0=undefined) to horizontal for a phase with exists with sanidine, but only at temperatures lower than those esti- non-stoichiometric Fe such as quartz (SiO2; Fe:Si=0/1=0). PER dia- mated for the Suswa lavas (b815 °C; Harms et al., 2004; Scaillet et al., grams for trachtyes and phonolites from Suswa and mafic to interme- 2008), which may explain the lack of phyric amphibole in the lavas diate rocks from the nearby Elmenteita, Ndabibi, and Lolonito–Akira– despite its presence in syenites. Experimental results also suggest that Tandamara volcanic fields (Macdonald et al., 2008a; this report) are the typical phyric assemblage of alkali feldspar+clinopyroxene+ presented in Fig. 9. Author's personal copy

98 J.C. White et al. / Lithos 152 (2012) 84–104

Fig. 8. MELTS models for conditions of generation of Suswa samples. Diamonds, maﬁc to intermediate compositions for volcanic ﬁelds adjacent to Suswa; circles, C1 compositions.

These figures strongly suggest that the mafic to intermediate volca- Within the Suswa suite, both the C1 and C2 suites follow identical nic fields evolved by similar mechanisms, most probably fractionation and overlapping trends, which suggest a similar origin and similar of an assemblage dominated by, and with approximately equal processes for within-suite major-element variations. Specifically, the amounts of, plagioclase and clinopyroxene. The mathematically pre- slope regressed through the plot of Na+K/Zr vs. Si+Al/Zr is 0.26, dicted slope for clinopyroxene fractionation ([Fe+Mg]:[Si+Al]= very close to the mathematically predicted slope for alkali feldspar 0.50) matches the slope regressed through the data (Fig. 9a), and the fractionation ([Na+K]:[Si+Al]=0.25), suggesting that variability average of the predicted slopes for clinopyroxene (Ca:[Si+Al]=0.50) of Na and K in the suite is the result of feldspar fractionation, resorp- and plagioclase ([Ca+Na+K]:[Si+Al]=0.25) is 0.38, which also tion, or accumulation. The slightly higher slope of Ca+Na+K/Zr and matches the slope regressed through the data (Fig. 9b). The much smaller slope of Fe+Mg/Zr suggest that clinopyroxene plays a dimin- lower slope regressed through the alkali data (Fig. 9c) is consistent ished, but still significant role, in the evolution of this suite. These ob- with fractionation of plagioclase with a composition of approximately servations accord with the petrography of both the maficto

An60, which is consistent with observed data. Although plots of Ca+ intermediate centers and the Suswa suite. They are also consistent Na+K/Zr and Fe+Mg/Zr are consistent with a fractional crystallization with the results of MELTS modelling for fractional crystallization of al- origin of the Suswa suite from similar alkali basalts (albeit with a dimin- kali basalt at ~400 MPa pressure, oxygen fugacities between FMQ and ished role for clinopyroxene, as suggested by decreased slopes in both FMQ-1, and initial water contents of 0.5–1.0 wt.%. For these plots, Zr plots), the plot of Na+K/Zr precludes fractional crystallization as the was estimated from the F-values calculated by MELTS, the concentra- sole mechanism responsible for the petrogenesis of Suswa trachyte, tion of Zr in the parental magma (76.4 ppm), and a bulk partition co- since the maﬁc-intermediate trend intersects the Suswa trend. efﬁcient for Zr of 0.16 (see below), such that Zr=76.4∙F-0.84. Author's personal copy

J.C. White et al. / Lithos 152 (2012) 84–104 99

Table 9

Summary of MELTS results of FC modelling at ~61 wt.% SiO2 normalized anhydrous.

fO2 buffer FMQ FMQ-1 Parent P (MPa) F T (°C) wt.% TA Fe/Al P.I. F T (°C) wt.% TA Fe/Al P.I.

Alkali basalt (ND64)

1.5 wt.% H2O 100 0.29 955 8.31 0.62 0.79 0.23 930 9.23 0.92 1.03 400 0.23 885 9.55 0.56 0.76 0.18 830 10.19 0.70 0.76 500 0.22 875 9.53 0.57 0.74 b0.20 b855 >10.25 b0.79 >0.72

1.0 wt.% H2O 100 0.27 965 8.45 0.56 0.83 0.25 930 9.10 0.90 0.99 400 0.18 910 10.38 0.55 0.87 0.17 895 10.27 0.78 0.92 500 0.17 905 10.26 0.57 0.85 b0.30 b880 >11.11 b1.03 >0.83

0.5 wt.% H2O 100 0.22 1000 8.73 0.64 0.96 0.18 960 9.76 0.95 1.21 400 0.14 965 10.93 0.63 1.07 0.12 925 11.65 0.87 1.27 500 0.13 950 11.30 0.62 1.07 b0.11 b920 >12.69 b0.95 >1.19

0.1 wt.% H2O 100 0.15 1040 9.15 0.77 1.23 0.12 995 10.50 1.70 1.65 400 0.08 1020 12.20 0.75 1.58 b0.07 b990 >11.95 >1.45 >1.86 500 0.06 1005 12.85 0.79 1.76 b0.06 b990 >12.77 >1.65 >1.79

Table 10 QUILF95 results.

Spinel Olivine Augite Results

Sample: N-Ti N-Mg N-Mn Fo La En Wo P (MPa) T (°C) log fO2 ΔFMQ aSiO2(qtz) 870156 Input 0.690 0.059 0.075 0.175 0.010 0.239 0.414 100 Calc 0.034 0.213 885 −13.80 −0.99 0.86 870156 Input 0.690 0.059 0.075 0.175 0.010 0.239 0.414 400 Calc 0.035 0.210 919 −13.12 −1.20 0.75 IS118 Input 0.594 0.021 0.070 0.232 0.010 0.283 0.447 100 Calc 0.041 0.260 858 −14.17 −0.84 0.65 IS118 Input 0.594 0.021 0.070 0.232 0.010 0.283 0.447 400 Calc 0.042 0.257 890 −14.17 −1.07 0.56

Entries in italics in the “Input” column signify values that were set as trial values; the values calculated by QUILF95 are in normal font in the sujacent “Calc” column.

6.2. Results of trace-element modeling (1997), respectively, with inputs (XAn, T for plagioclase; melt chemistry, P, and T for clinopyroxene) derived from MELTS results; in both 6.2.1. Origin of the pre- and syn-caldera suite (S1-5) cases, bulk partition coefficients were similar for incompatible trace Mass-balance and MELTS models suggest that ne-trachyte parental elements, but Sr and Ba were systematically higher for the calculated to the C1 suite (represented by sample KS/94/33) may be the result of values compared to the “static” values of Caroff et al. (1993). The re- 93% (mass-balance) or 82–88% (MELTS) fractional crystallization of an sults of trace-element modelling of the petrogenesis of alkali basalt to assemblage of plagioclase and clinopyroxene with subordinate oliv- ne-trachyte are presented as spiderdiagrams (Thompson, 1982)in ine, magnetite, and apatite from a Ndabibi-like alkali basalt (repre- Fig. 10a (ND64 to KS/94/33) and 10b (KS/94/33 to KS/94/29). These sented by sample ND64) at relatively high pressure (400 MPa) and results are in agreement with those derived from major-element low oxygen fugacity (FMQ to FMQ-1). Mass-balance models also sug- modelling, suggesting ~90% fractional crystallization for Stage 1 and gest that compositional variation within the C1 suite may be further ~60% fractional crystallization for Stage 2 for a total of ~96% attributed to an additional 60% fractional crystallization of an assem- crystallization. blage dominated by alkali feldspar with subordinate clinopyroxene, olivine, magnetite, and apatite from KS/94/33. These models are fur- 6.2.2. Origin of the post-caldera suite (S6-7) ther tested with trace-element modelling. Although mass-balance modelling suggests that ne-trachyte pa- Bulk partition coefficients are calculated for trace element model- rental to the C2 suite (represented by KS/94/03) may lie on the ling from the results of mass balance modelling (Table 8, Stages 1 and same liquid line of descent (LLOD) as KS/94/03 as the result of a sligh- 2) and mineral/melt partition coefficients from Caroff et al. (1993) for ter (~2% less) degree of fractional crystallization (Table 8, Stage 3), all minerals except apatite (Fujimaki, 1986) for Stage 1 and mineral/ MELTS models indicate no such relationship. This is, of course, consis- melt partition coefficients from Villemant et al. (1981), Villemant tent with the mineralogical evidence presented above that strongly (1988), Mahood and Stimac (1990), White et al. (2003), and recommends a role for open-system processes in the petrogenesis Dawson and Hinton (2008) for Stage 2. Selected bulk partition coeffi- of the C2 suite, and especially in the RTG. Additional evidence for cients for plagioclase and clinopyroxene were also calculated using this may be provided by trace-element geochemistry, notably the el- the formulations of Bindeman et al. (1998) and Wood and Blundy evated values of LREE/HREE and Ce/Y in the C2 suite compared to the C1 suite. We postulate that the fractional crystallization of C1 ne- trachyte to phonolite would have left behind a syenitic residue simi- Table 11 lar to xenoliths described from the Kilombe volcano (Ridolfi et al., Results of linear mixing regressions versus Rb (C2, X =mC2,Rb +b). 2006), and that a later intrusion of alkali basalt (again similar to mBR2 ND64) may have assimilated this syenitic material during fractional

Sr −2.7176 545.07 0.92 crystallization which resulted in the origin of the C2 suite. This postu- Y 0.7040 −15.15 0.91 lation is tested with trace-element models of assimilation and frac- Zr 6.3291 −210.33 0.96 tional crystallization (DePaolo, 1981) presented in Figs. 11 and 12. Nb 3.0810 −105.69 0.85 All model parameters (parent, bulk partition coefﬁcients) are identi- Ba −17.1930 2828 0.97 cal to those used to model the origin of C1, with an average of Author's personal copy

100 J.C. White et al. / Lithos 152 (2012) 84–104

of the post-caldera suite (S6-7). Potential values of Sr, Y, Zr, Nb, and Ba for the other end-member can be constrained by determining the equations of linear regressions of these versus Rb through the RTG samples plus KS/94/22 (Table 11) and then calculating these values over a series of values for Rb calculated for the other end member as a function of proportion of mixing by re-arranging the linear mixing formula to:

¼ ðÞ– þ α =α; C2 CH C1 C1

where C2 is the composition of the hypothetical other end member, C1 is the composition of the starting magma composition (taken to be KS/94/22), CH is the composition of hybrid magma (taken to be KS/94/44, the sample of RTG with the lowest Rb, 123 ppm), and α is

the proportion of C2 relative to C1. These results for Sr and Ba are presented in Fig. 13, along with the calculated LLOD for fractional crystallization of the post-caldera (S6-7) suite and the curve for the composition of the corresponding solid residue. All ﬁve curves (including those not presented) intersect the solid residue curve at approximately the same point and proportion of mixing; although the curves shown also seem to indicate that one of two samples of Lolonito basalt could also be a potential end-member, this is not suggested by the incompatible elements. Therefore, it seems most likely that compositional variation within the RTG is due to ~45% resorption of an early-crystallized fraction of the feldspar-dominated assemblage previously formed during fractional crystallization of the post- caldera suite or perhaps of the RTG suite prior to resorption. We propose that resorption was driven by thermal disequilibrium: reheating of the RTG magma chamber (which was otherwise evolving along a LLOD similar to the S6-7 suite) due to a likely intrusion of basaltic magma.

Fig. 9. Pearce (1968) element ratio diagrams for Suswa and adjacent ﬁelds. See text for further discussion.

“normal” Kilombe syenite analyses taken to represent the assimilant. Fig. 11 clearly shows that assimilation of such a syenite will result in melts with elevated Ce/Y and similar Zr/Nb ratios compared to the C1 suite. The ne-trachyte parental to C2 (KS/94/03) could be the result of 80% crystallization of alkali basalt coupled with a small degree (Ma/Mc=0.1) of syenite. AFC model results are also presented on a spiderdiagram (Thompson, 1982; Fig. 12), which conﬁrm these results. Lower amounts of fractional crystallization coupled with assimilation of silica-undersaturated syenite would thus result in melts similar to C1, but with lower SiO2, lower incompatible trace element concentrations, and higher Na2O, Sr, and Ba as observed. Composi- tional variation within the C2 suite is the result of both closed- and open-system processes (Fig. 13): fractional crystallization of alkali feldspar is responsible for most of the variation within the post- caldera (S6–7) portion of the C2 suite, but mineralogical evidence precludes this from also being true of S8.

6.2.3. Origin of the Ring Trench Group (S8) Mineralogical and trace element data provide compelling evidence for a significant role for magma mixing and feldspar resorption within the Ring Trench Group. Anorthoclase crystals in RTG lavas ex- hibit both normal and reverse zoning, compatible trace elements concentrations are high compared to the rest of the C2 suite (>190 ppm Sr, >1000 ppm Ba), and trace element diagrams that plot Rb versus compatible elements (Sr, Ba) form strongly linear trends (Fig. 13). If we postulate from this evidence that compositional variation within these lavas is the result of mixing processes, then it is a relatively triv- Fig. 10. Multi-element variation (following Thompson, 1982). (a) Degree of fraction- ial exercise to determine the chemical characteristics of the end- ation required to generate C1 ne-trachyte (KS/94/33) from mafic rock (ND64). F is members involved in these processes. One likely end-member is fraction of melt remaining. Bulk partition coefficients used in modeling are indicated represented by sample KS/94/22, one of the more evolved magmas on X-axis. (b) Degree of fractionation required for C1 ne-trachyte to C2 (KS/94/29). Author's personal copy

J.C. White et al. / Lithos 152 (2012) 84–104 101

Fig. 11. Trace element models for assimilation/fractional crystallization in post-caldera Fig. 13. Trace element models for fractional crystallization and mixing in Suswa sam- (C2) samples. El, Elmenteita; Nd, Ndabibi; LAT, Lolonito-Akira-Tandamara. TB ashes are ples. Symbols and acronyms as in Fig. 10. stratigraphically intercalated with S2 syn-caldera tuffs.

7. Conclusions

1. The initial shield-building trachyte and ne-trachyte lavas (PLG, S1) that built the Suswa volcanic ediﬁce were the result of protracted (90–96%) fractional crystallization of an Ndabibi-like basalt with

relatively low initial H2O concentrations (b1 wt.%) under reducing conditions (FMQ-1) at relatively high pressure (400 MPa) (Fig. 14, part 1). This pressure corresponds to a depth of approximately 14 km, which corresponds to the boundary between the upper and middle crust in the southern CKPP (Mooney and Christensen, 1994) that may have served as a regional magma trap. This conclusion implies that if Ndabibi basalts are also parental to silica- oversaturated trachytes and rhyolites that comprise the Greater Olkaria volcanic center (Macdonald et al., 2008b; Marshall et al., 2009), then they must have differentiated at a much more shallow level—a conclusion supported by experimental data that suggest that GOVC magma was stored at around 100 MPa (Scaillet and Macdonald, 2001). This is consistent with the mechanism described by LeMasurier et al. (2003, 2011)) for coexisting trachyte, rhyolite, and phonolite in the Marie Byrd Land, western Antarctica rift system. 2. Petrogenesis of the syn-caldera units, the eruption of which resulted in the formation of the ﬁrst Suswa caldera, was clearly similar to the petrogenesis of the pre-caldera lavas as evidenced by identical major- and trace-element concentrations and intra- suite variation. However, the presence of a greater amount of pyro- clastic activity, the contemporaneous eruption of trachybasaltic ashes, and the eruption of carbonate-rich trachytes strongly indi- cates a greater degree of complexity. A detailed study of the syn- caldera units will be necessary to resolve these issues. 3. Fractional crystallization of pre- and syn-caldera trachytes resulted Fig. 12. Spiderdiagram (Thompson, 1982) for assimilation/fractional crystallization models. in the formation of a syenitic residue. We propose that recharge of Author's personal copy

102 J.C. White et al. / Lithos 152 (2012) 84–104

Fig. 14. Schematic picture of the evolution of open-system behavior at Suswa following Skilling (1993). Part 1: Protracted fractional crystallization of a Ndabibi-like basalt results in the formation of trachyte lavas that erupt to build a large shield volcano, which subsequently undergoes caldera collapse. Fractional crystallization of pre- and syn-caldera trachytes resulted in the formation of a syenitic residue. Part 2: Recharge and crystallization of basalt into the magma chamber resulted in the assimilation of this syenite, resulting in the formation of more strongly silica-undersatuarated trachyte. Fractional crystallization of anorthoclase from post-caldera ne-trachyte resulted in the formation of phonolite lavas that built the Ol-Doinyo Nyoke cone. Part 3: A later injection of basaltic magma appears to have created thermal disequilibrium within the phonolitic magma chamber which resulted in the resorption of these anorthoclase phenocrysts, and the formation of the Ring Trench Group lavas.

Ndabibi-like basalt into a warm, crystallized or partially crystal- the Department of Geological Sciences at the University of Texas at lized syenitic magma chamber provided a ready assimilant for El Paso. this basalt (Fig. 14, part 2). Modelling suggests that a relatively small (Ma/Mc=0.1) amount of assimilation coupled with a lesser degree (80–90%) of fractional crystallization compared to the ear- References

lier suites resulted in the formation of lower-silica, more silica- Andersen, D.J., Lindsley, D.H., Davidson, P.M., 1993. QUILF: a PASCAL program to asses undersaturated trachyte that characterizes the post-caldera suite. equilibria among Fe-Mg-Mn-Ti oxides, pyroxenes, olivine, and quartz. Computers 4. Fractional crystallization of anorthoclase from post-caldera ne- and Geosciences 19, 1333–1350. trachyte resulted in the formation of phonolite lavas that built Andújar, J., Costa, F., Martí, J., Wolff, J.A., Carroll, M.R., 2008. Experimental constraints on pre-eruptive conditions of phonolitic magma from the caldera-forming El the Ol-Doinyo Nyoke cone. A later injection of basaltic magma ap- Abrigo eruption, Tenerife (Canary Islands). Chemical Geology 257, 173–191. pears to have created thermal disequilibrium within the phonolitic Andújar, J., Costa, F., Martí, J., 2010. Magma storage conditions of the last eruption of Teide volcano (Canary Islands, Sparin). Bulletin of Volcanology 72, 381–395. magma chamber which resulted in the resorption of these fi “ ” Asimov, P.D., Ghiorso, M.S., 1998. Algorithmic modi cations extending MELTS to calcu- anorthoclase phenocrysts, which is seen in a linear overprinting late subsolidus phase relations. American Mineralogist 83, 1127–1131. of post-caldera fractional crystallization (Fig. 14, part 3). Baer, G., Hamiel, Y., Shamir, G., Nof, R., 2008. Evolution of magma-driven earthquake 5. Although fractional crystallization was a dominant mechanism in swarm and triggering of the nearby Oldoinyo Lengai eruption, as resolved by InSAR, ground observations and elastic modeling, East African Rift, 2007. Earth the petrogenesis of the Suswa volcano, open system processes and Planetary Science Letters 272, 339–352. driven by recharge of basaltic magma, including assimilation and Bailey, D.K., Macdonald, R., 1975. Fluorine and chlorine in peralkaline liquids and the feldspar resorption were clearly important processes—especially need for magma generation in an open system. Mineralogical Magazine 40, 405–414. with respect to the origin of the post-caldera and ring trench Baker, B.H., Henage, L.F., 1977. Compositional changes during crystallization of some groups. peralkaline silicic lavas of the Kenya Rift Valley. Journal of Volcanology and Geo- thermal Research 2, 17–28. Baker, B.H., Mitchell, J.G., Williams, L.A.J., 1988. Stratigraphy, geochronology and Acknowledgements volcano-tectonic evolution of the Kengdon-Naivasha-Kinangop region, Gregory Rift Valley, Kenya. Journal of the Geological Society of London 145, 107–116. The authors would like to thank Ray Macdonald and an anony- Barberi, F., Ferrara, G., Santacroce, R., Treuil, M., Varet, J., 1975. A transitional basalt- mous reviewer for their reviews and suggestions that helped improve pantellerite sequence of fractional crystallization, the Boina Centre (Afar Rift, Ethiopia). Journal of Petrology 16, 22–56. this manuscript, as well as special volume editor Gregor Markl, Berndt, J., Holtz, F., Koepke, J., 2001. Experimental constraints on storage conditions Michael Marks, and the Institut für Geowissenschaften, Universität in the chemically zoned phonolitic magma chamber of the Laacher See volcano. – Tübingen, for sponsoring the 2011 PERALK-CARB symposium. A Contributions to Mineralogy and Petrology 140, 469 486. Biggs, J., Anthony, E.Y., Ebinger, C.J., 2009. Multiple inflation and deflation events at grant from the University Research Committee at Eastern Kentucky Kenyan volcanoes, East Africa Rift. Geology 37, 979–982. University provided funding for some whole rock analyses and field Bindeman, I.N., Davis, A.M., Drake, M.J., 1998. Ion microprobe study of plagioclase- work for JCW and NSF INT 0096863 and 0332956 provided funding basalt partition experiments in natural concentration levels of trace elements. Geo- chimica et Cosmochimica Acta 62, 1175–1193. to EYA, VVE, and PAO. Funding for electron probe microanalyses Boynton, W.R., 1984. Cosmochemistry of the rare earth elements: meteorite studies. was provided by National Science Foundation grant EAR-0116660 to Rare Earth Element GeochemistryElsevier, Amsterdam. Author's personal copy

J.C. White et al. / Lithos 152 (2012) 84–104 103

Bryan, W.B., Finger, L.W., Chayes, F., 1969. Estimating proportions in petrographic peralkaline rhyolite suite, the Greater Olkaria Volcanic Complex, Kenya rift valley. mixing equations by least-squares approximation. Science 163, 926–927. Journal of Petrology 49, 1515–1547. Caroff, M., Maury, R.C., Leterrier, J., Joron, J.L., Cotton, J., Guille, G., 1993. Trace element Macdonald, R., Bagiński, B., Leat, P.T., White, J.C., Dzierżanowski, P., 2011a. Mineral sta- behavior in the alkali basalt-comenditic trachyte series from Mururoa Atoll, French bility in peralkaline silicic rocks: Information from trachytes of the Menengai vol- Polynesia. Lithos 30, 1–22. cano, Kenya. Lithos 125, 553–568. Carracedo, J.C., Perez Torrado, F.J., Ancochea, E., Meco, J., Hernan, F., Cubas, C.R., Macdonald, R., Bagiński, B., Ronga, F., Dzierzanowski, P., Lustrino, M., Marzoli, A., Rodriguez-Badiola, E., Ahijado, A., 2002. Cenozoic volcanism II: the Canary Islands. Melluso, L., 2011b. Evidence for extreme fractionation of peralkaline silicic The Geological Society of London 18, 439–472. magmas, the Boseti volcanic complex, main Ethiopian rift. Mineralogy and Petrology. Clarke, M.C.G., Woodhall, D.G., Allen, D., Darling, G., 1990. Geological, volcanological doi:. and hydrogeological controls on the ocurrence of geothermal activity in the area Mahood, G.A., Stimac, J.A., 1990. Trace-element partitioning in pantellerites and tra- surrounding Lake Naivasha, Kenya, Ministry of Energy. 138 pp. chytes. Geochimica et Cosmochimica Acta 54, 2257–2276. Davies, G.R., Macdonald, R., 1987. Crustal influences in the petrogenesis of the Naivasha Marshall, A.S., Macdonald, R., Rogers, N.W., Fitton, J.G., Tindle, A.G., Nejbert, K., Himton, basalt-comendite complex: combined trace element and Sr-Nd-Pb isotope con- R.W., 2009. Fractionation of peralkaline silicic-magmas: Greater Olkaria volcanic straints. Journal of Petrology 28, 1009–1031. complex, Kenya Rift Valley. Journal of Petrology 50, 323–359. Dawson, J.B., Hinton, R.W., 2008. The composition of anorthoclase and nepheline in Mooney, W.D., Christensen, N.I., 1994. Composition of the crust beneath the Kenya rift. Mount Kenya phonolite and Kilimanjaro trachyte, and crystal-glass partitioning Tectonophysics 236, 391–408. of elements. The Canadian Mineralogist 46, 1455–1464. Nash, W.P., Carmichael, I.S.E., Johnson, R.W., 1969. The mineralogy and petrology of DePaolo, D.J., 1981. Trace elements and isotopic effects of combined wall rock assimi- Mount Suswa, Kenya. Journal of Petrology 10, 409–439. lation and fractional crystallization. Earth and Planetary Science Letters 53, Noble, D.C., 1967. Sodium, potassium and ferreous iron contents of some secondarily 189–202. hydrated natural silicic glasses. American Mineralogist 52, 280–286. Espejel-García, V., 2006. Magma dynamics inferred from feldspar zoning in post- Omenda, P.A., 1997. The geochemical evolution of Quaternary volcanism in the south- caldera phonolites, Suswa volcano, Kenya Rift, Africa. University of Texas at El central portion of the Kenya Rift. Doctoral dissertation, University of Texas at El Paso. 175 pp. Paso, 218 pp. Fitton, J.G., 1987. The Cameroon line, west Africa: a comparison between oceanic and Panter, K.S., Kyle, P.R., Smellie, J.L., 1997. Petrogenesis of a phonolite-trachyte succes- continental alkaline volcanism. The Geological Society of London, Special publica- sion at Mount Sidley, Marie Byrd Land, Antarctica. Journal of Petrology 38, tions 30, 273–291. 1225–1253. Foland, K.A., Landoll, J.D., Henderson, C.M.B., Chen, J., 1993. Formation of cogenetic Pearce, T.H., 1968. A contribution to the theory of variation diagrams. Contributions to quartz and nepheline syenites. Geochimica et Cosmochimica Acta 57, 697–704. Mineralogy and Petrology 19, 142–157. Freise, M., Holtz, F., Koepke, J., Scoates, J., Leyrit, H., 2003. Experimental constraints on Potter, L.S., 1996. Chemical variation along strike in feldespathoidal rocks of the eastern the storage conditions from the Kerguelen Archipelago. Contributions to Mineralogy alkalic belt, Trans-Pecos Magmatic Province, Texas and New Mexico. The Canadian and Petrology 145, 659–672. Mineralogist 34, 241–264. Freise, M., Holtz, F., Nowak, M., Scoates, J.S., Strauss, H., 2009. Differentiation and crys- Randel, R.P., Johnson, R.W., 1991. Geology of the Suswa area, Mines and Geological tallization conditions of basalts from the Kerguelen large igneous province: an ex- Department (Nairobi). 41 pp. perimental study. Contributions to Mineralogy and Petrology 158, 505–527. Ren, M., Anthony, E.Y., Omenda, P.A., White, J.C., Macdonald, R., Bailey, D.K., 2006. Appli- Frost, B.R., Lindsley, D.H., Andersen, D.J., 1988. Fe-Ti oxide-silicate equilibria assem- cation of the QUILF thermobarometer to the peralkaline trachytes and pantellerites of blages with fayalitic olivine. American Mineralogist 73, 727–740. the Eburru Volcanic Complex, East African Rift, Kenya. Lithos 91, 109–124. Fujimaki, H., 1986. Partition coefficients of Hf, Zr, and REE between zircon, apatite and Ridolfi, M., Renzulli, A., Macdonald, R., Upton, B.G.J., 2006. Peralkaline syenite autoliths liquid. Contributions to Mineralogy and Petrology 94, 42–45. from Kilombe volcano, Kenya Rift Valley: evidence for subvolcanic interaction with Ghiorso, M.S., Sack, R.O., 1995. Chemical mass transfer in magmatic processes IV. A re- carbonitic fluids. Lithos 91, 373–392. vised and internally consistent thermodynamic model for the interpolation and Rogers, N.W., Evans, P.J., Blake, S., Scott, S.C., Hawkesworth, C.J., 2004. Rates and time- extrapolation of liquid–solid equilibria in magmatic systems at elevated tempera- scales of fractional crystallization from 238U-230Th-226Ra disequilibria in tra- tures and pressures. Contributions to Mineralogy and Petrology 119, 197–212. chyte lavas from Longonot Volcano, Kenya. Journal of Petrology 45, 1747–1776. Harms, E., Gardner, J.E., Schmincke, H.-U., 2004. Phase equilibria of the lower Laacher Russell, J.K., Nicholls, J., 1988. Analysis of petrologic hypothesis with Pearce element ra- See tephra (East Eifel, Germany): constraints on pre-eruptive storage conditions tios. Contributions to Mineralogy and Petrology 99, 25–35. of a phonolitic magma reservoir. Journal of Volcanology and Geothermal Research Sack, R.O., Carmichael, I.S.E., Rivers, M., Ghiorso, M.S., 1980. Ferric-ferrous equilibria in 134, 135–148. natural silicate liquids at 1 bar. Contributions to Mineralogy and Petrology 75, Heumann, A., Davies, G.R., 2002. U-Th disequilibrium and Rb-Sr age constraint on the 369–376. magmatic evolution of peralkaline rhyolites from Kenya. Journal of Petrology 43, Scaillet, B., Macdonald, R., 2001. Phase relations of peralkaline silicic magmas and pet- 557–577. rogenetic implications. Journal of Petrology 42, 825–845. Johnson, R.W., 1969. Volcanic geology of Mount Suswa, Kenya. Philosophical Transac- Scaillet, B., Pichavant, M., Cioni, R., 2008. Upward migration of Vesuvius magma cham- tions of the Royal Society of London 265, 383–412. ber over the past 20,000 years. Nature 455, 216–220. Jutzeler, M., Schmincke, H.-U., Summita, M., 2010. The incrementally zoned Miocene Scott, S.C., Bailey, D.K., 1984. Coeruption of contrasting magmas and temporal varia- Ayagaures ignimbrite (Gran Canaria, Canary Islands). Journal of Volcanology and tions in magma chemistry at Longonot volcano, Central Kenya. Bulletin of Volca- Geothermal Research 196, 1–19. nology 47, 849–873. Le Bas, M.J., Le Maitre, R.W., Streckeisen, A., Zanettin, B., 1986. A chemical classification Scott, S.C., Skilling, I.P., 1999. The role of tephrachronology in recognizing synchronous of volcanic rocks based on the total alkali-silica diagram. Journal of Petrology 27, caldera-forming events at the Quaternary volcanoes Longonot and Suswa, south 745–750. Kenya Rift. The Geological Society of London, Special publications 161, 47–67. Leat, P.T., Macdonald, R., Smith, R.L., 1984. Geochemical evolution of the Menengai Sigmundsson, F., 2006. Magma does the splits. Nature 442, 251–252. Caldera Volcano, Kenya. Journal of Geophysical Research 89, 8571–8592. Skilling, I.P., 1988. The geological evolution of Suswa Volcano, Kenya. Doctoral disser- LeMasurier, W.E., Futa, K., Hole, M., Kawachi, Y., 2003. Polybaric evolution of phonolite, tation, University of Lancaster. trachyte, and rhyolite volcanoes in eastern Marie Byrd Land, Antarctica: controls Skilling, I.P., 1993. Incremental caldera collapse of Suswa volcano, Gregory rift Valley, on peralkalinity and silica saturation. International Geology Review 45, Kenya. Journal of the Geological Society of London 150, 885–896. 1055–1099. Smith, M., 1994. Stratigraphic and structural constraints on mechanisms of active rifting LeMasurier, W.E., Choi, S.H., Kawachi, Y., Mukasa, S.B., Rogers, N.W., 2011. Evolution of in the Gregory Rift, Kenya. Tectonophysics 236, 3–22. pantellerite-trachyte-phonolite volcanoes by fractional crystallization of basanite Smith, P.M., Asimow, P.D., 2005. Adiabat_1ph: A new public front-end to the MELTS, magma in a continental rift setting, Marie Byrd Land, Antarctica. Contributions to pMELTS, and pHMELTS models. Geochemistry Geophysics Geosystems 6 art. No. Mineralogy and Petrology. doi:10.1007/s00410-011-0646-z. Q02004. Macdonald, R., Baginski, B., 2009. The central Kenya peralkaline province: a unique Stanley, C.R., Russell, J.K., 1990. Matrix methods for the development of Pearce element assemblage of magmatic systems. Mineralogical Magazine 73, 1–16. ratio diagrams. In: Russell, J.K., Stanley, C.R. (Eds.), Theory and Application of Macdonald, R., Scaillet, B., 2006. The Central Kenya Peralkaline Province: insights into Pearce Element Ratios to Geochemical Data Analysis, Geological Association of the evolution of peralkaline salic magmas. Lithos 91, 59–73. Canada Short Course Notes, No. 8, pp. 131–156. Macdonald, R., Davies, G.R., Bliss, C.M., Leat, P.T., Bailey, D.K., Smith, R.L., 1987. Geo- Thompson, R.N., 1982. Magmatism of the British Tertiary volcanic province. Scottish chemistry of high-silica peralkaline rhyolites, Naivasha, Kenya Rift Valley. Journal Journal of Geology 18, 49–107. of Petrology 28, 979–1008. Thorarinsson, S.B., Holm, P.M., Duprat, H., Tegner, C., 2011. Silicic magmatism associated Macdonald, R., Kjarsgaard, B.A., Skilling, I.P., Davies, G.R., Hamilton, D.L., Black, S., 1993. with Late Cretaceous rifting in the Arctic Basin-petrogenesis of the Kap Kane se- Liquid immiscibility between trachyte and carbonate in ash flow tuffs from Kenya. quence, the Kap Washington Group volcanic, north Greenland. Lithos 125, 65–85. Contributions to Mineralogy and Petrology 114, 276–287. Velador, J.M., Omenda, P.A., Anthony, E.Y., 2003. An integrated mapping and remote Macdonald, R., Rogers, N.W., Fitton, J.G., Black, S., Smith, M., 2001. Plume-lithosphere sensing investigation of the structural control for fumarole location in the Eburru interactions in the generation of the basalts of the Kenya Rift, East Africa. Journal Volcanic Complex, Kenya Rift. Geothermal Resources Council Transactions 27, of Petrology 42, 877–900. 639–642. Macdonald, R., Bagiński, B., Belkin, H.E., Dzierżanowski, P., Jezak, L., 2008a. REE parti- Villemant, B., 1988. Trace element evolution in the Phlegrean fields (Central Italy): tioning between apatite and melt in a peralkaline volcanic suite, Kenya Rift Valley. fractional crystallization and selective enrichment. Contributions to Mineralogy Mineralogical Magazine 72, 1147–1161. and Petrology 98, 169–183. Macdonald, R., Belkin, H.E., Fiton, J.G., Rogers, N.W., Nejbert, K., Tindle, A.G., Marshall, Villemant, B., Jaffrezic, H., Joron, J.L., Treuil, M., 1981. Distribution coefficients of major A.S., 2008b. The roles of fractional crystallization, magma mixing, crystal mush and trace elements; fractional crystallization in the alkali basalt series of Chaine remobilization and volatile-melt interactions in the genesis of a young basalt- des Puys (Massif Central, France). Geochimica et Cosmochimica Acta 45, 1997–2016. Author's personal copy

104 J.C. White et al. / Lithos 152 (2012) 84–104

Weaver, S.D., Sceal, J.S.C., Gibson, I.L., 1972. Trace-element data relevant to the origin of White, J.C., Ren, M., Parker, D.F., 2005. Variation in mineralogy, temperature, and oxygen trachytic and pantelleritic lavas in the east African rift system. Contributions to fugacity in a suite of strongly peralkaline lavas and tuffs, Pantelleria, Italy. The Mineralogy and Petrology 36, 181–194. Canadian Mineralogist 43, 1331–1347. White, J.C., 2003. Trace-element partitioning between alkali feldspar and peralkalic Wilson, M., Downes, H., Cebria, J.M., 1995. contrasting fractionation trends in coexisting quartz trachyte to rhyolite magma. Part II: Empirical equations for calculating continental alkaline magma series; Cantal, Massif Central, France. Journal of Petrology trace-element partition coefﬁcients of large-ion lithophile, high ﬁeld-strength, 36, 1729–1753. and rare-earth elements. American Mineralogist 88, 330–337. Wood, B.J., Blundy, J.D., 1997. A predictive model for rare earth element partitioning White, J.C., Urbanczyk, K.M., 2001. Origins of a silica-oversaturated quartz trachyte- between clinopyroxene and anhydrous silicate melt. Contributions to Mineralogy ryolite suite through combined crustal melting, magma mixing, and fractional and Petrology 129, 166–181. crystallization: the Leyva Canyon volcano, Trans-Pecos Magmatic Province, Texas. Wright, T.J., Ebinger, C., Biggs, J., Ayele, A., Yirgu, G., Keir, D., Stork, A., 2006. Magma- Journal of Volcanology and Geothermal Research 111, 155–182. maintained rift segmentation at continental rupture in the 2005 Afar dyking White, J.C., Holt, G.S., Parker, D.F., Ren, M., 2003. Trace-element partitioning between episode. Nature 442, 291–294. alkali feldspar and peralkalic quartz trachyte to rhyolite magma. Part I: Systematics Yirgu, G., Ayele, A., Ayalew, D., 2006. Recent seismovolcanic crisis in Northern Afar, of trace-element partitioning. American Mineralogist 88, 316–329. Ethiopia. Eos Transactions AGU 87 (33), 325–336.