Week 4 – Lava Geochronology (focusing on the shorter timescales)
Ultra high resolution methods for submarine eruption verification and within-eruption emplacement histories (weeks to months)
Methods for determining eruption frequency, eruption sequences and repose history over multiple eruption cycles (decades to millennia)
Observations of Deep Submarine eruptions Eruption Detection . Direct observation . Visual observation soon after the fact (happenstance) . seismic monitoring . T-phase = sound waves . Instruments stuck in a lava flow . water column signatures (particles, heat, gasses) Event Verification . how do I know if lava or tephra were emplaced on the seafloor?: take a picture, or collect a sample and 210Po date it
Microbial mats and charred tubeworms at 9° 50’N in Apr-May 1991 new and old lava on the N. AlvinGorda Photos Ridge, WHOI
GG 711, Fall 2011, Lect. 4 1 Eruption Age . 210Po- 210Pb dating (weeks to yrs resolution) . 210Pb- 226Ra dating (multiple yrs to decades) . 226Ra-230Th dating (multiple centuries resolution) . paleomagnetic intensity (decades to centuries resolution)
Preparing a sample for 210Po- 210Pb analysis in the SOEST Isotope lab
The list of known historical eruptions with (a) high resolution date and (b) identified lavas on the deep sea floor is small (most examples are in the table below) Event detection dating resolution Some references JdF Cleft Mounds 1985 Megaplumes Differential seabeam ± 2 years* Chadwick & Embley, 1991 17°S EPR 1990 serendipity Multiple (3He/Heat) ± 2 years* Auzende et al., 1996; Sinton et al., 2002
9°50’N EPR 1991-92 serendipity 210Po ± 2 mo. Haymon et al., 1993; Rubin et al., 1994
JdF CoAxial 1993 seismic seismic ± week* Dziak et al., 1995; Embley et al., 2000
N. Gorda 1996 seismic 210Po/seismic ± 2 mo. Chadwick et al., 1998; Rubin et al., 1998
Loihi Smt 1996 seismic 210Po ± 2 mo¶ Loihi Science Team, 1997 Axial Smt 1998 seismic Seismic, records from Days Chadwick et al., 1999; Dziak et al., 1999 stuck instruments 10°45’N EPR 2003 Serendipity 210Po ± 1 mo. McClain et al., 2004; van der Zander et al., 2004 NW Rota SMT 2004-on serendipity 210Po, direct observation ± 1 mo. Embley et al., 2005
9°50’N EPR 2005-06 serendipity 210Po, seismic (OBS) ± 2 mo. Tolstoy et al., 2006; Soule et al., 2007
210 NELSC 2008-9 Megaplumes Po, H2-heat tbd Rubin et al., in prep, Baker et al., in prep
W. Mata 2008-09 Megaplumes Direct observation, Days to Resing et al., in review 210 Po, H2-heat weeks Axial Smt 2011 serendipity 210Po, seismic (both tbd stay tuned underway) * fresh lava observed but no definite link to the event is not established ¶ seismicity occurred 3 to 4 months after lavas were erupted
GG 711, Fall 2011, Lect. 4 2 The first time someone made unexpected visual observation of a deep submarine eruption soon after it happened: 9 50N EPR in April 1991 Rachel Haymon, Chief scientist
Soon after the 1991-2 N-EPR “tubeworm BBQ” eruption
Another time someone made unexpected visual observation of a deep submarine eruption soon after it happened: 10° 45’N, the FIELD expedition (Nov 2003) Janet Voight, Chief scientist
Bacterial “Snowblower” vents
Lots of diffuse flow venting Bacterial mat on glassy flow surface
GG 711, Fall 2011, Lect. 4 3 10° 45’N, 210Po dated to 2003
9° 50’N, 210Po dated to 1991-92; second eruption dated to 2005-6
Axial
Co- Axial
GG 711, Fall 2011, Lect. 4 4 1996 Gorda and 1993 CoAxial Seismicity Compared (same website source) The general nature of the seismicity appears very similar to that observed from the CoAxial dike injection and eruption of June/July 1993. A rapid series of small earthquakes is observed without a large "foreshock". The following histogram displays the number of recorded events per hour. This level of activity is comparable to the level recorded from CoAxial segment in 1993. The apparent decline in activity midday on Julian Day 62 until late on Julian day 65 is likely due to the loss of the closest array. Data are grouped by days.
Loihi 1996 epicenters
GG 711, Fall 2011, Lect. 4 5 Loihi 1996
General comments Dating Techniques
What is the timescale for tracer change ? How is clock set ? Do all components "close" simultaneously? How big is the initial signature ? How accurate is analysis procedure ?
The system MUST remain closed.
GG 711, Fall 2011, Lect. 4 6 General comments
Resolution for a chronometer is defined as T/T. The more precisely we can determine time increments the higher the resolution Memory Resolution
Weeks
Typical See a doctor
days
hours Great
parts of hours
Last year a month ago yesterday this morning
General comments Resolution of 230Th-234U-238U radiometric dating of corals using high precision mass spectrometry
Different application but the shape of the error envelope is the same
t½ is 75ka. Notice that best ages are within ca 0.1x to 10x of the half life. Resolution is NOT the same thing as error.
Resolution depends on measurement precision but absolute errors can shift age dates and their resolution bands up or down
GG 711, Fall 2011, Lect. 4 7 Actinide decay chains 238U decaying to 206Pb There are 3 naturally occurring actinide isotope half life = 4.46 x 109 yrs radioactive decay chains that we exploit in Geochronology
235U decaying to 207Pb half life = 0.7 x 109 yrs
232Th decaying to 208Pb half life = 14.1 x 109 yrs
Age of Earth = 4.55 x 109 yrs
Systematics 238U → → → 234U → 230Th → 226Ra → 222Rn → → → 210Pb → → 210Po
t½: 75200 yr 1600 yr 3.8 day 22 yrs 138.4 days
All are incompatible in silicate minerals found in basalts and ultra mafics Rn is a noble gas Pb is moderately chalcophile Po is volatile above ca 100 C U and Ra aqueous solubility >> Th
- 238 2 t - 2 t The U series A 2 = A 2 e + A 1(1-e ) Excess o water soluble
Not water b cd soluble Excess o Gas 2 I 2 (A /A )-1 /A (A
a
Equilibrium volatile
0 100 10000 1000000 Time Since Fractionation (years)
a ( 210 Po/ 210 Pb) b ( 210 Pb/ 226 Ra) cd( 226 Ra/ 230 Th) ( 230 Th/ 238 U)
GG 711, Fall 2011, Lect. 4 8 6
5
4
3
2 Log years
1
0
( 230 Th/ 238 U) ( 226 Ra/ 230 Th) ( 228 Ra/ 232 Th) -1 ( 231 Pa/ 235 U) ( 210 Pb/ 226 Ra) ( 210 Po/ 210 Pb) Tracer
Surface Flow Chronology
First Cro-Magnons Helen abducted by Trojans Ice Man takes his last alpine hike Atilla the Hun invades Rome ridge axis Pele arrives at Kilauea sea floor Spanish Inquisition begins Kamehameha unifies Hawai'i Millard Filmore elected president Radium discovered by the Curies Elvis joins the army Disco Obama elected president Mantle & Crustal Processes mantle 100 years ? -X 0 X 10,000 years ? distance a ( 210 Po/ 210 Pb) b ( 210 Pb/ 226 Ra)
226 230 230 238 cd( Ra/ Th) ( Th/ U)
GG 711, Fall 2011, Lect. 4 9 210 Po-210 Bi- 210 Pb Systematics 210Po-210Pb dating t ½ =138.4 days clock set on eruption ( 210 Po/ 210 Pb) > 100
(T = T ) 210 210 sf ( Bi/ Pb) ~ 20 analysis: counting 210 Po, 210 Bi
( 210 Po/ 210 Pb) ~ 0
210 Pb 210Pb
Direct Degassing Behavior of Metals During Volcanism Subaerial Shallow Submarine Deep Submarine (extensive literature) (limited literature) (only Po data)
Po ~100% degassed ~100% degassed ~100% degassed
Pb ~1-2% degassed ~1-2% degassed ?
Bi ~20% degassed ? ?
Macdonald Seamount – Rubin and Macdougall, 1989 Single Sample Evolution 1.8
1.6 210 = 210 Po f Pb 1.4
1.2 210 Po Activity 1.0 (dpm/g) 0.8 clock set on eruption 0.6
0.4
0.2 210 = 2 error Po i 0 0 200 400 600 800 1000 Days Since Eruption (October 22, 1987)
Multiple analyses of 210Po with time in a single sample
GG 711, Fall 2011, Lect. 4 10 1.2 0.14 Time Since Zero-point: secular equilibrium 210 1.0 0.12 ( Pb) 5 4 1.9 yrs 0.1 0.8 3 1.5 yrs "zero-point" of 0.08 2 1.1 yrs ingrowth curve 0.6 210
( Po) 0.06 0.8 yrs 0.4 1 0.04 210 210
( Po/ Pb) 0.4 yrs 0.2
0.02 ingrowth t sample collection date 1/2 0 0 half lives 0 500 1000 0 500 1000
Time (days)
Time Since Eruption:
1.00 5 690 days 1.9 yrs 4
3 552 days 1.5 yrs 0.75 2 414 days 1.1 yrs
( 210 Po/ 210 Pb) 0.50 276 days 0.8 yrs 1
0.25 138 days 0.4 yrs
0 days 0
GG 711, Fall 2011, Lect. 4 11 eruption window collection date
( 210 Po) 100 % 75 % degassing degassing maximum age error in maximum age Method details • Po degasses upon eruption
• “grows in” to grandparent 210Pb
• time-series 210Po analyses in a lava
• regression to ingrowth curve
• “maximum” age is intercept
• Conservative minimum age is lower of sample collection date or 210Po = 0.25 * 210Pb
Sources of Error
0.15 a. Extent of Intial Degassing b. Regression Intercept Projection
0.10 asy mp t ot e: 210 21 0 ( Po t)=( Pb)
0.05 210
( Po) dpm/g maximum age (100% Po d egassin g ) maximum age mi nim um age minimum age (75 % Po de gass ing ) 0
0.15 c. Time between eruption and d. Chemical Analysis Error 1st Po analysis resolution loss
0.10
0.05 210 ( dpm/g Po) max imum age maximum age minimum age m i nimum age
-250 0500100005001000 Time (days)
GG 711, Fall 2011, Lect. 4 12 a. 9° 50'N EPR eruptions Dead tube worms & young lava (Alvin photo)
30 De c 25 F eb 2372-1 ingrowth curve 2392-9 1 24 Jan 22 Mar 2 Nov 29 De c ( Po/ Pb) Po/ (
2363-7 2504-1 210 210 repeated 25 Feb 23 April 7 J an 5 Mar sample 2372-1 2497-1b analyses 0.5 Time sample collection date J F M A M J J A S ONDJ F MA (months) (210 Po) = 0.25 x ( 210 Pb) 1991 1992 (year) 0 0 200 400 (days) 500 days 1000 days event discovery cruise R/V AII subsequent cruise R/V AII subsequ ent cruise R/V AII
seismic event
b. GORDA Ridge earthquakes d. LOIHI Seamount eruption P286-1
W9604-C3
P286-6 ATV-178-7R New and old pillows (TowCam photo) ATV-178-1R1 50 eruption windows defined 40 30 event response Time collection date 20 cruise (R/V KOK) growth 10 JFMAMJJASOND (months) 210 Po) in ( earthquakes 0 1996 (year) 100% 75% degassing degassing maximum age Time: 0100 200 300 (days) JFMAMJJASOND (months) regression error in maximum age event response cruise Jason samples 1996 (year) R/V Wecoma ATV samples 0100 200 300 (days)
c. 10° 44’N EPR Bacterial mat on a lobate lava flow erupted eruptions young lava in 2003 at 10 44N EPR (Alvin photo) 3936-R2 Sample collection dates 3935-R2 PRELIMINARY DATA 3937-R3 3935-R1 Time JFMAMJJASOND (months) 2003 (year) 0200100 300 (days) even t di scovery cruise R/V Atlantis - FIELD expedition
Lau Basin – Northeast Lau Spreading Center, Nov 2008
Preliminary 210Po eruption ages: Oct. to Dec. 2008
Age in May 2009: 5 mo 6 mo 7 mo
GG 711, Fall 2011, Lect. 4 13 see V51D-1720 - Michael et al. for major element chemistry 7.40
7.20
7.00
6.80
6.60 MgO content 6.40
6.20
6.00 Large range in 2-Oct-08 1-Nov-08 1-Dec-08magma 31-Dec-08chemistry preliminary eruption age Eruption appears to have evolved chemically with time
MgO content: 6.1-6.4% 7.1-7.3%
2005-2006 EPR - 210Po Age Distribution of lava (18!) samples and a possible eruption sequence >75%-85% of lava emplaced Group Mean ages: between late 28-Jun-05 ± 18 days 15-Oct-05 ± 21 days 20-Aug-05 ± 5 days 25-Jan-06 ± 60 days June and Early Sept 2005
Lava flow map modified After Soule et al., 2007
Remember, this was the eruption June-July 2005 with the trapped seismometers Aug-Sept 2005 Oct-Nov 2005 Jan 2006
Rubin et al., in prep
GG 711, Fall 2011, Lect. 4 14 Coincidence of 210Po eruption ages with earthquakes measured in situ by ocean-bottom seismometers
10-day low-pass
The data suggest that the Po-ages are better resolved than the our conservative estimate of 1 month resolution based on analysis and regression errors. Rubin et al., in prep
Coincidence of vent fluid temperature time- M-vent high T logger before and after the eruption.
series data and eruption ages as deployed, Mar. 28, 2004 High-T probes were deployed in 3 hydrothermal Chimneys with at temperature recorded at 36 minute intervals Von Damm, Fornari, Bryce, et al., in prep.
>1m 210Po age groups 1 2 3 4
~1 m thick new lava flow surrounds and buries the base of this now inactive chimney
June 2006 (after eruption)
GG 711, Fall 2011, Lect. 4 15 Other dating methods
These are for determining longer-term volcanic histories at a site, such as dating eruption intervals and volcanic repose.
• Geophysical methods, such as paelomagnetic intensity • Radiometric methods, such as Pb-Ra and Ra-Th geochronology
Paleomagnetic field intensity dating (aka “paleointensity dating”) Archeomagnetic data model
• Different than paleomagnetic dating, which is based on field reversals. • This method is based on secular
variation in the intensity of the field intensity field • Clock is set when lava cools through the curie temperature for the magnetic mineral(s) in the rock
• Fairly simple variation over past 1 stacked sedimentary records ka, more complicated before that • Need local field curve for dating at a site field intensity field intensity
Bowles et al., 2006, G-cubed (see background reading if interested)
GG 711, Fall 2011, Lect. 4 16 9-10° N EPR
• Paleointensity is indicated by color. • Circles represent samples collected via Alvin submersible • squares represent samples collected via rock core. • Size of symbol corresponds to uncertainty in the site mean.
1991 field was 35.8 μT Bowles et al., 2006, G-cubed
Notice the asymmetry in young volcanism across axis at these two sites
Paleointensity versus distance from axis for samples falling inside the sample regions at 9°31′N ≥~150 ybp?? and 9°50′N
Data are shown only for modern samples for which 1σ < 3 μT and N > 1
GG 711, Fall 2011, Lect. 4 17 Slightly longer-lived radiometric methods clocks NOT set on eruption 226 Ra dating (basalts)
210 Pb dating (basalts) clocks set after melt is isolated from matrix (?)
(T = T tr + T mc + T sf ) analysis: counting mass spec.
magma chamber residence
( 226 Ra/ 230 Th) decrease or no change ( 210 Pb/ 226 Ra) decrease melt transport ( 226 Ra/ 230 Th) > 1 ( 210 Pb/ 226 Ra) < 1 melting Before melting: ( 226 Ra/ 230 Th) = 1 Depleted residue ( 210 Pb/ 226 Ra) = 1
(210Pb/ 226Ra) disequilibria in OIB and MORB
210Pb deficits
210Pb excesses
Berlo and Turner (2010) EPSL
GG 711, Fall 2011, Lect. 4 18 210Pb-226Ra - Rubin et al., Nature, 2005 (226Ra/ 230Th) 132 210Pb-226Ra disequlibria last ~120 yrs 226Ra-230Th disequlibria last ~8000 yrs Axial R2 = 0.6 1.1 i a) R 226
/ 1.0 n -M
b O Loihi RB P decay 210 ( 0.9 210Pb- 226 Ra secular equilibrium
N -M O RB Deficits Deficits Excesses
Larger disequilibria N-MORB (210Pb/ 226Ra) and (226Ra/230Th)
are correlated for Pacific ridges 226 230 Ra-Th secular equilibrium Different magnitude (210Pb/ 226Ra) and (226Ra/230Th) variations rule out primary 226Ra addition to an equilibrium system.
210Pb deficits in MORB are made in the mantle 210 226 Pb- Ra secular equilibrium Data for individual eruptions ts Excesses i c i ef D (226Ra/ 230Th) Mg# Larger disequilibria 132 50 60 70
2 Axial R2 = 0.6 Axial R = 0.8 1.1 1.1 i i )
a JDFR Ra) R
6 SEPR 22 226 1.0 n 1.0 NEPR -M
b/ O Loihi RB Pb/ P n -M 210
210 O
( R ( B 0.9 0.9
3.0 3.0 R2 = 0.8 )
h 2.5 2.5 Th) T 230 230 / Axial Individual eruptions a 2.0 2.0 Ra/ R 22 6 226 ( 1.5 ( 1.5 Loihi Loihi 1.0 1.0 50 60 70 (230Th/ 280 U) Mg# Rubin et al. (2005) Nature
GG 711, Fall 2011, Lect. 4 19 210Pb- 226Ra disequilibria get smaller by decay as differentiation progresses (
1.1 1.1 210 210 226
Pb/ N-MORB ( Pb/ Ra) Ra)
226 initial ratios correlate 226 1.0 R 2= 0.9 R 2= 0.8 1.0 strongly with compatible a ( Ra) Pb/ 0 element ratios such that 21 ( 0.9 0.9 210Pb deficit decreases with Ni/V ol ± cpx and plag removal 0.1 0.2 0.3 0.4 0.5 0.5 1.0 1.5 co-variations with
1.1 1.1 210 compatible elements do not Pb/ Ra) Ni/Co follow their sulfide 226 1.0 R 2= 0.9 1.0 226 compatibility sequence Ra) Pb/ (Ni>Cu>Co>V), from which 2 210
( R = 0.8 0.9 0.9 greatest deficits would Ni/Cu Th/U occur at lowest Ni/X ratios.
01232.0 2.5 3.0 3.5 6 There is no correlation with 1.1 Th/U (a source “enrichment” 5
Ra) indicator) Sr/Y 2 226 1.0 R = 0.8 4 Sr/Y vs Ni/V correlation Pb/
210 3 indicates plag involvement ( 0.9 n-MORB in differentiation R 2= 0.6 Sr/Y Ni/V 2
Rubin et al. (2005) Nature
Example melting + mixing scenario:
Distribution coefficients are favourable for making 210Pb- 226Ra-230Th in both spinel and garnet stability fields for a wide decay range of model conditions.
One has 2-3 half lives to a) transport magma and b) reside it in the crust prior to eruption to preserve the observed average signatures. decay But what about the end members? There is a trade off in composition vs time for the extremes
Rubin et al. (2005) Nature
GG 711, Fall 2011, Lect. 4 20 210Pb-226Ra Pb/Ra age ranges of magmas in compound lava flows
(assumes initial magma ratio was equal and the difference between samples represents a maximum allowable time spread of magma residence time in the crust)
JDFR N Cleft sheet flow Does magma generally 11586 days take 2 to 3 decades to 32 yrs accumulate between Axial Smt 1998 eruptions at 4986 days intermediate to fast 14 yrs spreading rate ridges? Tubeworm bbq 9 50N EPR 1991-92 6391 days 18 yrs An alternate explanation is that all the aldo-kihi (SEPR) magmas are formed from variable mixing 7208 days proportions of an old magma and a new on, 20 yrs so that there is no “age” information per se.
9° 50N EPR Comparing Pb-Ra disequilibria in successive eruptions at the same site Short-lived U-series signatures are generally consistent with closed-system decay of 1991-92 magmas to produce 2005-06 lavas
So not every MOR eruption requires new magma input from the mantle - or are both these eruptions part of the same magma input cycle to the ridge?
Weird rock up Weird rock up north - erupted north - erupted (off axis) (off axis)
With an estimate of magma chamber volume change over this time interval, we can estimate magmatic heat flux and thus heat source for the local hydrothermal system
GG 711, Fall 2011, Lect. 4 21 Another Volcanic stratigraphy over the similar example: least 400 yrs
Mapped young Lavas at 17° S on the N2 superfast spreading SEPR N1
Bergmanis, et al., 2007
N1 lava
Lower MgO (cooler)
higher MgO (hotter)
Rubin et al., (2009) Nat. Geosc. With data from Bergmanis, Sinton, Rubin 2007
GG 711, Fall 2011, Lect. 4 22 Two magma compositions in N2 N1 N1 and N2 lavas
• indistinguishable 210Pb-226Ra and overlapping Pb isotopic characteristics • both probably represent mixes of the same magma batches in the crust:
• one hot, younger and more enriched • the other cold, older and less enriched
Bergmanis, Sinton, Rubin 2007
GG 711, Fall 2011, Lect. 4 23