The Processes and Timescales That Produce Zoning and Homogeneity in Magmatic Systems

1
The Processes and Timescales That Produce Zoning
and Homogeneity in Magmatic Systems

• George Bergantz, Olivier Bachmann and Philipp
  Ruprecht
• University of Washington

2
How to Link Observations Across Scales?

• How to expand our toolbox for magma forensics?
• What are the dynamic templates that produce large
  scales?
• How are they reflected at the crystal scale?

3
Three types of zoning patterns thatcommonly
occur in ignimbrites
4
Mechanisms to produce compositional gaps and
gradients
5
Gradients in ignimbrites (See Table 1 in text)
Type of gradient Abrupt Linear (monotonic) Not measurable
Archtypal Examples Crater Lake, Aniakchak, Toconao-Atana, Katmai (Payne et al., V21C-2122), Chaitén (Lowenstern et al., V43D-2180) Bishop Tuff, Huckleberry Ridge Tuff, Bandelier Tuff Monotonous Dacites (Fish Canyon Tuff, Lund Tuff, Cerro Galan) Rhyolites (Taupo)
6
Compositional Gap (Daly Gap)

• Fig. 2 from paper

7
CF-induced Daly Gap
Same P-T, isotopic ratios Trace element
concentration crystal fractionation Interstitia
l melt in mafic (crystal-rich) end-member
compositionally similar to silicic end-member
(Crustal melting unlikely)
8
Interstitial melt expulsion from crystal-rich
mushes

1. Crystal-melt separation time within longevity of
   magma chambers
2. Melt expulsion enhancers (gas-driven
   filter-pressing, earthquake fluidization)

9
Gradients in ignimbrites (See Table 1 in text)
Type of gradient Abrupt Linear (monotonic) Not measurable
Archtypal Examples Crater Lake, Aniakchak, Toconao-Atana, Katmai (Payne et al., V21C-2122), Chaitén (Lowenstern et al., V43D-2180) Bishop Tuff, Huckleberry Ridge Tuff, Bandelier Tuff Monotonous Dacites (Fish Canyon Tuff, Lund Tuff, Cerro Galan) Rhyolites (Taupo)
10
(Hildreth and Wilson, 2007)
11
Gradients require mixing- what do we need?
Stretching Folding Circulation (many scales of
strain)
Mixing requires a 1) a magma chamber 2)
paddle, thermal plumes, crystal plumes, bubble
plumes, compositional effects 3) an energy
source- some change in the environment to produce
kinetic energy
12
Well, What Dictates the Dynamic Template?

• The Reynolds number

• Most of us know that this number delimts three
  regimes
• Re ltlt 1, laminar flow, neglect inertia
• Re gt 104, fully turbulent, self-similar flow
  MIXING TRANSITION
• 104 gt Re gt1 chaotic advection, both inertia and
  viscosity important

13
Demonstrate dripping crystal plumesSee paper by
Bergantz and Ni, 1999 cited in chapter
14
Mixing Efficiency

• For system-wide mixing caused by vertical
  transport, e.g. some flavor of plume, Jellinek
  and others proposed the concept of mixing
  efficiency.
• BUT be very careful about this concept- it is
  really a measure of STRATIFICATION

15
Bringing together types of zoning into a common
framework

• Formation of a cap by escape from sill-like mush
  (instead of from the walls)
• Unzoned cap

What happens in the cap? Top cooling and
assimilation Bottom T-buffered mush
below Convection in cap but weak, low-Reynolds
number
16
Gaps and zoning- no big deal after all!
17
Processes that Produce Complexity in a Crystal
Cargo

• Mixing
• In-situ hyper-solidus recycling dynamic mush
• Concurrent melting, assimilation and deformation

What are links to the dynamic templates?
18
Simulations of gas driven overturn with smart
crystals

• Movies from
• Modeling of gas-driven magmatic overturn
  Tracking of phenocryst dispersal and gathering
  during magma mixing Ruprecht, Bergantz and
  Dufek, G3, v. 9, no. 7, 2008

19
Conclusions from simulations

• For 2x105 crystals report back
• A single overturn is sufficient to gather
  crystals onto a thin-scale from as much as a 100
  m initial separation. Continued choatic stirring
  can increase these distances, in accord with
  natural examples.

20
But what do crystals really remember?

• Depends on rate of travel through regions of
  distinct chemical potential vs. rate at which
  crystals can record to changes
• Damköler number
• If Da ltlt 1, kinetics dominate
• If Da gtgt 1, equilibrium assumption okay

21
Crystals as recorders of events in real-time

• For rapid, e.g., gas driven overturn, crystal
  growth will lag and only record an echo of the
  process (Da ltlt 1), but dissolution may reach Da
  1
• For slower processes rate-limited by heat
  transfer, both growth and dissolution will have
  Da 1 or more

22
Gradients in ignimbrites (See Table 1 in text)
Type of gradient Abrupt Linear (monotonic) Not measurable
Archtypal Examples Crater Lake, Aniakchak, Toconao-Atana, Katmai (Payne et al., V21C-2122), Chaitén (Lowenstern et al., V43D-2180) Bishop Tuff, Huckleberry Ridge Tuff, Bandelier Tuff Monotonous Dacites (Fish Canyon Tuff, Lund Tuff, Cerro Galan) Rhyolites (Taupo)
23
Homogeneity

• Mostly in large, crystal-rich magmas with
  intermediate (dacitic) composition (Monotonous
  Intermediates)
• Also true for large granodioritic batholiths
  (main upper crustal building block)
• How to reach homogeneity on large volumes of
  viscous crystal-rich magmas?
• Low Re convection inevitably leads to
  gradients????
• How to retain homogeneity on large volumes?
• New magma recharge will inevitably occur???

24
New mass injections limited to similar
compositions?

• Once a critical crystallinity is reached, silicic
  mushes act as density filter, buffer for T, C
• But crystals often very strongly zoned

25

• Spectacular small-scale disequilibrium in FCT, a
  homogeneous intermediate
• Reflects a long history of overturn

(Charlier et al., 2007)
26
Time scales have dual nature homogeneity at the
large scale, heterogeneity at the small scale

• Toba chem oscillations in allanites gt .4 M.y.
  before eruption cycling of crystals through
  hyper-solidus domains (Reid et al.)
• Bandelier Tuff reheating prior to eruption
  (Wolff et al.)
• Fish Canyon reverse mineral zoning, complex
  crystal compositions (Bachmann, Charlier et al.)
• Tuolumne Intrusive Suite complexly zoned
  zircons,
• Spirit Mtn., Mojave system complex rejuvenation
  of intrusive sheets, zoned zircon (Miller et al.)

27
Lengthscale-dependent mixing

• Some bulk mixing must occur
• Crystals record a changing environment- not just
  change in intensive variables
• Zoning patterns different in juxtaposed crystals
• Homogeneous at hand sample scale

28
Large silicic system are NOT just strips of
rhyolite- geophysical evidence Long Valley
Caldera. Very different from Mt. St. Helens.
New injections of basalt or intermediate magma
common
29
Unzipping

• Sluggish convection regime
• Gradients induced by crystal plumes,
  assimilation, mixing
• As system grows, assimilation and mixing become
  more transparent
• Lock-up from floor as crystal accumulation
  reaches 50 vol
• Cooling slows down (at least by a factor of 2)
• New magmas can not mix in gt Heat plate
• Unzipping

30
Thanks