Outline for our third lecture

1
Outline for our third lecture

• Motivation To evaluate proxy-reconstructed
  changes within a dynamical framework to advance
  our understanding of climate dynamics.
• Choosing a model
• The default model for climate variability from
  the instrumental (review)
• Exotic Models/Dynamics possible scenarios.
• 5) An examination of the climate beasts in the
  proxy record using this framework a sampler and
  the rest of this course

2
1. Motivation

• Vision To evaluate proxy-reconstructed changes
  within a dynamical framework.
• Goal Establish framework for when changes in
  circulation matters WHICH MEANS when do we have
  to invoke a change (style or magnitude) in the
  basic circulation regimes, in order to reconcile
  dynamics with a particular proxy?
• ? Can we use current dynamics/understanding of
  dynamics to explain a proxy? Or
• ? Do the proxies compel us to invoke a change in
  circulation (style or magnitude).
• Both improve our understanding of climate
  history Only the latter advances the
  understanding of climate dynamics.

3
2. Choosing a modelSo you have your
instrumental or proxy record. How do you
understand what it is trying to tell you?
A recipe for choosing a model

• Choose a default model, show the data is
  inconsistent with it.
• Issues to think about Change in mean vs. change
  in variability tricky thing with proxies tend
  to reflect extremes, which doesnt help!
• Identify/dream-up scenarios consistent with data
• - Identify sign and amplitude
• Test with a) models b) data. Can this be done??
• Are the data and/or models adequate to truly
  falsify a hypothesis?

4
3. The default model for climate variability A
review of lectures 1 2

• In the midlatitudes, climate variability is due
  to storm-track dynamics that are intrinsic to the
  atmosphere
• The patterns of climate variability (the modes)
  are colocated with the jets, which are determined
  by gross boundary conditions (e.g., rotation
  rate, orography, gross land-sea contrasts, etc).
• The dynamical time scale associated with these
  modes is 5 days, and is set by the interactions
  between the storms and the jets that form them.

5
Where does variability come from?
Storms owe their existence to the jets. Over
their life cycle, storms both take and give
momentum to the jet
Storm Decay
Storm Growth
A Pattern is the expression the changes in the
jets due to storms.
6
Examples from the modern climate
The NAO
The PNA
The PDO
0 0.1 0.2 0.3 0.4 0.5
Frequency (/yr)
Frequency (/yr)
7
Where does the memory come from?

• The redness of spectrum (presence of enhanced
  decadal variability is guaranteed because of the
  interaction with the thermodynamic ocean (time
  scale 6-12 months.
• 2) The amplitude of the spectrum and the spatial
  pattern can be nudged by changes in geometry
• e.g., ice sheets, continent configuration,
  orography, sea ice, sea level, insolation
  gradient/CO2.
• 3) On decadal and longer time scales, the memory
  for reddening may come from other physics e.g.,
• 10 years gyre circulations, arctic sea ice,
• 100 years intermediate water masses ice streams
• 1000 years ice sheets, deep oceans

8
The ENSO Mode
The modes of the coupled atmosphere-ocean system
are found by linearizing the atmosphere and ocean
equations about the observed mean state where
is the vector of state variables (SST, 3-D
flow fields, thermocline depth, etc) and B is the
cyclostationary (due to the mean fields)
dynamical matrix. Solving (1), we
have (2) where the propagator matrix R is
defined using data. Hence, (3) Thus,
the yearly propagator is . (4)
Thompson and Battisti 2000
9
The ENSO Mode (cont)

• The eigen (Floquet) analysis of Ryear are the
  true thermo-dynamic modes of the coupled
  atmosphere system.
• The eigenvalues of Ryear yield the frequency and
  growth rate of the modes, with structures given
  by the eigenvectors.
• The leading (least damped) mode is the ENSO mode,
  so ENSO is a true model of the coupled
  atmosphere/ocean system
• . The ENSO mode
• Has a rich frequency structure, but a strong
  spectral peak at 3-4 years, and e-folding decay
  time of two years
• (cf, the leading patterns of stormtrack
  variability with intrinsic time scales of
  storm/mean flow interaction (5 days) and
  reddening due to thermodynamic interaction with
  the ocean (6-12 months)
• ENSO requires a mean state with a strong
  east-west temperature gradient.

Spectrum of the pure ENSO mode
10
The ENSO mode and Observations

• The ENSO mode shares many features of the
  observed (composite) ENSO cycle
• Bjerknes local atmosphere/ocean feedbacks in
  equatorial zone.
• A buildup of warm water in the equatorial wave
  guide that precedes the warming in the eastern
  Pacific Ocean (ie, the warm phase of ENSO)
• The spatio-temporal evolution of ocean heat
  content (via adiabatic dynamics) throughout the
  onset, growth and decay of a warm event.
• e.g., Kessler 1990 Bigg and Blundell 1992
  Mantua and Battisti 1994 Wakata and Sarachik
  1991 Rosati et al. 1995 Schneider et al. 1995.
• The ENSO mode is also consistent with the
  following observations
• ENSO is peaks at the end of the calendar year
  because of the annual cycle in the background
  state of the coupled system.
• El Nino events (warm phase of ENSO) last about
  one year.
• The Indian Ocean does not support ENSO
  variability (homogeneous basic state).

11
ENSO as a linear, stochastic system
1st EOF from Model
1st EOF from Data
Observed
Model
Results from a linear coupled atmosphere-ocean
model forced by white noise -- reproduces
essentially all the robust observations
12
4. Exotic Models/Dynamics

• Impulsive forcing (volcano, ice sheet
  disintegrating (Heinrich melt-water pulse)
• echo of forcing around the globe, but no change
  in dynamical regime, or new dynamic regime?
• new dynamics/circulation regime
• A switch between two different dynamical states
  (i.e., stadial/interstadial)?????
• stochastic resonance.
• Internal dynamical cycles
• ENSO, Quasi-Biennial Oscillation (a tropical
  stratospheric phenomenon w/ energy at 2-3 years),
  etc
• Intraseasonal Oscillation or Madden Julian
  Oscillation
• influence on hurricane tracks and Indian
  monsoonal rainfall
• Bond cycles??
• Slip-stick cycles in ice sheets
• Exotic physics

13
5. The climate beasts in the proxy record do
they justify an exotic dynamical model?

• Tropical Atlantic Variability (TAV)
• Atlantic Multidecadal Oscillation (AMO)
• Little Ice Age
• Millennial variability in the last glacial period
• Antarctica
• Greenland Dansaard/Oescheger Events
• Bipolar See-saw
• Heinrich Events and 8.2 kyr event
• Bond Cycles
• Milankovitch (ice age cycles)

14
Ice core d18O records
Antarctica
Greenland
0ka
90ka
Time

• Best dated, (and cross dated) records available
• d18O a more direct reflection of climate than
  other proxies

15
Case for the prosecution I The clock
Rahmstorf (2003)

• Abrupt events happen paced by a 1500-yr cycle.
• GISP2 record suggests a 23 cycle phase memory!

16
-80kyr
-20kyr
Time
17
Byrdindistinguishable from red noise
18
(No Transcript)
19
(No Transcript)
20
The GISP2 record is clearly different from the
default picture of climate variability.
21
Assymetry in GISP2 cannot be explained by the
default picture
22
(No Transcript)
23
(No Transcript)
24
Is there a Relationship between Antarctica and
Greenland?
One possibility
Byrd lagging behind -1 x GISP2 by 400 years
Another possibility
Byrd leading -? GISP2 dt by 200 years
25
Cross spectrum between GISP2 and Byrd - frequency
by frequency what is the correlation between the
records and it is significant?
0.5
1.0
0
0.1
Frequency ?
(2000 yrs)
(1000 years)
(10,000 yrs)
Large, long excursions are clearly connected
between hemispheres, It is not clear that that
abrupt D/O events are
26
Low
jet
Storm
High
Storms owe their existence to the jets. Over the
life cycle of the storm they both take and give
momentum to the jet
Low
Low
The Patterns are expression the changes in the
jets due to storms
High
High