Download Modeling of the 100kry cycle - Atmospheric and Oceanic Sciences

The 100 Kyr cycle
Agnes Barszcz
Who wants to be a
climatologist?
2 teams:
The Aphelions
 The Perihelions

Rules
Joker
What I am going to talk
about
Various hypotheses for 100 kyr cycle
Why Milankovich is wrong
Suggest a new hypothesis
See if it is reasonable
Look at it’s sensitivity to different parameters
Its flaws
What is the right theory
FUN: Giving out the price!!!
What is the 100 kyr
cycle?
What is causing it?
Many hypothesies:
The milankovitch cycle
 Isostatic adjustmets of the litosphere under
the weight of the glacier
 Feedback between atmosphere ocean and
Co2
 …..
 WE ARE STILL MISSING A SOLUTION!

Question 1:
Who can tell me in less then 2 minutes
what the Milankovich theory is ?
Milankovich and and
why it is wrong!
3 components
Eccentricity
 Precession
 Axial tilt

Which one was key for
us?
Milankovich and and
why it is wrong!
3 components: Eccentricity would
be the one we are interested in!
-> Top five reasons that we should
NOT……
Eccentricity changes are small
Orbital calculation when caried out with greater
presision show a major cycle of 400 Kyr
Well dated climate proxies show a 100 kyr cycle
only over the last million of years
Double peak in frequency domain
Causality problem
Spectrum of 100-kyr glacial cycle: Orbital inclination,not eccentricity. By RICHARD A. MULLER* AND
GORDON J. MACDONALD
http://www.pnas.org/cgi/reprint/94/16/8329.pdf , 1995
Suggested solution:
Changes in the orbital inclination
Find a suitable
solution…
Methods
Use simple or complex models
Pros and cons
Complex models: Global circulation
models
Pro: They take into account many
parameters. They are more realistic
 Cons: Require large computers $$$$$
 -> Used to simulate equilibrium responses
to various initial conditions

Pros and cons
Simple models:
Pro: Require less computer power, and run
faster..
 Con: Yield less realistic results. We do not
see the influence of all the small
parameters that we have neglected

What was used
A simple model
Claim
Hezi Gildor and Eli triperman say the
100 kyr cycle is NOT related to the
milankovich cycle!!!!!!!!
The answer!
Their hypothesis is:
The variation of the ice-albedo between
glacial and interglacial periods
 Variant of the precipitation-temperature
feedback

What the autors used
A simple zonaly averaged box
The components of the
model
Ocean meridional thermohaline
circulation
Atmospheric temperature-humidity
feedback
Land glaciers
Sea ice
What,where,how?
The ocean model

4 surface boxes (400 m)

2 polar:





Water may be covered with sea ice of variable extent
Land may be covered with land ice of variable extent
2 midlatitude boxes
4 Deep water boxes (4000 m)
** Important to note that the ablation rate of glacier
stays constant with time….
What,where,how?
The atmosphere model:
Each box can have 4 types of lower surfaces:
Land
Ocean
Land Ice
Sea ice
-> All have different albedos
The technicalities
What is a leapfrog scheme?
Leapfrog
Why leapfrog?
Time reversible
Assures energy conservation
A better accuracy
http://www.lifelong-learners.com/pde/SYL/s2node4.php
http://einstein.drexel.edu/courses/CompPhys/Integrators/leapfrog/
The crux of the problem
How it goes…
How it goes:
Ocean is ice free
Temperatures are mild
More precip then melting and carving
->->
Land ice sheet grows
How does that affect
the temperature?
As ice sheet grows
slowly
The albedo of the earth decrease
The sea temperature are below zero
only in the polar boxes.
->->-> At year 90 kyr, the global SST
reaches zero degrees
What’s special about the
moment that the SST
reaches zero?
Ice sheet rapidly grows
As the SST attained a critical value, sea
ice grows very rapidly!
In 20 year all polar box is covered in
sea ice.
->->-> Sea ice switch is ON
When the switch is on..
Sea Ice Stop growing!!!
Why????
Self-limitation of sea ice
The sea cools by giving out heat to the
atmosphere
When it is covered by ice, there is no
more exchange
The warm midlatitude waters mix with
the polar waters
->->-> No more sea ice growth
The ice age!
The glacier is at it’s maximum
The atmosphere is at it’s coldest
How do we get out of an ice age??
Moisture feedback
There is less moisture captured by cold
air, so less northward moisture transport
Because of the ice cap, there is also
less evaporation in the polar region
->->-> Less moisture present in polar
regions
We are loosing the
icesheet
The precipitaion rate is reduced by ½
The ablation rate stays constant
->->-> The glaciers retract
On the road to warm
times
As the ice sheet retract, the albedo is
decreasing.
Atmospheric and ocean temperatures rise
slowly.
As long as there are ice sheets in polar
region the ocean temperature in the region in
below zero :sea ice is present.
-> How does that affect the land ice sheet?
Further down, on the
road to warm times..
As long as the ice sheet is there, the
land glaciers retreat fast
The sea ice, does melt by sloooowly…

The meltdown is slow because the SSt is
close to zero..
-> What is causing it to melt anyways?
Close to interglaciation..
The sea ice melts down because of the
heat advected and diffused by the
ocean, that is coming from the
midlatitudes.
->What will cause the abrupt
acceleration of the melt down of sea
ice?
Deep Ocean
Because the deep ocean heats up:
But Why?
The deep ocean
Because of the melt down of the land
ice sheet
The switch is OFF
Switch is “off”
All the ice melts down in about 40 years
Atmosphere and ocean temperatures
rise again
And back…
The temperatures are maximum
The ice sheet is minimum
The amount of precipitation is at it’s
maximum
->We are back at the starting point!
The Results
Simulated years from 170 kyr to 70 kyr
Ok so we have a
scheme…
Theoretically it makes
sense…
Does it practically?
So what about the
THC?!?
THC = Thermohaline circulation
It was included in the model
It is rather an effect the a cause of the
100 kyr cycle
What changes the strength of the THC?
The THC strength
When glaciers melt, there is a lot of
fresh water relised: minimal THC (12Sv)
At interglacial periods, there is a lot of
evaporation: Salier water: strong
THC(16Sv)
When glaciers form: less evaporation:
weaker THC(13Sv)
Is it an ok resolution to
have only 4 boxes?
How many did we use in the model we
made in class??
 It is only the upper part of the ocean that
has to cool significantly for glaciation. (The
lower part’s role is to provide delayed
responses to various forcing)
 It was already demonstrated in previous
papers that this can be achieved in only a
few tens of years

Does it make sense to
have a 100 kyr period?
Let’s do some simple math!!!!!
Volume of land glaciers:
V= V max –V min
=2.4 e16 m3
M-Accumulation rate
maximum = 0.09 e6 m3/s
minumum = 0.03 e6 m3/s
S-Ablation rate
Sensitivity to different
parameters
Land–Ice sink term
Reduce by 4 %
Reaches faster the critical value at which
the ice starts growing
 Slower deglaciation

-> Cycle is longer by 10 kyr
-> Less saw-tooth like shape
Land–Ice sink term
Increase by 20 %

Sink term always exceeds the precipitation
rate
->>Glacier disappears!!!
Albedo
Increase both land ice and sea ice
albedo
What will happen??
Albedo
Increase both albedos
Shorter cycle
 Less land ice needed to reduce SST under
the threshold value
 Glaciers need to become smaller then
before to enable a temperature increase
that results in ice melting
-> Shorter growth/Longer meltdown
->Shorter but more symmetric cycle

Albedo
Very significant increase in sea ice
albedo
-> Permanent sea ice cover
->No land Ice
Albedo
Very significant decrease in Sea ice
albedo
-> Makes the sea-ice mechanism
less effective as a switch
Albedo
Increase in Land-ice albedo
->Reduces the amplitude of the oscillation and
shortens the time scale
Converse is true
Emissivity
The long wave emissivity represents:






Cloud cover
Humidity
Land cover
Topography
Aerosols
CO2
How will the emissivity evolve between a glacial
minimum and maximum?
Emissivity
The emissivity will increase between a
glacial minimum and a glacial maximum
because there is a smaller water vapor
concentration
How about the CO2?
Pelletier and Marshall:

P’=0.015*ln( CO2 / CO2 ref)
So for a 30% CO2 change

P changes by - 0.003
Induces an increase in the time scale of
the oscillation of about 9 kyr.
-> Why longer?
How about the CO2?
Why longer?
Whole system is warmer,
 We will need larger glaciers to turn the
switch on!!

How about CO2?
If we change P by +0.003

Cycle is longer by 12 kyr
If we change P by + 0.005

Cools the climate enough so there is a
permanent sea-ice cover
Ice sheet thickness?
Ice sheet thickness
Doesn’t change the cliamte!

Ice sheet play there role by
the albedo
 Insulating properties

Freezing temperature
Similar effect as with the emissivity
Lowering the freezing T by 0.5o

Time scale is longer by 20 kyr
Increasing the freezing T by 0.5o

Time scale is shorter by 20 kyr
Problems
The model is
Quite simple
 Highly idealized
 Neglecting any zonal variations
 Limited set of feed-backs

Problems….
It is not synchronous with observations
Fails to predict the synchronous
deglaciation of the southern hemisphere
BUT
Take home message
The aim of a such a simple model is
NOT to be accurate with observations
but just to explore potentially
IMPORTANT effects and feedback of
the climate.