Transcript ClimateModels2x - Alan Robock

Climate Models
Alan Robock
Department of Environmental Sciences
Rutgers University, New Brunswick, New Jersey USA
[email protected]
http://envsci.rutgers.edu/~robock
What is a model?
Alan Robock
Department of Environmental Sciences
What is a climate model?
Alan Robock
Department of Environmental Sciences
Rotating tank (dishpan)
Alan Robock
Department of Environmental Sciences
Why do climate modeling?
Alan Robock
Department of Environmental Sciences
Why do climate modeling?
•
We can’t bring the entire atmosphere,
ocean, land, and biosphere into the
laboratory.
•
We can’t do experiments in the real world.
Alan Robock
Department of Environmental Sciences
Types of experiments:
•
simulations of current climate – first test
of any model
•
simulations of past
•
•
ice ages
last 165 years (historical runs)
•
sensitivity to different forcings
•
sensitivity to different feedbacks
•
predictions
•
test different scenarios of future
Alan Robock
Department of Environmental Sciences
Types of climate models:
Determined by:
• spatial resolution and representation
• time step size, or steady-state
• portion of climate system that is included
Classification:
• 0-dimensional  3-dimensional
• atmosphere only (with fixed sea surface
temperatures), or
mixed-layer ocean, or
complete dynamic ocean
• inclusion of aerosols, chemistry,
biosphere, detailed stratosphere ...
Alan Robock
Department of Environmental Sciences
Types of climate models:
Energy-balance models
•
Zero-dimensional, steady-state:
S0
1     TE4
4
•
Zero-dimensional, time-dependent:
Ts S0
C
 1     TE4
t
4
Alan Robock
Department of Environmental Sciences
Types of climate models:
Energy-balance models
Ts S0
C
 1     TE4
t
4
Disadvantages of this model:
• no predictive value as sensitivity and C
cannot be evaluated here
• describes only global average
Advantages of this model:
• easily understood
• can be used to interpret more complex
models
Alan Robock
Department of Environmental Sciences
Model for the Assessment of Greenhouse Gas Induced Climate
Change (MAGICC),
an upwelling-diffusion energy-balance model
http://www.magicc.org/
Alan Robock
Department of Environmental Sciences
What is an RCM?
Alan Robock
Department of Environmental Sciences
What is an RCM?
1. Radiative-Convective Model, also now called a
Single Column Model (SCM)
or fractional cloudiness
Advantages: can look at
• radiation parameterizations
• cloud parameterizations and feedbacks
• water vapor feedbacks
• lapse rate feedbacks
• surface interactions and feedbacks
• effects of different atmospheric composition, including CO2
• O3 at different levels
• aerosols at different levels
Disadvantages:
• cannot look at horizontal distributions, albedo feedbacks, dynamics
Alan Robock
Department of Environmental Sciences
What is an RCM?
2. Regional Climate Model
Advantages:
• can look in detail at specific locations
• takes less computer time than a global simulation
Disadvantages:
• complicated to connect to boundaries (spectral nudging helps)
• reproduces in detail the climate determined by the boundary
conditions, but cannot change the basic climate
The Weather Research & Forecasting Model (WRF)
http://www.wrf-model.org/index.php
There are other models and regional strategies, such as stretched
grids.
Alan Robock
Department of Environmental Sciences
Example of WRF domains from:
Sertel, Elif, Alan Robock, and Cankut Ormeci, 2010: Impacts of land
cover data quality on regional climate simulations. Internat. J.
Climatology, 30, 1942-1953, doi:10.1002/joc.2036.
Alan Robock
Department of Environmental Sciences
General Circulation Models (GCMs)
Sometimes people now say GCM stands for global
climate model, but that is not actually correct, as even
a simple energy balance model is a global climate model
and what makes a GCM distinctive is that it explicitly
models the general circulation of the atmosphere (or
ocean).
All climate models are energy-balance models.
All climate models are global climate models.
Alan Robock
Department of Environmental Sciences
Evolution of processes
included in state-ofthe-art climate models
FAR: First IPCC
Assessment Report
SAR: Second Report
TAR: Third Report
AR4: Fourth Report
IPCC AR4, Chapter 1
Alan Robock
Department of Environmental Sciences
AOGCM = Atmosphere-Ocean General
Circulation Model.
ESM = Earth System Model
Alan Robock
Department of Environmental Sciences
IPCC AR4, Chapter 1
Alan Robock
Department of Environmental Sciences
Alan Robock
Department of Environmental Sciences
Alan Robock
Department of Environmental Sciences
General Circulation Models (GCMs) 1
Basic Physical Laws:
Conservation of energy (First law of thermodynamics)
Conservation of momentum (Newton’s second law of motion)
Conservation of mass (Continuity equation)
Conservation of moisture
Hydrostatic equilibrium
Gas law
Alan Robock
Department of Environmental Sciences
Alan Robock
Department of Environmental Sciences
General Circulation Models (GCMs) 2
Physical Processes That Must or Can Be Included:
Wind
Sea ice
Radiation
Snow
Precipitation
Glaciers
Soil moisture
Vegetation
Ground water
Ocean biota
Aerosols
Clouds, convective and large-scale
Air-sea exchanges of moisture, energy, and momentum
Air-land exchanges of moisture, energy, and momentum
Chemistry, particularly O3 and CO2
Ocean temperature, salinity, and currents
Alan Robock
Department of Environmental Sciences
Real World
vs.
Model World
Alan Robock
Department of Environmental Sciences
How to ConstructModel
a Climate Model
Theory of Climate
Development
How to Construct a Climate Model
Actual Climate
Model Development
Regular and
frequent
observations
Previous
knowledge
Programatic
objectives
Sparse and
infrequent
observations
Interpretation
of observations
Theoretical
understanding
Preconceived
notions
Management
directives
Theoretical
misunderstanding
Programmatic
objectives
Sophisticated
computer model
Computer code
development
Reasonable
assumptions
Computer model results
Further
understanding
Further refinement of
important details
Agreement
between theory
and observations
Publication
Incorrect interpretation
of observations
Management
directives
Sophisticated
computer model
Code
errors
Unrealistic
assumptions
Confusion
Further
misunderstanding
Further refinement of
unimportant details
Coincidental agreement
between theory
and observations
Publication
Alan Robock
Department of Environmental Sciences
Typical grid spacing of a GCM is now 1°x1°
latitude-longitude by 1 km in the vertical.
Each time the horizontal resolution is
increased by a factor of 2, the time
needed to run the model goes up by a
factor of 8.
When the vertical resolution is doubled the
time required doubles in general, but can go
up by more, if winds become faster.
Alan Robock
Department of Environmental Sciences
To include all the processes in a climate model which are of
a scale smaller than is resolved by the model, they must be
“parameterized.”
One of the most important and difficult climate elements
to parameterize is cloudiness. Clouds have a much smaller
spatial and temporal scale than a typical GCM grid box.
Usually, we consider separately 2 types of clouds, layer
clouds and convective clouds. There is no fundamental
prognostic equation for clouds (no conservation of clouds
principle); rather they form when condensation takes place
and dissipate due to precipitation and evaporation.
Alan Robock
Department of Environmental Sciences
Rows and flows of angel hair
And ice cream castles in the air
And feather canyons everywhere;
I’ve looked at clouds that way.
But now they only block the sun.
They rain and they snow on everyone.
So many things I would have done
But clouds got in my way.
I’ve looked at clouds from both sides now.
From up and down, and still somehow
It’s cloud illusions I recall.
I really don't know clouds at all.
— Joni Mitchell
Both Sides Now, 1967
Alan Robock
Department of Environmental Sciences
Model Intercomparison Projects (MIPs)
First was AMIP: Atmospheric Model Intercomparison
Project, using specified sea surface temperatures, and
running from 1979 through 1988.
CMIP (Coupled Model Intercomparison Project)
CMIP3 used for IPCC AR4 and CMIP5 used for AR5.
CMIP6 is now being organized.
There are also MIPs for just parts of the climate
system, like the Program for Intercomparison of Landsurface Parameterization Schemes (PILPS)
http://www.wcrp-climate.org/modelling-wgcm-mipcatalogue/modelling-wgcm-mips-2
Alan Robock
Department of Environmental Sciences
Taylor et al. (BAMS, 2012)
Alan Robock
Department of Environmental Sciences
Climate Model
Intercomparison
Project 6
(CMIP6)
design proposal
Meehl, G. A., R. Moss, K. E. Taylor, V. Eyring, R. J. Stouffer, S. Bony and B. Stevens, 2014: Climate
Alan Robock
Department of Environmental Sciences
model intercomparisons: Preparing for the next phase, Eos, 95, 77-78, doi:10.1002/2014EO090001.
Meehl, G. A., R. Moss, K. E. Taylor, V. Eyring, R. J. Stouffer, S. Bony and B. Stevens, 2014: Climate
Alan Robock
Department of Environmental Sciences
model intercomparisons: Preparing for the next phase, Eos, 95, 77-78, doi:10.1002/2014EO090001.
Meehl, G. A., R. Moss, K. E. Taylor, V. Eyring, R. J. Stouffer, S. Bony and B. Stevens, 2014: Climate
Alan Robock
Department of Environmental Sciences
model intercomparisons: Preparing for the next phase, Eos, 95, 77-78, doi:10.1002/2014EO090001.
I’m involved in GeoMIP and VolMIP, which both were
formally endorsed two weeks ago.
Meehl, G. A., R. Moss, K. E. Taylor, V. Eyring, R. J. Stouffer, S. Bony and B. Stevens, 2014: Climate
Alan Robock
Department of Environmental Sciences
model intercomparisons: Preparing for the next phase, Eos, 95, 77-78, doi:10.1002/2014EO090001.
Model Intercomparison Projects (MIPs)
All find that models are different from each other and
different from observations, so what is the point?
- models are tested in a controlled regime, and
modelers find errors in models when comparing to
observations
- estimation of range of confidence or uncertainty in
models
- identification of outliers
- development and dissemination of data sets that can
be useful to all
Alan Robock
Department of Environmental Sciences
Model Intercomparison Projects (MIPs)
Most have found that:
- no one model is best at everything
- no one test evaluates all aspects of models
- after excluding outliers with serious errors, the model
consensus outperforms any individual model
Alan Robock
Department of Environmental Sciences
AOGCM is an Atmosphere-Ocean General Circulation
Model.
ESM means Earth System Model, and includes more
components of the climate system.
Alan Robock
Department of Environmental Sciences
Alan Robock
Department of Environmental Sciences
Alan Robock
Department of Environmental Sciences
Earth System Models of Intermediate
Complexity (EMICs)
Do not explicitly model atmospheric dynamics, so
can run much faster and include more processes
to study long-time period processes.
Alan Robock
Department of Environmental Sciences
Alan Robock
Department of Environmental Sciences
Model for the Assessment of Greenhouse Gas Induced Climate
Change (MAGICC),
an upwelling-diffusion energy-balance model
http://www.magicc.org/
Alan Robock
Department of Environmental Sciences
Go to live.magicc.org, register, and then conduct the following
experiments. For a description of the emission scenarios, see
http://wiki.magicc.org/index.php?title=Online_Emission_Scenarios.
a. For emissions scenario RCP8.5 (representative concentrations pathway
with +8.5 W/m2 radiative forcing by 2100):
i. Describe the fossil CO2 time series for the 21st Century.
ii. On the next page, choose Standard run mode, CMIP3: DEFAULT
for the Climate Parameters and C4MIP: DEFAULT for the Carbon Cycle
settings. Click Advance Settings and report the Climate
Sensitivity. (This is one place where you can change parameters for your
own experiment - see d-g and j below).
iii. Click Next to run the model. What is the global average
temperature in 2100?
iv. Click on the file symbol to download the output files. Download
the DAT_SURFACE_TEMP.OUT file and rename it for this experiment.
b. Repeat the same experiment in a. for the RCP3-PD (which as well goes
by the name RCP2.6) scenario.
c. Using Excel, Matlab, or by hand, plot on the same graph the globalmean temperature for the two experiments above for 1765-2100. How
Alan Robock
do they compare? Why are they different?
Department of Environmental Sciences
Go to live.magicc.org, register, and then conduct the following
experiments. For a description of the emission scenarios, see
http://wiki.magicc.org/index.php?title=Online_Emission_Scenarios.
d.
e.
f.
g.
h.
Repeat a. with a climate sensitivity of half of the standard one.
Repeat a. with a climate sensitivity of twice the standard one.
Repeat b. with a climate sensitivity of half of the standard one.
Repeat b. with a climate sensitivity of twice the standard one.
Add the results of experiments d-g to the graph you did in part
c. This will result in a graph with 6 curves. Create another version
of the graph with time only going from 2000 to 2100, so as to see
the results more clearly. Include both graphs with your
assignment. How do the results compare? Why are they
different? Are the differences linear? (That is, is the climate
response proportional to the forcing and to the sensitivity?)
i. Repeat a. and b. with the Probabilistic, Multi-model ensemble
option. Do a screen capture and present each of these results. Is
the spread between models as large as the differences caused by the
different scenarios?
j. Design and carry out your own experiment to elucidate the climate
response. Describe what you did, why you did it, and your results.
Alan Robock
Department of Environmental Sciences