Actions

Difference between revisions of "Summer School on Global Sustainability-Projects & Working Groups"

From Santa Fe Institute Events Wiki

Line 4: Line 4:
==GROUP ASSIGNMENTS==
==GROUP ASSIGNMENTS==
===ENGLAND & NAKICENOVIC: Bush, Cullenward, Frisch, Nguyen, Nkem, Robinson<br/>===
===ENGLAND & NAKICENOVIC: Bush, Cullenward, Frisch, Nguyen, Nkem, Robinson<br/>===
07/15/09 - Climate models
An overview was first given about the definition, classification, characteristics and advantages of a climate model. The process of building a model extends to verification, validation and assessment. What is a climate model? A representation of reality predictive or diagnostic (zooms into reality in the past). Iconic = physcial version, analogue = reproduce behavior and mathematical = analytical relationships that govern events
Computer models are advantageous in terms of expense, comprehensiveness, logic, accessiblity, flexibility (capable of examining, sensitivity analysis, parameter tuning etc).
There are 6 governing equations in climate models: for ocean there are 3 of momentum, conservation for heat, salt, and mass and the same is for the atmosphere with humidity instead of salt. They have as many equations as unknowns.
Forcing conditions represent only processes at the top of the atmosphere for full blown climate models. For running submodels or subcomponents of the Earth it requires many more forcing conditions that need to be addressed. Hence it is harder to run submodels rather than a whole model.
It would be interesting to track down the progress, quantitatively and qualitatively, of the scientific modeling community.
For the most part models are run on FORTRAN90. The language is better in picking out errors in floating points.
What kind of grid is employed determines what kind of dynamics it will represent effectively.
About resolution, in the late 80s, the resolution is (4 degrees by 4 degrees) 400km by 400km. Now it is about 1 degree by 1 degree.
When we can't get what we want, we go down to a finer resolution by nesting a subgrid of fine scale on top of a coarse grid. Sometimes we have to change the model or some models give us the option of where to zoom in. The option of nesting different resolutions allows for adding in the affects of cities, etc that are quite small but can effect local weather/climate significantly.
The vertical vs. horizontal resolution is another issues. A spectral grid is such that the lines are parallel along the y axis but bend along the x axis.
1D e.g. that about an air column in the atmosphere and 2D models e.g. Henry Stommel's 2-boxes models still have good uses on their own. 3D models employ spectral grids which minimize numerical erros in atmospheric models.
How much time it would normally take for a model to be ready for use?
The process of running a model to get its output and spatial visualization are seperate.
Having models is one thing, having interfaces where researchers will simply walk in and use it is another option and there are some good ones.
Archives of observations are very important to test our models. But sometimes the archives are in very complicated formats even to view e.g. the Jet Propulsion Lab prefers .hdf - why do they make it so inaccessible?
There are good models and fake ones that look like real. This is not a parade. Open access is important.
CFC invade the ocean in very specific way and they act as tracers in ocean circulation experiment. As these were introduced, the presence/levels of CFCs in deep water provides an excellent sense of movement and mixing.
This can be used to doublecheck the accuracy of ocean models, being more precise than temperature and salinity checks...
To assess uncertainties in existing models, lland ice is a big uncertainty (1.6m is likely sealevel rise, 85cm from land ice)
Another is sulfite areosols–a pollutant that has helped to keep things cooler. as we clean up our combustion, this will change–the impacts of this are uncertain
===RUSSELL & GRÜBLER:  Bamutaze, Borgeson, Engler, Ona, Pasqualini, Zellner===
===RUSSELL & GRÜBLER:  Bamutaze, Borgeson, Engler, Ona, Pasqualini, Zellner===
[[Media:Russel_summary.doc]] Here is the summary of Joellen Russel's climate modeling talks.
[[Media:Russel_summary.doc]] Here is the summary of Joellen Russel's climate modeling talks.

Revision as of 17:55, 24 July 2009

Summer School on Global Sustainability


GROUP ASSIGNMENTS

ENGLAND & NAKICENOVIC: Bush, Cullenward, Frisch, Nguyen, Nkem, Robinson

07/15/09 - Climate models An overview was first given about the definition, classification, characteristics and advantages of a climate model. The process of building a model extends to verification, validation and assessment. What is a climate model? A representation of reality predictive or diagnostic (zooms into reality in the past). Iconic = physcial version, analogue = reproduce behavior and mathematical = analytical relationships that govern events Computer models are advantageous in terms of expense, comprehensiveness, logic, accessiblity, flexibility (capable of examining, sensitivity analysis, parameter tuning etc). There are 6 governing equations in climate models: for ocean there are 3 of momentum, conservation for heat, salt, and mass and the same is for the atmosphere with humidity instead of salt. They have as many equations as unknowns. Forcing conditions represent only processes at the top of the atmosphere for full blown climate models. For running submodels or subcomponents of the Earth it requires many more forcing conditions that need to be addressed. Hence it is harder to run submodels rather than a whole model.

It would be interesting to track down the progress, quantitatively and qualitatively, of the scientific modeling community.

For the most part models are run on FORTRAN90. The language is better in picking out errors in floating points.

What kind of grid is employed determines what kind of dynamics it will represent effectively. About resolution, in the late 80s, the resolution is (4 degrees by 4 degrees) 400km by 400km. Now it is about 1 degree by 1 degree. When we can't get what we want, we go down to a finer resolution by nesting a subgrid of fine scale on top of a coarse grid. Sometimes we have to change the model or some models give us the option of where to zoom in. The option of nesting different resolutions allows for adding in the affects of cities, etc that are quite small but can effect local weather/climate significantly. The vertical vs. horizontal resolution is another issues. A spectral grid is such that the lines are parallel along the y axis but bend along the x axis. 1D e.g. that about an air column in the atmosphere and 2D models e.g. Henry Stommel's 2-boxes models still have good uses on their own. 3D models employ spectral grids which minimize numerical erros in atmospheric models.

How much time it would normally take for a model to be ready for use? The process of running a model to get its output and spatial visualization are seperate. Having models is one thing, having interfaces where researchers will simply walk in and use it is another option and there are some good ones. Archives of observations are very important to test our models. But sometimes the archives are in very complicated formats even to view e.g. the Jet Propulsion Lab prefers .hdf - why do they make it so inaccessible?

There are good models and fake ones that look like real. This is not a parade. Open access is important.

CFC invade the ocean in very specific way and they act as tracers in ocean circulation experiment. As these were introduced, the presence/levels of CFCs in deep water provides an excellent sense of movement and mixing. This can be used to doublecheck the accuracy of ocean models, being more precise than temperature and salinity checks... To assess uncertainties in existing models, lland ice is a big uncertainty (1.6m is likely sealevel rise, 85cm from land ice) Another is sulfite areosols–a pollutant that has helped to keep things cooler. as we clean up our combustion, this will change–the impacts of this are uncertain

RUSSELL & GRÜBLER: Bamutaze, Borgeson, Engler, Ona, Pasqualini, Zellner

Media:Russel_summary.doc Here is the summary of Joellen Russel's climate modeling talks.

EDENHOFER & PAUL: Bottrill, Brelsford, Doshi, Geddes, Gong, Zaks

DASGUPTA & LOVINS: Clewlow, Dangerman, Hunt von Herbing, Kane, Sammeth, Wolf

HARGADON & MEADOWS: Gourdji, G. Jones, McInnis, Morgan, Mueller

RUBBIA & KUTSCHER: Gonzales, Hagerman, A. Jones, Krakauer, LaCerva, Lacroix

Participation

Active, assertive participation by students is an intrinsic part of this event. Attendance at all program sessions is mandatory. Participants have been divided into six groups; each group designed to be as topically and geographically diverse as possible. Chairperson and rapporteur roles may rotate within each group.

Group Tasks

Groups have several tasks. One function is to discuss each day’s lectures among themselves and to be prepared to actively participate in the group discussion sessions each day. In addition, each group will be tasked to write a summary of the main points of the two lecturers, one from the first week and one from the second week. Responsibilities may be divided (that is, each of the six students might be responsible for producing a lecture transcription, editing duties could be divided, etc). The lecture summaries should include comments by the group on the broader implications of the lectures, critical analysis of the research area covered, and resonance to other presentations. Each speaker will be available to meet with the groups covering his/her lectures. This material may be edited and produced in book form.

Research Agenda

Additionally, each group will be asked to produce a research agenda as a product of the school. This would not be limited to the lectures each group covered, but rather focus on the entire two-week agenda and identify synergies, areas of disagreement, and gaps in our knowledge that can be resolved by future research. The final day of the school will include a plenary session during which each group’s rapporteur presents the research agenda. Finally, we will draw on these agendas to produce an open letter to President Obama or Science Advisor John Holdren.