Difference between revisions of "Summer Modeling School/Notes/vanderVeen Aug4.rtf"

From Interactive System for Ice sheet Simulation
Jump to: navigation, search
m
m
 
Line 12: Line 12:
  
 
Kees has been studying ice sheets / mass balance for 25 years, but just now there is focused international interest in it.  Why?
 
Kees has been studying ice sheets / mass balance for 25 years, but just now there is focused international interest in it.  Why?
 +
  
 
IPCC Summary for Policy Makers (2007): Dynamical processes related to ice flow not included in current models but suggested by recent observations could increase the vulnerability of the ice sheets to waring, increasing future sea level rise.  Understanding of these processes is presently limited, and important.
 
IPCC Summary for Policy Makers (2007): Dynamical processes related to ice flow not included in current models but suggested by recent observations could increase the vulnerability of the ice sheets to waring, increasing future sea level rise.  Understanding of these processes is presently limited, and important.
 +
  
 
Challenges for glaciologists:
 
Challenges for glaciologists:
Line 19: Line 21:
 
- Develop quantitative prognostic models ( = models to predict future response)
 
- Develop quantitative prognostic models ( = models to predict future response)
 
- Incorporate small-scale processes into whole ice-sheet models
 
- Incorporate small-scale processes into whole ice-sheet models
 +
  
 
Focus on the important question for society: sea level rise!  18-59 cm by 2100 is the present (conservative?) estimate: mountain glaciers, thermal expansion, and the present bit of melting from Greenland / Antarctica.  It does *not* include unstable rapid changes observed in Greenland / Antarctica.  We should try to put an upper limit on SLR from those changes.
 
Focus on the important question for society: sea level rise!  18-59 cm by 2100 is the present (conservative?) estimate: mountain glaciers, thermal expansion, and the present bit of melting from Greenland / Antarctica.  It does *not* include unstable rapid changes observed in Greenland / Antarctica.  We should try to put an upper limit on SLR from those changes.
 +
  
 
History of ice sheet modelling
 
History of ice sheet modelling
----------------------------------------
 
 
- Mahaffy model, 1976 - the first "3D" model (not truly 3D, it was depth-averaged).  The Model A.
 
- Mahaffy model, 1976 - the first "3D" model (not truly 3D, it was depth-averaged).  The Model A.
 
- Jenssen, 1977 - included temperature, so model is actually 3D in a sense - x,y,T.  Temperature is important to dynamics - viscosity is highly temperature-dependent, and basal processes (sliding / frozen on) depend on basal temperature.  The Model T.
 
- Jenssen, 1977 - included temperature, so model is actually 3D in a sense - x,y,T.  Temperature is important to dynamics - viscosity is highly temperature-dependent, and basal processes (sliding / frozen on) depend on basal temperature.  The Model T.
 
- Glimmer (Rutt, Hagdom, Hulton, Payne) 2009. The Ford Mustang - has a sleeker body but the same fundamental engine.
 
- Glimmer (Rutt, Hagdom, Hulton, Payne) 2009. The Ford Mustang - has a sleeker body but the same fundamental engine.
 +
- Frank Pattyn (pat-TAYNE) model 2003 in JGR - Toyota Prius
  
 
The engine of ice sheet models:
 
The engine of ice sheet models:
Line 39: Line 43:
 
- myriad others
 
- myriad others
  
Frank Pattyn (pat-TAYNE) model 2003 in JGR - Toyota Prius
+
 
  
 
Model Applications
 
Model Applications
--------------------------
+
 
 
Slow physics (glacial orbital cycles)
 
Slow physics (glacial orbital cycles)
 
Feedback between elevation & mass balance: higher surface = more snowfall
 
Feedback between elevation & mass balance: higher surface = more snowfall
Line 49: Line 53:
 
The new paradigm?
 
The new paradigm?
 
Jakobshavn Isbrae since 1851... retreat & speedup
 
Jakobshavn Isbrae since 1851... retreat & speedup
 +
  
 
Addressing the IPCC 4AR
 
Addressing the IPCC 4AR
Line 57: Line 62:
 
Workshop in July 2008 in St. Petersburg, to begin the hard part. 40-45 participants brainstorming, creating document.
 
Workshop in July 2008 in St. Petersburg, to begin the hard part. 40-45 participants brainstorming, creating document.
 
Document is still floating around somewhere, lacks full consensus.
 
Document is still floating around somewhere, lacks full consensus.
 +
  
 
Overarching questions from St. Petersburg
 
Overarching questions from St. Petersburg
---------------------------------------------------------
+
 
 
- What can we achieve within 5-10 years?  Reasonable timeline is important to address global climate change.
 
- What can we achieve within 5-10 years?  Reasonable timeline is important to address global climate change.
 
- Will climate change lead to irreversible (rapid nonlinear) ice sheet response?  (i.e. is Jim Hanson right?)  Is there a threshold / tipping point?
 
- Will climate change lead to irreversible (rapid nonlinear) ice sheet response?  (i.e. is Jim Hanson right?)  Is there a threshold / tipping point?
 
- Do rapid changes at the margins lead to large mass changes?  How coupled are the margins to the interior?
 
- Do rapid changes at the margins lead to large mass changes?  How coupled are the margins to the interior?
 
- Are observed rapid ice sheet changes "natural variability", response to recent warming (ice shelf breakup), or results of a basal switch?
 
- Are observed rapid ice sheet changes "natural variability", response to recent warming (ice shelf breakup), or results of a basal switch?
 +
  
 
Integrated approach
 
Integrated approach
 +
 
- englacial processes
 
- englacial processes
 
- surface forcing (atmos scientists)
 
- surface forcing (atmos scientists)
Line 72: Line 80:
 
- numerical issues (computing / applied math)
 
- numerical issues (computing / applied math)
 
Members of these different glaciological communities must talk to each other!!
 
Members of these different glaciological communities must talk to each other!!
 +
  
 
A high-resolution full Stokes model is *not* enough.
 
A high-resolution full Stokes model is *not* enough.
 +
 
- Yes, we can and should - but as a long-term objective
 
- Yes, we can and should - but as a long-term objective
 
- Remain aware of limitations
 
- Remain aware of limitations
Line 82: Line 92:
 
more
 
more
 
Boundary conditions
 
Boundary conditions
 +
 +
  
 
What ice sheet models need (1)
 
What ice sheet models need (1)
-------------------------------------------
+
 
 
- Better understanding of physical processes, and which processes are important to include
 
- Better understanding of physical processes, and which processes are important to include
 
subglacial lakes in Antarctica (see Byrd glacier on EAIS / Ben Smith) - are these short-term changes important in a large model?
 
subglacial lakes in Antarctica (see Byrd glacier on EAIS / Ben Smith) - are these short-term changes important in a large model?
Line 91: Line 103:
 
- Acceptable parameterizations of physics
 
- Acceptable parameterizations of physics
 
- What do you mean by physics? says Charles Jackson.  In atmos sci, there is "physics", "dynamics", "chemistry" all separate.  Olga say: in glaciology, physics means everything.
 
- What do you mean by physics? says Charles Jackson.  In atmos sci, there is "physics", "dynamics", "chemistry" all separate.  Olga say: in glaciology, physics means everything.
 +
  
 
What happens under the ice? (super-nice-looking slide!)
 
What happens under the ice? (super-nice-looking slide!)
 +
 
Subglacial morphology
 
Subglacial morphology
 
- channels and trenches
 
- channels and trenches
Line 98: Line 112:
 
- geothermal flux - northeast Greenland's fast-moving ice stream may be due to a geothermal hot spot there (Fahnestock paper)
 
- geothermal flux - northeast Greenland's fast-moving ice stream may be due to a geothermal hot spot there (Fahnestock paper)
 
- bed topography at grounding line
 
- bed topography at grounding line
 +
  
 
Grounding line stability
 
Grounding line stability
 +
 
Helheim glacier and bedrock topography (Ian Howat) - will retreat down a downslope until it reaches an upslope, where it regains stability
 
Helheim glacier and bedrock topography (Ian Howat) - will retreat down a downslope until it reaches an upslope, where it regains stability
 
- Topography
 
- Topography
Line 107: Line 123:
  
 
What ice sheet models need (2 & 3)
 
What ice sheet models need (2 & 3)
------------------------------------------------
+
 
 
- Subglacial lakes
 
- Subglacial lakes
 
water storage
 
water storage
Line 120: Line 136:
 
water storage
 
water storage
 
sediment strength
 
sediment strength
 +
  
 
Ice shelves and other (near-) floating peripherals
 
Ice shelves and other (near-) floating peripherals
 +
 
- Mechanisms for breakup
 
- Mechanisms for breakup
 
thermal limit of viability (general) - it'd be nice if we had a more quantitative explanation / model.  Ted Scambos has put forward some ideas
 
thermal limit of viability (general) - it'd be nice if we had a more quantitative explanation / model.  Ted Scambos has put forward some ideas
Line 130: Line 148:
  
 
What ice sheet models need (4 & 5)
 
What ice sheet models need (4 & 5)
------------------------------------------------
+
 
 
- Calving "law" that works for all scales of icebergs and bits
 
- Calving "law" that works for all scales of icebergs and bits
 
one of the more important boundary conditions for your model
 
one of the more important boundary conditions for your model
Line 136: Line 154:
 
it is extremely difficult with current calving laws to get a glacier to advance - be careful that your calving law is not automatically building in a retreat / instability that may not be realistic
 
it is extremely difficult with current calving laws to get a glacier to advance - be careful that your calving law is not automatically building in a retreat / instability that may not be realistic
 
- Better mesoscale models for Greenland & Antarctica (need atmos science input)
 
- Better mesoscale models for Greenland & Antarctica (need atmos science input)
 +
  
 
Model validation
 
Model validation
 +
 
- model inter-comparisons
 
- model inter-comparisons
 
- data validation
 
- data validation
Line 143: Line 163:
  
 
What ice sheet models need (6)
 
What ice sheet models need (6)
-------------------------------------------
+
 
 
- Data for calibration (knob-turning) & validation (compare model's results to other observations)
 
- Data for calibration (knob-turning) & validation (compare model's results to other observations)
 
what data?
 
what data?
Line 149: Line 169:
 
compatible formats
 
compatible formats
 
easy to use
 
easy to use
 +
  
 
What are we supposed to do here at the Summer Modeling School?
 
What are we supposed to do here at the Summer Modeling School?
 +
 
- Student / teacher line should be blurred, or nonexistent
 
- Student / teacher line should be blurred, or nonexistent
 
- Make Dave Holland happy: come up with a number to limit SLR from dynamical processes
 
- Make Dave Holland happy: come up with a number to limit SLR from dynamical processes
 
- Ian Rutt suggests thinking about *how* to test your code every time you are writing code
 
- Ian Rutt suggests thinking about *how* to test your code every time you are writing code
 
- Ken Jezek wants us to better incorporate a measure of confidence we have in a model, but Charles Jackson says this is very hard to know.
 
- Ken Jezek wants us to better incorporate a measure of confidence we have in a model, but Charles Jackson says this is very hard to know.
 +
  
 
Bad polar bear limerick
 
Bad polar bear limerick
-------------------------------
+
 
 
A thin polar bear called Onassis
 
A thin polar bear called Onassis
 
Moaned "Why cant you get off your asses?
 
Moaned "Why cant you get off your asses?

Latest revision as of 12:07, 4 August 2009

Kees van der Veen, University of Kansas

August 4, 2009

Portland Summer Modeling School

Ice Sheet Mass Balance and Sea Level (ISMASS): St. Petersburg and Beyond


notes by Kristin Poinar


Kees has been studying ice sheets / mass balance for 25 years, but just now there is focused international interest in it. Why?


IPCC Summary for Policy Makers (2007): Dynamical processes related to ice flow not included in current models but suggested by recent observations could increase the vulnerability of the ice sheets to waring, increasing future sea level rise. Understanding of these processes is presently limited, and important.


Challenges for glaciologists: - Improve understanding of processes - Develop quantitative prognostic models ( = models to predict future response) - Incorporate small-scale processes into whole ice-sheet models


Focus on the important question for society: sea level rise! 18-59 cm by 2100 is the present (conservative?) estimate: mountain glaciers, thermal expansion, and the present bit of melting from Greenland / Antarctica. It does *not* include unstable rapid changes observed in Greenland / Antarctica. We should try to put an upper limit on SLR from those changes.


History of ice sheet modelling - Mahaffy model, 1976 - the first "3D" model (not truly 3D, it was depth-averaged). The Model A. - Jenssen, 1977 - included temperature, so model is actually 3D in a sense - x,y,T. Temperature is important to dynamics - viscosity is highly temperature-dependent, and basal processes (sliding / frozen on) depend on basal temperature. The Model T. - Glimmer (Rutt, Hagdom, Hulton, Payne) 2009. The Ford Mustang - has a sleeker body but the same fundamental engine. - Frank Pattyn (pat-TAYNE) model 2003 in JGR - Toyota Prius

The engine of ice sheet models: - Laminar flow (convenient and easy to use, not necessarily truthful) - Basal sliding - Temperature calculation Available extra options: - Ice shelves - Basal hydrology - Isostasy - Calving (Kees's favorite) - myriad others


Model Applications

Slow physics (glacial orbital cycles) Feedback between elevation & mass balance: higher surface = more snowfall Adjustment of earth crust to ice loading

The new paradigm? Jakobshavn Isbrae since 1851... retreat & speedup


Addressing the IPCC 4AR SCAR report, November 2007: A need for more realistic ice sheet models - "mea culpa", we are aware of the problems and realize the need for better models (Scientific Committee on Antarctic Research) Easy part: make a list of things that are missing from the models Hard part: add them Workshop in July 2008 in St. Petersburg, to begin the hard part. 40-45 participants brainstorming, creating document. Document is still floating around somewhere, lacks full consensus.


Overarching questions from St. Petersburg

- What can we achieve within 5-10 years? Reasonable timeline is important to address global climate change. - Will climate change lead to irreversible (rapid nonlinear) ice sheet response? (i.e. is Jim Hanson right?) Is there a threshold / tipping point? - Do rapid changes at the margins lead to large mass changes? How coupled are the margins to the interior? - Are observed rapid ice sheet changes "natural variability", response to recent warming (ice shelf breakup), or results of a basal switch?


Integrated approach

- englacial processes - surface forcing (atmos scientists) - basal conditions (hydrologists) - marine margins (oceanographers) - numerical issues (computing / applied math) Members of these different glaciological communities must talk to each other!!


A high-resolution full Stokes model is *not* enough.

- Yes, we can and should - but as a long-term objective - Remain aware of limitations - It excludes some processes necessarily: shear margins (too small-scale: 3-4 km is subgrid spacing) subglacial valleys weak basal layers more Boundary conditions


What ice sheet models need (1)

- Better understanding of physical processes, and which processes are important to include subglacial lakes in Antarctica (see Byrd glacier on EAIS / Ben Smith) - are these short-term changes important in a large model? flow law (strainrate vs. stress) - can we nail it down better? should we? - On what scales should ice sheets me modeled? - Acceptable parameterizations of physics - What do you mean by physics? says Charles Jackson. In atmos sci, there is "physics", "dynamics", "chemistry" all separate. Olga say: in glaciology, physics means everything.


What happens under the ice? (super-nice-looking slide!)

Subglacial morphology - channels and trenches - sediment vs. hard beds - soft sediments allow ice to move faster. Important on WAIS. - geothermal flux - northeast Greenland's fast-moving ice stream may be due to a geothermal hot spot there (Fahnestock paper) - bed topography at grounding line


Grounding line stability

Helheim glacier and bedrock topography (Ian Howat) - will retreat down a downslope until it reaches an upslope, where it regains stability - Topography - Sediment deposition - Mathematics of the transition zone (Schoof) may be less important than these two issues


What ice sheet models need (2 & 3)

- Subglacial lakes water storage importance of drainage events - Interaction of till and subglacial water - Addition of supraglacial meltwater - Subglacial water budget of different hydrologic systems - Coupling to ice flow model water transport & storage - Sliding laws basal pressure water storage sediment strength


Ice shelves and other (near-) floating peripherals

- Mechanisms for breakup thermal limit of viability (general) - it'd be nice if we had a more quantitative explanation / model. Ted Scambos has put forward some ideas - Role of sea ice - Subshelf circulation and melting - Restraint on interior ice - how important are they?


What ice sheet models need (4 & 5)

- Calving "law" that works for all scales of icebergs and bits one of the more important boundary conditions for your model feedback between geometry & calving it is extremely difficult with current calving laws to get a glacier to advance - be careful that your calving law is not automatically building in a retreat / instability that may not be realistic - Better mesoscale models for Greenland & Antarctica (need atmos science input)


Model validation

- model inter-comparisons - data validation - capability of simulating past and current changes

What ice sheet models need (6)

- Data for calibration (knob-turning) & validation (compare model's results to other observations) what data? open data access compatible formats easy to use


What are we supposed to do here at the Summer Modeling School?

- Student / teacher line should be blurred, or nonexistent - Make Dave Holland happy: come up with a number to limit SLR from dynamical processes - Ian Rutt suggests thinking about *how* to test your code every time you are writing code - Ken Jezek wants us to better incorporate a measure of confidence we have in a model, but Charles Jackson says this is very hard to know.


Bad polar bear limerick

A thin polar bear called Onassis Moaned "Why cant you get off your asses? "It's clear climate change "Is restricting my range "So you must make big cuts to bad gases!"