After finding a nuber of shortcomings in the forcing data for the baseline (B0) simulation, a proposal was made to address this:
COLA
will continue with a multi-model analysis based on the B0 simulations,
to be called the GSWP-2 Multi-Model Analysis, Version 1.0
Current
sensitivity studies will continue based on the B0 baseline, since
sensitivity studies are differences, and thus not as sensitive to
systematic errors.
IIS and FRCGC will produce, with COLA an improved baseline forcing data set B1, for a re-run by the modeling groups.
A
Multi-model analysis based on B1 will be "Version 2.0", probably using
GPCC/ERA40 hybrid precipitation with improved gauge undercatch
correction, ISCCP radiation (including spin-up), and ERA40 meteorology.
Koster: Suggests we do a soil moisture index study a la standard
normal deviates, that could allow for "translation" of GSWP soil wetness to initialize other inidividual models.
The ISCCP Landflux project may be more consist with B1 than B0 if use
ISCCP radiation for forcing.This
caveat shold be communicated to Bill Rossow and the GHP group, since
they are looking to GSWP as a source for surface latent and sensible
heat flux estimates.
Sensitivity study ideas:
Shift soils to a coarser grade (one category) to account for macropores, fractures, etc., as was done in ISLSCP Initiative I.
Data distribution within GSWP-2 will be done by password-protected access on the ICC website. Monthly tar files from each model available this way until results are published.
GSWP-3 Ideas:
Time span: 100y? 50y? Satellite era (1979-)?
Extend down to groundwater (Andy & Bart: coupling to common groundwater
model)
New science question needed to motivate GSWP2, not
just producing a data product (which a lot of GEWEX just wants) maybe deeper water
is it.But stick to water and leave carbon
to the IGBP community.
More widespread obs for groundwater than soil
wetness.New acronym?
Action Items:
GSWP participants = Pursue current GSWP-2 timeline with respect to sensitivity studies, evaluation and analysis
Oki = lead effort to generate and distribute B1 forcing data
Dirmeyer = will have a brief meeting at AMS/San Diego to update status and set timeline for new runs
Peters-Lidard = Find accomodation for a 2-hour GSWP meeting at AMS.
Dirmeyer = draft overview article with multi-model analysis in time for AMS, goal to publish in time for IPCC deadline (BAMS).
White paper is a start certainly a kernel for a GEWEX News
article.Maybe not beyond grey literature at this stage.
Multi-criteria
experiments (Papers 1 and 2 from Liu & Gupta et al.)
These papers distributed great motivator for
LOCO should be used as an engine for the paper
Shows that the uncoupled problem is no substitute for the coupled.
Discussion
of planning
Want an outline of a larger roadmap (revisit
Friday)
See 3 or 4 possible items already:
GABLS based PBL coupling leverage off their
stable nocturnal boundary layer work, but add in LSSs under the SCMs
Could expand into comparison of coupling in strong
versus week Omega regions (leverage GLACE) & do coupled vs uncoupled in those locations. Can we derive a Local Coupling Index (vdHurk is keen on
dimensionless approach to normalize across models)
Houser get into impact of horizontal heterogeneity on
grid-box vertical transport and physics, clouds, etc.Goes on to collaboration with GMPP & superparmeterizations.Big diurnal cycle potential with local circulations
here.
The original idea = coupled versus uncoupled DAS and
problems therein (Liu papers motivate this as well)
LoCo Action items:
vdHurk, Peters-Lidard, Houser = outline of SIP for discussion by Friday AM
vdHurk & Peters-Lidard (with Houser and Dirmeyer) = lead transformation of white paper to GEWEX News Article.
End of day
Thursday
GLACE
Progress
report (Koster)
Future
plans and possibilities
Cover more seasons. Not terribly groundbreaking, but possibly
interesting to SH concerns. Tough to motivate as many models
Look at IC anomaly maintenance, as well as the
feedback (G1)
Can we go with a multi-model prediction problem with realistic surface initial conditions?
Schlosser: what about separating out the snow signal?
Gedney: Can we diagnose LSS coupling strength offline, like in GSWP, and compare to coupled systems (another LoCo motivator)?
Gedney: Compare climatologies (model skill) vs. hot spot
strength - does land feedback improve, or merely exist?
GLACE Action items:
Koster = Float idea of other seasons to participants.
Koster = How to measure omega? Randy has an ergotic PDF
shift metric pursue its applicability to this problem.
Panel
wonders whether there are problems with the level of support
for Bastidas to run this project (perhaps so busy with faculty
responsibilities?) There have been problems with communication
(emails to Bastidas going unanswered - many months between contacts).
Not take revised sched too seriously, esp after
Steps 1 and 2 (it will take more time to get modeling groups up and running with MC
code.
Can Liu spend time on this step-3
assistance?Note to Hoshin as well.And what happened to Nijssen?
Will asses at AMS meeting how its going sort
of a probational period until then.
Worst case scenario postpone if resources
cannot be found to get this back on a reasonable track (cant string out the
early adopters too long).
PILPS-SP Action Items:
Dirmeyer = send an email to Luis (cc Pitman) asking him
about his resources to run SP, does he need help, offer Pitman (via PILPS lists) to
nag, be the "bad cop", to get things rolling and expand number of participants.
Pitman = will write a GLASS/PILPS list reminder, with
a clarifier about nothing mandatory after steps 1&2, & reiterating the selling
points.
Bastidas = Add Yongjui Dai- CLM to list of models - collect more participants.
Mostly analysis and writing going on this year - and a small extension of the integrations
PILPS-C1 Action Items:
Dirmeyer = Encourage Viovy not to delay on papers can get in IPCC
relevance perhaps they can get 2003 integration done is like what future may be (drought - extreme)
Pitman = Will push information to R. Betts to be sure IPCC knows
of PILPS-C1.
Heavy water isotopes can mark partitioning between evap and transp, and even affect CO2 isotopes.
Implications for measuring recycling/ sources, processes, etc.
A solid proposal has been presented - field of isotope models is already growing by leaps and bounds.
First
phase will use regional model (REMO) derived forcings from 3 sites
(Europe, Australia, S. America) to drive iLSSs and test/validate
using PILPS/ALMA standard methods.
See write-up for proposed details.
Most measurements are daily or less, but MDBWBP has a diurnal isotope bit for GMPP Diurnal Focus?
FTIR gives a vert flux profile on the cheap - useful for LoCo element?!?
PILPSy Scatter plot in here with GSWP2 same
spread (AHS thinks Siblings are right about LH/SH)
Koster offered some data input improvement ideas.
Peters-Lidard heterogeneity non-colocation of all
measurements could be a problem
Oki Thailandisotope
measurements suggest large-scale circulation signal dominates local
variations over seasonal time-scales. So maybe not such a problem
after all
vdHurk if one worries about soil carbon part, then one needs
centuries of spin-up. But the proposal is for a 10-day test period??
Dirmeyer is on iLEAPS SSC, and is GEWEX liason to iLEAPS
iLEAPS is approximately the IGBP cousin to GLASS, but more biogeochemistry oriented (Andreae)
Potential interface points include GSWP (some BAHC heritage) PILPS-C#, and AMMA
IPCC
land evaluation (Pitman, Dirmeyer)
IPCC is in need of some expert commentary on the land
surface component of the climate change models and projections.
Should GLASS volunteer to provide it?
How can we get our hands on the data via
PCMDI.Do they have enough land-relevant
output?
For results to get into the assessment, must be in
press by end 2005.
Action Items:
Deadlines?Pitman will check on that in Trieste
Anyone already tackling this (maybe Roads?) Pitman
will check.
Pitman will express our potential interest to Trieste.
Dirmeyer =Double check with Tom Phillips about data access.
Houser volunteers to look at global water bal. comp
to GSWP2, e.g.
Suggestion for a multi-model land use-climate impact study
with rigorous statistics to determine magnitude of LUCC impacts over
20th century (as compared to CO2).
Too late for IPCC, but could at least produce a review paper as a lead-in to the experiment in time for IPCC.
Do we (GLASS/Pitman) do this, or pass onto C20C crowd,
with a push to the foregone groups (who claim LUCC relevance without proper
statistics, e.g. Foley, DeFries, Pielke & Chase) to participate (is it too
much for them bacause of the more demanding C20C integration structure).
Pitman's proposal for integrations is more tractable than C20C, but requires starting from scratch.
R. Betts is doing both C20C and snapshot, to see
how experimental structure affects results.
Action Items:
Gedney will tap Folland, Dirmeyer = Kinter, about land cover
change plans, and introduce to Pitman. (cc emails across parties)
Mostly informational briefing - Climate Observation and
Prediction of the Earth System (COPES) is to be a pan-WCRP
synthesizer. Shukla and Hoskins at the forefront.
Best and Viterbo have done baseline runs to generate soil
wetness time series - they still need to swap the soil wetness data
sets (as BCs for each others' models) to show lack of direct
transferability of soil wetness between models
Koster was to write a the background for such an
experiment, to be published in BAMS, to expose the community to this
underappreciated issue
Action Items:
Gedney = push on Best and Viterbo to finish up integrations (they are meeting
this week anyway)
Koster
= has been waiting for results to start
writing will start writing background sections now, and bother
Best and Viterbo for model descriptions to get the project moving again.
ISLSCP Initiative II did not deliver fully on
data consistency not at all without a great deal of pressure and assistance
from GSWP.
GLASS does not see great utility in the path
ISLSCP has chosen with its recent and planned data initiatives. Modelers must be involved in production, meaning we'll end up producing the data sets anyway, with or without ISLSCP.
Data sets more suitable to the needs of land
surface modelers than those planned in II3 already exist, and will be improved
before ISLSCP gets there.
The need still exists for long-term global
consolidated data sets, but ISLSCP does not appear to have the capability or
support to deliver them at a high enough level.
Decision: GLASS is
not interested in the staus quo path for future data sets as proposed
by ISLSCP. A major reworking would be necessary for GLASS to be a
major customer of future data products.
Follows PILPS evolution start with easy
synthetic data, then go to obs, and then more sophisticated.
Somewhat independent of GABLS early on as
presented interfacing with them may help.
Action Items:
vdHurk = will lead completion of white paper, for
GEWEX News submission get LoCo on the map.
vdHurk, Peters-Lidard, Houser = Flesh out draft plan into a science and
implementation plan by end of year.Too
ambitious a timeline for experiments, but good to get things rolling (expect
adjustments later)
Dirmeyer = Deliver draft plan to GMPP/WGNE meeting Exeter - 11-14 Oct.
Wrap-up & Next Meeting
No
membership changes perhaps need a global change person?
Next
meeting try to go joint with GABLS, maybe in Netherlands.Bart has offered DeBilt. Final plan TBD.
5th
GLASS Panel