These minutes have been copied unedited from the Huddle site
Meeting Notes 26 January 2010
Apologies sent: Andrew Woolf
1) Update on actions from 12-Jan-2010
-- [A1/Andrew Woolf] (Subversion repository) Has been set up by OGC staff but with wrong name (metec.dwg); currently changing it and should be available soon.
DONE - svn repository containing empty EA model available at:
NEW Action A8/Dominic Lowe to write and publish EA/SVN integration documentation
-- [A2/Andrew Woolf] (users of geometry/topology as classification axis) Action still open.
DONE - published at: https://my.huddle.net/huddleworkspace/task.aspx?workspaceid=7612824&taskid=7904078
-- [A3/Chris Little] CL will post details of his use cases.
DONE - published at:
-- [A4/Aaron Braeckel] Will send details of Nextgen/NNEW use cases
DONE - published at: http://external.opengis.org/twiki_public/pub/MetOceanDWG/MetOceanUseCases/NextGen_Net-Enabled_Weather_IT_CONOPS-Dec_8_2009.doc
-- [A5/Jeremy Tandy & Andrew Woolf] Create top-level link for collecting use cases for development over next two weeks.
DONE by Aaron
-- [A6/all] Add use cases.
-- [A7/Jeremy Tandy] Will make a post on huddle once the change to wiki structure is made
DONE (by Aaron)
2) Discussion of use-cases to drive development of conceptual model -- work in progress (verbal discussion sought to identify whether these use-cases are representative of the information interoperability challenges)
Mostly informative discussion followed on each of the use cases:
-- 2.1) International Polar Year
This is a use case that allows us to understand how a reseach community collects, exchanges and maintains information from a long campaign of scientific research and observations.
This use case nicely crosses the boundaries between research and operational users.
Action A9/Andrew Woolf to flesh out this use case further using real example from polar year.
Action A10/Chris Little to send email to Øystein Godøym, Norway for further information/ideas for polar year use case.
-- 2.2) Winter maintenance of highways infrastructure
This use case is based on the way the Met Office works with highways agencies in the UK to keep roads gritted/open in winter weather conditions.
The use case starts with observations from roadside sensors along the road network measuring parameters such as surface temperature and condition, ice, ice or water layer thickness along with other sensors measuring parameters such as visibility, windspeed, humitidy, possibly camera images.
Some of these observations are exchanged under commercial licence - this depends(in UK) who maintins the intrastructure. So there are different licensing terms for different observations which needs to be considered in this use case.
These observations plus others are assimilated into a numerical weather prediction model which is then downscaled to a to 1.5 km uk meso model run 4 times daily. Because obs are assimilated into the model, we can apply quailtiy control flags to observations that come out of the meso model.
The output is then post processed to produce information at 2km resolution.
CL) Note that some of the obs may not come in e.g. if sensor is frozen over. Forecasters may add synthetic obs to estimate these values.
At particular sensor sites the model is downscaled to a local model and input into a road temperature model to estimate anticipated road surface temperatures.
Operational forecasters use this information to advise highways orgs whether to grit etc. taking into account anticiapted rainfall (which affects whether the grit works), snow etc.
This use case covers downscaling, analysis and nowcasting, IP rights and publication into 3rd party decision support tools.
-- 2.3) Flood forecasting (more generalized to Civil Contingencies environmental hazard management)
Alternatively could have a fire forecasting use case?
Some of the same principles apply.
(JS?) With fire especially one challenge is the need for a thin client - the pipe is small to the end user e.g. maybe using a PDA over phone network. Any data is modelling challenged by verbosity issues.
(CL) 2 operational domains (hydro/met) interoperating. (may not be the case in fire forecasting.
In the US greater runoff means rivers can swell rapidly, need to evacuate on strength of forecast.
Could we run with 2 use cases? fire, flood?
Action A11/Jeremy Tandy: Jeremy to take offline and see whether to merge (or split) hazard usecases (including hurricane 2.5 below).
-- 2.4) THORPEX / TIGGE
THORPEX is a 10-year WMO international research and development programme to accelerate improvements in the accuracy of one-day to two-week high impact weather forecasts for the benefit of society, the economy and the environment.
TIGGE is sub activity within THORPEX
Model output from different operational centres is brought together to create multi model ensembles.
These models may have different grids, resolutions, parameters levels. Creates many interoperability challenges.
All availble contributions are collected in near real time.
Where there is developing uncertainty in various areas of world this is looked at in more detail. As a response to this uncertainty extra observations and targetting mesoscale (limnite area) models are used to assess what is happening in particular location.
So the use case is:
1) multi-model ensemble
2) trying to do steering with outputs from this model
(AB)Suggest we could generailse this to a multi-model ensemble.
Someone commented that Ben Domenico posted an 'early LEAD example' using ensembles for high-resolution weather forecast model:
-- 2.5) Hurricane Forecasting & alerting
See use case on twiki:
(JT) Particularly interested in notifying the user (the public) at the end of the chain (sms, mobile etc). Again, small pipeline, as for fire forecasting.
(JT) Different products for different users, need considering in the use cases.
(JS) National weather service bringing in probablistic aspects to weather forecasting to provide a range of likely outcomes. In hurricane forecasting - cost benefits of evacuation are important.
-- 2.6) Climate impact assessment - water stress in sub-saharan africa (in 10 20 50 years time)
Based on UK Met Office work, but could be generalised.
This is a classic climate security question: migration patterns, land use issues.
In terms of data it uses climate datasets, ensembles, probablistic outputs. also has linkages to gis with demographics, cadastral boundaries etc. (interoperability across thematic domains).
(AB) We could/should talk generally about climate impact, then specific example.
-- 2.7) Subsistence farming in sub-saharan africa
Similar to 2.6 but more focused on dissemination of results: particularly to large number of people with poor or no network connections.
(CL) 2.6 and 2.7 could be part of the same use case
-- 2.8) Regulated aviation - terminal aerodrome forecasting
This is some thing we do today - an operation forecasting paradigm where we create particular data products conforming to a particular standard.
(JT) In terms of new projects something that sits between current opmet and where aaron is going with Ceasar and nextgen?
(JT) Forsee that we need to maintain backward compatability with TAF and METAR exchanges for at least 10 years.
Therefore products need to be automatiable. So can exchange old or new products according to requirement of bilateral agreements.
-- 2.9) NNEW 4D weather cube for aviation
(AB) This could be more generally written as a weather cube for aviation.
General discussion on use cases:
(AC) - What is the lifespan of these usecases?
(JT) -- In society we think these use cases will be around in 5 years - they are representative of the things people are doing or starting to do now. In the Metocean WG the goal is to create a canditate model to demonstrate the right direction then do encoding and field testing.
We anticipate ~ 2 further years of development before it's used in an operational environment.
(someone?) It's easy to focus on operational communities but we should also think about stuff outside within the wider environement - how it can be exploited across domain (e.g sub saharan africa use case)
Discussion about whether usecases could merge with wms usecases. General opinion was no at this time.
(AB?) Is there any area we haven't covered.
(CL) we could generalise flood forcastign to include tidal forecasting which couples met and ocean,
(JT) In toulouse we agreed to focus on met use cases first.
Decision D1: We will focus on Meterological Use Cases as agreed in Toulouse.
3) Any other business
None - no time.
4) Arrangements for next meeting
Action A12/ Marie-Francoise to organise next telecon in 2 weeks.
Summary of Open Actions and New Decisions:
-- Action A8/Dominic Lowe to write and publish EA/SVN integration documentation
-- Action A9/Andrew Woolf to flesh out this use case further using real example from polar year.
-- Action A10/Chris Little to send email to Øystein Godøym, Norway for further information/ideas for polar year use case.
-- Action A11/Jeremy Tandy: Jeremy to see whether to merge (or split) hazard usecases (including fire/flood/hurricane etc).
-- Action A12/ Marie-Francoise to organise next telecon in 2 weeks.
-- Decision D1: We will focus on Meterological Use Cases as agreed in Toulouse.
- 09 Feb 2010