Talk:Integrated Global Dataset

TKeating:TF HTAP: Seeking Advice on Funding for Geneva Follow-up

Dear TF HTAP Contributors, André and I have both been able to secure a small amount of funding (200K Euros and $200K, respectively) to invest in follow up efforts to the TFHTAP/WMO/GEO workshop in Geneva. We both need to make some quick decisions about how to spend these funds. I am writing to solicit your thoughts on what investments would be most useful in supporting the cooperative modeling and analysis under the TF HTAP?

One of the recommendations of the Geneva workshop was to develop a database (perhaps distributed) of observations (surface, aircraft, satellite, ...) suitable for evaluating models of intercontinental transport and hemispheric pollution. This database could also be part of a larger "model evaluation testbed" that would include protocols and tools for comparing models to observations.

We recognize that there are a number of completed or ongoing efforts that provide a piece of what is needed. The challenge that we are facing is identifying the most useful next steps. From a practical standpoint, we are not able to combine these funds and will need to fund our combined effort as two projects. We are both likely to need to award the funds through open competitions. André's funds are likely to be available in the summer and will need to be awarded before the end of this year. My funds will be available sometime after October 2007, hopefully!

We need to identify specific products that we would like to produce, and the various connections to other efforts that may need to be built into the projects. One initial division of effort may be to have Andre's funds focus on surface observations and to have my funds focus on satellite observations, with aircraft and other data fitting into either project where it is most appropriate. This division is suggested because there may be different needs associated with the different types of data and my impression that there may already be infrastructure for dealing with the different types of data that we can build on in Europe and North America respectively.

A quick reply with wise thoughts would be appreciated! (Not asking for much!) If you would like to discuss any of this by phone, please feel free to call. Thanks for your help, Terry, 22 March 2007 (EDT)

....RudolfHusar:TF HTAP: Seeking Advice on Funding for Geneva Follow-up

Hello Terrry, Andre, HTAP: This is a quick response to your request for comments on the possible followups to the Geneva HTAP meeting.

The goal of creating a robust observational database suitable for HTAP model evaluation seems still appropriate, timely and rewarding in the long run. However, creating such an integrated database should involve three equally important activities to make it suitable for model-data comparison: (1) data collection from the providers, (2)further quality assurance and (3) integration. In other words, focus only on (1) would probably not yield the desired quality and robustness.

At the breakout session of the Integration Group (Chapter 6), we agreed that the IGACO framework for data integration is a good starting point for the HTAP. I particularly liked the fact that in IGACO, Quality Assurance is an explicit component of the framework. A version of the IGACO Framework adopted/annotated for HTAP is attached for illustration. At this time refrained from cluttering the slide with IT issues with data flow and processing. Hope this helps a bit. Rhusar 22:57, 25 March 2007 (EDT)

....HajimeAkimoto:TF HTAP: Seeking Advice on Funding for Geneva Follow-up

It would be good to spend the funds for the database which will be used for model validation. Meanwhile, I think it would be a lot important to draw a grand design of model intercomparison how to quantify intercontinental and hemispherical transport. In my understanding, even though the ensemble mean of many global CTMs agree pretty well with observational data of i.e. surface ozone, the deviation between models are substantial. Nevertheless, the cause for the disagreement between models, and some models and observational data has never been identified in the past model intercomparison study in general.

Although I can easily imagine the difficulty of asking participating each modeler to identify the reason of disagreement and to tune the model, agreement between models, and between models and observations are critically important for the issue of evaluating fluxes of intercontinental transport/hemispherical pollution since the issue is a kind of trans-boundary air pollution between countries and regions. Some part of funds should be spent for drawing such a grand design, and discussion of useful protocol of such intercomparison before compiling observational data under TF-HTAP. Hajime Akimoto, March 28, 2007

....MartinSchultz:TF HTAP: Seeking Advice on Funding for Geneva Follow-up

Dear Terry et al., Thanks for the message. We would be interested to obtain a small contribution (e.g. 30000 €) in order to fund the extension of the Juelich data server and some service support of this (installation of software, development of a web interface). Martin Schultz, March 28, 2007

....RudolfHusar:TF HTAP: Seeking Advice on Funding for Geneva Follow-up

Martin, thanks for breaking the ice and speaking out on specific € requests. A very, very refreshing followup on the open request for ideas by our leaders. We at CAPITA are herewith requesting $40k (30K €) support for HTAP-specific activities: collaboration with the HTAP group on standard data interfaces; participation in data integration, IT support and some international travel. As in the case for most of us in HTAP, the design and implementation labor associated with data access, homogenization and delivery is financed by other projects.

As we discussed at the Saturday workshop in Geneva, dynamically linking the Juelich, DataFed and other HTAP data providers/mediators through standard interfaces would be an effective way to pursue the HTAP data integration. Also, it would serve as a practical demonstration of GEO-like systems-of-systems integration, as confirmed by Brandon Kelly in this thread.

Please note that I took the liberty of transferring the e-mail thread on the HTAP Integrated Dataset to an open discussion page on the ESIP wiki. Its the same wiki where both Christiane Textor's CF Naming and the AeroCom discussions, lead by Michael Schulz are conducted. If this is a problem for anyone, please let me know, I will remove the sensitive items. Rhusar 00:28, 31 March 2007 (EDT)

....BrendanKelly:TF HTAP: Seeking Advice on Funding for Geneva Follow-up

Hi everyone, First of all, thanks to André and Terry for securing this funding. From the point of view of the GEO process, the development of a comprehensive, integrated database of observations (surface, aircraft, satellite, etc) suitable for evaluating models of intercontinental transport and hemispheric pollution would be an excellent initiative. In particular, if it involved federating major regional datasets and ensuring they were interoperable, this would be a major achievement. Given the challenge involved in relation to the resources available, it might be appropriate to begin with a subset of the data considered a priority by the modelling community, as a demonstration case. The benefits highlighted by such a demonstration, as well as the experience gained in carrying out this process, could be a stimulus for other investments to further progress the initiative. Best regards, Brendan, March 30, 2007.

....OliverWild:TF HTAP: Seeking Advice on Funding for Geneva Follow-up

Dear All, The construction of a database of observations suitable for testing models of intercontinental transport would certainly be very valuable. The collation and quality control of observational data are useful for the modelling community, but are not of themselves sufficient for testing models thoroughly. Two issues we should consider:

(1) We need to identify key process weaknesses in our models, and this would be a lot easier if we had access to measurements where individual processes could be isolated (as the CCMVal project attempted to do in the stratosphere). Examples of this might be identification of clear LRT episodes at surface sites, aircraft measurements in conditions when convection or frontal processes dominated vertical lifting, or clear examples of mid-tropospheric transoceanic advection of NOx/aerosol/etc from satellite. A process-oriented evaluation would be much more informative than the crude testing we do against observations (and their variability) at the moment.

(2) Rudy highlighted quality assurance, but equally important is the appropriateness of the data for the purpose in hand (here HTAP) - not all data is equally useful. Identification of the temporal and spatial scales over which particular measurements are representative is important. Surface data is affected by PBL mixing and deposition, aircraft data by sparsity/representativeness, satellite data by little vertical information. Characterization or selection of data based on their appropriateness for LRT would be extremely useful.

These projects bridge the divide between observations and modelling, as some degree of analysis is required to select data that might be appropriate for process-oriented evaluation. Note that this approach would also make a valuable contribution to the AC&C initiative, where there is some intention to focus on model representation of specific processes. Cheers, Oliver, April 2, 2007