DAOS WG ACTIVITIES

(Submitted by Pierre Gauthier and Roger Saunders)

1.Introduction

The Data Assimilation and Observing Systems (DAOS) WG formed in 2008 is now well established and the group met in Montreal on 8-9 July 2010 to consolidate its activities. The main objectives of the DAOS-WG are:

  • Address data assimilation issues including the development of improved understanding of the sources and growth of errors in analyses and forecasts
  • Promote research activities that lead to a better use of observations and the understanding of their value
  • Provide input and guidance for THORPEX regional campaigns for the deployment of observations to achieve scientific objectives

The working group activities have been presented at a number of conferences and a report on the activities of the group has been published in Rabier et al. (2008).A more recent summary of the activities in which DAOS members have been involved in since the last meeting of the WWRP in Sept 2009 is given here.

2.Meetings

The activities of the DAOS-WG have been evident in several presentations at conferences and meetings during 2009/10. Of particular note were the following meetings:

a)5th WMO Symposium on Data Assimilation, Melbourne Oct 2009

Several members of the DAOS-WG were part of the scientific committee for the 5th WMO Symposium on Data Assimilation that took place on 5-9 October, 2009, in Melbourne, Australia. It was evident that data assimilation has made great advances since the first WMO symposium in 1990. Over 250 attendees were registered with 15 invited talks, 41 oral presentations and 196 posters.Highlights relevant to THORPEXwere:

  • Plans for data assimilation for the next few years are hybrid systems and/or long assimilation windows
  • Continuing improvements from better use of satellite data (e.g. use of radiances over cloud and land)
  • Problems still remain of retaining the beneficial impacts of the data in the short-range forecast

b)25th WGNE, Offenbach Nov 2009

Three members of the DAOS-WG (Gauthier, Hamill, Rabier) are members of WGNE and report on data assimilation activities within THORPEX. They attended the 25th meeting at Offenbach on 2-6 Nov 2009 in parallel with the ICSC-8 meeting.

c)3rd DAOS-WG, montreal July 2010

The main objective of the Third DAOS-WG workshop, held in Montréal, Canada in July 2010, was to review the use of adaptive observations in view of recent results and to assign tasks to write a review paper on the value of adaptive observations for NWP. A draft manuscript to be submitted to BAMS is now available and being reviewed by the group. A summary of the WG discussions on adaptive observations is summarized in the next section. For more details of the meeting visit where the presentations are available.

d)26th WGNE, tokyo Oct 2010

Three members from the DAOS-WG, P.Gauthier, F. Rabier and T. Hamill, attended this meeting and presented a report on DAOS activities. F. Rabier presented an overview of recent developments in ensemble data assimilation and hybrid systems with results sent in by several centres. She also gave a brief summary of results from recent observation campaigns (i.e. AMMA and Concordiasi), and the current status in the preparation of HyMex. She also presentedthe use of IASI cloud affected radiances which have been successfully inter-compared by comparing observation impacts fromdifferent NWP centres. The results will be published in a special issue of the Royal Meteorological SocietyQuarterly Journal. Consideration is now being given as to whether an independent truth can be obtained.

3.Activities

a)Adaptive observations

During the first half of Thorpex, the emphasis has been on the evaluation of the impact of observations, including targeted observations, based on results from field experiments (ATReC, AMMA, IPY, T-PARC), OSEs and OSSEs. The main outcomes from these impact studies may be summarised as follows:

  • The value of extra-tropical targeted data has been found to be positive but small on average
  • Observations taken in sensitive areas have more value than observations deployed randomly
  • Past experiments do not provide evidence of a major impact obtained from just a few observations (when averaged over a large sample of cases)
  • There are limitations to the current assimilation methods to be able to detect the small signal associated with precursors to the development of meteorological systems
  • The methods employed to identify sensitive areas do not appear to be a major problem
  • Additional observations around tropical cyclones have proven to be useful

These studies also suggest that additional benefit may be obtained from:

  • Regional (vs highly localized) and systematic targeting during low predictability flow regimes on a continuous basis (periods of days to weeks)
  • Adaptive processing and data selection of satellite data (e.g. Bormann and Bauer, 2010)

b)Higher resolution regional data assimilation

Of the studies reviewed by the group it is clear that global data assimilation, to provide lateral boundary conditions and background fields for example, is an essential ingredient of any regional forecasting scheme. Currently downscaled 4D-Var analyses provide better results than mesoscale data assimilation systems employing 3D-Var and it is clear that more work needs to be carried out to assess what is really required for mesoscale data assimilation. However, there are potential benefits:

  1. Higher resolution gives better representation of high-impact weather
  2. Higher resolution allows better assimilation and forecast of observed detail
  3. Affordable timely forecasts can be made with regional systems with the most recent observations.
  4. Regional systems provide a basis for tailored numerical weather prediction

There is still much research to be carried out into the optimal configurations for data assimilation in regional and mesoscale NWP models.

c)Intercomparison of Observation Impacts

The DAOS-WG promoted an intercomparison experiment to evaluate the robustness of new adjoint sensitivity tools to measure the impact of observations in NWP forecasts. The results indicated that these new approaches provide more detailed information on the impact of observations which is extremely valuable in the evaluation of the global observing systems. It was also shown that these tools are complementary to OSEs and permit the evaluation of the influence of other observations on the impact of a particular observation type. A paper has been published inMonthly Weather Review which presents the results from this intercomparison to which three centers took part (Gelaro et al.,2010). Other centres are now starting to use the same tools and the DAOS intercomparison is being used as the reference dataset to expand the number of participants in this comparison.

4.Future directions

WGNE acknowledged the role of the THORPEX DAOS WG as the leading expert group for atmospheric data assimilation research in WMO and a key element of WGNE for data assimilation issues that are so important for model development. There are however a number of other issues that are not part of the DAOS mandate which are covered in other programs. SPARC has an interest in data assimilation and so has the GLASS working group which has an interesting project, PILDAS, on surface data assimilation. Mesoscale data assimilation is also an important area and will raise a number of scientific challenges as well. These should notbe included in the mandate of the DAOS but linkages should be made so that the data assimilation issues can have a forum across all interests.

5.References

Bormann and Bauer, 2010Estimates of spatial and inter-channel observation-error characteristics for current sounder radiances for numerical weather prediction. I: Methods and application to ATOVS data. Quarterly Journal of the Royal Meteorological Society, 136: 1036–1050. doi: 10.1002/qj.616

Gelaro, Ronald, Rolf H. Langland, Simon Pellerin, Ricardo Todling, 2010: The THORPEX Observation Impact Intercomparison Experiment. Mon. Wea. Rev., 138, 4009–4025.

Rabier, F., Gauthier, P., Cardinali, C., Langland, R., Tsyrulnikov, M., Lorenc, A., Steinle, P., Gelaro, R., Koizumi, K., 2008: An update on THORPEX-related research in Data Assimilation and Observing Systems. Nonlinear Processes in Geophysics, 15, 81-94, 2008.

Third ThorpexDAOS working group meeting:all presentation have been posted on a website located at

6.Annex A: Current DAOS-WG membership

Pierre Gauthier, co-chair
UQAM, Canada / Roger Saunders, co-chair
Met Office, UK
Carla Cardinali
ECMWF / Ron Gelaro,
NASA, USA / Tom Hamill
NOAA, USA
Tom Keenan
CAWCR, Australia / Michael Tsyroulnikov
HydroMet Centre, Russia / Rolf Langland
NRL, USA
Andrew Lorenc
MetOffice, UK / Bertrand Calpini
MeteoSwiss, Switzerland / Florence Rabier
Météo-France, France
Chris Velden
Univ Wisconsin-CIMSS, USA / *Prof. Bin Wang, Chinese Academy of Sciences, China

* New group member to be formally invited