Ronald Gelaro (PI)
NASA Goddard Space Flight Center
ron.gelaro@nasa.gov

Atmospheric Data Assimilation Development

The GMAO proposes to develop a next-generation atmospheric data assimilation capability to meet NASA’s goals of maximizing the use of satellite observations to advance our understanding of processes related to climate variability and change, improve our modeling and prediction of the earth system and define the future earth observing system.  The overall thrust of the techniques we propose aims to increase the use and effectiveness of satellite observations with high temporal and spatial resolution in the next generation system, with an emphasis on observations related to key physical processes such as clouds and precipitation.  The proposal will encompass work in both data assimilation and model development directly related to the improved use of these data types.  The proposal will take as a starting point the GEOS-5 data assimilation system, comprised of the GMAO atmospheric model and the grid-space statistical interpolation (GSI) analysis system.  The GSI in its present form is a 3-dimensional variational (3D-VAR) analysis system being developed jointly by NCEP and GMAO for operational implementation at both institutions during the early part of 2005.  The GSI provides an extremely flexible initial framework from which the proposed next generation system will be developed.  This framework will be extended to incorporate both time dependent observation operators (4D-VAR) and flow-dependent background errors.  The former is important for improving the use of satellite observations with high temporal resolution, while the latter is important for improving the projection of the observed information onto the proper spatial scales and multivariate patterns in the analysis.   The proposed system will evolve in a controlled manner to ensure that NASA (i) has a continually maintained world class assimilation system for credible scientific investigations by internal and external partners, (ii) is able to meet its ongoing commitments to produce high quality analysis and forecast products for NASA instrument teams and other customers and (iii) extracts the maximum benefit from the NASA-NOAA collaboration in data assimilation by ensuring that contributions from both agencies can be quickly and directly transitioned between them.   The proposal will be organized in terms of three major work components. Part I will address the required scientific support for maintaining and extending the current assimilation system infrastructure to include time dependence, model bias correction, flow-dependent covariance models and new satellite data types.  Part II will focus on the experimental testing and development work required to incorporate new satellite data types, including the development of observation operators, bias correction and quality control.  In Part III we will develop an advanced suite of tools for diagnosing system performance and sensitivity. The use of adjoint versions of both the forecast model and analysis system will be key aspects of this effort.

+ Back to Participant Listing