NAME Modeling and Data Assimiliation R&D Meeting

yellow leaf trees
Jun. 6, 2003

6:00 pm MDT

College Park, MD
Main content

Executive Summary


A NAME Modeling and Data Assimilation R & D workshop was held on June 6, 2003 in College Park, Maryland. The goals of this meeting were (a) to revise, add to and generally complete the Science and Implementation Plan for NAME Modeling and Data Assimilation Research and Development; (b) to establish a time line of specific activities tied to specific individuals / groups. To achieve these objectives, deficiencies and uncertainties of current regional and global models were identified and examined. The path to improved warm season precipitation prediction was discussed

Suggestions and recommendations to improve the “white paper” were made. The revised white paper is attached as Appendix 3. 


1.     Objective and summary of the white paper

The objective of the white paper is to develop a strategy for accelerating progress on the fundamental modeling issues pertaining to the NAME science goals of providing improved understanding and prediction of

  • Warm season convective processes in complex terrain;
  • Intraseasonal variability of the monsoon;
  • The response of warm season atmospheric circulation and precipitation patterns to slowly varying, potentially predictable oceanic and continental surface conditions;
  • The life cycle of the North American monsoon system and its variability.

The guiding principals for the white paper are

  • To take maximum advantage of NAME enhanced observations, and to provide  model-based guidance to the evolving multi-tiered NAME observing program;
  • To maintain a multi-scale approach in which local processes are embedded in,   and are fully coupled with larger-scale dynamics.

The meeting was organized to reflect the strategic approach given in the white paper:

  • Identify the uncertainties and deficiencies of the models;
  • Multi-scale model development;
  • Multi-tier synthesis and data assimilation;
  • Prediction and global scale linkages.


2. Identify the uncertainties and deficiencies of the models

       The NAMAP activity provides an opportunity to assess the performance of current models, in particular

  • To identify and describe inter-model consistencies and differences;
  • To tentatively suggest physical explanations for differences;
  • To provide measurement targets for the NAME 2004 field campaign; and
  • To examine effects of core monsoon (Tier I) convection differences on the larger-scale (Tier II) circulation.

There were 7 groups participating in NAMAP to simulate or forecast monsoon development in 1990 June-September. While all models are able to produce rainfall over the core monsoon regions, the details differ. The models produce very different rainfall structure, rainfall maximum/minimum, seasonal cycle and diurnal cycle,

It was recommended that the NAME observations be used to address the following large model uncertainties:

  • The diurnal cycle (precipitation, low-level circulation);
  • The timing of late-day convection and amount of nocturnal rainfall;
  • The structure and location of the Gulf of California low-level jet;
  • Surface quantities (T, LH, SH fluxes);


3. Multi-scale model development

In the monsoon region, there are many mesoscale convective systems. They move in different directions with different speeds. The boundary layer depths over some regions can exceed 3000 m. Observations point to the complexity and multi-scale nature of the systems involved.

     Most models have problems capturing the seasonal evolution of the monsoon and the mesoscale features of rainfall and regional circulation anomalies. These are related to the convection schemes and microphysics, which needs to be improved in the models.

One possibility is to improve parameterizations of convection in current models. Another is to model the convective cloud systems explicitly. This will provide the basic understanding of the dynamical system.  However, the spatial scale for multi scale models is about 1km. Therefore, more work and resources are needed to apply this method to seasonal climate forecasts.


  • Multi scale modeling should be tested for the NAME period;
  • Data assimilation using a mesoscale mode with 10-20 km resolution is needed to provide large scale circulation for multi scale modeling;
  • Model development should focus on land/atmosphere interactions in the presence of complex terrain and ocean-atmosphere interactions.


4. Multi-tier synthesis and data assimilation

     During the NAME period, data impact studies will be performed with and without the NAME data for CDAS (T62 resolution), GDAS (T256 resolution) and EDAS by the NCEP. These products provide a benchmark for other assimilation and model forecasts.

     The data assimilation products depend on the model and the data assimilation scheme used. For example: the 12-36h precipitation forecasts during the data assimilation cycle for models with different convective schemes show very different rainfall amounts and patterns.


  • Various groups (with different models) are encouraged to perform data assimilation using the NAME data.  The model output should be evaluated based on key features of the warm season precipitation regime (diurnal cycle, MCS’s, moisture surges and boundary layer depths);
  • This is a good opportunity to test new assimilation methods such as the ensemble Kalman filtering method. 


5.  Prediction and global scale linkages 

     Outside the Tier I region, the boundary and surface forcing become very important.  Both sea surface temperature anomalies (SSTAs) in the Tropical Pacific and in the North Pacific have influences on the development of monsoon. In addition to the SSTAs, seasonal forecast skill increases when accurate soil moisture is provided at the initial time.

     Key elements of the monsoon system (including tropical cyclones) are modulated by intraseasonal oscillations (such as the Madden Julian Oscillation).  The inability for models to capture the intraseasonal oscillations has a negative impact on seasonal precipitation forecasts.


  • Realistic soil moisture and temperature information is needed to test the impact  of soil conditions on seasonal forecasts;
  • Improve simulations of tropical intraseasonal oscillations and their impact on the monsoon.


6. Other issues and follow-ups

1) The NAME modeling and data assimilation groups should work closely with other national and international programs like CEOP, PILPS and the subseasonal prediction program organized  by GSFC.

2) Investigators interested in NAME modeling should respond to the  OGP/PACS/GAPP  call for  proposals for the coming summer. This will link the specific issues in the NAME modeling  strategic plan to specific groups and activities;

3) Establish performance measures of the NAME modeling activity;

4) Better communication between modeling groups and observational  groups;

5) The “white paper” will be inserted in the NAME Science and Implementation Plan, but it should continue to be revised;

6) The NAME modeling and data assimilation activities should be reviewed during the  SWG5 meeting in Puerto Vallarta, MX.