Back to AMWG Homepage

Draft Convection Development Plan

Version 1.0 [19 June 2006]

Contents

Motivation Timeline Standards Objectives Simulation Requirements Process

1) Motivation for this document:

• To provide the climate modeling community with a pathway for contributing to the development of the next generation Community Climate System Model
• To set clear requirements for the community for contributing to the CCSM development process. The requirements encompass time constraints, scientific issues, and coding/software engineering matters.

We strongly recommend that the developers contact the AMWG liaison (Richard Neale) to arrange for a design review of their parameterization. This will help identify any possible issues in its interaction with and integration into CAM before extensive development.

2) Timeline for the development of CCSM4:

CAM4 time schedules are controlled in part by the overall development of CCSM4 (which is still under discussion). Here is our best guess at the timeline.

• Summer 2006: requirements document and metrics discussed at CCSM workshop
• Spring 2007: candidates for new parameterizations must be submitted to AMWG in a form that meets technical requirements under (3)Standards
• Summer/Fall 2007: Preliminary evaluation of paramerizations released by AMWG.
• Fall 2007: shortlist candidates for development
• Spring 2008 AMWG: decision on convective scheme for CAM4

Back to Contents

3) Standard Parameterization requirements:

a) Conservation: All convective (and other) parameterizations must conserve:

• Mass
• Total water
• Energy

b) Numerical: All parameterizations must show:

• Numerical stability with respect to variations in time step.
• Numerical stability with respect to variations in vertical resolution.
• Convergence with respect to increased vertical resolution
• Linear stability with respect to perturbations in input state.

c) Technical:

• The parameterization should be implemented in CAM using an interface conforming to the physics/chemistry interface design adopted since CAM3.1. (Likely 3.3.X)

• Must run in lahey with debug options. Must restart appropriately, must be stable and bit-for-bit with multiple processors, single processors, etc. Must be stable for multiple resolutions and dynamical cores and scale for large numbers of processors.

• CMS needs to be able to reproduce the runs by dropping your modified routines into a ëmymodsí directory and firing off the standard CAM/CCSM run scripts, reproducing your output.

• The intent is that CMS staff need to be able to reproduce the runs by dropping modified routines into a ëmymodsí directory and running standard CAM/CCSM run scripts, reproducing your output on multiple configurations on multiple platforms.

• The code is documented in such detail that can be reproduced/rewritten easily. Documentation includes equations as well as comments in the code.

• It is likely that there may be requirements for Output Variables from the scheme (such as sub-grid vertical velocity).

d) Documentation:

The new parameterization must be accompanied by a complete mathematical description of the equations and closure suitable for inclusion in the Technical Report describing CAM4. (eventually) it should appear in peer-reviewed literature describing the parameterization and its effects on the CAM and CCSM simulations.

Back to Contents

4) Scientific objectives:

The new convective parameterizations (deep and/or shallow) should have a sound theoretical basis and should help alleviate existing biases in CAM3/CCSM 3 related to convective parameterizations. Current model biases related to convective parameterizations in CAM3/CCMS3 are:

• Double ITCZ structure in the Tropics.
• Inadequate seasonal meridional migration of tropical precipitation.
• Inadequate longitudinal and vertical distribution of precipitable water.
• Anomalously cold tropical tropopause.
• Inadequate precipitation distribution over the United States.
• Poor representations of MJO and ENSO.

The resulting simulations will be evaluated using an overall list of metrics defined below. In addition the simulations will be evaluated using criteria specific to convective parameterizations.

a) Overall simulation metrics: This will be a part of the metrics document. It is likely they will include such items as:

• ENSO
• MJO
• US diurnal cycle

The intent is not to require that the scheme 'succeed' in reproducing these quantites, but that the metrics for them be evaluated against other schemes.

b) Additional evaluation criteria for convective parameterizations:

• Variability: Wheeler and Kiladis diagram + analysis of Ricciarduli and Garcia
• Constituent transport
• Conservation of ëwater tracersí
• momentum transport/sub-grid vertical velocity

Back to Contents

5) Simulation requirements:

New parameterizations must be evaluated against the defined CAM and CCSM control simulations and selected observations (metrics)

• CAM control simulation = FV (1.9 x 2.5o), 26 level CAM 3.2 (available shortly after CCSM workshop)
• CCSM control simulation = FV (1.9 x 2.5o), 26 level CAM 3.2 (This simulation can be carried out in collaboration with CMS once the CAM control simulation is complete).
• IOP cases for ARM summer & tropics

Back to Contents

6) Selection process:

Candidate parameterizations will be evaluated first by the AMWG subject to objective criteria and tests described in this document.

Back to Contents

Back to AMWG Homepage