***************************
			 **   THE GEM MESSENGER   **
			 ***************************
						     Volume 10, Number 36
						     September 1, 2000

---------------------------
A GGCM Metrics Solicitation
---------------------------
From: George Siscoe <siscoe at bu.edu>

During the GGCM Steering Committee meeting at the 2000 GEM workshop in
Snowmass, Rich Behnke sent through Kile Baker a behest mandating that
GEM's GGCM project obtain a calibration of the various models within its
purview against the newly formulated, NSF-commissioned metrics for
magnetospheric modeling.  This is an open invitation soliciting all
magnetospheric modelers interested to participate in the magnetospheric
modeling calibration exercise.

The Metrics Committee, which NSF charged with recommending a set of
metrics that cover areas of research essential to the National Space
Weather Program, put as its highest priority metric for magnetospheric
modeling the transpolar potential.  Accepting this as the target metric,
the GGCM Steering Committee, wishing also to make this first exercise
direct and simple, propose to implement the following plan.  For
specified dates and times of events to be furnished (by criteria and
means explained below), each participating modeler will compute the
electric field vector in 100 km averages along the trajectory of the
DMSP satellite (to be provided), measurements from which will serve as
ground truth.  Models will asked to give separately the polar cap
potential defined as the difference between the minimum and maximum
potentials along the DMSP trajectory.  An absolute score will be
computed from the differences between model-computed averages and
empirical averages.  Some normalization algorithm, yet to be devised,
will convert the absolute differences into a skill score.  (From the
point of view of the National Space Weather Program, it would be good if
the initial skill scores that this exercise will yield were low, so that
there would be something for the NSWP to improve.  Of course, each
modeler wants a high skill score, so there is an unavoidable divergence
of interest here.  If the skill scores are high, it means that the
chosen metric is insufficiently challenging.  The exercise is,
therefore, in part a test of the metric.)

AFRL has agreed to select candidate events to be used in the exercise,
and to collect and 'scrub' both the input data and the DMSP data.  The
input data will include solar wind and indications of the prior state of

the magnetosphere.  Events must be able to provide input data needed by
participating models.  All modelers will have access to all data.  From
the candidate events, a number will be selected by consensus among
participating modelers.  The objective is to identify for this first
exercise uncomplicated events that epitomize ideal situations, such as
long periods during which the IMF holds fixed, pointing in a cardinal
direction.  High priority must, of course, be given to the quality of
the data available during an event.

CCMC has agreed to act as the repository and distribution center for the
data pertaining to the events, and it will compute absolute scores and
skill scores.

Rich Behnke would like to see the results of the comparisons at next
year's Space Weather Week at SEC, which is the week of May 1, 2001.  The
following time schedule is intended to allow us to meet this deadline.
Now-Nov. 1. AFRL Preliminary event selection, meet with modelers (or
some combination of e mail, fax, and telecom), final event selection.
Nov. 1-Dec. 1: AFRL collect data files, event scrubbing, and delivery to

CCMC
Dec. 1-Jan. 15: CCMC preparation of data for distribution
Jan. 15-April 15: Model runs
April 15: delivery of model results to CCMC
April 15-May 1: CCMC analysis of results
May 1: Presentation of results at Space Weather Week

Modelers interested in participating in the calibration exercise please
contact siscoe at bu.edu

+-------------------------------------------------------------------------+
|To add name to the mailing list or for a message to the GEM community    |
| please contact: editor at igpp.ucla.edu                                    |
|                                                                         |
|URL of GEM Home Page: http://www-ssc.igpp.ucla.edu/gem/Welcome.html      |
|Please update your e-mail address.                                       |
+-------------------------------------------------------------------------+