ntim on changes to Eta model

• To: nws-changes@xxxxxxxxxxxxxxxx
• Subject: ntim on changes to Eta model
• From: "Unidata User Support" <support@xxxxxxxxxxxxxxxx>
• Date: Thu, 08 Sep 1994 16:08:35 -0600

ZCZC WSHPNSNMC ALL
TTAA00 KWSH DDHHMM

NATIONAL TECHNICAL INFORMATION MESSAGE 94-41
NATIONAL WEATHER SERVICE HEADQUARTERS
210 PM EDT THU SEP 8 1994

TO:         USERS OF NWS' Eta MODEL

FROM:       FRED ZBAR
            ACTING CHIEF, SERVICES DEVELOPMENT BRANCH

SUBJECT:    CHANGES TO THE OPERATIONAL Eta MODEL

(THE FOLLOWING INFORMATION WAS PROVIDED BY ERIC ROGERS, NMC
DEVELOPMENT DIVISION)

At 1200 UTC on September 7, 1994 a series of changes to the
analysis component of the operational Eta model were implemented.
In addition, the Eta model topography was modified.  A brief
description of these changes follows.

1) ANALYSIS CHANGES

The optimum interpolation analysis is now performed directly on
the Eta model grid points rather than on a latitude/longitude
grid, avoiding unnecessary interpolations.  Surface observations
within 25 mb of the model terrain are now included in the
analysis.  Previously, only those surface observations at or
above the model terrain were used.  The sensitivity of the
analysis near the surface was improved to allow for better
treatment of these data.  The treatment of upper-level wind
increments in the analysis was modified by allowing for tighter
geostrophic coupling with the analyzed geopotential height
increments.  Forty-six pieces of data (up from 30) are now used
in the optimum interpolation analysis at each model grid point. 
Finally, the analysis was modified to give more weight to
observations in closer proximity to the analysis grid point.

2) TOPOGRAPHY CHANGES

The terrain specification in the Eta model was modified to fit
more closely with the observed topography.  The new method
produces a more realistic gradient in the surface elevation,
especially in the regions adjacent to locally high terrain
(Colorado, Sierra Nevada, and the Wasatch Range near Salt Lake
City, for example).

3) IMPACT

These changes produced small positive impact on model performance
in all aspects during parallel testing.  With the improvements in
the analysis of surface observations and the model terrain, the
greatest impact was seen in  the analysis of low-level parameters
(temperature, moisture), especially in cases of cold air damming.
More modest improvement was observed in precipitation skill
scores and in forecast verification scores against rawinsonde
data.

If you have any questions regarding these changes, please call
Eric Rogers at (301)-763-8161.

END