NOTE: The nws-changes mailing list is no longer active. The list archives are made available for historical reasons.
ZCZC WSHPNSNMC ALL
TTAA00 KWSH DDHHMM
NATIONAL TECHNICAL INFORMATION MESSAGE 94-08
NATIONAL WEATHER SERVICE HEADQUARTERS
145 PM EDT MON APR 11 1994
TO: ALL NWS OFFICES, DOMESTIC DATA SERVICE SUBSCRIBERS
FROM: MARY C. NEWTON
ACTING CHIEF, SERVICES DEVELOPMENT BRANCH
SUBJECT: NEW MAX/MIN TEMPERATURE EQUATIONS TO BE IMPLEMENTED
IN THE MRF-BASED STATISTICAL FORECAST MESSAGES
(THE FOLLOWING INFORMATION WAS PROVIDED BY JOHN JENSENIUS, JR. OF
THE TECHNIQUES DEVELOPMENT LABORATORY.)
Beginning on April 14, 1994, during the 0000 UTC cycle, new
equations to predict maximum (max) and minimum (min) temperature
from the medium-range forecast (MRF) model will be implemented.
These equations will be used to produce the max/min forecasts
that appear in the MRF-based statistical forecast messages which
are distributed on AFOS under the FMR category and disseminated
to external users as the FOXCxx, FOXExx, FOXSxx, FOXWxx, and
FOUS20 messages, where xx is a two digit number identifying each
message. The min/max forecasts appear in the line labelled
"MN/MX" in the message.
The new forecasts are based on a calibrated MOS approach in
contrast to the calibrated perfect prog approach used to derive
the original equations. In the older equations, the 1000-500 mb
thickness was the main thermal predictor; consequently, the older
equations did not perform well in situations of low-level cold
air or low-level warm air. In contrast, the new equations
contain low-level thicknesses and temperatures as predictors and
are much more responsive to the low-level thermal structure of
the MRF model. The new equations also use observed surface
temperature data as predictors at the earlier projections.
Recent comparisons of the two sets of forecasts have shown that
differences of 10 to 15 degrees F between forecasts generated by
the old and new methods are not uncommon in certain synoptic
situations. Independent test results for the contiguous U.S.
from the winter of 1992-93 showed that the new equations averaged
about 1.6 degrees F better (in terms of mean absolute error) than
the older equations at the 24-h projection. In general, after
about 96 hours, there is little difference in the overall
accuracy between the old and new equations, however, the new
equations continue to be more responsive to the MRF's low-level
thermal structure. In comparison, results for the 1992-93 winter
also showed that new MRF-based max/min equations were roughly
equivalent in accuracy to new NGM-based max/min equations (not
yet operational) at the 36-h and 48-h projections and better than
the NGM-based equations at the 60-h projection.
With the implementation of the new equations, max/min guidance
will become available for Arcata, CA (ACV); Annette Island, AK
(ANN); Caribou, ME (CAR); Harrisburg, PA (CXY); and Long Beach,
CA (LGB). Due to a lack of developmental data, max/min forecasts
will no longer be made for Barter Island, AK (BTI). These
stations are contained in the collectives listed below:
AFOS PIL OF WMO HEADER OF
STATION COLLECTIVE COLLECTIVE
Arcata, CA cccFMREKA FOXW42 KWBC
Annette Is., AK N/A FOUS20 KWBC
Caribou, ME cccFMRPTX FOXE56 KWBC
Harrisburg, PA cccFMRCTP FOXE59 KWBC
Long Beach, CA cccFMRLOX FOXW50 KWBC
Barter Is., AK N/A FOUS20 KWBC
Technical Procedures Bulletin No. 411 describes the MRF-based
statistical guidance message and contains a description of the
line containing the max/min temperature guidance. If you do not
have a copy of this bulletin and would like a copy, please
contact Chris Alex of the Office of Meteorology at
(301)-713-0462. If you need further information about the
max/min temperature guidance or about the medium-range
statistical forecast messages, please call John Jensenius of the
Techniques Development Laboratory at (301)-763-8151.
END
nws-changes archives: