From: 	Bruneau,Richard [CMC]
Sent: 	March 6, 1998 12:09 PM
To: 	genot_metsis_changes [majordomo list]
Cc: 	Richard.Bruneau@eccmc3.cmc.ec.gc.ca
Subject: 	genot nocn02 no.005

NOCN02 CWAO 061930
GENOT FAX NO.005

GENOT NO. 006 EST LA VERSION FRANCAISE DE GENOT NO.005.

MODIFICATIONS TO THE SNOW ANALYSIS SCHEME AT CMC...

AS OF WEDNESDAY MARCH 11 1998 AT 1200 UTC, IMPROVEMENTS
TO THE OBJECTIVE SNOW DEPTH ANALYSIS SCHEME WILL BE
IMPLEMENTED AT THE CANADIAN METEOROLOGICAL CENTRE. THE
MODIFICATIONS TO THE NEW SCHEME WILL BE:

1.  INCLUSION OF INTERMEDIATE HOUR OBSERVATIONS FROM
     SYNOPTIC REPORTS, METAR AND SA REPORTS AS AVAILABLE.

2.  THE ANALYSES WILL BE DONE EVERY 6 HOURS (INSTEAD OF
     EVERY 12 HOURS).

3.  THE FIRST GUESS USED TO PERFORM THE ANALYSES (THE
     PREVIOUS SNOW DEPTH ANALYSIS) WILL BE MODIFIED BY ADDING A
     SNOWFALL ESTIMATE BASED ON A 6-HOUR FORECAST FROM THE GLOBAL
     MODEL.

4.  IMPROVEMENT OF THE SNOW MELT ALGORITHM.

5.  INCLUSION OF AN ESTIMATE OF SNOW PACK DENSITY.

6.  THE RETURN TO CLIMATOLOGY WILL BE ELIMINATED IN THE NEW
     ANALYSIS. THE COMBINATION OF THE SNOW MELT ALGORITHM WITH
     THE SNOWFALL ESTIMATE FROM THE MODEL PROVIDES BETTER RESULTS
     THAN THE RETURN TO CLIMATOLOGY.

7.  HORIZONTAL RESOLUTION WILL BE IMPROVED TO 0.33 DEGREE
     (INSTEAD OF THE PREVIOUS 0.9 DEGREE).

8.  TO IMPROVE THE ANALYSES IN MOUNTAINOUS AREAS, A
     VERTICAL CORRELATION BASED ON THE DIFFERENCE IN ELEVATION
     BETWEEN REAL STATIONS AND MODEL TOPOGRAPHY WILL BE INCLUDED.
     THIS VERTICAL CORRELATION WILL ALSO BE USED IN THE
     PREPARATION OF THE TEMPERATURE FIELD USED IN THE SNOW
     ANALYSES (SNOWMELT AND SNOWFALL ALGORITHMS).

THIS NEW SNOW ANALYSIS HAS BEEN RUN IN TEST MODE FOR
THE LAST TWO WINTERS AND ALSO DURING A ONE MONTH
PARALLEL RUN. THE NEW SNOW DEPTH ANALYSES, AND
FORECASTS FROM THE GEM MODEL INITIALIZED FROM A
REGIONAL 3D-VAR ANALYSIS SYSTEM WHICH INCLUDED THE NEW
SNOW ANALYSES HAVE BEEN ASSESSED BOTH OBJECTIVELY, AND
SUBJECTIVELY BY CMC OPERATIONAL METEOROLOGISTS.

OVERALL, THE NEW SNOW ANALYSIS HAS BEEN FOUND TO BE
MUCH MORE DETAILED AND REALISTIC THAN THE CURRENT
OPERATIONAL SNOW ANALYSIS; THE INCREASE OF RESOLUTION,
THE USE OF THE MODEL SNOWFALL ESTIMATE, AND THE
INTRODUCTION OF THE VERTICAL CORRELATION PROVIDE A MORE
DETAILED SNOW DEPTH ANALYSIS AND MUCH MORE REALISTIC
DEPTHS OF THE SNOW PACK OVER MOUNTAINOUS AREAS.

FORECASTS FROM THE GEM MODEL INITIALIZED FROM AN
ANALYSIS SYSTEM WHICH INCLUDED THE NEW SNOW ANALYSES
ARE QUITE SIMILAR TO THE ONES PRODUCED BY THE MODEL
USING THE CURRENT OPERATIONAL SNOW ANALYSIS. THE ONLY
DIFFERENCES ARE NOTICED TO BE IN THE VERTICAL
TEMPERATURE STRUCTURE IN THE BOUNDARY LAYER IN REGIONS
WHERE SNOW DEPTHS BETWEEN THE TWO ANALYSES WERE
DIFFERENT ENOUGH TO INDUCE SIGNIFICANT DIFFERENCES IN
THE SURFACE ALBEDO FIELD USED BY THE MODEL TO CALCULATE
SURFACE FLUXES. IN THESE REGIONS, IT IS THE FORECAST
SURFACE TEMPERATURE FROM THE MODEL THAT EXHIBITED THE
GREATEST SENSITIVITY TO THE NEW ANALYSIS. SINCE THE NEW
ANALYSIS IS MORE REALISTIC, MOST OF THE TIME THE
DIFFERENCES IN THE FORECASTS WERE TO THE ADVANTAGE OF
THE PARALLEL SYSTEM. HOWEVER, SINCE THERE ARE VERY
LITTLE DIFFERENCES IN THE FORECASTS OF THE MASS FIELDS,
THE IMPACT ON THE STATISTICAL FORECAST PRODUCTS USING
THE CMC OPERATIONAL MODELS AS DRIVING MODELS SHOULD BE
MINOR.

MCBEAN/ADMA/TORONTO