NOCN03 CWAO 201830
GENOT NO.  007

LE GENOT NO. 008  EST LA VERSION FRANCAISE DU GENOT NO 007.

IMPROVEMENTS TO THE GLOBAL AND REGIONAL DATA ASSIMILATION SYSTEMS AT 
CMC.

AS OF TUESDAY SEPTEMBER 21 2004 AT 1200 UTC, THE FOLLOWING CHANGES TO 
THE GLOBAL AND REGIONAL DATA ASSIMILATION SYSTEMS WILL BE IMPLEMENTED 
AT THE CANADIAN METEOROLOGICAL CENTRE. 

SUMMARY OF CHANGES

NEW OBSERVATIONS WILL BE INCORPORATED IN THE 3D-VAR (3-DIMENSIONAL 
VARIATIONAL) DATA ASSIMILATION SYSTEM. FIRST, IR RADIANCE DATA FROM 
THE 6.7 MICRON CHANNEL FROM THE GOES-12 (EAST) SATELLITE WILL BE 
INCORPORATED, IN ADDITION TO THE GOES-10 (WEST) DATA WHICH IS
ALREADY USED. ALSO, DATA FROM THE AMSUA INSTRUMENT ONBOARD THE NASA
AQUA SATELLITE WILL ALSO BE ASSIMILATED. THIS WILL COMPLEMENT THE 
AMSUA DATA ALREADY USED FOR THE NOAA-15 AND NOAA-16 SATELLITES AND
WILL PARTLY COMPENSATE FOR THE LOSS OF THE AMSUA INSTRUMENT FOR 
NOAA-17 WHICH OCCURRED IN LATE OCTOBER 2003. THERE ARE ALSO SOME
CHANGES TO THE ATOVS (AMSUA AND AMSUB) THINNING PROCESS AND THE 
VARIATIONAL QUALITY CONTROL WILL ALSO BE APPLIED TO ATOVS DATA. 
A THIRD CHANGE IS THE INCORPORATION OF SATELLITE WINDS IN THE
POLAR REGIONS OBTAINED FROM THE MODIS INSTRUMENT ONBOARD THE NASA
AQUA AND TERRA SATELLITES. MODIFICATIONS TO THE DATA SELECTION 
AIMING AT INCREASING THE NUMBER OF SATELLITE WINDS FROM THE GOES-P 
SATELLITE ARE ALSO INCLUDED. FINALLY, DATA FROM SELECTED USA WIND 
PROFILERS WILL ALSO BE ASSIMILATED.

THESE NEW DATASETS WILL BE INCORPORATED IN BOTH THE GLOBAL AND 
REGIONAL DATA ASSIMILATION SYSTEMS. THERE IS NO OTHER CHANGE TO THE 
NWP MODELS OR TO THE DATA ASSIMILATION SYSTEM. ADDITIONAL TECHNICAL 
INFORMATION ABOUT THE DATA AND THE ASSIMILATION PROCEDURES WILL BE 
FOUND ON THE MSC WEB SITE
HTTP://WWW.SMC-MSC.EC.GC.CA/CMC/OP_SYSTEMS/RECENT_E.HTML (IN LOWER 
CASE).

PERFORMANCE EVALUATION

THIS INCORPORATION OF NEW DATASETS IN THE DATA ASSIMILATION SYSTEMS 
HAS BEEN TESTED ON A LARGE NUMBER OF CASES. DATA ASSIMILATION CYCLES 
FOR THE 2 MONTH PERIOD OF 11 DECEMBER 2003 TO 10 FEBRUARY 2004 HAVE 
BEEN PERFORMED. THE NEW SYSTEM HAS ALSO BEEN RUNNING IN PARALLEL 
SINCE LATE JUNE. ANALYSES AND FORECASTS FROM THE PARALLEL GLOBAL
AND REGIONAL SYSTEMS HAVE BEEN VERIFIED AGAINST RADIOSONDE DATA 
OVER SEVERAL OBSERVATIONAL NETWORKS AS WELL AS AGAINST ANALYSES,
AND EVALUATED BY CMC OPERATIONAL METEOROLOGISTS.

ASSIMILATION

THIS CHANGE REPRESENTS A SMALL BUT POSITIVE STEP TOWARD THE 
ASSIMILATION OF NEW DATASETS. IMPACTS ON THE ANALYSES ARE SMALL AND 
OPERATIONAL METEOROLOGISTS DID NOT NOTICE ANY SUBSTANTIAL DIFFERENCE 
BETWEEN THE OPERATIONAL AND PARALLEL ANALYSES. HOWEVER, THE ADDITION 
OF THESE NEW DATASETS IMPROVES THE ROBUSTNESS OF THE ASSIMILATION 
SYSTEM IN THE EVENT OF THE LOSS OF SOME OPERATIONAL DATA. THE ABOVE 
CONCLUSIONS APPLY FOR BOTH WINTER AND SUMMER SITUATIONS.

GLOBAL MODEL FORECASTS

OBJECTIVE VERIFICATION HAS SHOWN SMALL BUT CLEAR IMPROVEMENTS IN THE 
MASS AND WIND FIELDS, PARTICULARLY OVER NORTH AMERICA. THE 
IMPROVEMENT WAS SOMEWHAT LARGER BEYOND DAY 3. THE REGIONS SHOWING
THE MOST IMPROVEMENTS WERE THE ARCTIC AND THE PACIFIC AREAS. THESE
ARE THE AREAS WHERE THERE ARE LESS CONVENTIONAL OBSERVATIONS AND
WHERE THE ADDITION OF SATELLITE DATA CAN CONTRIBUTE THE MOST.
VERIFICATION OF QPF DID NOT INDICATE ANY SIGNIFICANT DIFFERENCE UP 
48 HOURS.

OPERATIONAL METEOROLOGISTS COMPARED THE TWO MODELS DURING THE 
PARALLEL RUN PERIOD. GENERALLY, THEY INDICATED THAT THE TWO SYSTEMS 
ARE OFTEN (65 PERCENT) SIMILAR BUT THEY PREFERRED THE PARALLEL MORE 
OFTEN THAN THE OPERATIONAL. IN THE ARCTIC AND PACIFIC AREAS, THEY
PREFERRED THE PARALLEL 3 TIMES OUT OF 5 CASES SHOWING DIFFERENCES.
THEY ALSO NOTED THAT THE RESULTS WERE MORE NEUTRAL OVER THE CONTINENT
AND THE EAST COAST WHERE BOTH SYSTEMS WERE PREFERRED THE SAME
NUMBER OF TIMES. FOR QPF, THE EVALUATION ALSO INDICATED A SLIGHT
PREFERENCE FOR THE PARALLEL IN CASES SHOWING DIFFERENCES.

OVERALL, THE CMC OPERATIONAL METEOROLOGISTS JUDGED THE PARALLEL 
CONFIGURATION BETTER.

REGIONAL MODEL FORECASTS

ALTHOUGH THE NEW DATASETS HAVE ALSO BEEN INCORPORATED INTO THE 
REGIONAL DATA ASSIMILATION SYSTEM, THE SHORT CUT-OFF TIME OF THE 
REGIONAL GEM FORECAST RUN (T+1H35) RESULTS IN MUCH LOWER DATA 
AVAILABILITY, ESPECIALLY FOR SATELLITE DATA. THEREFORE THE IMPACT 
IS SOMEWHAT SMALLER THAN FOR THE GLOBAL SYSTEM. NONETHELESS, 
OBJECTIVE VERIFICATION DID SHOW A SMALL BUT POSITIVE IMPACT ON THE
MASS FIELD, AGAIN WITH MORE IMPACT OVER THE ARCTIC AND PACIFIC 
AREAS. FOR QPF, OBJECTIVE VERIFICATION DID NOT SHOW ANY SIGNIFICANT 
DIFFERENCE.

THE OPERATIONAL METEOROLOGISTS EVALUATED THE PARALLEL CONFIGURATION 
MASS FIELDS SLIGHTLY BETTER BEYOND 24 HOURS, PARTICULARLY IN THE 
ARCTIC AND PACIFIC AREAS. IN THE ARCTIC AREAS, THE PARALLEL WAS
PREFERRED TWICE OUT OF 3 CASES SHOWING DIFFERENCES. NO CLEAR
CONCLUSION COULD BE DRAWN FROM THE QPF ASSESSMENT.

OVERALL, SINCE THE MASS AND WIND FIELD FORECASTS OF BOTH SYSTEMS ARE 
VERY SIMILAR, IMPACTS ON STATISTICAL GUIDANCE FOR BOTH GLOBAL AND 
REGIONAL SYSTEMS ARE VERY SMALL, AND THEREFORE THERE IS VERY LITTLE 
IMPACT ON SCRIBE.

EVERELL / ADMA / SMA-SMC TORONTO