Meteorological Service of Canada Status Report

1. 17th NOAM-Europe Data Exchange Meeting Meteorological Service of Canada Status Report Montréal, Canada, May 26-28, 2004 Gilles Verner Chief, Data Assimilation and Quality Control Canadian Meteorological Centre

2. Outline • Current status: • Canadian Forecast System • Telecoms and Computers • NWP System at CMC • MSC Observing Networks and Canadian AMDAR • Changes to NWP System since last meeting • Results from AMSU-B / GOES radiances • Results from 15-km regional GEM • Results from new OSE’s • Future Plans and Requirements for Data • Part II presented by Godelieve Deblonde

3. MSC Forecast Offices and Satellite Reception Facilities Each MSC Office has a GOES Direct Readout Station. CMC has 3 GOES Readout Stations. MSC operates 3 HRPT Stations: Halifax, Gander and Edmonton. Data from a non-MSC Northern station is also available (Resolute). CMC has DOMSAT system for AVHRR-GAC (TeraScan)

4. MSC Forecast Offices and Satellite Reception Facilities Each MSC Office has a GOES Direct Readout Station. CMC has 3 GOES Readout Stations. X X X X MSC operates 3 HRPT Stations: Halifax, Gander and Edmonton. Data from a non-MSC Northern station is also available (Resolute). CMC has DOMSAT system for AVHRR-GAC (TeraScan) X X X X

5. CMC Computer Center Configuration

6. Telecommunications • GTS link CMC - NWS: upgraded from 64 kbps to T1 during 2003 • Upgrade was done for exchange of NWP model output (for EPS) and Canadian Radar data • Former NAIL link between Ice Centres (CIS - NIC) removed • Link UKMO-CMC still active (Meteosat - GMS - GOES) • Eumetsat - CMC link for EARS • Internal MSC links upgraded (radar, hrpt, NWP) • Internet still used for access to Nesdis data; CMC internet access upgraded during 2003 (100 mbps and 2 independent ISP) • Funding for high speed link (T1+) with Nesdis approved

7. global assimilation cycle 3D Var Analysis (data collection and assimilation) Models, 6-hour forecast (trial field) surface analysis GEM model Global (100km) : 0-10/15 days Regional (15km) : 0-48 hours HIMAP (10km) : 0-24 hours ensemble forecasts (16 members) monthly and seasonal forecasts Environmental Emergency Transport model (CANERM) Trajectory model Wave model (WAM) air quality (CHRONOS) OPERATIONAL RUNS: MAIN COMPONENTS regional data assimilation

8. CMC Operational Models • 3D-Var assimilation on model η levels (T108) • Background errors from 24-48 method • Observations QC with BG check and QC-VAR • GEM model (global, regional, meso) Global Model Regional Model • Uniform grid • Resolution of .9º (~100 km) • 28 eta levels • Kuo convection scheme • Sundqvist stratiform scheme • Force-restore surface module • with climatogical soil moisture • 10 day forecasts at 00Z • and 6 day forecasts at 12Z. • Cut-off of T+3h00 • Variable resolution grid • Resolution of .1375º (~15 km) • 58 eta levels • Kain-Fritsch scheme • Sundqvist stratiform scheme • ISBA surface module with • soil moisture pseudo-analysis • (error feedback, no data) • 48-hour forecasts (00Z -12Z) • Cut-off of T+1h35

9. Regional model 575 X 641 grid (66% in 15-km uniform area)

10. CMC Operational Models (cont.) Global ensemble Mesoscale model • GEM variable grid (HiMAP) • Resolution of .09º (~10 km) L35 • Kain-Fritsch convection scheme • Sundqvist condensation scheme • ISBA surface module • No assimilation cycle • (started from 6hr regional) • 30 hour forecasts at 06/18Z • Perturbed analyses from • perturbed assimilation cycles • Resolution of ~150 km (1.2º • for GEM and T149 for SEF) • 16 members • Multi model approach with • different physics options • 10 day forecasts at 00Z

11. Observations for 3D-Var analyses

12. Network now completed!

13. Canadian AMDAR Programme • Business Case and Workshop March-April 2000 • Canadian AMDAR Program Implementation Team: • TORs (EC, NC, airlines, WMO, NOAA, Sypher) • Canadian AMDAR s/w’s Specifications - enhanced ARINC 620-4 • Data Ownership and Process & Communication Studies • AMDAR Data Control System’s Specifications (ongoing) • Completion of DASH-8’s AMDAR s/w by UASC (Unilink) • Lobbying airlines to join, including major PR actions • Air Canada Jazz began reporting July 2002: • 21 Air Nova DASH-8s and 10 CRJs presently reporting data • probable fleet expansion to Air BC and Air Ontario (2004-05), JAZZ now planning ~65 CRJ and 60 DASH-8 by March 2006 • reception and processing of data by CMC; C-ADAS software • Partnering with First Air for an alternative AMDAR system

14. Canadian AMDAR coverage

15. Canadian AMDAR coverage - soundings

16. Monitoring of Canadian AMDAR - DASH-8 Temp. August 02

17. Monitoring of Canadian AMDAR - DASH-8 Wind August 02

18. Monitoring of Canadian AMDAR - CRJ Wind/TT

19. Canadian AMDAR - Status • Current sensor on DASH-8 is flush mounted on wing, next to hydraulic lines (non standard) • AC Jazz Temp probes to be replaced (2 probes) on forward part of fuselage, should fix the Temp bias problem: target is end of 2004 • Monitoring by CMC helped JAZZ to explain a long standing issue related to their FMS generating unneeded steering commands in light wind situations • First Air: have selected the NASA TAMDAR system for provision of data, test flight with TAMDAR to occur this Summer • If evaluation of TAMDAR is conclusive, 8 First Air aircraft to start reporting in early to mid 2005. • All (good) data to be on GTS: IUAA01 CWAO and IUAB01 CWAO, beginning soon with data from CRJ.

20. First Air Route(http://www.firstair.ca/images/FA-ROUTE.gif)

21. Other Canadian observations • Canadian AMDAR on GTS soon, using BUFR code • “Synoptic” observations from CTBTO network, using SYNOP MOBIL code, header: SNCN19 CWAO. Observations every 10-min, 10 stations for now but should increase, received at CMC by e-mail. • Canadian radar data - will be available centrally at CMC, once NRP (National Radar Processor) becomes operational, in about 1 year • Forestry and Road weather stations in British Columbia - data received at CMC, in old SA format. Data redistribution restrictions currently apply…should eventually be available in BUFR • Co-operative network in Quebec (Province, Hydro-Quebec, etc), received at CMC, also redistribution restrictions…eventually in BUFR • Montreal mesonet, including research wind profiler: under installation but operated by Universities, data not yet available and redistribution restrictions may apply…should be provided to USA profiler hub

22. Changes to CMC NWP System since last meeting Major conversion to IBM machine during 2003 – completed by January 04 New satellite data, implemented June 19, 2003: AMSUB + GOES • 4 AMSU-B channels (ch 2-5) • GOES 6.7 micron channel (tested GOES 8/10 but 8 replaced by 12, culd implement GOES-10 only) • last level of moisture analyses raised from 300 to 70 hPa • impacts mostly on moisture fields and precipitation • implementation in global system only, due to IBM conversion • note: loss of NOAA-17 AMSUA data on October 28, 2003

23. Average 250 mb (T-Td) analysis increments for the 22 Dec 01 to 31 Jan 02 assimilation period GOES - CONTROL AMSUB - CONTROL • AMSUB and GOES have similar average (quite large!!) corrections at 250 mb for T-Td

24. ATOVS available for Reg ATOVS available for Glob GOES available for Reg

25. AMSU-B+GOES OPE Results:48h forecast verificationover S. HemisphereAug-Sep 2002

26. Verification of 24-48 hr precipitation NOAMDec 02-Jan 03

28. Cross Validation with SSM/I - IWV

29. Operational WMO Verification AMSU-B GOES rad 3D-Var T,ps 3D-Var h levels 3D-Var GEM TOVS 1d TOVS 1b

30. Changes to CMC NWP System since last meeting New GEM regional model evaluated in parallel during last 3 months • MPI code • 15 km, 58 levels (575 x 641 grid) • Minimum resolution of 350 km • Sundqvist condensation scheme • Kain-Fritsch convection scheme • Moist TKE (Turbulence en air saturé) • Ktrans Shallow convection • Very good results for mass/wind and PCPN • 8x more cpu than previous regional model (24km, 28 levels) • Incorporates AMSUB and GOES radiance data (as in global) • Implemented May 18, 2004

31. Results fromPAR run:Oper 24-km vs new 15-km GEM;NOAMmid-feb to mid-apr 04 GZ500 UV250

32. Results fromPAR run:Oper 24-km vs new 15-km GEM;NOAMmid-feb to mid-apr 04

33. Implementations planned 2004-05 • Assimilation: new datasets • AMSUA (AQUA), GOES-12, Profilers, MODIS windsSpring 2004 • Stratos AMSU-A channels + AMSU-B (raised model top) 2005 + • IR channels (GOES 10-12; METEOSAT, MTSAT and AIRS) • IWV from SSM/I and SSM/IS (eventually from sfc-GPS) • Sfc winds (QuikScat, SSM/IS) • Global 4D-Var Fall 2004 • Improved GEM Global model Mid 2005 • EnKF (using 3D/4D operators) for EPS and flow dependent statisticsMid 2004

34. Impact of MODIS Polar Winds Difference in RMS of 24-hours Forecast errors, verified against Control analyses done with all the observations: No Polar Winds – Polar Winds exp., 500 hPa wind vector (m/s), 2003110800-2004012512

35. 17 dec 01 – 27 jan 02 | 8 nov – 31 dec 03 Anomaly correlation, GZ 500 hPa Northern Hemisphere: Control, No SAT and NOUA New OSE’s OSE without AMSUB/GOES OSE with AMSUB/GOES/MODIS Winter 03 exp Winter 02 exp

36. 17 dec 01 – 27 jan 02 | 8 nov – 31 dec 03 Anomaly correlation, GZ 500 hPa Southern Hemisphere: Control, No SAT and NOUA OSE without AMSUB/GOES OSE with AMSUB/GOES/MODIS Winter 03 exp Winter 02 exp

37. 17 dec 01 – 27 jan 02 | 8 nov – 31 dec 03 Anomaly correlation, GZ 500 hPa North America: Control, No SAT and NOUA OSE without AMSUB/GOES OSE with AMSUB/GOES/MODIS Winter 03 exp Winter 02 exp

38. Global 4D-Var assimilation cycle • 6h window with 45 min time step. • 4D data thinning. Spatial data selection is done for each 45 min time bin. • Incremental analysis performed at T108. • GEM tangent-linear and adjoint models on 240 x 120 grid. • 2 outer loops: • 40 inner loops : adiabatic physics (vertical diffusion) • 30 inner loops : improved physics (SGO, large scale condensation, convection)

39. Recent 4D Results: 20 Jun – 28 Aug 03 4D-Var OPS-3D

40. Improved Global GEM • Physics • Constant thermodynamic roughness over water surfaces • No vertical smoothing on net radiational fluxes • Bougeault-Lacarrere mixing length for the vertical diffusion • MoisTKE for boundary-layer clouds • Shallow convection with Kuo Transient • Deep convection with Kain-Fritsch • Grid-scale condensation with a modified Sundqvist scheme (consun) • Numerics / Dynamics • 800 x 600 x L58 (i.e., 33 km at 49o) • Timestep of 15 min • Numerical poles at geographical poles • Ñ6 horizontal diffusion on momentum variables • Modified sponge layer at model top • Filtered orography from USGS

41. Priorities for Satellite Data • ATOVS: Improved access to Level 1B data (new link with Nesdis), Canadian HRPT data (processing with AAPP) for regional model (cut-off of T+90 min.) and EARS • SSM/IS • QuikScat and Envisat (AATSR), both now received, tests underway • MODIS Products (Winds, SST, LAI, Cloud), winds now tested. A more “operational like” status of NOAA wind product would be very useful. • METEOSAT and GOES-9 Imagery (MTSAT) (CSR product) • High Density AMV for GOES9/MTSAT, GOES 9 soon in parallel • AIRS (assimilation now being tested) • METOP (IASI,…)