1. Environment Canada Monthly and Seasonal Forecasting Systems
Hai Lin
Meteorological Research Division
Environment Canada

2. Current EC Seasonal+monthly Forecasting System
Main features:
Multi-model ensemble approach
4 AGCMs: GEM-CLIM, SEF, GCM2 and GCM3
10 members for each model, lagged initial conditions
2-tier system
Persistent SST anomaly
4-month integration
1st month  monthly forecast
Seasonal forecasting system has been replaced on December 1, 2011
To assess the contribution of moisture and convections to LFV, especially the MJO.
Does tropical intraseasonal variability (TIV) exist in a dry model?
Global view of the intraseasonal variability.
Implication to the MJO

3. Next EC Seasonal and Monthly Forecasting Systems
2 separate systems
Monthly system based on GEPS, expected to be in operation in early 2012
Seasonal system uses 2 coupled models (1-tier system), expected in operation December 2011
To assess the contribution of moisture and convections to LFV, especially the MJO.
Does tropical intraseasonal variability (TIV) exist in a dry model?
Global view of the intraseasonal variability.
Implication to the MJO

4. Monthly System Two components 1) Real time forecasting system
GEPS based
Two components
1) Real time forecasting system
2) Hindcast (model climate and statistics) 4

5. Real time forecasting system
Extend GEPS to 32 days once a week
(00Z Thursday)
Persistent SST anomaly added to time-evolving SST climatology
GEM 4.4, 0.6x0.6L40
Perturbed physics
21 members, ensemble Kalman Filter 5

6. Hindcasts To generate GEPS model climatology
For the same date, past 15 years
4 members each year, 60 members for each date
Use 3 weeks centered at the date of the current Thursday, total of 180 members
Can be done any time before the real time forecasting, when the computer is available 6

7. Proposed products
Forecast products of weekly average T2m and PR over Canada
Ensemble mean anomaly maps
Probability maps for above, below and near normal 7

8. Evaluation Summer (July and August)
Winter (January and February)
Summer (July and August)
Past three years (24 forecasts for each season)
Correlation skill for ensemble mean 500 hPa geopotential height and T2m
Categorical forecast score for T2m and PR
The following maps are based on previous GEPS version GEM x0.9L28 8

9. Correlation skill
Winter Z500
30-day average 9

10. Correlation skill
Winter T2m
30-day average 10

11. for categorical forecasts
Percent Correct
for categorical forecasts
Winter T2m
30-day average 11

12. for categorical forecasts
Percent Correct
for categorical forecasts
Winter PR
30-day average 12

13. for categorical forecasts
10-day average T2m
Over Canada
Percent Correct
for categorical forecasts 13

14. New Seasonal System AGCM3 (T63L31) -> CanCM3
2-coupled models (CCCma models)
AGCM3 (T63L31) -> CanCM3
AGCM4 (T63L35) -> CanCM4
coupled with CanOM4 (1.41°× 0.94° / L40) ocean model
1-tier system
10 members for each model
1-12 month forecast issued every month 14