www-txt.gif (7823 bytes)

Example of reporting templates

To help in the evaluation of the clarity and accuracy of the proposed reporting templates, the following information should be taken into account:

  • The Canadian Meteorological Centre (CMC) long-range forecast (LRF) system is two-fold: two dynamical models are used at lead zero and an empirical model is used at lead 2, 3 and 4 seasons. The dynamical component consists of a twelve member ensemble prediction system (Derome et al., 2000) based on 6 runs of a General Circulation Model (GCM) (McFarlane et al., 1992) and six runs of a Numerical Weather Prediction (NWP) model (Ritchie, 1991). The empirical model is based on Canonical Correlation Analysis (CCA) (Shabbar and Barnston, 1996). The forecast parameters are surface air temperature anomaly (SAT) and precipitation anomaly both in three equi-probable categories (below, near and above normal).

  • In the examples below, results are presented for lead zero SAT over Canada in Summer (June-July-August) based on dynamical model outputs. The dynamical model outputs were combined using a simple arithmetic average. The verification data set comes from the station observation of Vincent and Gullett (1999).

  • Additionnally, results for the probabilistic geopotential height anomaly at 500 hPa are presented to demonstrate the completion of the templates for probabilistic forecasts. This new product is not disseminated yet by the CMC. Results are for winter (December-January-February) over the Northern Extra-tropics with a zero lead. The probabilistic forecasts are generated by counting the number of members of the ensemble in each of the three equi-probable category. The ROC score has been computed using Mason (1982) approach with 10% probability bins. The forecasts were compared with the corresponding NCEP reanalyses (Kalnay et al., 1996).

References:

Derome J., G. Brunet, A. Plante, N. Gagnon, G. J. Boer, F.W. Zwiers, S.J. Lambert, J.Sheng H.Ritchie, 2000: Seasonal Predictions Based on Two Dynamical Models. Submitted to Atmosphere Ocean.

Kalnay, E., M. Kanamitsu, R. Kistler, W. Collins, D. Deaven, L. Gandin, M. Iredell, S. Saha, G. White, J. Woollen, Y. Zhu, M. Chelliah, W. Ebisuzaki, W. Higgins, J. Janowiak, K. C. Mo, C. Ropelewski, J. Wang, A. Leetmaa, R. Reynolds, Roy Jenne, and Dennis Joseph, 1996: The NMC/NCAR 40-Year Reanalysis Project". Bull. Amer. Meteor. Soc., 77, 437-471.

Mason, I., 1982: A model for assessment of weather forecasts. Aust. Meteor. Mag., 30, 291-303.

McFarlane, N.A., G. J. Boer, J.-P. Blanchet and M. Lazare, 1992: The Canadian Climate Centre second-generation circulation model and its equilibrium climate. J. Climate, 5, 1013-1044.

Ritchie, H., 1991: Application of the semi-Lagrangian method to a multilevel spectral primitive-equations model. Quart. J. Roy. Meteor. Soc., 117, 91-106

Shabbar, A. and A. G. Barnston, 1996: Skill of seasonal climate forecasts in Canada using Canonical Correlation Analysis. Mon. Wea. Rev., 124, 2370-2385.

Vincent, L.A., and D. W. Gullett, 1999: Canadian historical and homogeneous temperature datasets for climate change analyses. International Journal of Climatology, 19,1375-1388.

Template used for LRF system description:

Identification

Country:

1 CANADA

Meteorological Centre:

2 CANADIAN METEOROLOGICAL CENTRE

LRF system identification:

3 CMC-DYN

 

Description of Long-Range Forecast (LRF) System

Status of LRF system:

information:

4 Yes

dissemination:

5 Disseminated trough the Meteorological Service of Canada regional offices, media and Internet.

guidance:

6 Historical (hindcast) percent correct map is provided with the forecast.

Type of LRF system:

numerical:

7 Yes

empirical:

8 No

hybrid:

9 No

coupled:

10 No

statistics:

11 Blend of the model 1000-500 hPa thickness using an arithmetic average and forecast SAT using a Perfect Prog simple linear regression with 1000-500 hPa hPa anomaly as predictor.

Type of forecasts:

deterministic:

12 Yes

probabilistic:

13 No

LRF output products:

parameter:

14 Surface air temperature anomaly and accumulated precipitation anomaly.

categories:

15 Below (< -0.43 times the inter-annual standard deviation), above (>0.43 times the inter-annual standard deviation) and near normal (otherwise).

frequency:

16 4 times a year

forecast period:

17 Season (3 months)

lead time:

18 Zero

forecast area:

19 Canada

Model description:

20 Two numerical models are used : 1) GCM version 2 (McFarlane et al., 1992) and 2) Global NWP model (Ritchie, 1991).

Model resolution:

horizontal:

21 GCM=T32
      NWP=T63

vertical:

22 GCM=L10
      NWP=L23

Bias correction:

23 No

Ensemble forecasting:

ensemble size:

24 12 members (6 for each model)

initialisation:

25 Initial conditions for each run are lagged by 24 hours.

SST specification:

26 Anomaly of the month prior to the start of the runs is prescribed through the 3 months forecast period.

Sea-Ice specification:

27 Climatology for the GCM and monthly mean of the month prior to the start of the runs is relaxed to climatology during the first 15 days.

Snow specification:

28 Initial snow line from observations. The GCM has a prognostic scheme while the NWP model uses a prescribed 10 day snow anomaly for half of the simulation afterwards the system reverts to climatology.

Soil temperature:

29 Climatology

Soil moisture:

30 Climatology

Hindcast evaluation:

hindcast period:

31 1969 to 1994

cross-validation:

32 No

particularities:

33 Initial conditions are lagged by 6 hours (instead of 24 as in the operational set-up).

 

Template used for exchange of verification scores for deterministic forecasts:

Identification

Country:

1 Canada

Meteorological Centre:

2 Canadian Meteorological Centre

LRF system identification:

3 CMC-DYN

 

Long-Range Forecast (LRF) verification results - deterministic forecasts

Verified parameter:

4 Surface air temperature anomaly (degree Celsius)

Forecast period:

5 June-July-August

Forecast lead time:

6 Zero

Verification area:

7 Canada

Verification period:

8 1969 to 1994 (hindcast)

Verification data set:

9 Gridded Canadian stations data (Vincent and Gullet, 1999).

Climatology data set:

10 Average of 1969 to 1994 data (both for model and observation).

Persistence:

11 Anomaly of the season prior to the verified season.

 

RMS-forecast

RMS-persistence

RMS-climatology

RMSSS-persistence

RMSSS-climatology

12 0.9135

13 0.8517

14 1.1278

5 -7.2567

16 19.0074

 

Binary event:

17 Above normal (anomaly greater than 0.43 times the inter-annual standard deviation)

Contingency Table:

 

observations

forecasts

occurrences

non-occurrences

occurrences

18 605

19 633

non-occurrences

20 542

21 1780

Kuipers Score:

22 0.630

   

 

Template used for exchange of verification scores for deterministic forecasts:

Identification

Country:

1 Canada

Meteorological Centre:

2 Canadian Meteorological Centre

LRF system identification:

3 CMC-DYN

 

Long-Range Forecast (LRF) verification results - deterministic forecasts

Verified parameter:

4 Surface air temperature anomaly (degree Celsius).

Forecast period:

5 June-July-August

Forecast lead time:

6 Zero

Verification area:

7 Canada

Verification period:

8 1969 to 1994 (hindcast)

Verification data set:

9 Gridded Canadian stations data (Vincent and Gullet, 1999).

Climatology data set:

10 Average of 1969 to 1994 data (both for model and observation)

Persistence:

11 Anomaly of the season prior to the verified season.

 

RMS-forecast

RMS-persistence

RMS-climatology

RMSSS-persistence

RMSSS-climatology

12 0.9135

13 0.8517

14 1.1278

5 -7.2567

16 19.0074

 

Binary event:

17 Near normal (anomaly lesser than 0.43 times the inter-annual standard deviation and greater than -0.43 times the inter-annual standard deviation)

Contingency Table:

 

observations

forecasts

occurrences

non-occurrences

occurrences

18 466

19 705

non-occurrences

20 758

21 1630

Kuipers Score:

22 0.518

   

 

Template used for exchange of verification scores for deterministic forecasts:

Identification

Country:

1 Canada

Meteorological Centre:

2 Canadian Meteorological Centre

LRF system identification:

3 CMC-DYN

 

Long-Range Forecast (LRF) verification results - deterministic forecasts

Verified parameter:

4 Surface air temperature anomaly (degree Celsius)

Forecast period:

5 June-July-August

Forecast lead time:

6 Zero

Verification area:

7 Canada

Verification period:

8 1969 to 1994 (hindcast)

Verification data set:

9 Gridded Canadian stations data (Vincent and Gullet, 1999).

Climatology data set:

10 Average of 1969 to 1994 data (both for model and observation).

Persistence:

11 Anomaly of the season prior to the verified season.

 

RMS-forecast

RMS-persistence

RMS-climatology

RMSSS-persistence

RMSSS-climatology

12 0.9135

13 0.8517

14 1.1278

5 -7.2567

16 19.0074

 

Binary event:

17 Below normal (anomaly less than -0.43 times the inter-annual standard deviation)

Contingency Table:

 

observations

forecasts

occurrences

non-occurrences

occurrences

18 620

19 531

non-occurrences

20 568

21 1840

Kuipers Score:

22 0.653

   

 

Template used for exchange of verification scores for probabilistic forecasts:

Identification

Country:

1 CANADA

Meteorological Centre:

2 CANADIAN METEOROLOGICAL CENTRE

LRF system identification:

3 CMC-DYN

 

Long-Range Forecast (LRF) verification results - probabilistic forecasts

Verified parameter:

4 Geopotential height at 500 hPa anomaly.

Forecast period:

5 December-January-February

Forecast lead time:

6 Zero

Verification area:

7 Northern Extra-Tropics

Verification period:

8 1969-1994 (hindcast)

Verification data set:

9 NCEP reanalyses

Climatology data set:

10 Average of 1969 to 1994 data (both for model and analyses).

 

Binary event:

11 Above normal (anomaly greater than 0.43 times the inter-annual standard deviation)

Contingency Table:

 

observations

probability intervals

occurrences

non-occurrences

12 0-9%

13 1786

14 7021

15 10-19%

16 1832

17 4426

18 20-29%

19 1943

20 4406

21 30-39%

22 2026

23 3701

24 40-49%

25 2014

26 3196

27 50-59%

28 3222

29 3920

30 60-69%

31 1044

32 1195

33 70-79%

34 780

35 651

36 80-89%

37 453

38 294

39 90-100%

40 1234

41 946

ROC score:

42 0.6353  

 

Template used for exchange of verification scores for probabilistic forecasts:

Identification

Country:

1 CANADA

Meteorological Centre:

2 CANADIAN METEOROLOGICAL CENTRE

LRF system identification:

3 CMC-DYN

 

Long-Range Forecast (LRF) verification results - probabilistic forecasts

Verified parameter:

4 Geopotential height at 500 hPa anomaly.

Forecast period:

5 December-January-February

Forecast lead time:

6 Zero

Verification area:

7 Northern Extra-Tropics

Verification period:

8 1969-1994 (hindcast)

Verification data set:

9 NCEP reanalyses

Climatology data set:

10 Average of 1969 to 1994 data (both for model and analyses).

 

Binary event:

11 Near normal (anomaly lesser than 0.43 times the inter-annual standard deviation and greater than -0.43 times the inter-annual standard deviation)

Contingency Table:

 

observations

probability intervals

occurrences

non-occurrences

12 0-9%

13 860

14 2206

15 10-19%

16 1669

17 3867

18 20-29%

19 2660

20 6071

21 30-39%

22 3134

23 6981

24 40-49%

25 2656

26 5713

27 50-59%

28 2558

29 5107

30 60-69%

31 241

32 535

33 70-79%

34 60

35 156

36 80-89%

37 7

38 24

39 90-100%

40 67

41 180

ROC score:

42 0.5111  

 

Template used for exchange of verification scores for probabilistic forecasts:

Identification

Country:

1 CANADA

Meteorological Centre:

2 CANADIAN METEOROLOGICAL CENTRE

LRF system identification:

3 CMC-DYN

 

Long-Range Forecast (LRF) verification results - probabilistic forecasts

Verified parameter:

4 Geopotential height at 500 hPa anomaly.

Forecast period:

5 December-January-February

Forecast lead time:

6 Zero

Verification area:

7 Northern Extra-Tropics

Verification period:

8 1969-1994 (hindcast)

Verification data set:

9 NCEP reanalyses

Climatology data set:

10 Average of 1969 to 1994 data (both for model and analyses).

 

Binary event:

11 Below normal (anomaly less than -0.43 times the inter-annual standard deviation)

Contingency Table:

 

observations

probability intervals

occurrences

non-occurrences

12 0-9%

13 1416

14 6977

15 10-19%

16 1690

17 4660

18 20-29%

19 1968

20 4319

21 30-39%

22 1943

23 3649

24 40-49%

25 2043

26 3014

27 50-59%

28 3547

29 4563

30 60-69%

31 1101

32 1128

33 70-79%

34 817

35 555

36 80-89%

37 409

38 331

39 90-100%

40 1226

41 887

ROC score:

42 0.6542  

 


 

WMO Front, About WMO, WWW Front, Library, International Weather