Predicting Fire-weather Severity using Seasonal Forecasts

1. Predicting Fire-weather Severity using Seasonal Forecasts Kerry Anderson Peter Englefield Richard Carr Canadian Forest Service

2. Introduction

3. Introduction The Canadian Forest Fire Danger Rating System (CFFDRS) has been an important part of forest protection operations in Canada since 1970. The ability to forecast fire-weather conditions associated with the CFFDRS is critical to operational decisions and thus is a routine part of fire management planning. This paper presents a method to predict fire-weather severity for a fire season. Predictions are based on Environment Canada's seasonal forecasts and information contained in the Canadian Wildland Fire Information System (CWFIS).

4. Introduction Canadian Wildland Fire Information System The Canadian Wildland Fire Information System (CWFIS) is a computer-based fire management information system that monitors fire danger conditions across Canada. Daily weather conditions are collected from across Canada and used to produce fire weather and fire behaviour maps based on the Canadian Forest Fire Danger Rating System (CFFDRS).

5. Introduction Canadian Forest Fire Danger Rating System The fire-weather severity outlook is based on the forecasted seasonal severity as calculated by the CFFDRS. The two principal models within the CFFDRS are the Canadian Forest Fire Weather Index (FWI) System and the Canadian Forest Fire Behaviour Prediction (FBP) System. This report focuses on components of the FWI system.

6. Introduction Daily Severity Rating The Daily Severity Rating (DSR) and its time-averaged value are extensions of the FWI System. The DSR is a transformation of the daily FWI value The DSR can be accumulated over time as the cumulative DSR (or CDSR), or averaged over an entire fire season as the seasonal severity rating (SSR). DSR = 0.0272 FWI 1.77

7. Introduction Climate and Weather Outlooks The Canadian Meteorological Centre (CMC) of Environment Canada has been producing seasonal outlooks for Canada since 1995. Based on numerical weather prediction (NWP) models, these outlooks predict the temperature and precipitation anomalies for the next 12 months.

8. Introduction Ensemble Approach Currently, CMC produces a probabilistic forecast for the next four months using an ensemble approach while the 12 month extended-range forecast is based on statistical techniques. These come as GriB data for every month and every ensemble member at 2.5o resolution.

9. Introduction Ensemble Approach The ensemble approach uses six independent runs of two models: the Global Environmental Multiscale (GEM) model and the Second Generation Atmospheric General Circulation Model (AGCM2) . Through the total 12 members, the ensemble approach produces a probability and confidence of above, below and normal anomalies.

10. Methodology

11. Methodology Summary • Average daily weather conditions are calculated for 62 weather stations across Canada. • Monthly anomaly predictions are added to the average weather values. • Forecasted fire weather conditions are mapped and compared to average conditions as a ratio.

12. Methodology • Precipitation was accumulated and applied on every seventh day. • Fire weather conditions were calculated on the average weather. Average Weather Conditions • Average temperature, humidity, wind speed and precipitation per day per station over a 30 year period.

13. Methodology Monthly Anomaly Predictions • Anomalies were interpolated from monthly ensemble members to the study weather stations and applied to the daily values. • Fire weather conditions were recalculated for each ensemble member thus providing an ensemble of fire weather predictions.

14. Methodology Forecasted Conditions • Forecasted and average severity ratings were mapped for Canada using inverse distance weighting. • The forecast map was then compared to the average map as a ratio, showing the regions that are above and below average. • These maps were produced for the individual months (MSRs) as well as the three month period (SSR).

15. Methodology Monthly Severity Ratings Maps Average and forecast Monthly Severity Ratings alone show where fire-weather conditions are high or low.

16. Methodology Forecasted Conditions The ratio of the two maps show the regions where fire-weather conditions are above or below average. Hotspots indicate areas where fire activity was high.

17. Methodology Validation Study A validation study was conducted using data available from Environment Canada. This included 7 data sets of 5 years for the GEM and the GCM model. Above or below-average predictions were compared to observed conditions at each of the stations for each day of the forecast.

18. Methodology Skill Scores Standard skill scores were used to measure the skill of the model. Bias = PC = POD = FAR = CSI = Hits Hits + Misses Forecast Above Below Hits + Zeroes Total Observations Observed Below Above Hits Hits + Misses False Hits + False Hits Hits + Misses + False

19. Results

20. Results GCM GCM GEM GEM GEM GCM GEM GCM

21. Results GCM GEM

22. Results GCM GCM GEM GEM GCM GEM GEM GCM

23. Results GCM GEM

24. Conclusion

25. Conclusion This paper presents a method to predict fire-weather severity for a fire season. Predictions are based on Environment Canada's seasonal forecasts and information contained in the Canadian Wildland Fire Information System (CWFIS). While the results are not strong, they show there is some skill in this long-range outlook, suggesting there is value in the approach presented in this paper. This system is still in development and it is our intention is to conduct more detailed validations with more data years.

26. The End