Dr Andrew Watkins - Improved seasonal forecast service
•
0 likes•310 views
Dr Andrew Watkins from the Bureau of Meteorology talks about climate outlooks now and into the future.
Recommended
More Related Content
Similar to Dr Andrew Watkins - Improved seasonal forecast service
Similar to Dr Andrew Watkins - Improved seasonal forecast service (20)
More from Department of Housing and Public Works (Queensland)
More from Department of Housing and Public Works (Queensland) (6)
Recently uploaded
Recently uploaded (20)
Dr Andrew Watkins - Improved seasonal forecast service
- 1. Dr. Andrew Watkins (abw@bom.gov.au) MCV Climate Week 17 November 2015 Improved Seasonal Forecast Service Climate Outlooks now and in the future
- 2. The Bureau's Climate Prediction service • ENSO Wrap Up • ENSO Tracker • Model Summary • Weekly Tropical Climate Note • Tropical Cyclone outlook • Northern Rainfall Onset • Seasonal Outlooks • Videos • Briefings/engagement/ministerials/high level advice
- 3. • 5% of GDP ($58B) exposed to annual climate variability • Bureau (and climate science can influence Australia’s ability to respond effectively • Potential value of climate forecasts is >$1.5B • Departures from “normal” are increasing due to global warming The Bureau's Climate Outlook service
- 4. 1996 The Climate Outlook service – Seasonal Outlook http://www.bom.gov.au/climate/outlooks/ 1989-now 1989 2015
- 5. The Climate Outlook service – Seasonal Outlook http://www.bom.gov.au/climate/outlooks/ • Temperature and rainfall • Can add more variables • Model is reliable • Engaging and intuitive • Large user base • Operationally supported • State of the art/science
- 6. Skill of the model – seasonal (rainfall) Lead time 10 days [Multiple lead times possible 0-50days, 5 day increments] Hindcast Forecast
- 7. Skill of the model – monthly (rainfall) Lead time 10 days [Multiple lead times possible 0-50days, 5 day increments] Hindcast Forecast
- 8. Skill of the model – fortnightly (rainfall) Lead time 0 days Hindcast Forecast
- 9. Skill of the model – seasonal (rainfall) vs statistical Lead time 10 days [Multiple lead times possible 0-50days, 5 day increments] System Period BSS REL RES PC Statistical 1981-2010 0.3% 0.0028 0.0028 51.7% (45.4,58.1) POAMA-lagged (10 day) 1981-2010 5.2% 0.0015 0.014 54.0% (53.0,69.2) Statistical 2000-2011 2.7% 0.0022 0.0078 58.3% (47.8,68.7) POAMA-lagged (10 day) 2000-2011 5.0% 0.0049 0.017 64.0% (47.5,78.7) Statistical 1950-1979 1.3% 0.00035 0.00338 51.2% (45.5-56.2) Statistical 1950-1999 0.55% 0.00095 0.00228 50.3% (45.9-54.8) Statistical 1980-1999 -0.77% 0.00385 0.00173 49.1% (41.9-56.7) Charles et al., (2015)
- 10. Areas for improvement… • Coarse 250km grid resolution • Limited compatibility with decision support models • No explicit climate change signal • Gap between days and months • Skill remains modest • Model differs from weather model
- 11. SEASONAL OUTLOOKSBETTER Finer model detail More outlook periods Higher outlook skill World class service Bigger user returns Moving from 250 km to 60 km resolution meaning more localised information by accounting for local conditions Australia:120 to 2000 grid points Seamless: filling the gap between 7-day and monthly outlooks Outlooks updated weekly Season Month Fortnight Week Likely 10% improvement in outlook accuracy meaning the best outlooks for Australia of all international models meaning information is clear, concise and available when and where you need it + Not only rainfall and temperature More intelligence possible: • Evaporation • Humidity • Wind • Drought • Extremes • Tropical Cyclones Reduce losses: agricultural production lost from 2010-11 La Niña: More than $2 billion Potential value of improved seasonal forecasts: More than $1 billion per year ABARES Centre for International Economics 2014
- 12. Improved resolution POAMA-2 ACCESS-S Australian topography • Resolution improves from 250 km to 60 km • Resolves the Great Dividing Range, Tasmania, WA Darling Ranges, Pilbara (Tom Price)
- 13. Improved resolution • Resolution improves from 250 km to 60 km • Resolves the Great Dividing Range, Tasmania, WA Darling Ranges, Pilbara (Tom Price) Metress
- 14. Better model climate • Able to provide more realistic climate patterns • Link to decision support tools (e.g., fire models, crop models etc) August Mean Rainfall POAMA-2 Observations ACCESS-S mm/dayy
- 15. Better model climate • Able to provide more realistic climate patterns • Link to decision support tools (e.g., fire models, crop models etc) August Mean Rainfall POAMA-2 Observations ACCESS-S mm/day
- 16. Better model climate • Able to provide more realistic weather sequences/climate patterns • Link to decision support tools (e.g., fire models, crop models etc) mm/day
- 17. More accurate outlooks • Early testing shows improved accuracy for rainfall • Better predictions of El Niño / La Niña NINO3.4: all start months El Niño forecast accuracyRainfall forecast accuracy New model
- 18. Heat extremes case studies September 2013 Monthly temperature forecast Source: ABC online
- 20. Where to next? • 2015: Obtain feedback on current service and priorities for improvement → Develop service solutions • 2016: Test deployment of new model • 2017: First deployment of new outlook service (including multi-week) • 2018: Further upgrade to model (physics, initial conditions) • 2019: New outlooks service deployed