reanalysis needs for climate monitoring craig s long arun kumar and wesley ebisuzaki cpc noaa...
TRANSCRIPT
Reanalysis Needs for Climate Monitoring
Craig S LongArun Kumar and Wesley Ebisuzaki
CPC
NOAA Climate Test Bed (CTB) Meeting – November 9-10, 2015 - NCWCP
Reanalysis Needs for Climate Monitoring
• Outcomes from the May 2015 NCRTF Workshop.
• Topics to cover:– What are the differences between:
• reanalysis for monitoring and • reforecast initializations?
– What reanalysis research advances can be transitioned to improve operational climate-reanalysis?
– How to coordinate the two reanalysis efforts in a resource-constrained environments?
NOAA Climate Test Bed (CTB) Meeting – November 9-10, 2015 - NCWCP
Why the Need for Climate Reanalysis
• Monitoring:– Monitoring real-time climate anomalies requires placing them in
a historical context, and hence, the need for a climate reanalysis
• Attribution:– and an increasing demand to provide explanations for extreme
climate events requires access to physically consistent climate reanalysis data sets
• Applications:– Climate reanalysis data sets are used in a wide array of societal
applications, e.g., decision making in the context of infrastructure development
NOAA Climate Test Bed (CTB) Meeting – November 9-10, 2015 - NCWCP
Why the Need for Climate Reanalysis
• Climate reanalysis data sets:– provide base climatology relative to which climate forecasts are
issued – are needed to verify hindcasts and real-time forecasts and
provide skill information to the users – are needed to bias correct and calibrate real-time forecasts – provide initial conditions for hindcasts and real-time climate
forecasts – ocean, land, sea ice
NOAA Climate Test Bed (CTB) Meeting – November 9-10, 2015 - NCWCP
Review
• Review of what we’ve got:– R1, R2, CFSR, 20CR
• What are key attributes / inadequacies of the above?
• What’s being done to improve things?
NOAA Climate Test Bed (CTB) Meeting – November 9-10, 2015 - NCWCP
NCEP/NCAR Reanalysis (R1)
• Circa 1995 model and data assimilation system • Atmospheric model – T62/L28 • 1948-present; maintained in real-time by CPC • Still widely used for the analysis of climate variability
NOAA Climate Test Bed (CTB) Meeting – November 9-10, 2015 - NCWCP
R1 Issues
• Relies on data assimilation system that is 20 years old • Portability and maintenance • Uses satellite retrievals (that are under constant threat to
be discontinued)
• Replacement for R1 preferably should :– Extend over the same period – Not have climate trends & discontinuities worse than R1
NOAA Climate Test Bed (CTB) Meeting – November 9-10, 2015 - NCWCP
Climate Forecast System Reanalysis
• Coupled: Atmosphere, Land, Ocean, Sea Ice• Atmosphere: T382/L64 : 1979-2010• Upgraded to T574/L64 in 2010 with other changes• Six Streams• One year spin-up for atmosphere, land and ocean• Used as IC for hindcasts and CFS forecasts• Time constrained development and testing
– 1979-2010 was completed in limited time of available computer– No time to go back and rerun.– Only had R1 and ERA40 to compare against
NOAA Climate Test Bed (CTB) Meeting – November 9-10, 2015 - NCWCP
CFSR Issues
Analysis during earlier period is an outlier
Multiple Streams – Zonal Avg. Resolution Change in 2010
NOAA Climate Test Bed (CTB) Meeting – November 9-10, 2015 - NCWCP
As a consequence…
• Continued CPC’s reliance for climate monitoring products using R1
• Developed a research effort to understand some of the issues with climate reanalysis efforts, result of which is…
• The NOAA Climate Reanalysis (NCR) effort – Develop a hierarchical approach for climate reanalysis – Investigate the impact to the time-varying quality and density of
the observing system and determine ways to reduce this impact – Deliverable: develop a prototype climate reanalysis as a
potential candidate for replacing R1
NOAA Climate Test Bed (CTB) Meeting – November 9-10, 2015 - NCWCP
NOAA Reanalysis Projects
• Continuing:– R1 – CFSR
• Current and Future:– AMIP (1850-present)– 20CRv3 (1850-present)– Conventional (1948-present)
• SFC + Radiosonde• Test results are promising
– Better analyses than R1
• Overlap of reanalyses should reveal biases– Increase certainty of trends
• Use common data assimilation infrastructure (EnkF) shared across NOAA
NOAA Climate Test Bed (CTB) Meeting – November 9-10, 2015 - NCWCP
Some Outstanding Research Issues
• Understanding reasons for discontinuities when new observational platforms come in
NOAA Climate Test Bed (CTB) Meeting – November 9-10, 2015 - NCWCP
Discontinuities due to changes in observational platforms
NOAA Climate Test Bed (CTB) Meeting – November 9-10, 2015 - NCWCP
No Trend
No Trend
Data Source Change
How to deal with transitions?
NOAA Climate Test Bed (CTB) Meeting – November 9-10, 2015 - NCWCP
No Trend
No Trend
Data Source Change
No Trend
No Trend
Transition
How to deal with transitions?
NOAA Climate Test Bed (CTB) Meeting – November 9-10, 2015 - NCWCP
Thoughts I
• Old and New sensors are looking at the same Earth!• Model reacts differently to new data versus old data
sources.– Why?
• More / sharper channels• Smaller errors• Higher density• Schemes developed for new data, not old data
– Solution?• Assuming new data is better….• Can we use new to bias adjust (train) the old
– How does this affect the Historical Context?• TOVS vs ATOVS
NOAA Climate Test Bed (CTB) Meeting – November 9-10, 2015 - NCWCP
Thoughts II
• What will model improvements provide?– Higher Horizontal– Higher/more vertical– SL vs Eularian– Model climatology– New Physics– New Parameterizations
• How to transition from density poor to density rich?
• Is assimilation of more data sources better? • Or is fewer, but higher quality better?• Above requires lots of testing!
– Testing on same model version to be used for next reanalysis?– Any benefit of testing on older model version?
NOAA Climate Test Bed (CTB) Meeting – November 9-10, 2015 - NCWCP
What still needs to be done?
• What are barriers to getting things done:– Model architecture
• NEMS, NGGPS to the rescue?
– Funding for:• Manpower• Computers
– Testing– Running– Archive
– Coordination between centers• Assemble• Test• Evaluate
NOAA Climate Test Bed (CTB) Meeting – November 9-10, 2015 - NCWCP
Discussion?!
NOAA Climate Test Bed (CTB) Meeting – November 9-10, 2015 - NCWCP