Data assimilation as a tool for C cycle studies Collaborators: P Stoy, J Evans, C Lloyd, A Prieto Blanco, M Disney, L Street, A Fox (Sheffield) M Van Wijk (Wageningen), E B Rastetter (MBL), G Shaver (MBL) www.abacus-ipy.org Mathew Williams, University of Edinburgh
www.abacus-ipy.org. Data assimilation as a tool for C cycle studies. Mathew Williams, University of Edinburgh. Collaborators: P Stoy, J Evans, C Lloyd, A Prieto Blanco, M Disney, L Street, A Fox (Sheffield) M Van Wijk (Wageningen), E B Rastetter (MBL), G Shaver (MBL). - PowerPoint PPT Presentation
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Data assimilation as a tool for C cycle studies
Collaborators: P Stoy, J Evans, C Lloyd, A Prieto Blanco, M Disney, L Street, A Fox (Sheffield)
M Van Wijk (Wageningen), E B Rastetter (MBL), G Shaver (MBL)
www.abacus-ipy.org
Mathew Williams, University of Edinburgh
Transferring information across scales
The upscaling problem and data assimilation An Arctic C cycle application REFLEX – a comparison of DA approaches for
C flux estimation
Upscaling C fluxes
How do we cope with spatial variation? What are the critical feedbacks over longer
time scales? How can model/parameters be improved? How can multiple data be combined? How trustworthy are such combinations?
The Kalman Filter in theory
MODEL At Ft+1 F´t+1OPERATOR
At+1
Dt+1
Assimilation
Initial state Forecast ObservationsPredictions
Analysis
P
Drivers
SWEDEN
What is the carbon balance of an Arctic landscape?
How will C balance change in the future?What measurements should we take to
improve understanding and forecast skills?
A multiscale approach
Arctic Biosphere Atmosphere Coupling at multiple Scales
Observation operator: NDVI-LAI
Van Wijk & Williams, 2005
LAI harvest calibrates indirect measurement (NDVI)
Shaver et al. J. Ecol. (2007)
GPP Croot
Cwood
Clitter
CSOM/CWD
Ra
Ar
Aw
Cfoliage
Af Lf
Lr
Lw
Rh
D
Temperature controlled
5 model pools9 model fluxes9 unknown parameters2 data time series