Carbon, soil moisture and fAPAR assimilation

1. Carbon, soil moisture and fAPAR assimilation Wolfgang Knorr Max-Planck Institute of Biogeochemistry Jena, Germany 1 Acknowledgments: Nadine Gobron 2, Marko Scholze 3, Peter Rayner 4, Thomas Kaminski 5, Ralf Giering 5, Heinrich Widmann1 2 3 LSCE QUEST / IES/JRC FastOpt 4 5

2. Overview • CO2 – climate linkages • Satellite fAPAR as soil moisture indicator • Assimilation of fAPAR

3. Atmospheric CO2 Measurements CCDAS inverse modelling period ... and more stations in CCDAS

4. Carbon Cycle Data Assimilation System (CCDAS) Assimilated Prescribed Assimilated CO2 + Uncert. satellite fAPAR + Uncert. Phenology Hydrology CCDAS Step 2 BETHY+TM2 only Photosynthesis, Energy&Carbon Balance CCDAS Step 1 full BETHY Background CO2 fluxes* Optimized Params + Uncert. Diagnostics + Uncert. * * ocean: Takahashi et al. (1999), LeQuere et al. (2000); emissions: Marland et al. (2001), Andres et al. (1996); land use: Houghton et al. (1990)

5. ENSO … preliminary results from extended CCDAS run Terr. biosphere–atmosphere CO2 fluxes

6. … global drying and warming trend ENSO and global climate normalized anomalies ENSO temperature –precipitation

9. for more information see: http://www.CCDAS.org

10. Overview • CO2 – climate linkages • Satellite fAPAR as soil moisture indicator • Assimilation of fAPAR

11. ITOC ITOC fAPAR: [(ITOC+IS)–(ITOC+IS)] / ITOC canopy IS IS soil Remotely Sensed Vegetation Activity

12. SeaWiFS fAPAR archive • developed by Nadine Gobron, Bernard Pinty, Frédéric Melin, IES/JRC, Ispra • 3-channel algorithm taylored to SeaWiFS ocean color instrument (blue, red, near-infrared) • cloud screening algorithm • requires no atmospheric correction • starts 10/1997, continuing... • being extended by same product for MERIS

13. Precipitation – fAPAR precipitation gridded station data soil moisture BETHY simulations leaf area index fAPAR satellite observations BETHY simulations

14. 1-month lag r>0 r<0 4-month lag precipitation vs. fAPAR from SeaWiFS satellite obs. percent area with 99% significant correlation 0.5°x0.5°, ≥50% cloud free, ≥75% temporal coverage

15. 1-month lag r>0 r<0 BETHY simulated fAPAR precipitation vs. fAPAR: satellite and model SeaWiFS fAPAR percent area with 99% significant correlation

16. r>0 r<0 precipitation vs. fAPAR: satellite and model 4-month lag SeaWiFS fAPAR percent area with 99% significant correlation BETHY simulated fAPAR

17. 1-month lag r>0 r<0 BETHY simulated soil moisture precipitation vs. satellite fAPAR and simulated soil moisture SeaWiFS fAPAR percent area with 99% significant correlation

18. r>0 r<0 BETHY simulated soil moisture precipitation vs. satellite fAPAR and simulated soil moisture 4-month lag SeaWiFS fAPAR percent area with 99% significant correlation

19. ENSO – SeaWiFS fAPARlagged correlation 3-month lag

20. Overview • CO2 – climate linkages • Satellite fAPAR as soil moisture indicator • Assimilation of fAPAR

21. optimization mismatch satellite fAPAR fAPAR Assimilation PFT distribution* ecosystem model parameters Prescribed climate & soils data** BETHY model-derived fAPAR carbon and water fluxes

22. measurements model diagnostics error covariance matrix of measurements assumed model parameters a priori error covariance matrix of parameters a priori parameter values aim: minimize J(m) [for each grid point separately] The Cost Function Measure of the mismatch (cost function):

23. The Parameters vector of prior parameter values m0: represents: parameter vector m={m1,m2,m3}: m1 Tf shift of leaf onset/shedding temperature temperature limitation DTf=0 m2 wmax,0 (derived from FAO soil map) wmax maximum soil water holding capacity water limitation m3 fc fraction of grid cell covered with vegetation residual, unmodelled limitations (nitrogen, land use) fc,0 (function of P/PET and Temp. of warmest month)

24. T=5°C T=12°C for crops T=15°C ^ Prior Parameter 1 prior values: note: each 0.5°x0.5° has mixture of up to 6 PFTs map reflects presence of crops; red: unvegetated

25. bucket model: runoff =overflow precipitation =input full bucket: wmax current bucket: w wmax,0 [mm] Prior Parameter 2

26. ^ fc,0=Pannual/PETannual*W(Twarmest month)/ Prior Parameter 3

27. Prior Parameter Errors error covariance matrix of parameters Cm0: • off-diagonal elements assumed 0 here = no prior correlation between errors of different parameters

28. The Assimilated Data model diagnostics vector y={y1,y2,...,y12}: yi modelled fAPAR of month i satellite-derived diagnostics vector y0={y0,1,y0,2,...,y0,12}: y0,i SeaWiFS derived fAPAR of month i

29. Prior Errors of Measurements error covariance matrix of measurements Cy: i=j • off-diagonal elements again 0 = no prior correlation between errors of different months

30. local site Parameter 2 (regional) soil water-holding capacity prior optimized

31. remote sensing data Local Simulations Paragominas 3°S 48°W 63 m precipitation [mm/month] fAPAR 1992 no remote sens. data fAPAR prescribed evapotranspiration [mm/month] NPP [gC/(m2 month)] fAPAR assimilated - -

32. remote sensing data no remote sens. data fAPAR prescribed fAPAR assimilated Measured Soil Moisture Paragominas 3°S 48°W 63 m precipitation [mm/month] fAPAR 1992 0...8m depth 0...2m depth - - 1992 1992

33. evapotranspiration (regional) prior optimized mm/year mm/year

34. Parag. July soil moisture (regional, dry season) prior optimized mm mm

35. Conclusions • The carbon cycle is highly sensitive to climate fluctuations • Vegetation can be quantified reliably from space • fAPAR lags precipitation by ~1–4(?) months • seems to behave similar to soil moisture • assimilation of fAPAR can deliver valuable information on soil moisture status

36. Conclusions • Need to improve phenology model • Implement sequential 2-D var assimilation scheme • Assimilate fAPAR into coupled ECHAM5-BETHY model (hope not too distant) goal: make fAPAR what SST is for ocean-atmosphere interactions... and improve seasonal forecasts

37. Thank You For Your Attention!