fastopt camels a prototype global carbon cycle data assimilation system (ccdas) wolfgang knorr 1,...
DESCRIPTION
FastOpt CAMELS Top-down / Bottom-up atm. CO 2 data inverse atmospheric transport modelling net CO2 fluxes at the surface process model climate and other driving dataTRANSCRIPT
![Page 1: FastOpt CAMELS A prototype Global Carbon Cycle Data Assimilation System (CCDAS) Wolfgang Knorr 1, Marko Scholze 2, Peter Rayner 3,Thomas Kaminski 4, Ralf](https://reader035.vdocuments.us/reader035/viewer/2022062600/5a4d1b6b7f8b9ab0599b3308/html5/thumbnails/1.jpg)
FastOpt CAMELS
A prototype Global Carbon Cycle Data Assimilation
System (CCDAS)
Wolfgang Knorr1, Marko Scholze2, Peter Rayner3,Thomas Kaminski4, Ralf Giering4
1 2 3 4
![Page 2: FastOpt CAMELS A prototype Global Carbon Cycle Data Assimilation System (CCDAS) Wolfgang Knorr 1, Marko Scholze 2, Peter Rayner 3,Thomas Kaminski 4, Ralf](https://reader035.vdocuments.us/reader035/viewer/2022062600/5a4d1b6b7f8b9ab0599b3308/html5/thumbnails/2.jpg)
FastOpt CAMELS
Overview
• Top-down vs. bottom-up• Gradient method and optimisation• Results: Optimal fluxes• Uncertainties in parameters + results• Uncertainties in fluxes + results• Possible assimilation of flux data• Conclusions and Outlook
![Page 3: FastOpt CAMELS A prototype Global Carbon Cycle Data Assimilation System (CCDAS) Wolfgang Knorr 1, Marko Scholze 2, Peter Rayner 3,Thomas Kaminski 4, Ralf](https://reader035.vdocuments.us/reader035/viewer/2022062600/5a4d1b6b7f8b9ab0599b3308/html5/thumbnails/3.jpg)
FastOpt CAMELS
Top-down / Bottom-upatm. CO2 data
inverseatmospheric
transportmodelling
net CO2fluxes at the
surface
processmodel
climate and other driving data
![Page 4: FastOpt CAMELS A prototype Global Carbon Cycle Data Assimilation System (CCDAS) Wolfgang Knorr 1, Marko Scholze 2, Peter Rayner 3,Thomas Kaminski 4, Ralf](https://reader035.vdocuments.us/reader035/viewer/2022062600/5a4d1b6b7f8b9ab0599b3308/html5/thumbnails/4.jpg)
FastOpt CAMELS
Carbon Cycle Data Assimilation
Biosphere Model: BETHY
Atmospheric Transport Model: TM2
Misfit to Observations
Parameters: 58
Fluxes: 800,000
Station Conc. 6,500
Misfit 1 Forward modelling: Parameters –> Misfit
Adjoint or Tangent linear model:
Misfit / ∂ Parameters
parameter optimization
![Page 5: FastOpt CAMELS A prototype Global Carbon Cycle Data Assimilation System (CCDAS) Wolfgang Knorr 1, Marko Scholze 2, Peter Rayner 3,Thomas Kaminski 4, Ralf](https://reader035.vdocuments.us/reader035/viewer/2022062600/5a4d1b6b7f8b9ab0599b3308/html5/thumbnails/5.jpg)
FastOpt CAMELS
Figure taken from Tarantola '87
Space of m (model parameters)
First derivative (Gradient) of J(m) w.r.t. m (model parameters) :
–∂J(m)/∂m yields direction of steepest
descent
Gradient Method
Cost Function J(m)
![Page 6: FastOpt CAMELS A prototype Global Carbon Cycle Data Assimilation System (CCDAS) Wolfgang Knorr 1, Marko Scholze 2, Peter Rayner 3,Thomas Kaminski 4, Ralf](https://reader035.vdocuments.us/reader035/viewer/2022062600/5a4d1b6b7f8b9ab0599b3308/html5/thumbnails/6.jpg)
FastOpt CAMELS
Carbon Cycle Data Assimilation System (CCDAS)
CCDAS Step 2BETHY+TM2
only Photosynthesis, Energy&Carbon Balance
CO2
+ Uncert.
Optimized Params + Uncert.
Diagnostics + Uncert.
veg. indexsatellite +Uncert.
CCDAS Step 1full BETHY
PhenologyHydrology
AssimilatedPrescribedAssimilated
BackgroundCO2 fluxes*
* * ocean: Takahashi et al. (1999), LeQuere et al. (2000); emissions: Marland et al. (2001), Andres et al. (1996); land use: Houghton et al. (1990)
![Page 7: FastOpt CAMELS A prototype Global Carbon Cycle Data Assimilation System (CCDAS) Wolfgang Knorr 1, Marko Scholze 2, Peter Rayner 3,Thomas Kaminski 4, Ralf](https://reader035.vdocuments.us/reader035/viewer/2022062600/5a4d1b6b7f8b9ab0599b3308/html5/thumbnails/7.jpg)
FastOpt CAMELS
BETHY(Biosphere Energy-Transfer-Hydrology Scheme)
• GPP:C3 photosynthesis – Farquhar et al. (1980)C4 photosynthesis – Collatz et al. (1992)stomata – Knorr (1997)
• Raut:maintenance respiration = f(Nleaf, T) – Farquhar, Ryan (1991)growth respiration ~ NPP – Ryan (1991)
• Rhet:fast/slow pool resp. = wQ10
T/10 C fast/slow / fast/slow
slow –> infin.
average NPP = average Rhet (at each grid point)
<1: source>1: sink
t=1h
t=1h
t=1day
lat, lon = 2 deg
![Page 8: FastOpt CAMELS A prototype Global Carbon Cycle Data Assimilation System (CCDAS) Wolfgang Knorr 1, Marko Scholze 2, Peter Rayner 3,Thomas Kaminski 4, Ralf](https://reader035.vdocuments.us/reader035/viewer/2022062600/5a4d1b6b7f8b9ab0599b3308/html5/thumbnails/8.jpg)
FastOpt CAMELS
Optimisation
![Page 9: FastOpt CAMELS A prototype Global Carbon Cycle Data Assimilation System (CCDAS) Wolfgang Knorr 1, Marko Scholze 2, Peter Rayner 3,Thomas Kaminski 4, Ralf](https://reader035.vdocuments.us/reader035/viewer/2022062600/5a4d1b6b7f8b9ab0599b3308/html5/thumbnails/9.jpg)
FastOpt CAMELS
Optimised fluxes (1)
global fluxes
Major El Niño events
Major La Niña event
Post Pinatubo Period
![Page 10: FastOpt CAMELS A prototype Global Carbon Cycle Data Assimilation System (CCDAS) Wolfgang Knorr 1, Marko Scholze 2, Peter Rayner 3,Thomas Kaminski 4, Ralf](https://reader035.vdocuments.us/reader035/viewer/2022062600/5a4d1b6b7f8b9ab0599b3308/html5/thumbnails/10.jpg)
FastOpt CAMELS
Optimised fluxes (2)normalized CO2 flux and ENSO
ENSO and terr. biosph. CO2:correlation seems strong
lag correlation(low-pass filtered)
correlation between Niño-3 SST anomaly and net CO2 flux shows maximum at 4 months lag, forboth El Niño and La Niña states
![Page 11: FastOpt CAMELS A prototype Global Carbon Cycle Data Assimilation System (CCDAS) Wolfgang Knorr 1, Marko Scholze 2, Peter Rayner 3,Thomas Kaminski 4, Ralf](https://reader035.vdocuments.us/reader035/viewer/2022062600/5a4d1b6b7f8b9ab0599b3308/html5/thumbnails/11.jpg)
FastOpt CAMELS
net CO2 flux to atm.gC / (m2 month)
during El Niño (>+1)
Optimised fluxes (3)
lagged correlationat 99% significance
-0.8 -0.4 0 0.4 0.8
flux sites?
![Page 12: FastOpt CAMELS A prototype Global Carbon Cycle Data Assimilation System (CCDAS) Wolfgang Knorr 1, Marko Scholze 2, Peter Rayner 3,Thomas Kaminski 4, Ralf](https://reader035.vdocuments.us/reader035/viewer/2022062600/5a4d1b6b7f8b9ab0599b3308/html5/thumbnails/12.jpg)
FastOpt CAMELS
Figure taken from Tarantola '87
J(x)
Second Derivative (Hessian) of J(m):
∂2J(m)/∂m2 yields curvature of J,provides estimateduncertainty in mopt
Error Covariances in Parameters
Space of m (model parameters)
![Page 13: FastOpt CAMELS A prototype Global Carbon Cycle Data Assimilation System (CCDAS) Wolfgang Knorr 1, Marko Scholze 2, Peter Rayner 3,Thomas Kaminski 4, Ralf](https://reader035.vdocuments.us/reader035/viewer/2022062600/5a4d1b6b7f8b9ab0599b3308/html5/thumbnails/13.jpg)
FastOpt CAMELS
Error Covariances in Parameters
first guess optimized prior unc. opt.unc. Vm(TrEv) Vm(EvCn) Vm(C3Gr) Vm(Crop)µmol/m 2s µmol/m 2s % %
Vm(TrEv) 60.0 43.2 20.0 10.5 0.28 0.02 -0.02 0.05Vm(EvCn) 29.0 32.6 20.0 16.2 0.02 0.65 -0.10 0.08Vm(C3Gr) 42.0 18.0 20.0 16.9 -0.02 -0.10 0.71 -0.31Vm(Crop) 117.0 45.4 20.0 17.8 0.05 0.08 -0.31 0.80
error covarianceexamples:
Cost function (misift):
€
J ( r m ) =12[ r m −r m 0]Cm0
-1[ r m −r m 0]+12[r y ( r m )−r y 0]Cy
-1[r y (r m )−r y 0]model diagnostics
error covariance matrixof measurements
measurements
assumedmodel parameters
a priori covariance matrixof parameter + model errora priori
parameter values
Error covariance of parametersafter optimisation:
€
Cm = ∂2J∂mi, j
2⎧ ⎨ ⎩
⎫ ⎬ ⎭
−1
= inverse Hessian
![Page 14: FastOpt CAMELS A prototype Global Carbon Cycle Data Assimilation System (CCDAS) Wolfgang Knorr 1, Marko Scholze 2, Peter Rayner 3,Thomas Kaminski 4, Ralf](https://reader035.vdocuments.us/reader035/viewer/2022062600/5a4d1b6b7f8b9ab0599b3308/html5/thumbnails/14.jpg)
FastOpt CAMELS
Relative Error Reduction
![Page 15: FastOpt CAMELS A prototype Global Carbon Cycle Data Assimilation System (CCDAS) Wolfgang Knorr 1, Marko Scholze 2, Peter Rayner 3,Thomas Kaminski 4, Ralf](https://reader035.vdocuments.us/reader035/viewer/2022062600/5a4d1b6b7f8b9ab0599b3308/html5/thumbnails/15.jpg)
FastOpt CAMELS
Error Covariances in Diagnostics
Error covariance of diagnostics, y,after optimisation (e.g. CO2 fluxes):
€
Cy( r m opt) = ∂yi( r m opt)∂mj
⎛ ⎝ ⎜ ⎞
⎠ ⎟ Cm
∂yi(r m opt)∂mj
⎛ ⎝ ⎜ ⎞
⎠ ⎟
T
error covarianceof parameters
adjoint ortangent linear
model
![Page 16: FastOpt CAMELS A prototype Global Carbon Cycle Data Assimilation System (CCDAS) Wolfgang Knorr 1, Marko Scholze 2, Peter Rayner 3,Thomas Kaminski 4, Ralf](https://reader035.vdocuments.us/reader035/viewer/2022062600/5a4d1b6b7f8b9ab0599b3308/html5/thumbnails/16.jpg)
FastOpt CAMELS
Regional Net Carbon Balance and Uncertainties
![Page 17: FastOpt CAMELS A prototype Global Carbon Cycle Data Assimilation System (CCDAS) Wolfgang Knorr 1, Marko Scholze 2, Peter Rayner 3,Thomas Kaminski 4, Ralf](https://reader035.vdocuments.us/reader035/viewer/2022062600/5a4d1b6b7f8b9ab0599b3308/html5/thumbnails/17.jpg)
FastOpt CAMELS
Comparison shows impact of a (pseudo) flux measurement in the broadleaf evergreen biome on Q10 estimated by an inversion of SDBM:
Upper panel: only concentration data
Lower panel: concentration data +pseudo flux measurement(mean: as predicted sigma: 10gC/m^2/year)
Details:Kaminski et al., GBC, 2001
a posteriori mean/uncertaintiesa priori mean/uncertainties
![Page 18: FastOpt CAMELS A prototype Global Carbon Cycle Data Assimilation System (CCDAS) Wolfgang Knorr 1, Marko Scholze 2, Peter Rayner 3,Thomas Kaminski 4, Ralf](https://reader035.vdocuments.us/reader035/viewer/2022062600/5a4d1b6b7f8b9ab0599b3308/html5/thumbnails/18.jpg)
FastOpt CAMELS
• CCDAS with 58 parameters can already fit 20 years of CO2 concentration data
• Sizeable reduction of uncertainty for ~13 parameters
• terr. biosphere response to climate fluctuations dominated by ENSO
• System can test model with uncertain parameters, and deliver a posteriori uncertainties on parameters, fluxes
Conclusions
![Page 19: FastOpt CAMELS A prototype Global Carbon Cycle Data Assimilation System (CCDAS) Wolfgang Knorr 1, Marko Scholze 2, Peter Rayner 3,Thomas Kaminski 4, Ralf](https://reader035.vdocuments.us/reader035/viewer/2022062600/5a4d1b6b7f8b9ab0599b3308/html5/thumbnails/19.jpg)
FastOpt CAMELS
• explore more parameter configurations• include fire as a process with uncertainties• need more constraints, e.g. eddy fluxes –>
reduce uncertainties• however: needs to solve scaling problem
(satellites?)• approach can be regionalized easily• extend approach to ocean carbon cycle• projection of uncertainties into future
Outlook