An objective prior error quantification for regional atmospheric inverse applications

P. Kountouris, C. Gerbig, K.U. Totsche, A.J. Dolman, A.G.C.A. Meesters, G. Broquet, F. Maignan, B. Gioli, L. Montagnini, C. Helfter

Research output: Contribution to JournalArticleAcademicpeer-review

Abstract

Assigning proper prior uncertainties for inverse modelling of CO2 is of high importance, both to regularise the otherwise ill-constrained inverse problem and to quantitatively characterise the magnitude and structure of the error between prior and "true" flux. We use surface fluxes derived from three biosphere models - VPRM, ORCHIDEE, and 5PM - and compare them against daily averaged fluxes from 53 eddy covariance sites across Europe for the year 2007 and against repeated aircraft flux measurements encompassing spatial transects. In addition we create synthetic observations using modelled fluxes instead of the observed ones to explore the potential to infer prior uncertainties from model-model residuals. To ensure the realism of the synthetic data analysis, a random measurement noise was added to the modelled tower fluxes which were used as reference. The temporal autocorrelation time for tower model-data residuals was found to be around 30 days for both VPRM and ORCHIDEE but significantly different for the 5PM model with 70 days. This difference is caused by a few sites with large biases between the data and the 5PM model. The spatial correlation of the model-data residuals for all models was found to be very short, up to few tens of kilometres but with uncertainties up to 100 % of this estimation. Propagating this error structure to annual continental scale yields an uncertainty of 0.06 Gt C and strongly underestimates uncertainties typically used from atmospheric inversion systems, revealing another potential source of errors. Long spatial e-folding correlation lengths up to several hundreds of kilometres were determined when synthetic data were used. Results from repeated aircraft transects in south-western France are consistent with those obtained from the tower sites in terms of spatial autocorrelation (35 km on average) while temporal autocorrelation is markedly lower (13 days). Our findings suggest that the different prior models have a common temporal error structure. Separating the analysis of the statistics for the model data residuals by seasons did not result in any significant differences of the spatial e-folding correlation lengths.
Original languageEnglish
Pages (from-to)7403-7421
JournalBiogeosciences
Volume12
DOIs
Publication statusPublished - 2015

Fingerprint

uncertainty
autocorrelation
aircraft
folding
transect
eddy covariance
flux measurement
surface flux
inverse problem
biosphere
data analysis
statistics
France
modeling

Cite this

Kountouris, P. ; Gerbig, C. ; Totsche, K.U. ; Dolman, A.J. ; Meesters, A.G.C.A. ; Broquet, G. ; Maignan, F. ; Gioli, B. ; Montagnini, L. ; Helfter, C. / An objective prior error quantification for regional atmospheric inverse applications. In: Biogeosciences. 2015 ; Vol. 12. pp. 7403-7421.
@article{362d30c21457434e9e1ea988a05e8d34,
title = "An objective prior error quantification for regional atmospheric inverse applications",
abstract = "Assigning proper prior uncertainties for inverse modelling of CO2 is of high importance, both to regularise the otherwise ill-constrained inverse problem and to quantitatively characterise the magnitude and structure of the error between prior and {"}true{"} flux. We use surface fluxes derived from three biosphere models - VPRM, ORCHIDEE, and 5PM - and compare them against daily averaged fluxes from 53 eddy covariance sites across Europe for the year 2007 and against repeated aircraft flux measurements encompassing spatial transects. In addition we create synthetic observations using modelled fluxes instead of the observed ones to explore the potential to infer prior uncertainties from model-model residuals. To ensure the realism of the synthetic data analysis, a random measurement noise was added to the modelled tower fluxes which were used as reference. The temporal autocorrelation time for tower model-data residuals was found to be around 30 days for both VPRM and ORCHIDEE but significantly different for the 5PM model with 70 days. This difference is caused by a few sites with large biases between the data and the 5PM model. The spatial correlation of the model-data residuals for all models was found to be very short, up to few tens of kilometres but with uncertainties up to 100 {\%} of this estimation. Propagating this error structure to annual continental scale yields an uncertainty of 0.06 Gt C and strongly underestimates uncertainties typically used from atmospheric inversion systems, revealing another potential source of errors. Long spatial e-folding correlation lengths up to several hundreds of kilometres were determined when synthetic data were used. Results from repeated aircraft transects in south-western France are consistent with those obtained from the tower sites in terms of spatial autocorrelation (35 km on average) while temporal autocorrelation is markedly lower (13 days). Our findings suggest that the different prior models have a common temporal error structure. Separating the analysis of the statistics for the model data residuals by seasons did not result in any significant differences of the spatial e-folding correlation lengths.",
author = "P. Kountouris and C. Gerbig and K.U. Totsche and A.J. Dolman and A.G.C.A. Meesters and G. Broquet and F. Maignan and B. Gioli and L. Montagnini and C. Helfter",
year = "2015",
doi = "10.5194/bg-12-7403-2015",
language = "English",
volume = "12",
pages = "7403--7421",
journal = "Biogeosciences",
issn = "1726-4170",
publisher = "European Geosciences Union",

}

Kountouris, P, Gerbig, C, Totsche, KU, Dolman, AJ, Meesters, AGCA, Broquet, G, Maignan, F, Gioli, B, Montagnini, L & Helfter, C 2015, 'An objective prior error quantification for regional atmospheric inverse applications' Biogeosciences, vol. 12, pp. 7403-7421. https://doi.org/10.5194/bg-12-7403-2015

An objective prior error quantification for regional atmospheric inverse applications. / Kountouris, P.; Gerbig, C.; Totsche, K.U.; Dolman, A.J.; Meesters, A.G.C.A.; Broquet, G.; Maignan, F.; Gioli, B.; Montagnini, L.; Helfter, C.

In: Biogeosciences, Vol. 12, 2015, p. 7403-7421.

Research output: Contribution to JournalArticleAcademicpeer-review

TY - JOUR

T1 - An objective prior error quantification for regional atmospheric inverse applications

AU - Kountouris, P.

AU - Gerbig, C.

AU - Totsche, K.U.

AU - Dolman, A.J.

AU - Meesters, A.G.C.A.

AU - Broquet, G.

AU - Maignan, F.

AU - Gioli, B.

AU - Montagnini, L.

AU - Helfter, C.

PY - 2015

Y1 - 2015

N2 - Assigning proper prior uncertainties for inverse modelling of CO2 is of high importance, both to regularise the otherwise ill-constrained inverse problem and to quantitatively characterise the magnitude and structure of the error between prior and "true" flux. We use surface fluxes derived from three biosphere models - VPRM, ORCHIDEE, and 5PM - and compare them against daily averaged fluxes from 53 eddy covariance sites across Europe for the year 2007 and against repeated aircraft flux measurements encompassing spatial transects. In addition we create synthetic observations using modelled fluxes instead of the observed ones to explore the potential to infer prior uncertainties from model-model residuals. To ensure the realism of the synthetic data analysis, a random measurement noise was added to the modelled tower fluxes which were used as reference. The temporal autocorrelation time for tower model-data residuals was found to be around 30 days for both VPRM and ORCHIDEE but significantly different for the 5PM model with 70 days. This difference is caused by a few sites with large biases between the data and the 5PM model. The spatial correlation of the model-data residuals for all models was found to be very short, up to few tens of kilometres but with uncertainties up to 100 % of this estimation. Propagating this error structure to annual continental scale yields an uncertainty of 0.06 Gt C and strongly underestimates uncertainties typically used from atmospheric inversion systems, revealing another potential source of errors. Long spatial e-folding correlation lengths up to several hundreds of kilometres were determined when synthetic data were used. Results from repeated aircraft transects in south-western France are consistent with those obtained from the tower sites in terms of spatial autocorrelation (35 km on average) while temporal autocorrelation is markedly lower (13 days). Our findings suggest that the different prior models have a common temporal error structure. Separating the analysis of the statistics for the model data residuals by seasons did not result in any significant differences of the spatial e-folding correlation lengths.

AB - Assigning proper prior uncertainties for inverse modelling of CO2 is of high importance, both to regularise the otherwise ill-constrained inverse problem and to quantitatively characterise the magnitude and structure of the error between prior and "true" flux. We use surface fluxes derived from three biosphere models - VPRM, ORCHIDEE, and 5PM - and compare them against daily averaged fluxes from 53 eddy covariance sites across Europe for the year 2007 and against repeated aircraft flux measurements encompassing spatial transects. In addition we create synthetic observations using modelled fluxes instead of the observed ones to explore the potential to infer prior uncertainties from model-model residuals. To ensure the realism of the synthetic data analysis, a random measurement noise was added to the modelled tower fluxes which were used as reference. The temporal autocorrelation time for tower model-data residuals was found to be around 30 days for both VPRM and ORCHIDEE but significantly different for the 5PM model with 70 days. This difference is caused by a few sites with large biases between the data and the 5PM model. The spatial correlation of the model-data residuals for all models was found to be very short, up to few tens of kilometres but with uncertainties up to 100 % of this estimation. Propagating this error structure to annual continental scale yields an uncertainty of 0.06 Gt C and strongly underestimates uncertainties typically used from atmospheric inversion systems, revealing another potential source of errors. Long spatial e-folding correlation lengths up to several hundreds of kilometres were determined when synthetic data were used. Results from repeated aircraft transects in south-western France are consistent with those obtained from the tower sites in terms of spatial autocorrelation (35 km on average) while temporal autocorrelation is markedly lower (13 days). Our findings suggest that the different prior models have a common temporal error structure. Separating the analysis of the statistics for the model data residuals by seasons did not result in any significant differences of the spatial e-folding correlation lengths.

U2 - 10.5194/bg-12-7403-2015

DO - 10.5194/bg-12-7403-2015

M3 - Article

VL - 12

SP - 7403

EP - 7421

JO - Biogeosciences

JF - Biogeosciences

SN - 1726-4170

ER -