A global carbon assimilation system using a modified ensemble Kalman filter

Zhang, Shizhong; Zheng, Xunhua; Chen, J. M.; Chen, Z.; Dan, Bo; Yi, Xiaobin; Wang, Lu; Wu, Guorong

doi:10.5194/gmd-8-805-2015

Cited by 19 publications

(17 citation statements)

References 32 publications

(45 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…χ 2 values of GCAS range from 1 to 2.5 with a mean of 1.53 (Figure ). It is reasonable and consistent with the results from GCAS without error inflations (Zhang et al, ). A clear seasonality of χ 2 can be found in Figure .…”

Section: Resultssupporting

confidence: 90%

“…Chi‐square statistics ( χ 2 ) are usually used to test the error covariances or the innovations in a data assimilation system (Zhang et al, ; Zupanski & Zupanski, ). For the t th time step, it is defined as

χ^{2} = \frac{1}{N_{obs}} {(c_{t} - M (x_{t}^{m}))}^{T} {(\frac{1}{n - 1} M R_{t} {boldM}^{T} + O_{t} + P_{t})}^{- 1} ((), c_{t} - M ((), x_{t}^{m}))

…”

Section: Resultsmentioning

confidence: 99%

“…There are some ways to help reduce errors in a data assimilation system. Our previous results showed that the inflations on forecast errors and observation errors applied to GCAS can improve the estimation of error statistics (Zhang et al, ). Furthermore, more observations used or control variables optimized in a data assimilation system also can help reduce the errors (Zupanski & Zupanski, ).…”

Section: Resultsmentioning

confidence: 99%

“…As described by Zhang et al (, ), the ensemble Kalman filter was used in GCAS. The length of an assimilation window is set to 1 week.…”

Section: Gcasmentioning

confidence: 99%

“…This study aims at exploring spatiotemporal variations of TBM parameters using atmospheric CO 2 concentration data through the newly developed Global Carbon Assimilation System (GCAS). GCAS is developed for optimizing global land surface net carbon fluxes and TBM parameters at 1° resolution at weekly time intervals (Chen et al, ; Zhang et al, , ; Zheng et al, , ).…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Optimization of Terrestrial Ecosystem Model Parameters Using Atmospheric CO₂ Concentration Data With the Global Carbon Assimilation System (GCAS)

Chen

Zhang

et al. 2017

JGR Biogeosciences

View full text Add to dashboard Cite

The Global Carbon Assimilation System that assimilates ground‐based atmospheric CO2 data is used to estimate several key parameters in a terrestrial ecosystem model for the purpose of improving carbon cycle simulation. The optimized parameters are the leaf maximum carboxylation rate at 25°C ( Vmax25), the temperature sensitivity of ecosystem respiration (Q10), and the soil carbon pool size. The optimization is performed at the global scale at 1° resolution for the period from 2002 to 2008. The results indicate that vegetation from tropical zones has lower Vmax25 values than vegetation in temperate regions. Relatively high values of Q10 are derived over high/midlatitude regions. Both Vmax25 and Q10 exhibit pronounced seasonal variations at middle‐high latitudes. The maxima in Vmax25 occur during growing seasons, while the minima appear during nongrowing seasons. Q10 values decrease with increasing temperature. The seasonal variabilities of Vmax25 and Q10 are larger at higher latitudes. Optimized Vmax25 and Q10 show little seasonal variabilities at tropical regions. The seasonal variabilities of Vmax25 are consistent with the variabilities of LAI for evergreen conifers and broadleaf evergreen forests. Variations in leaf nitrogen and leaf chlorophyll contents may partly explain the variations in Vmax25. The spatial distribution of the total soil carbon pool size after optimization is compared favorably with the gridded Global Soil Data Set for Earth System. The results also suggest that atmospheric CO2 data are a source of information that can be tapped to gain spatially and temporally meaningful information for key ecosystem parameters that are representative at the regional and global scales.

show abstract

Section: Resultssupporting

confidence: 90%