TY - JOUR
T1 - A linked carbon cycle and crop developmental model: Description and evaluation against measurements of carbon fluxes and carbon stocks at several European agricultural sites
AU - Sus, O.
AU - Williams, M.
AU - Bernhofer, C.
AU - Beziat, P.
AU - Buchmann, N.
AU - Ceschia, E.
AU - Doherty, R.
AU - Eugster, W.
AU - Grunwald, T.
AU - Kutsch, W.
AU - Smith, P.
AU - Wattenbach, M.
PY - 2010/11/15
Y1 - 2010/11/15
N2 - Croplands are the largest biospheric source of carbon (C) to the atmosphere in Europe each year (similar to 0.3 GtC y(-1)). Moreover, the biological dynamics of managed landscapes affect the fluctuations of atmospheric CO2 levels on annual and interannual scales. The activity of croplands must therefore be included in efforts to quantify, understand and regulate the global C cycle. C flux data for croplands, existing mostly from similar to 2003 onwards, provide an opportunity to better understand the timing and magnitude of C exchanges in croplands and can help to improve crop C modelling on various spatiotemporal scales. A crucial process to be considered for C budget modelling is crop development, to which the C allocation pattern and senescence and remobilisation processes are directly linked.We developed, parameterised, and tested a new crop C mass balance model (SPAc), which included all major developmental stages (DS) and C remobilisation. Model results were compared against independent observations of C fluxes and biometric data for eight site years, at six European sites growing winter wheat and barley. We used a single parameterisation for all sites, varying only maximal carboxylation rate (V-cmax), maximal photosynthetic electron transport rate (J(max)), and C per leaf area (C-la) between wheat and barley. The model effectively described the seasonal dynamics of net ecosystem exchange (NEE) through simulation of DS. We found high model reliability for predicting daily NEE fluxes (mean R-2 = 0.83, mean RMSE = 1.47 gC m(-2) d(-1)). Results for cumulative NEE from sowing to harvest were of lower quality (mean R-2 = 0.27). The model tended to overpredict leaf area index (LAI, 160% of observed), and underpredict final yield (71% of observed). The introduction of a dead leaf C pool considerably improved at-harvest estimates of total leaf C mass. Simulating the shift from green to brown above-ground biomass is of potential significance for solving the surface energy balance and for future remote sensing studies through an improved simulation of crop canopy optical properties. (C) 2010 Elsevier B.V. All rights reserved.
AB - Croplands are the largest biospheric source of carbon (C) to the atmosphere in Europe each year (similar to 0.3 GtC y(-1)). Moreover, the biological dynamics of managed landscapes affect the fluctuations of atmospheric CO2 levels on annual and interannual scales. The activity of croplands must therefore be included in efforts to quantify, understand and regulate the global C cycle. C flux data for croplands, existing mostly from similar to 2003 onwards, provide an opportunity to better understand the timing and magnitude of C exchanges in croplands and can help to improve crop C modelling on various spatiotemporal scales. A crucial process to be considered for C budget modelling is crop development, to which the C allocation pattern and senescence and remobilisation processes are directly linked.We developed, parameterised, and tested a new crop C mass balance model (SPAc), which included all major developmental stages (DS) and C remobilisation. Model results were compared against independent observations of C fluxes and biometric data for eight site years, at six European sites growing winter wheat and barley. We used a single parameterisation for all sites, varying only maximal carboxylation rate (V-cmax), maximal photosynthetic electron transport rate (J(max)), and C per leaf area (C-la) between wheat and barley. The model effectively described the seasonal dynamics of net ecosystem exchange (NEE) through simulation of DS. We found high model reliability for predicting daily NEE fluxes (mean R-2 = 0.83, mean RMSE = 1.47 gC m(-2) d(-1)). Results for cumulative NEE from sowing to harvest were of lower quality (mean R-2 = 0.27). The model tended to overpredict leaf area index (LAI, 160% of observed), and underpredict final yield (71% of observed). The introduction of a dead leaf C pool considerably improved at-harvest estimates of total leaf C mass. Simulating the shift from green to brown above-ground biomass is of potential significance for solving the surface energy balance and for future remote sensing studies through an improved simulation of crop canopy optical properties. (C) 2010 Elsevier B.V. All rights reserved.
KW - Crop modelling
KW - Phenology
KW - Carbon budget
KW - Land management
KW - Eddy flux
KW - Yield
KW - NET ECOSYSTEM EXCHANGE
KW - WINTER-WHEAT CROP
KW - SIMULATION-MODEL
KW - PHENOLOGICAL DEVELOPMENT
KW - STOMATAL CONDUCTANCE
KW - SYSTEMS SIMULATION
KW - CLIMATE-CHANGE
KW - WATER
KW - RESPIRATION
KW - GROWTH
UR - http://www.scopus.com/inward/record.url?scp=78449247816&partnerID=8YFLogxK
U2 - 10.1016/j.agee.2010.06.012
DO - 10.1016/j.agee.2010.06.012
M3 - Article
SN - 0167-8809
VL - 139
SP - 402
EP - 418
JO - Agriculture, Ecosystems & Environment
JF - Agriculture, Ecosystems & Environment
IS - 3
ER -