Data assimilation of photosynthetic light-use efficiency using multi-angular satellite data: II Model implementation and validation

REMOTE SENSING OF ENVIRONMENT

卷:121

Article

Hilker, Thomas¹  Hall, Forrest G.¹  Tucker, Compton J.¹  Coops, Nicholas C.²  Black, T. Andrew³  Nichol, Caroline J.⁴  Sellers, Piers J.¹  Barr, Alan⁵  Hollinger, David Y.⁶  Munger, J. W.⁷ 

[1] NASA, Goddard Space Flight Ctr, Biospher Sci Branch, Greenbelt, MD 20771 USA

[2] Univ British Columbia, Fac Forest Resources Management, Vancouver, BC V6T 1Z4, Canada

[3] Univ British Columbia, Fac Land & Food Syst, Vancouver, BC V6T 1Z4, Canada

[4] Univ Edinburgh, Sch Geosci, Edinburgh EH9 3JN, Midlothian, Scotland

[5] Environm Canada, Saskatoon, SK, Canada

[6] US Forest Serv, No Res Stn, Durham, NH USA

[7] Harvard Univ, Cambridge, MA 02138 USA

关键词: Data assimilation;    Photosynthesis;    Eddy-flux;    Multivariate function;    Epsilon;    Epsilon max;    Global carbon cycle;    Carbon modeling;    Vegetation carbon cycle;    Downregulation;    CHRIS/Proba;    PRI';    Multi-angular;   

DOI  :  10.1016/j.rse.2012.02.008

来源: Elsevier

【 摘 要 】

Spatially explicit and temporally continuous estimates of photosynthesis will be of great importance for increasing our understanding of and ultimately closing the terrestrial carbon cycle. Current capabilities to model photosynthesis, however, are limited by accurate enough representations of the complexity of the underlying biochemical processes and the numerous environmental constraints imposed upon plant primary production. A potentially powerful alternative to model photosynthesis through these indirect observations is the use of multi-angular satellite data to infer light-use efficiency (epsilon) directly from spectral reflectance properties in connection with canopy shadow fractions. Hall et al. (this issue) introduced a new approach for predicting gross ecosystem production that would allow the use of such observations in a data assimilation mode to obtain spatially explicit variations in epsilon from infrequent polar-orbiting satellite observations, while meteorological data are used to account for the more dynamic responses of epsilon to variations in environmental conditions caused by changes in weather and illumination. In this second part of the study we implement and validate the approach of Hall et al. (this issue) across an ecologically diverse array of eight flux-tower sites in North America using data acquired from the Compact High Resolution Imaging Spectroradiometer (CHRIS) and eddy-flux observations. Our results show significantly enhanced estimates of epsilon and therefore cumulative gross ecosystem production (GEP) over the course of one year at all examined sites. We also demonstrate that epsilon is greatly heterogeneous even across small study areas. Data assimilation and direct inference of GEP from space using a new, proposed sensor could therefore be a significant step towards closing the terrestrial carbon cycle. (c) 2012 Elsevier Inc. All rights reserved.

【 授权许可】

Free