Estimation of CO2 exchange at regional scales using spatial correlations of flux fields

At regional to continental scales (> 1000 km), measurements obtained by eddy covariance flux towers are essentially samples at individual points within a flux field. A challenge is to estimate total flux exchange over this field from a limited number of sampling observations. This difficult task can be achieved by combining eddy covariance flux observations with spatial correlation structures of the flux field. It is well known that environmental conditions at different locations of a region are correlated. Remote sensing of gross primary production and net primary production also shows distinctive and coherent spatial structures. Therefore it is not surprising that CO2 flux exchanges at different locations of a region are correlated (either positively or negatively). This correlation provides powerful constraint for regional flux integration. Here a mathematical framework for conducting this integration is outlined. A preliminary analysis of statistical structures of flux fields based on MODIS products is presented. Directions for further research are suggested.