Pedro Coimbra

The need to reduce anthropogenic pressure on climate is now shared among most countries worldwide. A key challenge lies in predicting the evolution of greenhouse gases (GHG), notably CO2, by continental surfaces on a changing climate with land use and practices changes. Models are the tool for the job, but they need to be calibrated alongside satellite data themselves. Continuous and precise CO2 flux measurements on networks contribute to advances in this matter. Typically, CO2 fluxes are calculated over 30 minutes averaging periods using the eddy covariance method. FLUXNET and the regional carbon networks (e.g., ICOS, AsiaFlux, AmeriFlux) share this standard. In sum, this method segments each period and correlates the vertical component of the wind-speed with the gas concentration, both sampled at 20 Hz. Spectral approaches include Fourier transformed and wavelet spectral integrations that consist in decomposing the measured signals in frequency that are either localized in time (Wavelet) or not (Fourier). The Wavelet method is hence more adapted to non-stationary and does not destroy the frequency information, including lower spectra. It has been commonly used where time resolution matters the most, for instance, to assign the flux spatially during aircraft flux measurements or capture short-term turbulent events for methane. My research focuses on applying a wavelet approach to retrieve CO2 flux cospectra and apply it in an inversion model using fine resolution remote sensing. We expect with that to find a method for scaling turbulent flux towers measurements to the extent of some km?.