Recovery of sparse urban greenhouse gas emissions

Abstract. To localize and quantify greenhouse gas emissions from cities, concentrations are typically measured at a small number of sites then linked to emission fluxes using atmospheric transport models. Solving this inverse problem is challenging because the system equations often has no unique solution can be sensitive noise. A common top–down approach for solving Bayesian inversion with assumption multivariate Gaussian distribution as prior field. However, such an drawbacks when assumed spatial incorrect or not distributed. In our work, we investigate sparse reconstruction (SR), alternative method that achieve reasonable estimations without field by making sparse. We show generally true cities investigated use discrete wavelet transform helps make urban even more evaluate performance SR, created concentration data applying forward model CO2 inventories several major European cities. used SR locate sources compressed sensing theory compared results regularized least squares (LSs) methods. Our requires fewer measurements than LS methods better localizing quantifying unknown emitters.