Non-linear weak lensing forecasts

We investigate the impact of non-linear corrections on dark energy parameter estimation from weak lensing probes. We find that using halofit expressions, suited to LCDM models, implies substantial discrepancies with respect to results directly obtained from N-body simulations, when w(z)\neq-1. Discrepancies appear strong when using models with w'(z=0)>0, as fiducial models; they are however significant even in the neighborhood of LCDM, where neglecting the degrees of freedom associated with the DE state equation can lead to a misestimate of the matter density parameter \Omega_m.