How accurately can the deuterium abundance be determined?

By using a Reverse Monte Carlo technique aimed at the inverse problem in the analysis of the H+D Ly-a absorption from QSO spectra we have estimated the physical parameters of the gas and the large-scale line of sight velocity structure towards Q1009+2956 at z = 2.504 (observations by Burles and Tytler, astro-ph/9603070). We show that for a precise evaluation of the column densities, the detailed velocity field structure is to be determined simultaneously. Our study of the z = 2.504 system yields D/H = (3.75 +/- 0.85)*10^{-5} (2sigma) which gives good agreement between the standard BBN predictions and observational constraints on extra-galactic 4He mass fraction and 7Li abundances in the atmospheres of population II (halo) starts. We conclude that the discordance of D/H with the light element abundances discussed in the literature is a consequence of the assumption of microturbulence (i.e. completely uncorrelated bulk motions) in the analysis of line profiles. The generalized model accounting for a finite correlation length leads to the common eta-interval of (3.9-5.3)*10^{-10}, where eta is the baryon to photon ratio