“…The observed attenuation is so slow that one of the major tools for analyzing the dynamics of solitary wave-like solutions, namely approximations of the problem with the Korteweg-de Vries (KdV) (or similar) equations [12,30,15,6,29], is incapable of capturing the phenomena. Such approximations are valid over very long but nevertheless finite time intervals; the erosion is so subtle during the period of good approximation that it falls within the natural error bounds.…”