“…Even for uniform in longitudinal direction channels, it is not easy to find the solution for the nonlinear traveling waves (solitons, cnoidal waves) analytically. This problem requires solving the 2D Laplace equation in cross-section domain with curvilinear boundary (Peregrine, 1968(Peregrine, , 1969Fenton, 1973;Shen and Zhong, 1981;Das, 1985;Mathew and Akylas, 1990;Teng and Wu, 1992, 1997Caputo and Stepanyants, 2003). At the same time approximation of a rectangular cross-section does not correspond to natural narrow bays, which bottom configuration varies in both longitudinal and transversal directions.…”