A critical examination of the use of a two-dimensional turbulent profile family to represent three-dimensional boundary layers