An average-derivative optimal scheme for frequency-domain scalar wave equation

Forward modeling is an important foundation of full-waveform inversion. The rotated optimal nine-point scheme is an efficient algorithm for frequency-domain 2D scalar wave equation simulation, but this scheme fails when directional sampling intervals are different. To overcome the restriction on directional sampling intervals of the rotated optimal nine-point scheme, I introduce a new finite-difference algorithm. Based on an average-derivative technique, this new algorithm uses a nine-point operator to approximate spatial derivatives and mass acceleration term. The coefficients can be determined by minimizing phase-velocity dispersion errors. The resulting nine-point optimal scheme applies to equal and unequal directional sampling intervals, and can be regarded a generalization of the rotated optimal nine-point scheme. Compared to the classical five-point scheme, the number of grid points per smallest wavelength is reduced from 13 to less than four by this new nine-point optimal scheme for equal and unequal directional sampling intervals. Three numerical examples are presented to demonstrate the theoretical analysis. The average-derivative algorithm is also extended to a 2D viscous scalar wave equation and a 3D scalar wave equation.