Prolate spheroidal wavefunctions as an alternative to Chebyshev and Legendre polynomials for spectral element and pseudospectral algorithms (Q1883505): Difference between revisions

Prolate spheroidal functions of order zero are generalizations of Legendre functions which oscillate more uniformly than either Legendre or Chebyshev polynomials. This suggests that prolate functions yield more uniform spatial resolution and allow a longer stable timestep than Legendre polynomials. It is shown that these advantages are real and that it is trivial to modify existing pseudospectral or spectral element codes to use the prolate basis. The prolate basis is not likely to radically expand the range of problems but it is expected that there are improvements for applications, such as numerical weather prediction.