Computational Diffusion in Atmospheric Boundary Layer Models,
Abstract
A study is made of the truncation error and stability of finite difference advection/diffusion schemes. Their truncation error may be conveniently interpreted as an additional diffusion term and a dispersive phase velocity. Both quantities are functions of the grid spacing, time step, wavelength of the Fourier component and of the differencing scheme itself. The Taylor series determination of the computational or implicit diffusion coefficient yields the first term of a perturbation series. Retaining only the first term is invalid for short wavelengths. A simple, exact method is presented which calculates the computational diffusion coefficient for all wavelengths for equations of constant coefficients. When physical, exp;icit diffusion is introduced into an advection equation, the effects of computational and explicit diffusion are generally not additive except for scales about an order of magnitude larger than the grid interval. The schemes studied are intended for use in one dimensional atmospheric boundary layer models, but the analysis can be extended to higher dimensions. (Author Modified Abstract)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 01, 1972
- Accession Number
- AD0760122
Entities
People
- Paul E. Long
Organizations
- Drexel University