American Geophysical Union Become an AGU Member
Subscribe to AGU Journals		 AGU Home AGU Publications

Read Full Article    Cited by

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 100, NO. C10, PAGES 20,763–20,778, 1995

Upwind-weighted advection schemes for ocean tracer transport: An evaluation in a passive tracer context

Matthew W. Hecht

National Center for Atmospheric Research, Boulder, Colorado


William R. Holland

National Center for Atmospheric Research, Boulder, Colorado


Philip J. Rasch

National Center for Atmospheric Research, Boulder, Colorado


Abstract

Centered-in-space, centered-in-time integration has generally been used for the advection of scalars in ocean models. An assessment is made of the implications of centered leapfrog integration in the context of two-dimensional passive tracer advection within a Stommel (1948) gyre. Nonphysical ripples in the tracer field grow to alarming levels in purely advective integrations. Diffusive parameterizations of eddy mixing moderate these ripples, but it is found that Laplacian diffusion greatly reduces the peak amplitude of the tracer field, while biharmonic or weaker Laplacian diffusion allows ripples of large amplitude. Several forward-in-time, upwind-weighted schemes are found to provide better solutions. Smolarkiewicz's (1984) Multi-Dimensional Positive-Definite Advection and Transport Algorithm (MPDATA) scheme is slightly superior for an integration at moderate resolution within which the western boundary current is poorly resolved in typical fashion. Third-order, upwind-based schemes exhibit little sensitivity to the details of multidimensional treatment for this problem of passive tracer advection, with results nearly as good as for MPDATA.

Received 28 December 1994; accepted 13 June 1995.

Read Full Article    Cited by

Citation: Hecht, M. W., W. R. Holland, and P. J. Rasch (1995), Upwind-weighted advection schemes for ocean tracer transport: An evaluation in a passive tracer context, J. Geophys. Res., 100(C10), 20,763–20,778.