DieCAST FAQ

• For what kinds of applications is DieCAST best suited?
• When is DieCAST's rigid-lid boundary condition a good approximation?
• Can DieCAST accurately simulate density currents?
• Can DieCAST accurately simulate convection and convective adjustment?
• When is the DieCAST hydrostatic approximation valid?
• Are special ad hoc treatments recommended within the DieCAST model (such as instant convective adjustment)?
• Do I need to do any special treatments (e.g. filtering) of bathymetry before using DieCAST?
• Can I get data from DieCAST simulations to use as boundary conditions for my ocean model?
• What platforms  DieCAST can be running on?
• Is there a DieCAST manual
• How do I make a bug report? 
• DieCAST licensing

A: For applications that do not require simulation of a free-surface barotropic mode, the well-validated DieCAST ocean model uses the rigid-lid approximation. For high frequency, free-surface application, please use the Canadian version of DieCAST (CANDIE, http://www.phys.ocean.dal.ca/programs/CANDIE/) which is also well-validated.

For the "slow modes" that dominate the ocean climate, the rigid-lid approximation is a good approximation (see Tseng et al. (2005) for further explanation). High frequency "fast modes" are better treated using the shallow water equations, which can be one- or two-way coupled to the rigid-lid version. An alternate approach is semi-implicit approach used by the CANDIE model that uses the basic 4th-order-accurate, non-staggered-grid DieCAST numerical engine.

YES. DieCAST's recently published results show that high-order, non-dissipative numerics is required to accurately simulate density current using z-coordinate model (Dietrich et al, 2007; Tseng and Dietrich, 2006). See Tseng and Dietrich (2006) for further discussion.

YES. The non-hydrostatic version solves the full, non-hydrostatic primitive equations that govern convection (Dietrich and Lin, 2001; Tseng et al., 2005; Tseng and Breaker, 2007). Further, contrary to the ocean modeling literature, the important big scale convective adjustment can be accurately simulated using the hydrostatic approximation.

Modes having horizontal scales bigger than their vertical scales are quasi-hydrostatic; these dominate the ocean circulation due to its thin-shell, stratified nature except in the high latitudes and perhaps regions being forced by intense violent storms (e.g. hurricanes and east coast winter storms).

NO. Instant convective adjustment is tantamount to assuming infinite vertical diffusivity and is inappropriate for modeling important density currents in the framework of a purely z-level model.

NO. Bathymetry filters, contrary to those who must use them to keep their models from blowing up, inherently decreases the accuracy of any model. The DieCAST model is robust using raw, unfiltered bathymetry.

Data from a global model running a supercomputer multiple-grid version of DieCAST will be soon available. The nearly seamless multiple-grid approach to be used is well validated. The Harvard Ocean Modeling System is already using open boundary conditions from the DieCAST Mediterranean Sea model.