Claude Literature Review
Ill-Posedness and Unphysical High-Frequency
Growth in Geophysical and Climate Science Model Equations: A Survey of Known Examples
[Prepared by Claude as companion document to the two-phase flow dispersion analysis series]
March 20, 2026
The Unifying Mathematical Signature.
Every example in this survey shares the same mathematical fingerprint: a system of first-order partial differential equations (PDEs) whose linearized coefficient matrix has complex characteristics in some parameter regime, or whose initial-boundary value problem is ill-posed in the sense of Hadamard .
Hadamard ill-posedness means that no finite constant C exists such that the growth rate σ(k)≤C for all wavenumbers k. Instead, σ(k)→∞ as k→∞. Grid refinement invites faster growth, making convergence impossible by construction.
Critical Principle.
When ill-posedness is identified, the mathematically and physically correct response is to identify the genuine physical process that provides the missing stabilizing mechanism and to include it from first principles. Adding artificial dissipation, drag, sponge layers, or flux limiters solely to achieve numerical stability — without physical justification from the fundamental equations — suppresses a real mathematical pathology without curing it, and introduces a false length or time scale that contaminates all predictions at that scale.
This distinction between genuine physical regularization and what the two-phase thermal-hydraulics community has called “regularization by illegal addition” is the central concern of this survey.
Practical Ramifications.
The Lax Equivalence Theorem states that convergence of numerical solutions of discrete approximations to solutions of the continuous equations is ensured when (1) the finite difference approximations are consistent with the continuous equations, and (2) the numerical solution method is stable. The presence of complex characteristics ensures that stable numerical solution methods for consistent discrete approximations are not possible.
These aspects of ill-posed equation systems are summarized in the paper by Lyczkowski et al. (1978), in which citations to the basic literature are given. The papers by Gidaspow (1974) and Gidaspow et al. (1973) provide additional information about the fundamental problems.
Various regularizations are frequently used in attempts to rectify the outcome of ill-posed equation systems. These will sometimes make the finite approximations differ from a consistent approximation. Convergence to solutions of the continuous equations is not ensured.
A PDF file that has a summary of the literature serach is here.
No comments yet.
Models Methods Software 
Leave a comment