Probabilistic Stability of an Atmospheric Model to Various Amplitude Perturbations

• DOI: 10.1175/1520-0469(2002)059<2860:psoaam>2.0.co;2
• Publication Date: 2002-09-12T17:02:07Z
• Authors: Peter C. Chu; Leonid M. Ivanov; Tatyana M. Margolina; Oleg V. Melnichenko
• Keywords: Predictability; Robustness

Every forecast should include an estimate of its likely accuracy, as a measure predictability. A new measure, the first passage time (FPT), which is defined period when model error exceeds predetermined criterion (i.e., tolerance level), proposed here to theoretical framework developed determine mean and variance FPT. The low-order Lorenz atmospheric taken example show robustness using FPT quantitative for prediction skill. Both linear nonlinear perspectives errors are analytically investigated self-consistent Nicolis model. largely depends on ratio between twice maximum Lyapunov exponent (σ) intensity attractor fluctuations (q2), λ = 2σ/q2. Two types predictability found: > 1 referring low < high can be represented by e-folding timescales in low-predictability range, but not high-predictability range. transition two ranges caused variability characteristics along reference trajectory.