Complex oscillatory dynamics in a three-timescale El Niño Southern Oscillation model

The paper defines the three-timescale ENSO model, its critical manifold M1 = MP ∪ MS with MP = {x=0}, MS = {x + y + c(1 − tanh(x+z)) = 0}, the loss-of-normal-hyperbolicity sets FP and L±, and the 2-critical manifold M2 with M2P and M2S; it then classifies parameter space into A1–A3 (via a(x) = x^2/d(x) and a±(c,k)) and V1–V6 regimes, proves/argues for singular Hopf onsets at a = a± + O(δ,ρ) (none in A3), and derives the plateau entry–exit integral W(zin,zout)=0 on MP, all of which the candidate solution reproduces with the same geometric mechanisms and references. See the model formulation and M1/FP definitions (equations (5), (9)–(18) in the paper) ; the M2S eigenvalue/folds/folded singularities q± and their coordinates (Section 2.3) ; the a(x)=x^2/d(x) map and a± thresholds (equations (55)–(58)) ; singular Hopf localization “a = a± + O(δ,ρ)” and the no-Hopf case in A3 (Section 3.1.3) ; and the way-in/way-out plateau integral W from Lemma 7 (equation (60)) . The regime-wise outcomes match V1–V6 in the paper (Section 3.2 and Conclusion) including plateau vs plateauless distinctions and SAOs near Hopf . Minor differences: the model claims an O(δ) Hopf-location uniform in ρ, whereas the paper states O(δ,ρ); and the model states “unique globally attracting equilibrium” in A3, while the paper asserts convergence to steady state and notes uniqueness of the equilibrium numerically. These are modest strengthenings, not contradictions.