Extra Invariant in Magnetohydrodynamics of Planets and Stars

Referee: Section 3: The dynamo argument requires the extra invariant to be conserved on the cascade timescale, but the invariant is exact only for the Padé approximant dispersion relation (Eq. 13), not for the exact dispersion relation (Eq. 8). The author acknowledges that the Taylor expansion (Eq. 12) 'quickly separates from the exact lowest frequency,' but no estimate is provided for the rate at which the invariant leaks due to the discrepancy between Eqs. (8) and (13). If the leakage rate is comparable to or faster than the inverse cascade rate, the constraint forcing energy toward the q-axis disappears. This is load-bearing for the dynamo claim and needs at least an order-of-magnitude estimate or a scaling argument.

Authors: The referee is correct that this is the load-bearing issue for the dynamo argument, and we agree that an estimate is needed. We will add a scaling argument in the revised manuscript. The key observation is that the Padé approximant (Eq. 13) matches the exact lowest-frequency root of Eq. (8) through O(k^7), so the relative discrepancy between the approximate and exact dispersion relations is O(k^4) for small k (since the leading term is O(k^3) and the correction is O(k^7)). The invariant is conserved exactly for three-wave resonances of the approximate dispersion relation; for the exact dispersion relation, the resonance conditions are perturbed, and the invariant leaks at a rate proportional to the discrepancy in the resonance manifold times the nonlinear interaction rate. If the nonlinear broadening of resonances (the resonance width) is δω, then the leakage rate scales as (δω_exact - δω_approx)/δω_approx ~ O(k^4). The inverse cascade rate, by contrast, scales as the nonlinear frequency shift divided by the frequency itself. For the dispersion relation (13), the frequency is O(k^3) (in the long-wave limit), while the nonlinear interaction time scales as 1/(εk^2) where ε is the wave amplitude. The ratio of leakage rate to cascade rate thus scales as O(k^4) times a factor involving the wave amplitude spectrum. For the long waves that participate in the inverse cascade (k ~ 1/R, where R is the system size), this ratio is small provided conditions (20a–b) are satisfied, which we already show holds for Earth's core. We will include this scaling argument explicitly in the revised Section 3, while acknowledging that a rigorous calculation of the leakage rate would require solving the kinetic equation with the exact dispersion relation, which is beyond the scope of this paper. revision: partial

Referee: Section 3, paragraph following Eq. (19): The cascade argument is qualitative. The paper inspects contour plots of the ratio φ̃/ω (Fig. 1) and argues heuristically that energy must accumulate near the q-axis because accumulating elsewhere would require 'too much' extra invariant. This is not derived from the wave kinetic equation—no stationary spectrum, no flux calculation, no demonstration that the cascade actually follows the predicted pathway. For the dynamo claim to be credible, at minimum a scaling argument for the cascade flux direction based on the invariant structure should be provided, or the claim should be substantially hedged.

Authors: We agree that the cascade argument as currently presented is qualitative and that a full kinetic equation analysis (stationary spectrum, flux calculation) is not provided. We will take both actions the referee suggests: we will add a scaling argument and substantially hedge the dynamo claim. The scaling argument is as follows. The extra invariant Ĩ = ∫ φ̃(k) N_k dk constrains the energy distribution because the ratio φ̃/ω diverges away from the q-axis (as shown in Fig. 1). If energy accumulates at large scales away from the q-axis (|p| ~ |q| ~ k → 0), the extra invariant per unit energy scales as φ̃/ω ~ 1/k^2 (from Eq. 18, since φ̃ = O(p^3) and ω = O(pk^2), so φ̃/ω ~ p^2/k^2 ~ O(1) near the q-axis but ~ O(1/k^2) away from it when p ~ q ~ k). This means that the extra invariant density grows without bound as k → 0 away from the q-axis, while the source injects a finite amount of extra invariant per unit energy. Therefore, the inverse cascade cannot deposit energy at large scales away from the q-axis without violating the invariant budget; it is funneled toward the q-axis. This is a scaling argument based on the invariant structure, not merely contour inspection. However, we acknowledge that this does not constitute a derivation from the kinetic equation. We will revise the language in Section 3 to describe this as a 'scaling constraint' rather than a derived cascade pathway, and we will explicitly state that a kinetic equation analysis is needed to confirm the predicted energy accumulation. The dynamo claim will be hedged accordingly: we will state that the invariant 'suggests' or 'is consistent with' energy accumulation in the toroidal magnetic field, rather than claiming it 'requires' dynamo action. revision: partial

Referee: Eq. (13) vs. Eq. (8): No quantitative comparison between the Padé approximant and the exact lowest-frequency root of Eq. (8) is provided. The author notes that the Taylor expansion (Eq. 12) diverges from the exact solution at large k, but does not show over what range of k the Padé approximant remains accurate. Since the invariant is inherited from the approximate dispersion relation, specifying the regime of validity in k-space is essential for assessing whether the invariant is meaningful for the waves that actually participate in the inverse cascade.

Authors: This is a fair and important point. We will add a quantitative comparison in the revised manuscript. Specifically, we will include a figure (or table) showing the relative error |ω_Padé - ω_exact|/|ω_exact| as a function of k = sqrt(p^2 + q^2) for representative parameter values (including the case α = 2, i.e., A = c, used in Fig. 1). The Padé approximant (Eq. 13) matches the exact root through O(k^7), so the relative error is O(k^4) for small k. For the parameter values relevant to Earth's core (m ~ 90 km, R ~ 3475 km), the long waves participating in the inverse cascade have k ~ 1/R ~ 3×10^{-4} km^{-1}, while m^{-1} ~ 1.1×10^{-2} km^{-1}, so kR_m ~ k/m^{-1} ~ 0.03. At these wavenumbers, the Padé approximant is accurate to within a fraction of a percent. We will show that the approximant remains accurate (say, within 10%) up to k ~ O(1/m), which covers the range of wavenumbers relevant to the inverse cascade under conditions (20a–b). We will also note that the Padé approximant has the correct large-k behavior (ω ~ k, as does the exact solution), unlike the Taylor expansion (Eq. 12) which grows as k^5, so the Padé approximant is qualitatively correct at all k even if quantitatively imprecise at large k. This comparison will be added to Section 2, between the current discussion of Eqs. (12) and (13). revision: yes