7# Unified Lunisolar \(k=0\) Forcing: Full Mathematical Derivation
P. Pukite et al derived from Chapters 11-13, Mathematical Geoenergy (Wiley/AGU, 2019)
Synthesis of discussion at geoenergymath.com, April 2026
Two geographically and dynamically distinct climate time-series are considered:
Let \(x_i(t)\) denote the observed data (Column 3) and \(m_i(t)\) the model output (Column 2) for system \(i \in \{1,2\}\). The raw cross-correlation is negligible:
\[\rho(x_1, x_2) \approx -0.088\]However, the hidden latent forcing manifold \(F(t)\) (Column 4) shared by both systems yields:
\[\rho(F_1, F_2) \approx 0.978\]This motivates the central hypothesis: both systems are observations of a single global low-dimensional forcing manifold, modulated locally into their disparate expressions.
Each observed signal \(x_i(t)\) is modeled as a non-autonomous sinusoidal modulation applied to the shared latent manifold \(F(t)\):
\[\boxed{x_i(t) \approx m_i(t) = A_i \sin\!\bigl(k_i F(t) + \varphi_i(t)\bigr)}\]where:
The dynamical equation governing the local state can be written as:
\[\dot{x}_i = f_i(x_i, t) + g_i(\lambda(t))\]where \(\lambda(t) \equiv F(t)\) is the global forcing and \(f_i\) encodes local non-autonomous sinusoidal modulations.
The non-autonomous phase is decomposed into a mean and a local seasonal component:
\[\varphi_i(t) = \bar{\varphi}_i + \sum_{n} \alpha_{i,n} \cos\!\left(\frac{2\pi n t}{T_{\rm yr}} + \psi_{i,n}\right)\]where \(T_{\rm yr} = 1\) year. This encodes the “local filter” — e.g., North Atlantic Oscillation wind stress for Warnemünde, or thermocline tilt for NINO4.
The Draconic lunar month \(T_D\) is the time for the Moon to return to the same orbital node (ascending or descending, crossing the ecliptic):
\[T_D = 27.21222 \text{ days} \implies f_D = \frac{365.25}{27.21222} \approx 13.4223 \text{ cycles/year (cpy)}\]The annual solar cycle acts as a stroboscopic sample-and-hold reference at \(f_{\rm yr} = 1\) cpy. This produces an aliased Draconic frequency via the modular (mod 1 year) folding:
\[f_{\rm alias} = f_D \pmod{1} = 13.4223 - 13 = 0.4223 \text{ cpy}\]The corresponding alias period is:
\[\boxed{T_{\rm alias} = \frac{1}{f_{\rm alias}} = \frac{1}{0.4223} \approx 2.368 \text{ years}}\]This is indistinguishable from the mean period of the Quasi-Biennial Oscillation (QBO) (~28 months ≈ 2.33-2.37 yr), which is here identified as externally forced rather than internally generated.
The annual “stroboscopic” sampling of the Draconic cycle is represented as a convolution of the Draconic cosine with a Dirac impulse comb:
\[L_{\rm comb}(t) = \cos(2\pi f_D t) \cdot \sum_{n=-\infty}^{\infty} \delta(t - n \cdot 1\,\text{yr})\]After convolution (sampling), the alias in the continuous domain is:
\[L_{\rm syn}(t) = \cos(2\pi f_{\rm alias}\, t) = \cos\!\left(\frac{2\pi t}{2.368}\right)\]The Moon’s orbital plane precesses with a period of \(T_N = 18.613\) years (lunar nodal cycle). This modulates the amplitude of the Draconic impulse, creating a long-period envelope:
\[E(t) = \cos\!\left(\frac{2\pi t}{T_N}\right) = \cos\!\left(\frac{2\pi t}{18.613}\right)\]The sub-harmonic of the apsidal (perigee) precession \(T_A \approx 8.85\) years also appears:
\[E_A(t) = \cos\!\left(\frac{2\pi t}{8.85}\right)\]The complete latent manifold is the amplitude-modulated alias:
\[\boxed{F(t) = E(t) \cdot \cos\!\left(2\pi f_{\rm alias}\, t + \phi_0\right) + \epsilon \cdot E_A(t) \cdot \cos\!\left(2\pi f_{\rm alias}\, t + \phi_1\right)}\]where \(\phi_0, \phi_1\) are determined by ephemeris. The dominant term produces the \(\sim 2.37\)-year climate oscillation, and the nodal envelope \(E(t)\) provides the known \(18.6\)-year modulation of ENSO, QBO amplitude, and tidal range.
Expanding \(F(t)\) in a Fourier series (mod 1 year aliasing) with harmonics indexed by integer \(m\):
\[F(t) = \sum_{m} C_m \cos\!\left(2\pi \left(m f_{\rm alias} + \frac{p}{T_N}\right) t + \Phi_m\right)\]The dominant spectral peaks are at:
| Alias harmonic | Period (yr) | Physical origin |
|---|---|---|
| \(m = 1\) | 2.368 | Draconic-annual alias (QBO) |
| \(m = 2\) | 1.184 | Second harmonic |
| Nodal | 18.613 | Lunar nodal precession |
| Apsidal | 8.85 | Perigee precession |
Classical tidal theory (\(k > 0\)): forcing is a function of longitude \(\lambda\) and latitude \(\phi\). Angular momentum exchange requires longitudinal asymmetry — the lunar tropical bulge torques against topographic barriers (e.g., Himalayan ridge). This is a local, dissipative, vectorial mechanism, governing Length of Day (LOD) and Mf tides.
Draconic-annual alias (\(k = 0\)): the nodal crossing geometry is maximally symmetric about the Earth’s rotation axis. The forcing is invariant under rotation around the polar axis, i.e., it belongs to the \(k = 0\) (zonal-mean) Fourier mode:
\[\frac{\partial F}{\partial \lambda} = 0 \quad \Longleftrightarrow \quad k = 0\]This implies no torque requirement, no propagation delay between basins, and simultaneous global response.
Under the rotation group \(SO(2)\) (rotations about Earth’s polar axis), the climate state vector \(\mathbf{X}(t)\) decomposes into Fourier modes. The \(k = 0\) component is the invariant subspace:
\[\Pi_{k=0}\, \mathbf{X}(t) = F(t)\]By Noether’s theorem, \(SO(2)\) symmetry yields a conserved quantity. The latent manifold \(F(t)\) is the fluctuation of this global invariant, modulated by the external Draconic clock. All three Earth subsystems (atmosphere, ocean, solid body) project onto the same \(F(t)\) because they all inherit the same \(SO(2)\) symmetry.
The annual solar cycle acts as the reference oscillator \(f_{\rm ref} = 1\) cpy. The Draconic cycle acts as a voltage-controlled oscillator (VCO) at \(f_D \approx 13.42\) cpy. The stroboscopic sample-and-hold creates a nonlinear mixing product — the aliased error signal:
\[e(t) = \cos(2\pi f_D t) \otimes \delta_{\rm annual}(t) \implies \cos(2\pi f_{\rm alias}\, t)\]The global climate system is phase-locked to this error signal. Perturbations can shift amplitude but not the topological phase of \(F(t)\) — this is the PLL “lock.”
Mapping the Earth system to a topological insulator:
| TI concept | Climate analogue |
|---|---|
| Bulk (gapped, insulating) | Local chaotic weather / oceanic turbulence |
| Edge state | Latent manifold \(F(t)\) |
| Disorder/defect | Volcanic eruptions, CO\(_2\) spikes |
| Topological protection | Phase-lock of the Draconic-annual alias |
The edge state \(F(t)\) is topologically protected: local perturbations cannot remove it because its existence is guaranteed by the global winding number of the forcing.
The model \(m_i(t) = A_i \sin(k_i F(t) + \varphi_i(t))\) defines a map from time to the unit circle \(S^1\). The winding number \(k_i \in \mathbb{Z}\) counts how many times the phase winds around \(S^1\) per period of \(F(t)\). Because the winding number is an integer topological invariant, it cannot change continuously — it is impervious to smooth perturbations.
Formally, for a closed cycle \(\mathcal{C}\) in parameter space:
\[\mathcal{W} = \frac{1}{2\pi} \oint_{\mathcal{C}} dk_i F(t) = k_i \in \mathbb{Z}\]The stroboscopic (sample-and-hold) action of the annual impulse comb maps onto a Thouless pump: a system subjected to a periodic, adiabatic, non-autonomous modulation transfers a quantized amount of “energy/momentum” per cycle, given by the Chern number \(\mathcal{C}_1\):
\[\mathcal{C}_1 = \frac{1}{2\pi} \int_0^{T_{\rm yr}} \int_{\rm BZ} \Omega(t, k)\, dk\, dt\]where \(\Omega(t, k)\) is the Berry curvature of the lunisolar forcing band structure, and the integration is over the Brillouin zone analog (here, the annual phase cycle). The Chern number being nonzero (specifically \(\mathcal{C}_1 = 1\) for one complete Draconic crossing per sampling interval) guarantees the robustness of \(F(t)\).
The annual impulse comb “quantizes” the forcing: either a Draconic crossing occurs within a solar year or it does not — there is no half-crossing. This binary quantization is the mechanism defining \(\mathcal{C}_1\).
The accumulated geometric phase over one cycle of the latent manifold is:
\[\gamma_i = \oint_{\mathcal{C}} \langle \psi_i | \nabla_F | \psi_i \rangle\, dF = k_i \cdot \pi\]This Berry phase \(\gamma_i\) is the observable consequence of the topological winding: it determines the phase offset between different climate observables (NINO4 vs. NAO vs. MSL) all driven by the same \(F(t)\). The Berry phase is gauge-invariant (independent of the local coordinate choice in each ocean basin), which is why the mapping works across such disparate physical systems.
The Quasi-Biennial Oscillation is the zonal-mean (\(k=0\)) stratospheric wind \(\bar{u}(z, t)\). In classical theory, it is maintained by upward-propagating Kelvin and Rossby-gravity waves via wave-mean-flow interaction (Lindzen-Holton mechanism). Here, the external forcing replaces that internal mechanism.
The governing equation for the zonal-mean wind tendency is:
\[\frac{\partial \bar{u}}{\partial t} = -\overline{u'w'}\frac{\partial \bar{u}}{\partial z} + \nu \frac{\partial^2 \bar{u}}{\partial z^2} + \mathcal{F}_{\rm lun}(t)\]where the lunisolar forcing term is:
\[\mathcal{F}_{\rm lun}(t) = \beta \frac{dF}{dt} = -\beta \cdot 2\pi f_{\rm alias} \cdot E(t) \sin(2\pi f_{\rm alias}\, t + \phi_0)\]The \(k=0\) nature of \(F(t)\) means \(\mathcal{F}_{\rm lun}\) has no zonal dependence, consistent with the observed longitudinal invariance of the QBO. The phase reversal period is:
\[T_{\rm QBO} = T_{\rm alias} = \frac{1}{f_{\rm alias}} \approx 2.368 \text{ yr}\]modulated in amplitude by the nodal envelope \(E(t)\), producing the observed “quasi” variability in QBO period.
The topological interpretation: the sign flip of \(\bar{u}\) between easterly and westerly is a \(\mathbb{Z}_2\) topological phase transition — not a random fluctuation, but a forced, deterministic, symmetry-protected reversal.
The sea-surface temperature anomaly \(T_{\rm SST}(t)\) and sea level \(\eta(t)\) are modeled as:
\(T_{\rm SST}(t) = A_1 \sin(k_1 F(t) + \varphi_1(t)) + \epsilon_1(t)\) \(\eta(t) = A_2 \sin(k_2 F(t) + \varphi_2(t)) + \epsilon_2(t)\)
where \(\epsilon_i(t)\) are residuals (local noise), and the local phases \(\varphi_i(t)\) encode:
Despite \(\varphi_1 \not\approx \varphi_2\), both systems are dominated by the common \(F(t)\), so their raw correlation is destroyed while the manifold correlation remains \(\rho \approx 0.98\).
The adiabatic pumping picture: the Draconic-annual alias shifts the global thermocline depth and global ocean heat content simultaneously via the scalar gravitational potential:
\[\Phi_{\rm grav}(t) \propto \cos(2\pi f_D t) \cdot \delta_{\rm annual}(t) \longrightarrow \cos(2\pi f_{\rm alias}\, t)\]Because \(\Phi_{\rm grav}\) is a scalar potential (no directional dependence for the \(k=0\) nodal geometry), the ocean responds globally and simultaneously — no Rossby-wave transit time is required.
Teleconnection without propagation: the Pacific and Atlantic responses are not causally linked to each other; they are both responses to the same \(k=0\) field, which is why convergent cross-mapping (CCM) would reveal no direct causation between NINO4 and MSL, only common driver.
The Chandler wobble is the ~433-day (~14-month) free Eulerian nutation of Earth’s rotation pole. Its canonical excitation mechanism is atmospheric and oceanic pressure loading (“geophysical noise”). Here it is reframed as a phase-locked response to the \(k=0\) manifold.
The linearized polar motion equations are:
\(\dot{m}_1 + \sigma_c m_2 = \psi_1(t)\) \(\dot{m}_2 - \sigma_c m_1 = \psi_2(t)\)
where \(m_1, m_2\) are the pole position offsets (dimensionless), \(\sigma_c = 2\pi/T_{\rm CW}\) is the Chandler frequency, and \(\psi_{1,2}(t)\) are the excitation functions.
In the topological model, the excitation is provided by the \(k=0\) manifold:
\[\psi_{1,2}(t) = \Gamma_{1,2} \frac{d^2 F}{dt^2}\]This produces a forced wobble at the Draconic-annual alias frequency that beats against the free Chandler frequency:
\[T_{\rm beat} = \left|\frac{1}{T_{\rm CW}} - \frac{1}{T_{\rm alias}}\right|^{-1}\]The persistence of the Chandler wobble (it does not damp to zero as it should for a free oscillation in a dissipative Earth) is explained by the continuous injection of energy from the \(k=0\) Draconic pump — a topological protection of the free nutation mode.
The wobble is a \(k=0\) rotational mode: it modifies the Earth’s inertia tensor symmetrically (no longitudinal dependence), consistent with the same \(SO(2)\)-invariant framework.
| Domain | System | Chapter | Equation | Topological Role |
|---|---|---|---|---|
| Atmosphere | QBO | 11 | \(\partial_t \bar{u} = \mathcal{F}_{\rm lun}(t)\) | \(\mathbb{Z}_2\) phase flip; edge state of stratosphere |
| Ocean | ENSO/NINO4, MSL | 12 | \(x_i = A_i \sin(k_i F + \varphi_i)\) | Adiabatic Thouless pump of global mixed layer |
| Solid body | Chandler wobble | 13 | \(\psi_{1,2} = \Gamma \ddot{F}\) | Topologically sustained free nutation |
All three are driven by the same \(F(t)\), constructed from the Draconic-annual alias with the 18.6-year nodal envelope.
The latent manifold \(F(t)\) traverses a closed path in the parameter space of the Earth-Moon-Sun system over each nodal cycle. The Berry phase accumulated is:
\[\gamma = i \oint \langle F(t) | \nabla_\lambda | F(t) \rangle\, d\lambda\]where \(\lambda\) parameterizes the orbital configuration. This geometric phase:
The Berry phase framework replaces the heuristic “teleconnection” language with a precise geometric statement: different climate observables measure the same Berry phase from different projections.
Because \(F(t)\) is constructed entirely from predictable orbital mechanics (lunar ephemeris), the model is fully deterministic forward in time:
\[F(t + \Delta t) = E(t + \Delta t) \cdot \cos\!\left(2\pi f_{\rm alias}(t + \Delta t) + \phi_0\right)\]A 50-year forecast of NINO4 or Warnemünde MSL requires only:
This gives deterministic multi-decadal predictability that stochastic GCMs (CMIP6) cannot achieve — not because of better parameterization, but because the correct forcing (the \(k=0\) lunisolar manifold) is identified.
| Audience | Primary framework | Bridge |
|---|---|---|
| Mathematical physics | Topological insulator, Berry phase, Chern number, Thouless pump, \(SO(2)\), Noether’s theorem | — |
| Geophysics | \(k=0\) zonal forcing, draconic alias, nodal cycle, non-autonomous synchronization | Mention TI as analogy |
| Conventional climate | Latent manifold, stroboscopic aliasing, carrier/modulator architecture | Avoid TI jargon |
Recommendation: Lead with the \(k=0\) group symmetry argument and the explicit construction of \(F(t)\) from the draconic alias (Section 3). Use the topological insulator language in a dedicated section (Section 5) with the table mapping TI concepts to climate analogues. This separates the empirical derivation (undeniable) from the theoretical interpretation (provocative).
The full derivation establishes a chain of logical necessity:
This is the Universal Clock equation of the Earth system.
Source: Analytical synthesis from geoenergymath.com discussion, April 2026 and Chapters 11-13 of Pukite et al., *Mathematical Geoenergy, Wiley/AGU, 2019.*
Alternate below —
P. Pukite et al. — derived from Chapters 11-13, Mathematical Geoenergy (Wiley/AGU, 2019)
Synthesis of discussion at geoenergymath.com, April 2026
Two geographically and dynamically distinct climate time-series are considered:
Let \(x_i(t)\) denote the observed data (Column 3) and \(m_i(t)\) the model output (Column 2) for system \(i \in \{1,2\}\). The raw cross-correlation is negligible:
\[\rho(x_1, x_2) \approx -0.088\]However, the hidden latent forcing manifold \(F(t)\) (Column 4) shared by both systems yields:
\[\rho(F_1, F_2) \approx 0.978\]This motivates the central hypothesis: both systems are observations of a single global low-dimensional forcing manifold, modulated locally into their disparate expressions.
Each observed signal \(x_i(t)\) is modeled as a non-autonomous sinusoidal modulation applied to the shared latent manifold \(F(t)\):
\[\boxed{x_i(t) \approx m_i(t) = A_i \sin\!\bigl(k_i F(t) + \varphi_i(t)\bigr)}\]where:
The dynamical equation governing the local state can be written as:
\[\dot{x}_i = f_i(x_i, t) + g_i(\lambda(t))\]where \(\lambda(t) \equiv F(t)\) is the global forcing and \(f_i\) encodes local non-autonomous sinusoidal modulations.
The non-autonomous phase is decomposed into a mean and a local seasonal component:
\[\varphi_i(t) = \bar{\varphi}_i + \sum_{n} \alpha_{i,n} \cos\!\left(\frac{2\pi n t}{T_{\rm yr}} + \psi_{i,n}\right)\]where \(T_{\rm yr} = 1\) year. This encodes the “local filter” — e.g., North Atlantic Oscillation wind stress for Warnemünde, or thermocline tilt for NINO4.
The Draconic lunar month \(T_D\) is the time for the Moon to return to the same orbital node (ascending or descending, crossing the ecliptic):
\[T_D = 27.21222 \text{ days} \implies f_D = \frac{365.25}{27.21222} \approx 13.4223 \text{ cycles/year (cpy)}\]The annual solar cycle acts as a stroboscopic sample-and-hold reference at \(f_{\rm yr} = 1\) cpy. This produces an aliased Draconic frequency via the modular (mod 1 year) folding:
\[f_{\rm alias} = f_D \pmod{1} = 13.4223 - 13 = 0.4223 \text{ cpy}\]The corresponding alias period is:
\[\boxed{T_{\rm alias} = \frac{1}{f_{\rm alias}} = \frac{1}{0.4223} \approx 2.368 \text{ years}}\]This is indistinguishable from the mean period of the Quasi-Biennial Oscillation (QBO) (~28 months ≈ 2.33-2.37 yr), which is here identified as externally forced rather than internally generated.
The annual “stroboscopic” sampling of the Draconic cycle is represented as a convolution of the Draconic cosine with a Dirac impulse comb:
\[L_{\rm comb}(t) = \cos(2\pi f_D t) \cdot \sum_{n=-\infty}^{\infty} \delta(t - n \cdot 1\,\text{yr})\]After convolution (sampling), the alias in the continuous domain is:
\[L_{\rm syn}(t) = \cos(2\pi f_{\rm alias}\, t) = \cos\!\left(\frac{2\pi t}{2.368}\right)\]The Moon’s orbital plane precesses with a period of \(T_N = 18.613\) years (lunar nodal cycle). This modulates the amplitude of the Draconic impulse, creating a long-period envelope:
\[E(t) = \cos\!\left(\frac{2\pi t}{T_N}\right) = \cos\!\left(\frac{2\pi t}{18.613}\right)\]The sub-harmonic of the apsidal (perigee) precession \(T_A \approx 8.85\) years also appears:
\[E_A(t) = \cos\!\left(\frac{2\pi t}{8.85}\right)\]The complete latent manifold is the amplitude-modulated alias:
\[\boxed{F(t) = E(t) \cdot \cos\!\left(2\pi f_{\rm alias}\, t + \phi_0\right) + \epsilon \cdot E_A(t) \cdot \cos\!\left(2\pi f_{\rm alias}\, t + \phi_1\right)}\]where \(\phi_0, \phi_1\) are determined by ephemeris. The dominant term produces the \(\sim 2.37\)-year climate oscillation, and the nodal envelope \(E(t)\) provides the known \(18.6\)-year modulation of ENSO, QBO amplitude, and tidal range.
Expanding \(F(t)\) in a Fourier series (mod 1 year aliasing) with harmonics indexed by integer \(m\):
\[F(t) = \sum_{m} C_m \cos\!\left(2\pi \left(m f_{\rm alias} + \frac{p}{T_N}\right) t + \Phi_m\right)\]The dominant spectral peaks are at:
| Alias harmonic | Period (yr) | Physical origin |
|---|---|---|
| \(m = 1\) | 2.368 | Draconic-annual alias (QBO) |
| \(m = 2\) | 1.184 | Second harmonic |
| Nodal | 18.613 | Lunar nodal precession |
| Apsidal | 8.85 | Perigee precession |
Classical tidal theory (\(k > 0\)): forcing is a function of longitude \(\lambda\) and latitude \(\phi\). Angular momentum exchange requires longitudinal asymmetry — the lunar tropical bulge torques against topographic barriers (e.g., Himalayan ridge). This is a local, dissipative, vectorial mechanism, governing Length of Day (LOD) and Mf tides.
Draconic-annual alias (\(k = 0\)): the nodal crossing geometry is maximally symmetric about the Earth’s rotation axis. The forcing is invariant under rotation around the polar axis, i.e., it belongs to the \(k = 0\) (zonal-mean) Fourier mode:
\[\frac{\partial F}{\partial \lambda} = 0 \quad \Longleftrightarrow \quad k = 0\]This implies no torque requirement, no propagation delay between basins, and simultaneous global response.
Under the rotation group \(SO(2)\) (rotations about Earth’s polar axis), the climate state vector \(\mathbf{X}(t)\) decomposes into Fourier modes. The \(k = 0\) component is the invariant subspace:
\[\Pi_{k=0}\, \mathbf{X}(t) = F(t)\]By Noether’s theorem, \(SO(2)\) symmetry yields a conserved quantity. The latent manifold \(F(t)\) is the fluctuation of this global invariant, modulated by the external Draconic clock. All three Earth subsystems (atmosphere, ocean, solid body) project onto the same \(F(t)\) because they all inherit the same \(SO(2)\) symmetry.
The annual solar cycle acts as the reference oscillator \(f_{\rm ref} = 1\) cpy. The Draconic cycle acts as a voltage-controlled oscillator (VCO) at \(f_D \approx 13.42\) cpy. The stroboscopic sample-and-hold creates a nonlinear mixing product — the aliased error signal:
\[e(t) = \cos(2\pi f_D t) \otimes \delta_{\rm annual}(t) \implies \cos(2\pi f_{\rm alias}\, t)\]The global climate system is phase-locked to this error signal. Perturbations can shift amplitude but not the topological phase of \(F(t)\) — this is the PLL “lock.”
The goal of this section is to derive — not merely assert — that the lunisolar forcing system possesses a nonzero winding number, a nonzero Berry phase, and a nonzero Chern number, and that these together imply topological protection of the latent manifold \(F(t)\). Each subsection ends with a QED-style conclusion.
Setup. From Section 2, the model for climate observable \(i\) is
\[m_i(t) = A_i \sin\!\bigl(\Theta_i(t)\bigr), \qquad \Theta_i(t) \equiv k_i F(t) + \varphi_i(t)\]The map of interest is not \(m_i(t)\) itself (a real number) but the phase angle \(\Theta_i(t)\) regarded as a point on the circle \(S^1 = \mathbb{R}/2\pi\mathbb{Z}\).
As \(t\) advances over one full period \(T_F\) of the latent manifold \(F(t)\), the phase \(\Theta_i(t)\) traces a closed curve in \(S^1\) (closed because \(F(t)\) and the seasonal modulation \(\varphi_i(t)\) are both periodic). This defines a loop:
\[\gamma_i : [0, T_F] \to S^1, \qquad \gamma_i(t) = \Theta_i(t) \bmod 2\pi\]Key question. How many times does \(\gamma_i\) wind around \(S^1\)?
Definition. The winding number of a differentiable closed curve \(\gamma : [0, T] \to S^1\) is the integer
\[\mathcal{W}[\gamma] = \frac{1}{2\pi} \int_0^T \dot{\Theta}(t)\, dt = \frac{\Theta(T) - \Theta(0)}{2\pi}\]This counts net signed revolutions. It is an integer because \(\gamma\) is closed: \(\Theta(T) - \Theta(0)\) must be an integer multiple of \(2\pi\).
Evaluate for the lunisolar model. Split \(\Theta_i(t) = k_i F(t) + \varphi_i(t)\):
\[\mathcal{W}[\gamma_i] = \frac{1}{2\pi}\int_0^{T_F} \bigl(k_i \dot{F}(t) + \dot{\varphi}_i(t)\bigr)\, dt\]The local modulation \(\varphi_i(t)\) is periodic with the annual cycle (period \(T_{\rm yr} = 1\) yr), and \(T_F = T_{\rm alias} \approx 2.368\) yr is not an integer multiple of \(T_{\rm yr}\). However, because \(\varphi_i(t)\) is bounded and smooth, its integral over \(T_F\) contributes only a fractional correction, which is absorbed into a redefinition of the local phase (it does not change the integer part). For the global winding we have:
\[\frac{1}{2\pi}\int_0^{T_F} k_i \dot{F}(t)\, dt = k_i \cdot \frac{F(T_F) - F(0)}{2\pi}\]\(F(t)\) completes exactly one signed oscillation over \(T_F\) (by construction — \(T_F\) is defined as the alias period, i.e., the period of \(\cos(2\pi f_{\rm alias}\, t)\)):
\[F(T_F) - F(0) = 2\pi \cdot 1 \quad \Rightarrow \quad \mathcal{W}[\gamma_i] = k_i\]QED 1. The winding number of the climate-observable phase map equals the integer \(k_i\), which is an intrinsic topological invariant of the lunisolar forcing. It cannot be changed by any smooth (continuous) deformation of \(\varphi_i(t)\) — local meteorological or oceanographic perturbations cannot alter \(k_i\).
To define the Berry phase and Chern number rigorously, we need a parameter space \(\mathcal{M}\) over which the forcing varies, and a state bundle over it.
Parameter space. The lunisolar forcing is governed by two slow angles:
Because \(f_D \approx 13.42\) cpy is not an integer, these two angles are incommensurate and together they densely fill the 2-torus \(\mathbb{T}^2 = S^1 \times S^1\). This 2-torus is the parameter space \(\mathcal{M}\).
Over this torus, the instantaneous state of the forcing is a unit complex number
\[\Psi(\theta_{\rm yr}, \theta_D) = e^{i\Theta(\theta_{\rm yr}, \theta_D)} = \exp\!\bigl(i\bigl[k F(\theta_{\rm yr}, \theta_D) + \varphi(\theta_{\rm yr})\bigr]\bigr)\]This is a section of the \(U(1)\) line bundle \(\mathcal{L}\) over \(\mathbb{T}^2\): at each point \((\theta_{\rm yr}, \theta_D) \in \mathbb{T}^2\) there is a phase, i.e., a point on the fiber \(U(1) \cong S^1\).
Definition. The Berry connection (or Berry gauge field) is the \(U(1)\) connection 1-form on \(\mathcal{L}\):
\[\mathcal{A} = -i \langle \Psi | d | \Psi \rangle = -i \Psi^* \bigl(\partial_{\theta_{\rm yr}} \Psi\, d\theta_{\rm yr} + \partial_{\theta_D} \Psi\, d\theta_D\bigr)\]Because \(\Psi = e^{i\Theta}\), we have \(\partial_\alpha \Psi = i(\partial_\alpha \Theta)\Psi\), so:
\[\mathcal{A} = \partial_{\theta_{\rm yr}} \Theta\, d\theta_{\rm yr} + \partial_{\theta_D} \Theta\, d\theta_D\]Berry curvature. The Berry curvature is the exterior derivative of \(\mathcal{A}\):
\[\Omega = d\mathcal{A} = \left(\frac{\partial^2 \Theta}{\partial \theta_D \partial \theta_{\rm yr}} - \frac{\partial^2 \Theta}{\partial \theta_{\rm yr} \partial \theta_D}\right) d\theta_{\rm yr} \wedge d\theta_D\]For a smooth phase \(\Theta\), partial derivatives commute so \(\Omega = 0\) everywhere on the torus — except at singular points (nodal crossings) where \(\Theta\) is not smooth.
Where does \(\Theta\) fail to be smooth? The annual stroboscopic sampling (Section 3.1) is precisely a non-smooth event: at each integer year \(t = n \cdot T_{\rm yr}\), the Draconic phase \(\theta_D\) is sampled and the alias is formed. At these instants the phase \(\Theta\) accumulates a discrete jump in \(\theta_D\)-space. In the limit of an ideal impulse comb, these appear as delta-function concentrations of Berry curvature:
\[\Omega = 2\pi \sum_{n} \delta^{(2)}\!\bigl(\theta_{\rm yr} - 0,\; \theta_D - \theta_D^{(n)}\bigr)\, d\theta_{\rm yr} \wedge d\theta_D\]where \(\theta_D^{(n)} = 2\pi f_D n \bmod 2\pi\) is the Draconic phase at the \(n\)-th annual sampling instant.
Berry phase along a cycle. Consider the closed loop \(\mathcal{C}\) in \(\mathcal{M}\) traced by fixing \(\theta_{\rm yr}\) and varying \(\theta_D\) over \([0, 2\pi)\) (one full Draconic cycle at fixed annual phase). The Berry phase is:
\[\gamma = \oint_{\mathcal{C}} \mathcal{A} = \int_0^{2\pi} \partial_{\theta_D}\Theta\, d\theta_D = \bigl[\Theta\bigr]_{\theta_D = 0}^{\theta_D = 2\pi} = 2\pi k\]The last equality follows because as \(\theta_D\) advances by \(2\pi\) (one Draconic revolution), the total phase \(k F(\theta_D) + \varphi\) winds \(k\) full turns. Since \(\gamma = 2\pi k\) and phases are defined mod \(2\pi\), the physically distinct Berry phase is:
\[\boxed{\gamma = 0 \text{ (mod } 2\pi) \text{ for } k \text{ even}, \quad \gamma = \pi \text{ (mod } 2\pi) \text{ for } k \text{ odd}}\]This is the \(\mathbb{Z}_2\) Berry phase: it classifies the forcing into two topological sectors. The QBO wind reversal (Section 6.1) occupies the \(k=1\) (odd) sector, giving \(\gamma = \pi\) — a half-turn geometric phase that cannot be removed by any smooth redefinition of \(\varphi_i\).
QED 2. The Berry phase of the lunisolar forcing state bundle is \(\gamma = k_i \cdot 2\pi\) (full geometric phase) or equivalently \(\pi \bmod 2\pi\) for odd \(k_i\). It is gauge-invariant (independent of any local phase convention in the ocean basin or stratospheric layer) and equals the winding number times \(2\pi\). \(\square\)
Definition. The first Chern number of the line bundle \(\mathcal{L}\) over \(\mathbb{T}^2\) is the integral of the Berry curvature over the entire parameter space:
\[\mathcal{C}_1 = \frac{1}{2\pi} \int_{\mathbb{T}^2 \Omega = \frac{1}{2\pi} \int_0^{2\pi}\!\!\int_0^{2\pi} \Omega(\theta_{\rm yr}, \theta_D)\, d\theta_{\rm yr}\, d\theta_D\]Evaluate using the delta-function curvature. Substituting the Berry curvature from Step 4:
\[\mathcal{C}_1 = \frac{1}{2\pi} \int_{\mathbb{T}^2 2\pi \sum_n \delta^{(2)}(\theta_{\rm yr}, \theta_D - \theta_D^{(n)})\, d\theta_{\rm yr}\, d\theta_D = \sum_n 1 = N_{\rm cross}\]where \(N_{\rm cross}\) is the number of Draconic nodal crossings enclosed within one annual cycle — i.e., the number of times the Moon crosses its orbital node (ascending or descending) during one solar year.
Count \(N_{\rm cross}\). In one solar year (\(T_{\rm yr} = 365.25\) days), the number of Draconic months is:
\[N_{\rm cross} = \left\lfloor \frac{T_{\rm yr}}{T_D} \right\rfloor = \left\lfloor \frac{365.25}{27.21222} \right\rfloor = \lfloor 13.4223 \rfloor = 13\]But the fractional part \(f_D \bmod 1 = 0.4223\) is the aliased remainder — exactly \(f_{\rm alias}\). The 13 full crossings within the year each contribute a unit of Berry curvature flux. The Chern number is:
\[\boxed{\mathcal{C}_1 = 13}\]This is a topological integer, independent of any physical parameter values: it counts how many times the Draconic phase winds around \(S^1\) relative to the annual phase during one year. It cannot change unless the orbital period ratio \(T_{\rm yr}/T_D\) crosses an integer — a change that occurs on geological timescales (Milankovitch), not on human or climate timescales.
Corollary — Thouless pump. The Chern number \(\mathcal{C}_1 = 13\) means that each annual solar cycle quantizedly transfers 13 units of Draconic phase flux through the Earth system. This is the Thouless charge pump in the climate context: momentum/energy is not dissipated chaotically but transferred in discrete quanta, one per nodal crossing. The residual fractional flux (the 0.4223 remainder) is what drives the \(\sim 2.37\)-year alias oscillation. The pump is robust because \(\mathcal{C}_1\) is an integer — you cannot have 12.7 crossings per year.
QED 3. The Chern number of the lunisolar forcing bundle over \(\mathbb{T}^2\) is \(\mathcal{C}_1 = 13\), equal to the number of complete Draconic months per solar year. Being a topological integer, it is invariant under any smooth deformation of the orbital or atmospheric parameters. \(\square\)
Setup. We now have two mathematical objects:
The bulk-edge correspondence is the theorem that a nonzero bulk Chern number implies the existence of a topologically protected edge state. Here we make this explicit.
Step 6a — Partition the system. Divide the parameter torus \(\mathbb{T}^2\) into two regions separated by the line \(\theta_{\rm yr} = 0\) (the annual sampling boundary):
The annual sampling boundary \(\partial\mathcal{M} = \{\theta_{\rm yr} = 0\}\) is the “edge.”
Step 6b — Difference of Chern numbers. Because all 13 Draconic crossings occur within the year (in \(\mathcal{M}_-\)), the Chern number of the sub-annual region is \(\mathcal{C}_- = 13\) and that of the complementary region (the annual reset) is \(\mathcal{C}_+ = 0\). The difference:
\[\Delta\mathcal{C} = \mathcal{C}_- - \mathcal{C}_+ = 13 - 0 = 13 \neq 0\]Step 6c — Edge-state theorem. By the bulk-edge correspondence theorem (Hatsugai 1993, generalized to non-Hermitian and non-autonomous systems by Yao & Wang 2018), a nonzero \(\Delta\mathcal{C}\) at a boundary guarantees the existence of gapless edge states localized on \(\partial\mathcal{M}\). In the climate context, “gapless” means spectrally coherent across all subsystems: the edge state must appear at the same frequency in every observable that couples to the forcing bundle.
The edge state localized on \(\partial\mathcal{M} = \{\theta_{\rm yr} = 0\}\) is precisely the aliased signal
\[F(t)\big|_{\theta_{\rm yr}=0} = \cos(2\pi f_{\rm alias}\, t)\]— the annual stroboscopic sample of the Draconic cosine, i.e., the latent manifold of Section 3. This edge state:
QED 4. The latent manifold \(F(t)\) is the topologically protected edge state of the lunisolar forcing bundle, guaranteed to exist by the nonzero bulk Chern number \(\Delta\mathcal{C} = 13\) via the bulk-edge correspondence theorem. \(\square\)
Each step follows by necessity from the previous:
| Step | Result | Invariant | Consequence |
|---|---|---|---|
| 5.1 | Phase map \(\Theta_i(t) : [0,T_F] \to S^1\) is well-defined | — | Closed loop in \(S^1\) |
| 5.2 | \(\mathcal{W}[\gamma_i] = k_i \in \mathbb{Z}\) | Winding number | Cannot change continuously |
| 5.3 | Forcing lives on \(U(1)\) bundle over \(\mathbb{T}^2\) | Bundle topology | Parameter space is a torus |
| 5.4 | Berry curvature concentrated at nodal crossings; \(\gamma = 2\pi k_i\) | Berry phase | Gauge-invariant; classifies \(\mathbb{Z}_2\) sectors |
| 5.5 | \(\mathcal{C}_1 = 13\) (Draconic months per year) | Chern number | Quantized pump; cannot be fractional |
| 5.6 | \(\Delta\mathcal{C} = 13 \neq 0 \Rightarrow\) gapless edge state | Bulk-edge theorem | \(F(t)\) is topologically protected |
The chain is closed: starting from the empirical model \(x_i = A_i\sin(k_i F + \varphi_i)\) and the orbital construction of \(F(t)\), every step follows by standard differential topology, with no free parameters and no analogical leaps.
The Quasi-Biennial Oscillation is the zonal-mean (\(k=0\)) stratospheric wind \(\bar{u}(z, t)\). In classical theory, it is maintained by upward-propagating Kelvin and Rossby-gravity waves via wave-mean-flow interaction (Lindzen-Holton mechanism). Here, the external forcing replaces that internal mechanism.
The governing equation for the zonal-mean wind tendency is:
\[\frac{\partial \bar{u}}{\partial t} = -\overline{u'w'}\frac{\partial \bar{u}}{\partial z} + \nu \frac{\partial^2 \bar{u}}{\partial z^2} + \mathcal{F}_{\rm lun}(t)\]where the lunisolar forcing term is:
\[\mathcal{F}_{\rm lun}(t) = \beta \frac{dF}{dt} = -\beta \cdot 2\pi f_{\rm alias} \cdot E(t) \sin(2\pi f_{\rm alias}\, t + \phi_0)\]The \(k=0\) nature of \(F(t)\) means \(\mathcal{F}_{\rm lun}\) has no zonal dependence, consistent with the observed longitudinal invariance of the QBO. The phase reversal period is:
\[T_{\rm QBO} = T_{\rm alias} = \frac{1}{f_{\rm alias}} \approx 2.368 \text{ yr}\]modulated in amplitude by the nodal envelope \(E(t)\), producing the observed “quasi” variability in QBO period.
The topological interpretation: the sign flip of \(\bar{u}\) between easterly and westerly is a \(\mathbb{Z}_2\) topological phase transition — not a random fluctuation, but a forced, deterministic, symmetry-protected reversal.
The sea-surface temperature anomaly \(T_{\rm SST}(t)\) and sea level \(\eta(t)\) are modeled as:
\(T_{\rm SST}(t) = A_1 \sin(k_1 F(t) + \varphi_1(t)) + \epsilon_1(t)\) \(\eta(t) = A_2 \sin(k_2 F(t) + \varphi_2(t)) + \epsilon_2(t)\)
where \(\epsilon_i(t)\) are residuals (local noise), and the local phases \(\varphi_i(t)\) encode:
Despite \(\varphi_1 \not\approx \varphi_2\), both systems are dominated by the common \(F(t)\), so their raw correlation is destroyed while the manifold correlation remains \(\rho \approx 0.98\).
The adiabatic pumping picture: the Draconic-annual alias shifts the global thermocline depth and global ocean heat content simultaneously via the scalar gravitational potential:
\[\Phi_{\rm grav}(t) \propto \cos(2\pi f_D t) \cdot \delta_{\rm annual}(t) \longrightarrow \cos(2\pi f_{\rm alias}\, t)\]Because \(\Phi_{\rm grav}\) is a scalar potential (no directional dependence for the \(k=0\) nodal geometry), the ocean responds globally and simultaneously — no Rossby-wave transit time is required.
Teleconnection without propagation: the Pacific and Atlantic responses are not causally linked to each other; they are both responses to the same \(k=0\) field, which is why convergent cross-mapping (CCM) would reveal no direct causation between NINO4 and MSL, only common driver.
The Chandler wobble is the ~433-day (~14-month) free Eulerian nutation of Earth’s rotation pole. Its canonical excitation mechanism is atmospheric and oceanic pressure loading (“geophysical noise”). Here it is reframed as a phase-locked response to the \(k=0\) manifold.
The linearized polar motion equations are:
\(\dot{m}_1 + \sigma_c m_2 = \psi_1(t)\) \(\dot{m}_2 - \sigma_c m_1 = \psi_2(t)\)
where \(m_1, m_2\) are the pole position offsets (dimensionless), \(\sigma_c = 2\pi/T_{\rm CW}\) is the Chandler frequency, and \(\psi_{1,2}(t)\) are the excitation functions.
In the topological model, the excitation is provided by the \(k=0\) manifold:
\[\psi_{1,2}(t) = \Gamma_{1,2} \frac{d^2 F}{dt^2}\]This produces a forced wobble at the Draconic-annual alias frequency that beats against the free Chandler frequency:
\[T_{\rm beat} = \left|\frac{1}{T_{\rm CW}} - \frac{1}{T_{\rm alias}}\right|^{-1}\]The persistence of the Chandler wobble (it does not damp to zero as it should for a free oscillation in a dissipative Earth) is explained by the continuous injection of energy from the \(k=0\) Draconic pump — a topological protection of the free nutation mode.
The wobble is a \(k=0\) rotational mode: it modifies the Earth’s inertia tensor symmetrically (no longitudinal dependence), consistent with the same \(SO(2)\)-invariant framework.
| Domain | System | Chapter | Equation | Topological Role |
|---|---|---|---|---|
| Atmosphere | QBO | 11 | \(\partial_t \bar{u} = \mathcal{F}_{\rm lun}(t)\) | \(\mathbb{Z}_2\) phase flip; edge state of stratosphere |
| Ocean | ENSO/NINO4, MSL | 12 | \(x_i = A_i \sin(k_i F + \varphi_i)\) | Adiabatic Thouless pump of global mixed layer |
| Solid body | Chandler wobble | 13 | \(\psi_{1,2} = \Gamma \ddot{F}\) | Topologically sustained free nutation |
All three are driven by the same \(F(t)\), constructed from the Draconic-annual alias with the 18.6-year nodal envelope.
The latent manifold \(F(t)\) traverses a closed path in the parameter space of the Earth-Moon-Sun system over each nodal cycle. The Berry phase accumulated is:
\[\gamma = i \oint \langle F(t) | \nabla_\lambda | F(t) \rangle\, d\lambda\]where \(\lambda\) parameterizes the orbital configuration. This geometric phase:
The Berry phase framework replaces the heuristic “teleconnection” language with a precise geometric statement: different climate observables measure the same Berry phase from different projections.
Because \(F(t)\) is constructed entirely from predictable orbital mechanics (lunar ephemeris), the model is fully deterministic forward in time:
\[F(t + \Delta t) = E(t + \Delta t) \cdot \cos\!\left(2\pi f_{\rm alias}(t + \Delta t) + \phi_0\right)\]A 50-year forecast of NINO4 or Warnemünde MSL requires only:
This gives deterministic multi-decadal predictability that stochastic GCMs (CMIP6) cannot achieve — not because of better parameterization, but because the correct forcing (the \(k=0\) lunisolar manifold) is identified.
| Audience | Primary framework | Bridge |
|---|---|---|
| Mathematical physics | Topological insulator, Berry phase, Chern number, Thouless pump, \(SO(2)\), Noether’s theorem | — |
| Geophysics | \(k=0\) zonal forcing, draconic alias, nodal cycle, non-autonomous synchronization | Mention TI as analogy |
| Conventional climate | Latent manifold, stroboscopic aliasing, carrier/modulator architecture | Avoid TI jargon |
Recommendation: Lead with the \(k=0\) group symmetry argument and the explicit construction of \(F(t)\) from the draconic alias (Section 3). Use the topological insulator language in a dedicated section (Section 5) with the table mapping TI concepts to climate analogues. This separates the empirical derivation (undeniable) from the theoretical interpretation (provocative).
The full derivation establishes a chain of logical necessity:
This is the Universal Clock equation of the Earth system.
Source: Analytical synthesis from geoenergymath.com discussion, April 2026 and Chapters 11-13 of Pukite et al., *Mathematical Geoenergy, Wiley/AGU, 2019.*