Feedback Loop in 2D: Formation of Zonal Jets and Suppression of Turbulence

Sergey Nazarenko, Warwick, UK
Drift waves in fusion devices
Rossby waves in atmospheres of rotating planets

Charney-Hasegawa-Mima equation
Baroclinic instability and ITG
Close analogy between the baroclinic instability and the ion-temperature gradient instability in tokamak.
GFD: two-layer model.
Plasma: Hasegawa-Wakatani model for plasma potential and density.

Low-to-High confinement transitions in fusion plasmas
LH transition discovered in the ASDEX tokamak (Wagner,1982) and now routinely observed in most tokamaks and stellarators.
Left: Heliac data (Shats et al, 2004)
ZF generation
DW suppression

LH transition paradigm
Small-scale turbulence causes anomalous transport, hence L-mode.
Negative feedback loop.
Suppressed turbulence →no transport →improved confinement & H-mode.

Barotropic governor in GFG
James and GrayÕ 1986

Zonal flow generation:
the local turbulence view.
CHM becomes 2D Euler equation in the limit β→0, kρ→°. Hence expect similarities to 2D turbulence.
Inverse energy cascade and direct cascade of potential enstrophy (Fj¯rtoftÕ53 argument).

Ubiquitous features in Drift/Rossby turbulence
Condensation into zonal jets in presence of β.

Rhines scale crossover
Nonlinear=linear → Rhines scale.
ÒLazy 8Ó separates vortex-dominated and wave-dominated scales (RhinesÕ75, Valis & MaltrudÕ93, HollowayÕ84)
Outside of lazy-8: KraichnanÕs isotropic inverse cascade.
Inside lazy-8 the cascade is anisotropic and dominated by triad wave resonances.

Weakly nonlinear drift waves with random phases→ wave kinetic equation (Longuet-Higgens &Gill, 1967)
Anisotropic cascades in drift turbulence
CHM has a third invariant (Balk, SN, Zakharov, 1990).
3 cascades cannot be isotropic.

Cartoon of nonlocal interaction
Eddy scale L decreases via shearing by ZF
Potential enstrophy Z is conserved.
=> Eddy energy E =Z L2 is decreasing
Total E is conserved, => E is transferred from the eddy to ZF
Wrong! Both smaller and larger LÕs are produced. The energy of the eddy is unchanged. (Kraichnan 1976).

Small-scale energy conservation
Energy in SS eddies is conserved if they are initially isotropic (Kraichnan 1976)
1. Dissipation: ellipse cannot get too thin.
2. Nonisotropic eddies: Modulational Instability (LorenzÕ72, GillÕ74, Manin, Nazarenko, 1994; Manfroi, Young, 1999; Smolyakov et al, 2000)
3. Breaking of the scale separation due to inverse cascade

Nonlocal 2D turbulence
Condensate forms – interaction of scales becomes nonlocal (Smith & YakhotÕ93, Maltrud &ValisÕ93, BorueÕ94, Laval, SN & DubrulleÕ99).
Small-scale spectrum changes to E~s-1ε k-1. (Kraichnan 1974, SN & Laval 2000; Connaugton et al 2007).

Condensate coupled with turbulence
 Instability forcing:                              ε(t) ~ γ(kf) E(kf) kf
Spectrum of small-scale turbulence:                              E(kf) ~ s-1ε kf-1
Condensate energy:                  Ec ~ Vc2/2 ~ s2L2 ~ ºε(t) dt,
1,2 =>
(i) E(kf) =0, - suppression of turbulence by jets;
(ii) s ~ γ(kf) – saturation of the jets.

Feedback loop in 2D turbulence
Instability generates small-scale turbulence.
Inverse cascade leads to energy condensation (into jets in presence of beta).
Jets kill small-scale turbulence and saturate.
LH transition: this is why ITER must work.
Barotropic governor and other GFD mechanisms.

Modulational Instability
Manin, Nazarenko, 1994; Manfroi, Young, 1999;
 Smolyakov et al, 2000;
Ongoing numerics: Connaughton, Nadiga, SN, Quinn.
Unstable if
3ky2 < kx2-2

Nonlinear development of MI:
narrow zonal jets
Formation of intense narrow Zonal jets.
Transport/mixing Barriers. Analog of LH transition in fusion plasmas.
Secondary instability preferentially breaks westward jets (consistent with linear condition β-uyy <0 ?).
 Irregular multiple jets with westward preference
Rhines spectrum: E ~ β2 k-5. Chekhlov et alÕ95.

Evolution in the k-space
Energy of WP is partially transferred to ZF and partially dissipated at large kÕs.
2 regimes: random walk/diffusion of WP in the k-space (Balk, Nazarenko, Zakharov, 1990),
Coherent wave – modulational instability (Manin, Nazarenko, 1994, Smolyakov et at, 2000).

Fast mode: modulational instability of a coherent drift wave.
Two component description Ψ = ΨLS.
Small-scale Rossby wave sheared by large-scale ZF.
Large-scale ZF pumped by RW via the ponderomotive force.

Evolution of nonlocal drift turbulence:
retain only interaction with small kÕs and Taylor-expand the integrand of the wave-collision integral; integrate.
Diffusion along curves
Ωk = ωk –βkx =conts.
S ~ZF intensity

Drift-Wave instabilities
Maximum on the kx-axis at kρ ~ 1.
 γ=0 line crosses k=0 point.

Initial evolution
Solve the eigenvalue problem at each curve.
Max eigenvalue <0 → DW on this curve decay.
Max eigenvalue >0 → DW on this curve grow.
Growing curves pass through the instability scales

ZF growth
DW pass energy from the growing curves to ZF.
ZF accelerates DW transfer to the dissipation scales via the increased diffusion coefficient.

ZF growth
Hence the growing region shrink.
DW-ZF loop closed!

Steady state
Saturated ZF.
Jet spectrum on a k-curve passing through the maximum of instability.
Suppressed intermediate scales (Dimits shift).
Balanced/correlated DW and ZF
 (Shats experiment).

Shats experiment
Suppression of inermediate scales by ZF
Scale separation
Nonlocal turbulence

Shats experiment
Instability scales are strongly correlated with ZF scales
Nonlocal scale interaction

Saturation of zonal flow
Different expressions for random 3-wave (low γ) and coherent (high γ) regimes
Intermediate range with Uzf ~ V*.
Only weak ZF damping dependence (important γ is at ρk~1).
 No oscillatory behaviour. ZF cannot fall below the crit value because itÕd be immediately pumped due to renewed instability.

Summary
Self-regulating DW-ZF system.
Drift turbulence creates ZF.
ZF kills drift turbulence and switches the forcing off (cf Dimits shift).
For large grad T small scales reappear because ZF gets KH unstable.
Predictions for the saturated ZF, scale separation, jet-like spectrum of drift turbulence.
Experimental evidence in Heliac. Tokamaks?

Breakdown of local cascades
Kolmogorov cascade spectra (KS) nk ~kxνx kyvy.
Exact solutions of WKE É if local.
Locality corresponds to convergence in WKE integral.
For drift turbulence KS obtained by Monin Piterbarg 1987.
All Kolmogorov spectra of drift turbulence are proven to be nonlocal (Balk, Nazarenko, 1989).
Drift turbulence must be nonlocal, - direct interaction with ZF scales

Coupled large-scale & small-scale  motions
(Dyachenko, Nazarenko, Zakharov, 1992)
Shear flow geometry
"North-Pacific zonal jets at 1000..."
North-Pacific zonal jets at 1000 m depth as seen in 58-year simulation with ECMWF climotological forcing (Nakano and Hasumi, 2005)

Co-authors and relevant publications
Kolmogorov Weakly Turbulent Spectra of Some Types of Drift Waves in Plasma (A.B. Mikhailovskii, S.V. Nazarenko, S.V. Novakovskii, A.P. Churikov and O.G. Onishenko) Phys.Lett.A 133 (1988) 407-409.
Kinetic Mechanisms of Excitation of Drift-Ballooning Modes in Tokamaks (A.B. Mikhailovskii, S.V. Nazarenko and A.P. Churikov) Soviet Journal of Plasma Physics 15 (1989) 33-38.
Nonlocal Drift Wave Turbulence (A.M.Balk, V.E.Zakharov and S.V. Nazarenko) Sov.Phys.-JETP 71 (1990) 249-260.
On the Nonlocal Turbulence of Drift Type Waves (A.M.Balk, S.V. Nazarenko and V.E.Zakharov) Phys.Lett.A 146 (1990) 217-221.
 On the Physical Realizability of Anisotropic Kolmogorov Spectra of Weak Turbulence (A.M.Balk and S.V. Nazarenko) Sov.Phys.-JETP 70 (1990) 1031-1041.
A New Invariant for Drift Turbulence (A.M.Balk, S.V. Nazarenko and V.E. Zakharov) Phys.Lett.A 152 (1991) 276-280.
On the Nonlocal Interaction with Zonal Flows in Turbulence of Drift and Rossby Waves (S.V. Nazarenko) Sov.Phys.-JETP, Letters, June 25, 1991, p.604-607.
 Wave-Vortex Dynamics in Drift and beta-plane Turbulence (A.I. Dyachenko, S.V. Nazarenko and V.E. Zakharov) Phys,Lett.A 165 (1992) 330-334.
Nonlinear interaction of small-scale Rossby waves with an intense large-scale zonal flow. (D.Yu. Manin and S.V. Nazarenko) Phys. Fluids. A 6 (1994) 1158-1167.

Ubiquitous features in Drift/Rossby turbulence
Drift Wave turbulence generates zonal flows
Zonal flows suppress waves
Hence transport barriers, Low-to-High confinement transition

Drift wave – zonal flow turbulence paradigm
Local cascade is replaced by nonlocal (direct) interaction of the DW instability scales with ZF.

Zonostrophy invariant
Extra quadratic invariant of CHM (Balk, Nazarenko, Zakharov, 1991).
Conserved by triad interactions.

"At high kÕs it is..."
At high kÕs it is essentially KraichnanÕs isotropic inverse cascade.
Anisotropy only occurs when Rhines scale is reached.
Weak/wave turbulence theory inside lazy-8: Triad wave resonances.

Plan
Dual-cascade behavior.
Condensation into zonal jets:
 Zonostrophy invariant,
modulational instability,
Rhines spectrum.
Transition to nonlocal interaction:
Suppression of turbulence by zonal jets,
Reduced turbulence transport,
LH transition.