Anomalous transport in tokamaks is generally attributed to turbulent fluctuations. Since a large variety of modes are potentially unstable, a wide range of short-scale fluctuations should be measured, with wavenumbers from kρ i ∼ 0.1 to kρ i >> 1. On the Tore Supra tokamak, a light scattering experiment has made possible fluctuation measurements in the medium and high-k domains where a transition in the k-spectrum is observed: The fluctuation level decreases much faster than usual observations, typically with a power law S(k) ≡ k −6 . A scan of the ion Larmor radius shows that the transition wavenumber scales with ρ i around kρ i ∼ 1.5. This transition indicates that a characteristic length scale should be involved to describe the fluctuation non linear dynamics in this range. The resulting very low level of fluctuations at high k does not support a strong effect of turbulence driven by electron temperature gradient. For this gyroradius scan, the characteristics of turbulence also exhibit a good matching with predictions from gyro-Bohm scaling: the typical scale length of turbulence scales with the ion Larmor radius, the typical time scales with a/c s ; the turbulence level also scales with ρ i , according to the mixing length rule.

Properties of the I-mode confinement regime on the ASDEX Upgrade tokamak are summarized. A weak dependence of the power threshold for the L-I transition on the toroidal magnetic field strength is found. During improved confinement, the edge radial electric field well deepens. Stability calculations show that the I-mode pedestal is peeling-ballooning stable. Turbulence investigations reveal strongly intermittent density fluctuations linked to the weakly coherent mode in the confined plasma, which become stronger as the confinement quality increases. Across all investigated structure sizes (k ⊥ ≈ 5 -12 cm −1 , with k ⊥ the perpendicular wavenumber of turbulent density fluctuations), the intermittent turbulence bursts are observed. Comparison with bolometry data shows that they move poloidally toward the X-point and finally end up in the divertor. This might be indicative that they play a role in inhibiting the density profile growth, such that no pedestal is formed in the edge density profile.

Backscattering of a microwave beam close to the cut-off allows for measurement of density fluctuations ñ( k ⊥ ) at a specified wave-number, selected by the scattering geometry k ⊥ = −2 k i , where k i is the beam wave-number at the reflection layer. On the Doppler reflectometry system installed on Tore Supra, both the scattering wave-number k ⊥ and the scattering localization (r/a) can be changed during the shot owing to the steppable probing frequency and the motorized antenna. Operating in O mode, the spatial and wave-number ranges depend essentially on density profile, typically probing 0.5 < r/a < 0.95 and 2 < k < 15 cm −1 . Wave number spectra are similar to those obtained with conventional scattering systems. The perpendicular fluctuation velocity in the laboratory frame is obtained from the Doppler shift of the frequency spectrum ω = k ⊥ v ⊥ . It is dominated by the plasma E r × B velocity. In the core, the latter is mainly due to the projection of the toroidal velocity, as this is shown by comparison with measurements by charge exchange recombination spectroscopy. In the set of analysed Tore Supra ohmic and ICRH plasmas, the observed rotation is consistent with a poloidal velocity in the electron diamagnetic direction and/or a toroidal velocity in the counter current direction. The detailed structure of the velocity profile, at the edge and in different plasma regimes, allows us then to get information on the radial electric field distribution. The dynamics of the fluctuation velocity can be studied from the time frequency analysis of the signal, for investigating intermittent behaviour and transient regimes.