Table of contents

This special issue of Plasma Physics and Controlled Fusion contains the invited and contributed papers from the 6th IAEA Technical Committee Meeting on H-Mode Physics. The sixth meeting in this series (San Diego 1987, Gut Ising 1989, Abingdon 1991, Naka 1993, Princeton 1995) was organized by IPP, Max-Planck-Institut für Plasmaphysik, Garching. It took place at Kloster Seeon, Germany, from Monday 22 September to Wednesday 24 September 1997.

The programme of the meeting covered traditional H-mode physics, but also the physics of other transport barriers such as internal transport barriers associated with weak or negative central shear. In order to give the atmosphere of the meeting a `workshop' spirit, and not that of a conference, invited papers were presented in sessions devoted to the topics

• Formation and structure of internal and edge transport barriers

• Theory of internal and edge transport barriers

• `Hidden' variables affecting the L/H transition (neutrals, profile shapes...)

• ELM physics and theory

• Transport in high-performance plasmas with a transport barrier

• Compatibility of improved confinement scenarios with reactor requirements.

Contributed papers on the topics were then presented in poster sessions.

At the workshop, there was a meeting of the scientific programme committee, during which some members were replaced. The new committee consists of:

• V Golant, Ioffe Institute, RF

• M Greenwald, MIT, USA

• R Groebner, GA, USA

• K Ida, NIFS, Japan

• Y Miura, JAERI, Japan

• E Synakowski, PPPL, USA

• D Ward, JET, UK

• H Zohm, Stuttgart University, Germany.

H Zohm acted as chairman of the committee for this meeting. It was decided to continue the series of workshops in 1999 at JET.

Kloster Seeon, as a former monastery, was a very nice place to inspire fruitful discussions; this special issue gives proof of the interesting physics discussed at the workshop. I hope that everyone enjoyed the workshop as much as I did. Finally, I would like to thank Mrs C Stahlberg for her enormous effort in dealing with all the organizational problems.

Hartmut Zohm

Guest Editor

Stuttgart, 8 April 1998

The phenomenology of ELMs is reviewed, emphasizing the conditions for their occurrence, their repetition frequencies, their impact on confinement and the properties of associated fluctuations, in order to understand the physics involved. Theories of ELMs are then briefly described and the extent to which they can successfully model ELMs is discussed.

Recent progress in experimental investigations, related to the determinant variables of the L-H transition, is reviewed in this paper. In particular, it is put forward that the `wall effect', which often produces spurious contributions to the various parametric dependences of the H-mode threshold power, may be an embodiment of the atomic processes, involving the edge neutrals as a `hidden' variable. In addition, relations between the edge plasma quantities and threshold scaling are emphasized, quoting the ostensibly different observations in various tokamaks.

The reactor physics requirements are outlined in terms of characteristic performance measures. Four reactor designs of different scope are taken as examples. ITER is based on the ELMy H-mode as an operating regime demonstrated over many energy confinement times. Designs for power producing reactors require higher power densities and bootstrap current fractions, and therefore higher values of normalized . The capability of different enhanced confinement regimes, combining internal transport barriers with the L- or H-mode edge to satisfy these demands, together with the requirements on helium pumping and divertor power load, is discussed.

A new era of tokamak transport physics research has opened up with the discovery of a variety of internal transport barrier regimes. With suppression of some instabilities in the plasma, subdominant transport processes can be observed for the first time. In this paper, the theory of internal and edge transport barriers is reviewed with an emphasis on what is still needed in order to have a predictive transport theory of tokamak plasmas with barriers.

The phenomenology of transport barrier formation is reviewed with a focus on physics that may be common to the edge and the core. To this end, the framework of velocity shear reduction of turbulence-induced fluxes, applied to the edge for some time, is studied in the light of measurements of core bifurcation dynamics and recent tests of causality. Also, the possible role of the magnetic shear structure in facilitating core barrier formation is examined. Experimental and theoretical challenges for developing a predictive capability for reactor-relevant conditions are highlighted by recent observations of spontaneous electric field shear generation far removed from edge effects, and efforts to characterize the plasma edge at and across the L-H transition.

In recent tokamak experiments several types of high-performance plasmas were obtained. One is the H (high)-mode triggered by the L (low) to H (high) transition. In this case the transport barrier is generated at the edge region. The obtained high confinement plasmas are characterized by gyro-Bohm-type transport from the scaling law. Another is the negative (central) shear or (enhanced) reverse shear mode with the transport barrier at the inner region close to the minimum point of the safety factor profile. With the internal transport barrier (ITB) the turbulent transport is suppressed significantly; however, the scaling study is primitive for the core plasma. To understand the negative shear configuration with the transport barrier, the high- plasmas are useful, since the ion thermal transport is improved without the reversal of magnetic shear. One key ingredient to obtain the above high-performance plasmas is the velocity shear which may suppress the turbulence and associated anomalous transport in general. Shafranov shift is also favourable. However, it is a difficult task to predict by what degree the anomalous transport is suppressed, particularly compared to L-mode plasmas. Some theoretical efforts into the understanding of high-performance plasmas are discussed.

Impurity poloidal rotation velocities have been measured in the core of TFTR plasmas using a new spectroscopic diagnostic. Two types of transition to enhanced confinement in reversed shear plasmas are examined. A bifurcation in carbon poloidal rotation is observed to occur before the transition to enhanced confinement for one of these types, while other measured plasma parameters remain constant. A narrow radial region with reversed poloidal rotation and rotational shear is established 60-100 ms before the transition and is associated with a large negative radial electric field.

The combination of additional heating of MW neutral beam injection and wall conditioning (boronization, titanium gettering and inter-shot helium-glow-discharge cleaning) has allowed the START spherical tokamak to achieve record beta values of . Here we report the more recent demonstration of a transition to a higher confinement mode.

Tritium and helium gas-puff experiments have been conducted on TFTR to investigate transport in reverse shear plasmas. Small amounts of tritium and helium gas are puffed into the equilibrium phase of these plasmas. Measurements of the 14 MeV neutron fusion product and the charge-exchange spectrum of helium reveal the transport details of tritium and helium, respectively. The inferred tritium and helium density profiles indicate the formation of a core transport barrier for each species in reverse shear plasmas with enhanced confinement.

In the CHS heliotron/torsatron, dynamic phenomena associated with transitions in radial electric field were observed during combined ECH + NBI heated plasmas. The observations with high temporal resolution confirmed a nonlinear relationship between radial electric field and radial current as the cause of these phenomena associated with electric field bifurcation.

In W7-AS the H-mode is characterized by an edge transport barrier localized in the first 3-4 cm inside the separatrix. In the ELMy H-mode preceding the quiescent state ELMs appear as a sudden breakdown of the edge transport barrier in coincidence with bursts of fluctuations. Between ELMs fluctuations are identical to those of the quiescent H-mode. The operational range of the quiescent H-mode is determined by narrow windows of the edge rotational transform and a threshold edge electron density. In contrast, ELM-like events are observed for a variety of plasma conditions by far exceeding the narrow operational windows for the quiescent state.

This paper summarizes the evidence for the role of shear flow in confinement improvement during electrode biasing experiments in TEXTOR-94: the transport barrier spatially overlaps the shear layer, the temporal causality is clearly established and the magnitude of shear at which the confinement occurs, agrees rather well with theoretical models. These results strengthen the conjecture that stabilization is a viable mechanism for improved confinement in tokamaks.

Reduction of electron thermal diffusivity was investigated by comparing the internal transport barrier (ITB) with that of high- discharges. In reversed shear discharges heated by neutral beams, reduction was significant when negative magnetic shear was strong. Radial location of ITB propagated outwards and then localized near the region of q-minimum for both the ion and the electron channels. Reduction in was also observed in the electron heating regime by ion cyclotron heating

Experiments using high-power heating during the current ramp-up phase of the discharge have obtained the highest D-D neutron rates in JET; neutrons , with , keV and keV. The best discharges ( MA and tesla) have peaked pressure profiles with a transport barrier located at = 0.55. The pressure peaking is limited by MHD modes and requires active input power control to achieve the best performance. Deuterium neutral beam injection into a tritium-rich target plasma has established internal transport barriers at power levels close to the lowest threshold for pure deuterium plasmas.

The role of rotation shear in the internal transport barrier formation is examined by means of a one-dimensional transport simulation based on the current-diffusive ballooning mode turbulence model. Delayed transition to the enhanced improved transport was reproduced in the medium heating power range. The threshold power is strongly reduced by the effect of rotation shear. Simulation including particle transport was also applied to NBI-heated plasmas and the steepness of density and temperature gradients is discussed.

A possible relation between the flow curvature and the reduction of electron thermal diffusivity is discussed. We show that unlike the flow shear, the flow curvature has no universal effects on the evolution of the two-point correlation function. Since various instabilities respond differently to the flow curvature, we can gain useful insights into the nature of the ambient turbulence responsible for the electron thermal transport by examining the profiles. Initial results from DIII-D and TFTR are discussed.

The influence of plasma flow inhomogeneity on the magnetic surface, which is driven by anisotropic hot ions, on the micro-turbulence of the low/negative magnetic shear tokamak is investigated. It is found that the poloidal inhomogeneity is effective in suppressing the current-diffusive ballooning mode turbulence which has a large nonlinear growth rate. This new mechanism of turbulence suppression provides the model of improved confinement associated with the reversed magnetic shear.

We demonstrate that Alfvén waves may be the convenient trigger for the formation of transport barriers in tokamaks. Kinetic or slow Alfvén waves, which have small radial localizations, can be used to create strongly sheared plasma flows at any point of the plasma minor radius to suppress the plasma turbulence and to obtain the locally reversed magnetic shear configuration.

The H-mode power threshold has a weak but positive dependence when the ion drift is away from the X-point, in contrast to the nearly linear dependence when the ion drift is toward the X-point. This indicates that geometry plays an important role in the H-mode power threshold scaling. A simple model of the drift effect failed to predict this behaviour, but successfully predicted the sign change of gas puffing and low X-point height on the power threshold. The difference between the threshold power required for sawtooth and non-sawtooth-triggered transitions can be substantial. This effect may contribute to the observed B scaling of the H-mode power threshold.

Measurements in DIII-D of edge ion and electron temperatures ( and ) just prior to the transition to H-mode are presented. A fitting model based on a hyperbolic tangent function is used in the analysis. The edge temperatures are observed to increase during the L-phase with the application of auxiliary heating. The temperature rise is small if the H-mode power threshold is close to the ohmic power level in the absence of auxiliary heating and is large if the H-mode threshold is well above the ohmic power level. The edge temperatures just prior to the transition are approximately proportional to the toroidal magnetic field , for the field either in the reversed or forward direction. However, for the reversed magnetic field, the temperatures are at least a factor of two higher than for the forward direction.

The effect of a non-uniform on ripple-trapped neutral beam ions is investigated using both the Monte Carlo method and Fokker-Planck approach. The increase in the neutral particle flux from ripple-trapped ions observed by the CX detector at an L-H transition in the ASDEX Upgrade tokamak is reproduced in both simulations by turning on near the plasma periphery.

The density of ripple-trapped ions is strongly influenced by negative radial electric fields that can compensate the grad B drift. This density can be monitored with a time resolution of about by the measurement of charge-exchange neutral particle fluxes. According to Monte Carlo and Fokker-Planck calculations these fluxes should respond with a time delay of less than , if the halfwidth of the electric field is reasonably large (a few centimetres). In the experiment, rise times of the fluxes after the L-H transition are longer or about 1 ms; jumps of the electric field in much shorter times can be excluded under such circumstances. The same is found for the H-L transition, where the flux measurement does not show any indication for a fast drop of the fields.

An edge temperature threshold for the L-H transition is found on the Alcator C-Mod tokamak. The critical temperature depends weakly on density and increases with and with unfavourable drift direction. at the H-L transition can be greater than or equal to the L-H threshold. Magnetic fluctuations are observed at some H-L transitions. Measured parameters are compared with H-mode theories, including nonlinear drift-ballooning code simulations.

The orbit loss current is calculated for the ASDEX Upgrade geometry using the 5D (3D in configuration space and 2D in velocity space) Monte Carlo code ASCOT. The balance between the obtained current and the analytical estimate of the return current shows an L-H transition for the normalized collision frequency as expected from analytic theory. The transition in ASDEX Upgrade, however, occurs at larger values of .

In TCV, L-mode to H-mode transitions have been observed in a wide variety of plasma configurations. Two statistical methods, discriminant analysis and a nonlinear system identification, have been used to localize the boundary between the L-mode and H-mode domains in the multidimensional operational space. For control purposes, an estimate of the shortest distance between the actual position in the operational domain and the boundary is defined, and the parameters to modify in order to reach the transition are identified.

A general equation for the radial current due to the ion-neutral friction and neoclassical viscosity is given. The equation models the effect of a possible poloidal non-uniformity of the neutral friction, and is valid in an arbitrary toroidally symmetric geometry. A strong dependence on the poloidal distribution of the neutral friction is found. With the help of Monte Carlo simulations the neutral friction in ASDEX Upgrade has been calculated. A large poloidal asymmetry in the neutral density exists, and for the studied shot the neutral friction in the edge is not large enough to compete with the neoclassical viscosity.

The `back' transition from H-mode to L-mode has been studied on DIII-D as a part of our investigation of the L-H transition power threshold scaling. Based on a density-dependent scaling for the H-mode power threshold, ITER will require substantial hysteresis in this parameter to remain in H-mode as rises. Defining the hysteresis in terms of the ratio of sustaining to threshold power, may need to be as small as 50% for ITER. Operation of DIII-D at injection powers slightly above the H-mode threshold results in an oscillatory behaviour with multiple forward-backward transitions in the course of a discharge. These discharges represent a unique system for studying various control parameters that may influence the state transition. Careful analysis of the power flow through the edge gives values for the sustaining power well below the corresponding threshold powers , indicating that substantial hysteresis can be achieved in this parameter. Studies of other control parameter candidates such as the edge temperature during the back transitions are less clear: the amount of hysteresis seen in these parameters, if any, is primarily dependent on the nature (ELMing, ELM-free) of the parent H-state.

The JT-60U divertor geometry was recently modified to a W-shaped structure in order to reduce the back flow of neutrals from the divertor. The H-mode threshold power decreased subtly compared with the open divertor case. As regards density dependence, the exponent of density was 0.5 to 0.75. In addition, we found that the outflux from the outer baffle plate should be considered for the evaluation of neutrals.

To assess the effect of edge neutrals on the low-to-high confinement transition threshold, a broad range of plasma discharges has been analysed. From this analysis, the transition power divided by the density, at constant magnetic field, appears to be a function of a single parameter measuring the neutrals' effect. This results suggest that there is a missing parameter linked to the neutrals in the power-threshold scaling laws.

An analysis of the complete JET D-D threshold database, from 1990 to 1997 inclusive is presented. A JET scaling is derived for the threshold power through the separatrix, estimated by considering also radiation losses from the bulk plasma. Such scalings are applied to obtain a first qualitative estimate of the amount of power above threshold that is needed to obtain steady-state high confinement (H89 > 1.8). The L-H and H-L transitions are also analysed in terms of local edge parameters. The first results on isotope scaling of the H-mode power threshold in D-T and T-T plasmas are also presented, confirming the scaling.

Global and local H-mode threshold analyses are presented and discussed. The density window was extended up to the density limit towards which the power threshold exhibits a dramatic increase. First results with the new divertor are compared to the previous threshold scaling. Local edge measurements at the L-H transition show that the temperature decreases slightly with density and increases with magnetic field, and that the ion collisionality at the plasma edge is always clearly above unity, between 5 and 15. Experiments at high power threshold exhibit improved L-mode confinement and suggest that two mechanisms are required for the L-H transition.

Recent high-power electron-cyclotron resonance heating (ECRH) H-mode experiments on COMPASS-D have used fundamental high-power ECRH to access high densities. Within the operating range set by the H-mode density threshold and ECRH accessibility, H-mode power thresholds are significantly higher than expected from ITER scalings, and show a contrary dependence on density. Energy confinement is somewhat lower than that suggested by the present ITER scalings, but correction for incomplete RF absorption may be needed. Complex H-mode and ELM behaviour is observed.

A statistical picture for the collapse is proposed. The physics picture of the crash phenomenon, which is based on the turbulence-turbulence transition, is extended to include the statistical variance of observables. The dynamics of the plasma gradient and the turbulence level are examined, with the hysteresis nature in the flux-gradient relation. Probabilistic excitation is predicted. The critical condition is described by the statistical probability.

Because of its high confinement properties, the H-mode provides good opportunities to achieve high beta values in a tokamak. In this paper the results of an experimental study of and limits in the H-mode, obtained in a circular cross section tokamak without auxiliary heating are presented. The experiments were performed in the TUMAN-3M tokamak. The device has the following parameters: (limiter configuration), . The stored energy was measured using diamagnetic loops and compared with W calculated from kinetic data obtained by Thomson scattering and microwave interferometry. Measurements of the stored energy and of the were performed in the ohmic H-mode before and after boronization and in the scenario with fast current ramp-down in ohmic H-mode.

A maximum value of of 2.0% and of 2.0 were achieved. The limit achieved reveals itself as a `soft' (non-disruptive) limit. The stored energy slowly decays after the current ramp-down. No correlation was found between beta restriction and MHD phenomena. Internal transport barrier (ITB) formation was observed in ohmic H-mode. An enhancement factor of 2.0 over ITER93H(ELM-free) was found in the ohmic H-mode with ITB.

The microwave reflectometry system on ASDEX Upgrade measures density profiles (in broadband swept operation) and plasma fluctuations (fixed-frequency operation) both at the high-field side (HFS) and low-field side (LFS). Densities up to can be probed. We analyse the evolution of turbulence at the L-H transition and during the ELMy phase of H-mode discharges. The detailed density profile evolution during type I ELMs is resolved and profile oscillations associated with ELM precursors are studied.

The edge pressure gradient (H-mode pedestal) for computed equilibria in which the current density profile is consistent with the bootstrap current may not be limited by the first-regime ballooning limit. The transition to second stability is easier for higher elongation, intermediate triangularity, larger aspect ratio, pedestal at larger radius, narrower pedestal width, higher , and lower collisionality.

The energy confinement time of high-current ELMy H-modes in JET is found to be lower than predicted by the scaling. An unprovoked transition to L-mode is often observed in ELMy H-modes without additional deuterium gas fuelling. This spontaneous H-L transition can be avoided by keeping the input power well above the L-H threshold power. The transition is analysed both in terms of global and local parameters.

Just above the H-mode threshold power at low density, small amplitude high repetition rate (1-5 kHz) type III edge-localized modes (ELMs) are observed. At higher input power, the H-modes become ELM-free. At high density and high input power (, MW), the plasma often enters a new H-mode regime now called the enhanced H-mode, which is a steady-state H-mode often without distinct ELMs but with high levels of emission. Enhanced H-modes are accompanied by persistent high-frequency broadband (50-400 kHz) density and magnetic fluctuations. Above about , the enhanced H-modes may also exhibit bursting or continuous low m, n coherent magnetic fluctuations with frequencies from 60-90 kHz. In the H-mode, steep edge gradients are observed with a nine-channel electron cyclotron emission (ECE) polychromator and a 38-channel, high-resolution, edge soft-x-ray array. The soft-x-ray pedestal width remains around 3 to 4 mm in the ELM-free H-mode and between type III ELMs. The pedestal collapses within during an ELM. Calculations indicate that the plasma in the H-mode is often near ideal ballooning first stability limits, but no type I ELMs have yet been identified. Edge micro-tearing modes destabilized by a steep temperature gradient are also considered as a possible explanation of the observed fluctuations during the enhanced H-mode.

Particle and energy losses due to type-I edge-localized modes (ELMs) and the effect of H-mode pedestal parameters on global confinement are studied for H-Mode discharges in ASDEX Upgrade. ELM frequency as well as particle and heat losses during ELMs are strongly affected by gas puffing, impurity injection, variation of plasma current and heating power. However the average power loss due to ELMs shows much less variation, and cannot account for the variations of stored energy encountered, e.g. in a plasma current scan. A good correlation between pedestal pressure and stored energy is found for type-I ELMy discharges.

By solving nonlinear fluid model equations, it is shown that the nonlinear interchange mode generates shear flow through Reynolds stress, which suppresses fluctuation amplitude substantially. The nonlinear interaction between vortex motion and shear flow produces a relaxation oscillation. This behaviour is also confirmed by a Lorenz type model with five ordinary differential equations. However, for the same Rayleigh number and Prandtl number , the period of relaxation oscillation becomes shorter and the amplitude of vortex motion becomes smaller.

Using the soft x-ray tomographic analysis, the rapid change processes of the plasma profile are investigated at the internal disruption and following intermittent bursts. A relation between the coherent structure with a low mode number (m = 2) and the local profile gradient is studied. It is found that there is a hysteresis relation between these. An explosive growth of the coherent mode and a functional dependence of the growth rate on the mode amplitude are also discussed

A simple physical model for toroidal flow shear stabilization of a hybrid dissipative trapped electron-ion temperature gradient mode (namely, the hybrid electron-ion drift mode) is presented. Coupling between the sheared flows and non-adiabatic electrons is proposed as the stabilization mechanism of the hybrid electron-ion drift mode. This implies that the confinement improvement can be achieved not only through the poloidal sheared rotation flow in the edge region of tokamaks for the L-H transition but also through the toroidal sheared rotation in the core region for further confinement improvement associated with the internal transport barrier and negative shear.

Several interesting new results have come from studies of ICRF-heated, H-mode plasmas in Alcator C-Mod. Dimensionless scaling studies have found gyro-Bohm-like transport similar to that reported on other devices; however, the dependence on collisionality was surprisingly strong, with . Despite high edge temperatures and strong edge pressure gradients, type I edge-localized modes (ELMs) are not observed in C-Mod. Instead we obtain a regime that we have dubbed enhanced (EDA) which is accompanied by high-frequency density fluctuations. For all H-modes, core gradients were found to increase linearly with edge temperature, suggesting the importance of critical gradient/marginal stability behaviour. Comparisons with the IFS-PPPL model have begun, showing quantitative agreement in some cases. Impurity particle transport was studied via the laser blow-off technique with impurity confinement found to be effectively infinite for ELM-free discharges but reduced into the range 0.1-0.2 s for the EDA plasmas.

The physics of the formation and termination of high ion temperature mode (high mode) is studied by controlling density profiles and radial electric field. High ion temperature mode is observed for neutral beam heated plasmas in heliotron/torsatron plasmas (Heliotron-E). This high mode plasma is characterized by a peaked ion temperature profile and is associated with a peaked electron density profile produced by neutral beam fuelling with low wall recycling. This high mode is terminated by flattening of the electron density caused either by gas puffing or by second-harmonic ECH (core density `pump out').

The improvement in H-mode confinement in the JFT-2M tokamak shows the difference between CO- and counter-neutral beam (CTR-NB) heating. The improvement in the CO is larger than that in the CO + CTR or CTR H-mode. The reason for is considered not to be the effect of the sheared toroidal rotation, , in the core. at the last closed flux surface shows a finite value (about 20-40 and the direction is dependent on the beam injection angle) in a divertor H-mode. The finite toroidal rotation at the edge plays a role in the observed . Since the radial electric field formed by the poloidal rotation is negative, the electric field shear at the edge is increased by the positive electric field of . In contrast, the electric field from is negative, which reduces the electric field shear. Therefore, the difference in the improvement between the CO and CO + CTR H-mode, which comes from the difference in the pedestal height, may be related to the electric field shear at the edge.

An empirical model for anomalous transport in tokamak, supplemented by the proper boundary conditions, has been developed and applied to four different types of JET discharges: ELM-free hot-ion H-mode, L-mode, ELMy H-mode and discharges with optimized magnetic shear. The results of the numerical modelling show that the model gives quite reasonable agreement with experiment in a wide range of plasma parameters and operational regimes. Moreover, they reveal some important generic features of plasma transport which are model independent.

Increased plasma confinement, temperature and fusion reactivity in negative central shear (NCS) discharges in DIII-D are accompanied by reduced core electrostatic microturbulence. The microturbulence reduces as the local radial electric field and shear increases, consistent with a theoretical model incorporating turbulence stabilization by shear radial electric field. Reduced turbulence and the associated anomalous transport reduction leads to further increased radial electric field shear via a steeper pressure gradient and reduced momentum transport. During the H-mode phase of a core transport barrier discharge, the microturbulence is virtually quenched in the core. Increasing evidence indicates that the transport barrier initially forms in the plasma interior when shear is large.

We show that the large negative radial electric fields which are measured in front of the separatrix in H-mode discharges are easily explainable on the basis of the rigorous `revisited' neoclassical theory, including finite Larmor radii and inertia effects that was published earlier (Rogister A 1994 Phys. Plasmas 1 619); the same theory naturally leads to sub-neoclassical energy transport and novel particle pinch terms. The calculation has so far been developed only in the high collisionality regime: step sizes comparable to gradient-scale sizes are therefore not required to explain observed properties! Based on the analysis, we conclude that the radial electric field profile develops a well in front of the separatrix when the plasma is unable to sustain ambipolar flows otherwise.

Computations of fluid drift turbulence treating ions and electrons on equal footing, including both temperatures, are conducted in a model toroidal geometry. The resulting `ion mixing mode' turbulence bears features of both electron drift-Alfvén and ion temperature gradient turbulence, and nonlinear sensitivity to the relative strengths of the density and temperature gradients provides a possible route to the bifurcation needed for the L-H transition.

Exploratory work on scaling the H-mode threshold power with edge non-dimensional quantities was first undertaken based on the JT-60U H-mode database. Predominant edge dependence as well as the strong negative contribution of edge were found. Obtained scaling is consistent with the global parameter dependences and also satisfies the non-dimensional constraint.

Transport simulations have been carried out for L- and H-mode plasmas in DIII-D and JET installed in the ITER database with four transport models, i.e. the current diffusive ballooning mode (CDBM) model, the mixed shear model and drift-wave turbulences called IFS/PPPL and multi-mode models. In the CDBM and the mixed shear models, simulation results show very similar characteristics in L- and H-mode plasmas on DIII-D and JET, although the adjustment of the diffusion coefficient might be necessary for both models. In comparison with L-mode plasmas, the H-mode plasmas should be simulated with a reduction of the transport coefficient by a factor of for electrons and for ions. In addition, a systematic difference is observed in the confinement performance between DIII-D and JET. Compared with the transport coefficient in JET, that in DIII-D should be reduced by a factor of for electrons and for ions. In the IFS/PPPL model with no shear stabilization, systematic differences are not observed in L-/H-modes and DIII-D/JET plasmas. The ion temperature profile seems to be very stiff to the critical ion temperature gradient, especially in H-mode plasmas in comparison with L-mode ones. By changing the density profile slightly and the edge pedestal temperature, the sensitivity of the ion temperature profile is examined. The ion temperature is, however, not so sensitive to these parameters. When the density profile is changed by a factor of with the form of , the change of stored energy is around -5-13%. In the multi-mode model, the agreement between experimental data and simulation results seems to be better, compared with other models.

We propose a set of unified models to describe H-mode operation of the core plasma, the edge plasma including the transport barrier, and the scrape-off layer plasma to extrapolate to ITER. These models are being incorporated into several tokamak transport codes for use in extrapolating to ITER. Edge pedestal temperatures of 3.5-4 keV for edge densities between 6 and are predicted for ITER, which is the pedestal temperature range required for good confinement with ITG-based core models. The edge operational space predicted for ITER is compatible with both the required core density and core confinement and exhibits a large window between H-mode threshold and the ballooning limit.

The H-mode confinement enhancement factor, H, is found to be strongly correlated with the height of the edge pressure pedestal in ITER shape discharges. In discharges with type I edge-localized modes (ELMs) the pedestal pressure is set by the maximum pressure gradient before the ELM and the width of the H-mode transport barrier. The pressure gradient before type I ELMs is found to scale as would be expected for a stability limit set by ideal ballooning modes, but with values significantly in excess of that predicted by stability code calculations. The width of the H-mode transport barrier is found to scale equally well with pedestal or . The improved H value in high- discharges may be due to a larger edge pressure gradient and wider H-mode transport barrier consistent with their higher edge ballooning mode limit. Deuterium puffing is found to reduce H, which is consistent with the smaller pedestal pressure which results from the reduced barrier width and critical pressure gradient. Type I ELM energy loss is found to be proportional to the change in the pedestal energy.

An offset nonlinear scaling is developed for the ELMy H-mode by analysing the ITER H-mode database ITERH.DB2. The offset part of the stored energy is determined by the MHD stability of ELM, and the incremental confinement time of a nonlinear function of heating power is determined by the gyro-Bohm-like transport of the core plasma. This scaling predicts a lower confinement time for ITER than that predicted by general power-law scaling.

Analysis of the latest ITER H-mode threshold database is presented. The power necessary for the transition to H-mode is estimated for ITER, with or without the inclusion of radiation losses from the bulk plasma, in terms of the main engineering variables. The main geometrical variables (aspect ratio , elongation and average triangularity ) are also included in the analysis. The H-mode transition is also considered from the point of view of the local edge variables, and the electron temperature at 90% of the poloidal flux is expressed in terms of both local and global variables.

1 - 3 June 1998 25th International Conference on Plasma Sciences Raleigh, NC, USA Contact: J G Gilligan, North Carolina State University, Box 7901, 147 McKinnon Center, Raleigh NC 27695-7901, USA. Tel +1-919-515-3639, Fax +1-919-515-7951, E-mail john_gilligan@ncsu.edu

7 - 11 June 1998 12th Topical Conference on High-Temperature Plasma Diagnostics Princeton, NJ, USA Contact: Robert Kaita, PPPL, PO Box 451, Princeton, NJ 08543-0451, USA. Tel +1-609-243-3275, Fax +1-609-243-2418, E-mail kaita@pppl.gov

7 - 11 June 1998 Embedded Topical Meeting on the Technology of Fusion Energy Nashville, TN, USA Contact: American Nuclear Society Meetings Department, 555 North Kensington Avenue, LaGrange Park, IL 60326, USA. Tel +1-708-352-6611, Fax +1-708-352-6464, E-mail meetings@ans.org, website http://www.ans.org/meetings/

29 June - 3 July 1998 1998 International Congress on Plasma Physics (Joint Conference of the 12th Kiev International Conference on Plasma Theory and 12th International Congress on Waves and Instabilities in Plasmas) and 25th Conference on Controlled Fusion and Plasma Physics Prague, Czech Republic Contact: ICPP 98, Institute of Plasma Physics, Za Slovankou 3, PO Box 17, 182 00 Prague 8, Czech Republic. Tel +420-2-688-3028, Fax +420-2-858-6389, E-mail plasma98@ipp.cas.cz

30 June - 1 July International Conference on Fusion and Industry Oxfordshire, UK Contact: Ms Sue Frye, Thomas Telford Conferences, One Great George Street, London SW1P 3AA, UK. Tel +44-171-665-2315, Fax +44-171-233-1743.

6 - 17 July 35th Culham Plasma Physics Summer School Abingdon, UK Contact: Geoff Maddison, UKAEA Fusion, Culham Science Centre, Abingdon, Oxon OX14 3DB, UK. Tel +44-1235-463-672, Fax +44-1235-464-192, E-mail geoff.maddison@ukaea.org.uk, Website http://www.fusion.org.uk/info/sschool98.htm

27 - 31 July International Conference on Open Magnetics Systems for Plasma Confinement Novosibirsk, Russian Federation Contact: Dr Peter Bagryansky, OS98, Budker Institute of Nuclear Physics, Prospect Lavent'eva 11, 630090 Novosibirsk, Russian Federation. Tel +7-3832-359-772, Fax +7-3832-352-163, E-mail os98@inp.nsk.su, Website http://www.inp.snk.su/OS98/

28 - 31 July Chapman Conference on Magnetic Helicity in Space and Laboratory Plasmas Boulder, CO, USA Contact: AGU Meetings Department, 2000 Florida Avenue NW, Washington, DC 20009, USA. Tel +1-202-462-6900 x215, Fax +1-202-328-0566, E-mail meetinginfo@kosmos.agu.org

7 - 11 September 20th International Symposium on Fusion Technology (SOFT) Marseille, France Contact: Mme V Poli - 20th SOFT, DRFC, CEA/Cadarache, F13108 Saint-Paul-lez-Durance Cedex, France. Tel +33-4-42-25-62-60, Fax +33-4-42-25-62-55.

18 - 24 October 17th International Conference on Fusion Energy Yokohama, Japan Contact: Conference Service Section, IAEA, PO Box 100, A-1400 Vienna, Austria. Tel +43-1-2060-21310, Fax +43-1-20607, E-mail official.mail@iaea.org