Version 1.1.0¶

API Changes¶

The freq_cutoff_params parameter for the Cpdr class has been changed from a list to a Gaussian object. A convenience method, gauss.from_gyrofrequency_params, has been added to help preserve the old functionality. This change has been motivated by the need to fix the wave frequency spectrum parameters (with respect to latitude) when computing bounce-averaged diffusion coefficients in bounce_averaged_diffusion_coefficients.py.

For bounce averaged calculations the wave frequency spectrum parameters are now fixed with respect to latitude (and are defined in terms of equatorial gyrofrequency in the input file).
The lower bound for the WhistlerFilter class in wavefilter.py is now defined as the maximum of the lower hyrbid frequency and the proton gyrofrequency. In addition there is a new constraint that the wave resonance cone angle is a real number.
For bounce averaged calculations the number density is now fixed with respect to latitude in bounce_averaged_diffusion_coefficients.py (and is defined via the ratio of the equatorial plasma frequency to the equatorial gyrofrequency in the input file).

Added a lower_hybrid_freq property to enable users to conveniently obtain the lower hybrid frequency for each particle in a plasma without the need to (re-)calculate this separately.

Added the mlat_cutoff parameter to the input specification, allowing users to set a maximum magnetic latitude beyond which it is assumed that no waves are present.

Used the newly added PlasmaPoint.lower_hybrid_freq property to set a lower bound (the lower hybrid frequency for protons) on the acceptable frequency range in WhistlerFilter. There is an implicit assumption in here that we are only looking at electron-proton plasmas, which will need to be addressed when we improve the support for more complicated plasma compositions in the future.

Cap the upper limit on the range of integration over the magnetic latitude to 0.9999 times the mirror point, to ensure we use a consistent approach for the integrable singularity at the mirror point. Previously, when calculating the integrand at the mirror point Bounce.get_bounce_pitch_angle would return either exactly pi/2 or a value very close to it (approaching from below), dependent on floating point precision. The pi/2 pitch angle gets skipped (for being singular) whereas anything else is included. Capping the magnetic latitude at 0.9999 times the mirror point ensures that we don’t have to deal with this ambiguity. A more robust long-term solution might be to re-examine the behaviour of Bounce.get_bounce_pitch_angle and look for a way to more accurately evaluate the integrand at the mirror point.

Added a warning message that is printed when the resonance cone angle is imaginary.