ttrace2equator produces excessive numbers of trace points, ttrace2iono sometimes "gets lost" with t01 model with actual solar wind parameters

[jameswilburlewis]

When tracing to the equator, the traces look reasonable until the get close to the equator...apparently when the 'refine' part of the algorithm kicks in, it starts spinning its wheels without really getting anywhere. Traces that should complete in 30-50 points are generating thousands more points, sometimes exceeding the 10,000 point limit and quitting before the end condition is reached. If the trace points are being saved, that results in a gigantic tplot variable -- as much as 350 Gbytes for tracing one day of THEMIS-A data when save to a .tplot file. That's too big to upload to Github, which interferes with the IDL vs Python tests I'd like to include in our Github Actions workflows.

This is definitely on the FORTRAN side of things, so I'm going to have to pass this along to Haje to handle. It should be pretty straightforward to distill the problem down to something that's easily reproducible with GEOPACK-only calls.

[jameswilburlewis]

ttrace2iono gives reasonable results with t89, t96, and t04s, when using actual solar wind parameters, but there are some problems when tracing with the t01 model. For some times/locations, the foot points are way out in space, instead of near the ionosphere. This happens with or without /geopack_2008, and with or without /refine. This is another case where we'll have to reproduce the issue with just GEOPACK calls, then ask Haje to look at it.