Data-optimized source modeling with the Backwards Liouville Test-Kinetic method

In this paper, we discuss the general applicability of the Backwards Liouville Test-Kinetic (BLTK) method to the modeling of time-dependent particle distributions in magnetospheric plasmas. We discuss a new particle tracing model, SHIELDS-PTM, and describe certain novel features of its implementation. Finally, we apply SHIELDS-PTM to highlight a powerful but oft-overlooked feature of the BLTK approach: the invariant trajectories are independent of the quantity being transported, so it is straightforward to optimize source models with undetermined parameters. We demonstrate this by using SHIELDS-PTM to determine the local-time distribution of particle fluxes around geosynchronous orbit using simple models for the electromagnetic fields and statistical models of geosynchronous electron fluxes. We find that optimized source distributions generally provide a far superior result than statistical models of the source region, improving the accuracy of modeled fluxes at some energies by more than an order of magnitude.