I'm reading this paper describing the capabilities of the ESA's DELTA space debris forecast model.
It's clear that this model uses a statistical, flux-based model derived from the MASTER debris collision risk assessment tool.
However, it's not immediately clear to me whether taking such a flux-based approach, rather than considering individual pieces of debris would qualify this as a stochastic one.
Are flux-based modelling techniques fundamentally stochastic?
I think the question in your title and the question posed in your body might have different answers.
DELTA itself contains a stochastic process for collision prediction (see page 2) and so is inherently stochastic itself (though the authors prefer to call it "semi-deterministic"):
The final source of space debris in DELTA are collision events. The collision event prediction is done by using a target centered approach, developed to stochastically predict impacts between all objects within the DELTA population [7, 8, 9].
There's also a number of Monte Carlo steps, which are going to necessarily be stochastic (starting on page 3).
I've struggled some with the answer to the body question and I think that there is a way to make even a flux-based model completely non-stochastic if you don't have steps like the one I've quoted about collision events, where there's a sort of collapse from statistics to an event by "rolling the dice." If you kept the representation "in statistics" — i.e., a "collision" is some kind of convolution and now there's a spread of probabilities of debris fluxes and the ultimate result is also some set of statistics — then I think it might not be a stochastic process.
Hopefully that's enough hedging on that. I stand by DELTA itself being stochastic though.
