What we didn't show in the paper was that if you do this for say a year, on average everything comes out to zero. That gave us confidence that we were on the right track. What we found and showed in the paper was a bit different. We were looking specifically at substorms, a type of space weather - kind of like a space tornado. One thought is that as we fill the magnetosphere up with energy, it eventually needs to be released and that's what drives a substorm. It was thought (and I think still is) that there is a growth phase where energy input into the magnetosphere is increased. The growth phases last for typically upwards of an hour. However, if there has been not much going on, it could be longer. When we look at these time scales, we found that there was a continuous loss of energy from the system. This tells us there are potentially a couple of different things happening:
- We are missing energy that is coming into the magnetosphere
- We aren't allowing for enough of a build-up before the substorm starts
- We are overestimating our loss terms.
In the paper, we focused more on looking at how the different terms affected the estimate. For example, while the Dst may be a proxy for loss from the ring current - it is really just measuring the change in the magnetic field (not an energy term). We considered that perhaps it's not the best proxy or needs to be better adjusted to reflect the energy loss.
This was really a fun project and paper to be involved in. It's a simple approach and provides quite a bit of insight into the system.