|Modelling the second wave
|Waves have become a political football, along with masks.
However, there is no question that in the last month there
has been both a change in exponential growth and the
infected population have apparently become a lot younger.
To test this idea I have used the model with time varying
infections per person, to create 2 independent overlapping
infections. The result for Texas can be seen to the right, the
sudden drop in cases in late May, signals that the first wave
has stopped growing, and starts to decay. The sharp rise in
mid June, signals that the second wave has taken over.
Overall the fit is encouraging. There are reports that
second wave has a much younger demographic, which
explains why the fatality rate in wave 2 is 1/3 of wave 1. I was
not able to find an archive of the demographic of cases to
test this idea.
I applied the same 2 wave approach to Florida, as shown to
the right, we similar results. I focused on the fit to cases,
which has really good. The death count during the transition
is a little underestimated. The fatality rate in waves 2 was
also lower, and anecdotally the population was younger.