As in the past I note that a reasonable a priori case can be made on the basis of propagation characteristics that somewhat different metrics in the G(Δ, n) series might be better representations of RBN coverage on some of the bands. However, rather than make this into a full-scale research project, I shall simply continue to use the G(15, 100) metric on the basis that it seems "good enough" on all bands.
RBN Posting Stations as a Function of Time
We begin by looking simply at how the number of per-band posters to the RBN has varied since the RBN's inception. (NB Throughout this post, we ignore posters for which the location is not recorded by the RBN; plots for which the abscissa is time show one datum per month.)
First, a plot of the total number of posters as a function of time:
G(15, 100) as a Function of TimeTurning now to the geographical distribution of the posting stations, we can display the mensal values of G(15, 100) in a similar manner:
As in prior years, this makes it depressingly obvious how much room remains for improvement in RBN coverage, and how slowly the situation is changing.
G(15, 100) as a Function of the Number of Posters
Finally, we can combine the mensal values of G(15, 100) and the number of posters. Firstly, including all bands:
The summary plot for these data is slightly different, as the ordinate is multi-valued for some values of the abscissa. So, in this summary plot, we take the mean value of G(15, 100) in bins of width equivalent to ten posters, and plot rectangles in the equivalent colours: