We have good records of impact levels on the Moon, Mars and Jupiter (although Jupiter is a little weird).

Ćirković, Sandberg & Bostrom refer to these briefly in the paper, but seem to think they're not adequate or as relevant:

For instance, the asteroidal and cometary impact history of the solar system is, in theory, easier to obtain for the Moon, where the erosion is orders of magnitude weaker than on Earth. In practice, this is still not feasible for obtaining the fair sampling of the impactors because: (1) precise dating of a large set of lunar craters is beyond our present capacities and (2) most of the large lunar craters are known to originate in a highly special epoch of the so-called Late Heavy Bombardment, ca. 4.0-3.8 billion years B.P., thus strongly skewing any attempt to plot the empirical distribution function of impacts for "normal" times. In practice, in the current debates about the rates of cometary and asteroidal impacts, it is the terrestrial cratering rates that are used as an argument for or against the existence of a dark impactor population, thus offering a good case on which the anthropic model bias can, at least potentially, be tested. [Footnote & endnote references omitted.]