Basically you'd need to replace those 1.66, 6.399 and requirement of 700 km. However from what I understand you'll actually need values such as magnitude at distance and Richter magnitude to make a formula. As for how to do it without that (like through comparison with current formula) I don't know.
For example 1.66 scales how strongly magnitude decreases with logΔ, but a bigger planet may have different decay slope. Asking these to DT will be better imo.
If you want the source to analyze better
here it is.(page 11) But there Richter C. F. 1958. Elementary seismology. San Francisco: W. H. Freeman seems to be cited (page 342). Unfortunately on the table only distances up to 600 km are shown.
But logic is that (not sure but someone can correct me if needed), at some point (700 km in this case) graph becomes more linear and likely matches linear function a * logΔ + b.
So if you fit a good line that'll relatively represent those data points, its slope will be approximately ≈1.66. This means that amplitude attenuation follows 1 / Δ^1.66. So 1.66 here is average that fits best the observed data and it accounts for all factors that reduce wave energy.
Other than that we need to take a value for b, here 6.399 is calibration constant, to match result of formula to values we have. Shortly, we figure slope of line and additive constant that'd match equation to the needed result.
So to make one, you'll prob need a whole dataset of how magnitude decreases with distance in OP planet, and figure out at what distance magnitude change becomes likely linear with log of Δ. Then based on that make constants yourself.