Accurate Determination of Local Magnitude for Earthquakes in the Western Canada Sedimentary Basin

In this study, we take a close look at the constituents of the Richter (1935) relation for calculation of local magnitude (M-L), which is the basis for magnitude determination by Natural Resources Canada (NRCan) in the western Canada sedimentary basin (WCSB). Using a comprehensive catalog of Wood-Anderson amplitudes from earthquakes in northeast British Columbia and western Alberta, we first compare the distance correction terms - log(A(0)) for the Richter magnitude scale previously obtained for WCSB and several other regions. We also formulate a new correction term specifically for NRCan's routine M-L calculation that better accounts for the attenuation of direct and refracted waves from events within WCSB. Based on a bilinear model for ground-motion attenuation, our -log(A(0)) is {0.7974 x log (R-hypo/100) + 0.0016 x (R-hypo - 100) + 3.0 R-hypo <= 85 km, -0.1385 x log(R-hypo/100) + 0.0016 x (R-hypo - 100) + 3.0 R-hypo > 85 km in which R-hypo is the hypocentral distance. Our - log(A(0)) results in lower M-L by an average of 0.29, 0.27, 0.12, and 0.34 units, respectively, compared with those obtained by Richter (1958; California), Hutton and Boore (1987; California), Brazier et al. (2008; Ethiopian plateau), and Bona (2016; Italy) over all distances, but gives higher M-L values than those obtained by Yenier (2017; WCSB), with an average of 0.12 unit over all distances. The difference between our M-L calculation and Yenier (2017) is more significant for R-hypo <= 50 km (0.27 unit) and varies slightly for larger R-hypo : 0.08 unit for 50 km < R-hypo <= 100 km, 0.12 for 100 km < R-hypo <= 200 km, and 0.10 for R-hypo > 200 km.