Magnitudes at close and very close distances in eastern Canada: issues and recommendations

Abstract

The Nuttli (MN) scale is the most frequently used magnitude scale in eastern Canada. It is based on the amplitude of the Lg phase and therefore is not appropriate for distances of less than 50 km where the Lg phase is not developed. The original Richter, ML, scale developed for use in California and known to be inappropriate for eastern North America, is used in eastern Canada only when the Lg phase is highly attenuated or non-existent, generally for earthquakes occurring in oceanic crust or small earthquakes that are not recorded at distances of greater than 50 km. This study focuses on the latter by establishing a magnitude relation between MN calculated at appropriate distances and MN or ML calculated at close distances. The magnitude relation would enable magnitudes for very small earthquakes to be calculated from a larger number of stations and a magnitude recurrence relation to be established over a wider magnitude range. It would also enable a direct comparison of earthquakes from a particular region recorded only locally with those recorded over a wider distance range. Data from earthquakes occurring within several regions of eastern Canada for which MN was reported as the preferred magnitude but where amplitude and period data from local stations were also measured and archived, ML and MN from stations at less than 50 km from the epicenter were calculated and compared to the published or event magnitudes. In most regions ML underestimates the magnitude by more than 1 magnitude unit whereas the MN (<50 km) values were only about 0.1-0.2 units smaller than the presumed magnitude of the earthquake indicated by MN calculated at appropriate distances. A series of conversion relations were developed and the effects of initial magnitude type, complexity of conversion relation and global vs. regional corrections were compared. The effect of using hypocentral distance instead of epicentral distance for earthquakes at less than 50 km was evaluated but except for the closest and deepest events, the effect is not significant. Additionally, because precise depths are rarely determined for small earthquakes in eastern Canada hypocentral distances for most regions would be approximations only. For distances from approximately 10 km to 50 km it is recommended that MN be calculated and that 0.11 be added to the value. These magnitudes can then be combined with MN magnitudes from distances of greater than 50 km to determine the event magnitude. At distances of less than 10 km the conversion relations are not reliable and there appears to be more regional variation. Unless there are no stations at distances of 10 km or greater, magnitudes should not be calculated from such close distances. When there is no alternative, the same procedure as for 10-50 km distances should be followed but it should be noted that the magnitudes might not be equivalents to regional MN. For routine seismicity analysis, stations at extremely close distances are rarely essential for magnitude determination. However, in special cases usually related to focused studies of a particular location data from these distances may be crucial. In conjunction with the study of magnitudes determined from close distances, the practice in eastern Canada of using only the MN equation for greater distances of the two equations proposed by Nuttli (1973) for all distances in eastern Canada was re-evaluated and found to be a better choice than the two-equation option. However, it was also demonstrated that attenuation relation is not ideal for eastern Canada as there is a discrepancy between magnitudes calculated at distances greater and less than 4°.