Laserfiche WebLink
within the study region (Figure 2) (partly due to poor seismographic coverage) and hence <br /> sparse data preclude this approach. Earthquake recurrence for the Colorado Rocky Mountain <br /> region, which stretches westward from the Front Range to 108°W and from 37°N to 41°N <br /> (excluding the Rio Grande rift province), was estimated based on a historical catalogue <br /> containing just over 250 earthquakes (Unruh et al., 1995a). <br /> The recurrence relationships were estimated following the maximum-likelihood procedure <br /> developed by Weichert (1980) and the estimated completeness intervals for the region. <br /> Dependant events, either foreshocks, aftershocks or smaller events within an earthquake <br /> swarm (the largest event is assumed to be a mainshock) were identified using empirical <br /> criteria for the size in time and space of foreshock-mainshock-aftershock sequences <br /> developed by Arabasz and Robinson (1976), Gardner and Knopoff (1974), and Urhammer <br /> (1986). If an event were identified as dependent by two of the three criteria, it was deleted <br /> from the catalogue. <br /> After adjusting the historical records for dependent events and incompleteness, the recurrence <br /> for the Rocky Mountain Region was calculated based on 22 earthquakes in the range ML 3.0 <br /> to 6.5. Although there is no specific requirement on the number of events necessary to <br /> constitute an adequate database which would justify the assumption of a Poisson process <br /> (hence, provide an accurate assessment of earthquake recurrence), the small number of <br /> events suggests that large uncertainties must be associated with any resulting estimates. The <br /> number of earthquakes was normalized on an annual basis and per km2. The area used for <br /> the Colorado Rocky Mountains region is approximately 110,000 km2. Regression was <br /> performed on the resulting data points as described by Weichert (1980). The recurrence <br /> relationship is of the form of the Gutenberg-Richter relationship: log N = a - bM, where <br /> N is the annual number of earthquakes of magnitude >_ M, the a-value is a constant and <br /> represents the log of the annual number of earthquakes for M >_ 0 and the b-value is the <br /> slope of the regression. <br /> A b-value of 0.90 f 0.14 (standard deviation) was determined for the Colorado Rocky <br /> Mountain region (Unruh et al., 1995a). In comparison, Presgrave (1977) computed a b- <br /> value of 0.86 for the Steamboat Springs and Elkhead Mountains areas based on the <br /> recordings of earthquakes at GOL from 1966 to 1973 and more recently McGuire (1993) <br /> computed a b-value of 0.83 for the whole of Colorado. The normalized a-value of-2.33 for <br /> HACONTRAMTENMELE115 15 M0412951500 <br />