Estimates of the probability of future earthquake activity are difficult to make in areas where historical seismicity may be low or absent, but where young fault scarps attest to recent or ongoing tectonism. Three non-Poisson models, a Weibull model, a Gaussian model and a lognormal model, are used to estimate the earthquake hazard for one such area, the northern Rio Grande Rift. This portion of the Rio Grande Rift displays numerous Holocene faults attesting to ongoing tectonism, but displays essentially no historical seismicity. The earthquake hazard for the Sangre de Cristo fault zone from Taos, New Mexico to Salida, Colorado calculated using these models is remarkably consistent (probability of at least one Mo = 7 earthquake in the next 50 years ∼ 2.5 × 10−3), with increased hazard for the Sangre de Cristo fault in north San Luis Valley (∼5.0×10−3) and near Taos (∼1.0×10−2) due to the long holding times along these segments.
