Probabilistic Modelling for Earthquake Forecasting in the Northwestern Part of Haryana State, India

Abstract

This study attempts to estimate the probability of the occurrence of large earthquakes (Mw ≥ 5.5) in the northwestern part of Haryana state, India, where a new nuclear power plant (NPP) is going to be constructed in the near future. First, an earthquake catalogue is developed for the period 1803–1986, and five stochastic models, namely lognormal, Weibull, gamma, Rayleigh, and double exponential, are then applied to past earthquake data. The performance of these models is checked using three statistical tests, and the lognormal, Weibull, and Rayleigh models are found to produce good approximations for this region, whereas the double exponential and gamma models yield intermediate and poor results. Hence, cumulative and conditional probabilities and related hazard curves for future earthquakes are estimated using the most suitable models. The cumulative probability of the occurrence of an earthquake (Mw ≥ 5.5) since the last event (1986) reached 0.95–0.98 as of 2018. The conditional probability of the occurrence of such an earthquake reaches 0.90–0.95 about 9–12 years from now (2027–2030), when the elapsed time will be 32 years (i.e., since 2018). The probability of earthquakes with different threshold magnitudes is then estimated, and based on the outcome of this investigation, earthquake magnitudes are classified from occasional (Mw ≤ 6.1) to very rare events (Mw ≥ 7.6) that this region may experience in the future. The findings of this study will be considered in seismic hazard assessment, liquefaction hazard assessment, and earthquake-resistant design of NPP components.