Header menu link for other important links
Implementation of Sugeno: ANFIS for forecasting the seismic moment of large earthquakes over Indo-Himalayan region
Published in Springer Netherlands
Volume: 90
Issue: 1
Pages: 391 - 405

The earthquake is known to be an unpredictable geophysical phenomenon. Only few seismic indicators and assumptions of earthquakes can be predicted with probable certainty. This study attempts to analyze the earthquakes over the Indo-Himalayan Border region including Bhutan, Bangladesh, Nepal, China and India during the period from 1995 to 2015. Bangladesh, Bhutan and China borders experience fewer earthquakes than Nepal and India border regions. However, Indo-China rim has inconsistency and vast range in its magnitude. Bangladesh though is a small country with respect to others, but it experiences earthquakes comparable to Bhutan. Nepal experiences highest number of earthquakes. In the last 20 years around 800 records have been observed with moment magnitude CloseSPigtSPi 4.0 Richter scale, while very few records (around 10–12) have been observed for large earthquakes having moment magnitude CloseSPigtSPi 6.0 Richter scale over the region. In this study adaptive neuro-fuzzy inference system has been implemented to assess the predictability of seismic moment associated with large earthquakes having the moment magnitude between 6.0 and 8.0 Richter scales using different combination of epochs, technique and membership functions. The Gaussian membership function with hybrid technique and 40 epochs is observed to be the reasonable model on the basis of the selected spatial and temporal scale. The forecast error in terms of root-mean-square error with the stopping criterion 0.001 has been observed to be 0.006 in case of large earthquakes (CloseSPigtSPi 6.5 Richter scale), that is, forecast accuracy of 99.4%. The model bias of 0.6% may be due to inadequate number of large earthquakes having moment magnitude CloseSPigtSPi 6.5 Richter scale over the region. © 2017, Springer Science+Business Media B.V.

About the journal
JournalData powered by TypesetNatural Hazards
PublisherData powered by TypesetSpringer Netherlands