EARTHQUAKE EARLY WARNING SYSTEM: SITE CLASSIFICATION, INTENSITY MAP AND ATTRIBUTES

Abstract

The Himalaya is one of the most seismically active regions of the world. Many Indian states are in the Himalayan region, and Uttarakhand is one of them. This region lies in seismic zones IV and V of the seismic zone map of India (IS 1893 (Part1), 2016) and is prone to seismic hazards. Seismic gap studies carried out of the Indian Himalayan region (Srivastava et al., 2015) identify a central seismic gap, namely Uttarakhand-Dharchula between Kaurik fault in Himachal-Pradesh and Main Central Thrust (MCT) near Dharchula, Nepal. The major part of the Uttarakhand falls in this seismic gap region. Many researchers have envisaged that a large earthquake is impending in this region. Therefore, an Earthquake Early Warning (EEW) System for this region can serve as a foremost tool to reduce losses due to future earthquakes. An EEW System for Uttarakhand has been developed by EEW System Laboratory, Centre of Excellence in Disaster Mitigation & Management (CoEDMM), Indian Institute of Technology (IIT) Roorkee. Now, this system is full-fledged operational. A further research study that needs to be addressed is required to be carried out. In the present thesis, an attempt has been made to fill the identified gaps. In this study, all the prevalent EEW attributes and their formulation are presented in tabular form. The empirical relationships to estimate magnitude using EEW parameters developed by different authors are also presented in a tabular form. Several authors have formulated GMPEs for the Indian Himalayan region, which are described in the literature review. Various site classification schemes used by different authors are described in the tabular form. Studies carried out by different authors in the preparation of PGA and intensity maps are explained. The developed regression relationships between Modified Mercalli Intensity (MMI) and PGA by various authors are presented in a tabular form. Finally, the research gaps are identified to address the objective to enhance the potential of existing EEW System and carried out unwinding of complexities of the earthquakes in general. There is no regional attenuation model available for the Uttarakhand region; therefore, an attempt has been made to develop a regional GMPE for this region. A dataset of 116 records of 9 earthquakes having magnitudes ranges from 5.0 to 6.8 of this region, recorded by ‘U.P. array’ (Shrikhande, 2001), Indian National Strong Motion Instrumentation Network (NSMIN) accelerographs (Kumar et al., 2012), and presently operating EEW System for Uttarakhand has been used to develop attenuation relationship for peak ground acceleration (horizontal components) for Uttarakhand region. The hybrid (i.e. two step-stratified) regression analysis has been used in the modelling of a new relationship (equation 1). 𝑙𝑜𝑔10 𝐴 (𝑔) = −1.091 + 0.3245𝑀 − 1.0632 𝑙𝑜𝑔10(𝑋 + 𝑒𝑥𝑝0.4561𝑀) (1) The developed relationship has been compared with other developed relationships for the whole Himalayan region (Singh et al., 1996; Sharma, 1998; Kumar et al., 2017; Joshi et al., 2013), which at present, could be used for the Uttarakhand region also. However, the newly developed relationship for Uttarakhand region shows good results despite having small data set used in the development of this relationship. Thus, the developed GMPE can be considered to be more reliable for this region. The root mean square error (RMSE) of the present relationship is found to be comparable with other relationships, but much superior to the relationship developed by Joshi et al. (2013). The instrumentation of the EEW System for Uttarakhand comprises 167 sensors installed in the hilly region of the state. After processing of the data, only 97 stations qualified the eligibility criteria. The qualification criteria are signal to noise ratio (SNR) and preset threshold of PGA of the records. The horizontal by vertical (H/V) spectral ratio curves of all the stations are created by both Fourier amplitude spectra (FAS) and 5% damped response spectra. The effects of depth, distance, and magnitude of earthquakes on the H/V spectral ratio curves are discussed. The effect of change in percentage of damping ratio in response spectra method is also discussed. Four methods have been used for classification. In the first method, the predominant period is estimated by the H/V spectral ratio curve of FAS of the data. In the second method, the predominant period is estimated by the H/V spectral ratio curve of 5% damped response spectra of the data. The third method is based on an empirical approach, and a site classification index uses cumulative distribution function (Zhao et al., 2006). The fourth method is based on Spearman’s rank correlation coefficient and uses the site index for classification (Ghasemi et al., 2009). The present work of classification uses only strong ground motion of earthquakes recorded by the instrumented sensors of EEW System for Uttarakhand. Two classification schemes, namely the NEHRP scheme (Building Seismic Safety Council, 2003) in the first method and Japan Road Association (1980) scheme in the second method have been used to classify sites. Conclusively, the site class, which has the highest frequency of occurrence amongst the four methods has been selected the final site classification of that station. In the present study, the aggregate H/V spectral ratio curve of each site class is compared with the standard H/V spectral ratio curves given by Zhao et al. (2006).A server of the EEW System for Uttarakhand continuously processes the streamed ground motion data coming from the distant installed sensors. The complete records of the earthquake recorded by the triggered sensors become available in real-time. A computer program using python programming language has been developed to create PGA and intensity maps through the EEW System for Uttarakhand. The reason to choose python is its open-source nature and availability of required numerous in-built modules and packages. The computer program is written to read ground motion data to implement GMPE and MMI scaling models. This program reads ground motion data (acceleration) from the server and performs pre-processing operations (e.g. baseline correction, filtering etc.). Before creating PGA and intensity maps, code adaptively creates new GMPE using the records of the past earthquakes of the region augmented with the new earthquake records, and when RMSE of new GMPE is found to be less than the existing regional GMPE, then it uses modified GMPE; otherwise, it uses existing GMPE of the region. An opensource image and graph drawing software, namely Generic Mapping Tools (GMT) is used to draw the spatial variation of varying seismic intensity (Wessel and Smith, 1991) and variation of PGA in the region. The GMT tool plots high-quality images, graphs, and maps. A colour palette is used to show geospatial variation in PGA and intensities on the maps. For a regional EEW System, fast determination of the magnitude of the impending earthquake on the basis of the first few seconds of the P-wave record is more complex than the determination of EEW parameters. The scaling relationships between magnitude and EEW parameters have been developed and tested for many regions of the world. In the present study, Pd based magnitude does a good agreement with catalogue magnitude. As data is very limited and the highest magnitude in the catalogue used in this study is 5.5; therefore, the developed model would give a good approximation to the magnitude range from 3 to 5. The derived scaling relationship (equation 2) with RMSE 0.252 is a good fit for the Uttarakhand region for small earthquakes. Magnitude is estimated by inverting the model given in equation (2). 𝑙𝑜𝑔 (𝑃𝑑) = −2.6826 + 0.52258 × 𝑀𝑤 − 1.2011 × 𝑙𝑜𝑔 (𝑋) (2) This thesis contributes the research work useful to the professionals of diverse fields. The developed maps would be very useful to disaster management authorities, land-use planners, and professionals. Classification of the sites and developed GMPE model would help in seismic hazard analysis of the region. Pd based magnitude estimation would help professionals who are involved in the development and maintenance of the EEW Systems.