Peak ground acceleration prediction for on-site earthquake early warning with deep learning

Rapid and accurate prediction of peak ground acceleration (PGA) is an important basis for determining seismic damage through on-site earthquake early warning (EEW). The current on-site EEW uses the feature parameters of the first arrival P-wave to predict PGA, but the selection of these feature parameters is limited by human experience, which limits the accuracy and timeliness of predicting peak ground acceleration (PGA). Therefore, an end-to-end deep learning model is proposed for predicting PGA (DLPGA) based on convolutional neural networks (CNNs). In DLPGA, the vertical initial arrival 3-6 s seismic wave from a single station is used as input, and PGA is used as output. Features are automatically extracted through a multilayer CNN to achieve rapid PGA prediction. The DLPGA is trained, verified, and tested using Japanese seismic records. It is shown that compared to the widely used peak displacement (Pd) method, the correlation coefficient of DLPGA for predicting PGA has increased by 12-23%, the standard deviation of error has decreased by 22-25%, and the error mean has decreased by 6.92-19.66% with the initial 3-6 s seismic waves. In particular, the accuracy of DLPGA for predicting PGA with the initial 3 s seismic wave is better than that of Pd for predicting PGA with the initial 6 s seismic wave. In addition, using the generalization test of Chilean seismic records, it is found that DLPGA has better generalization ability than Pd, and the accuracy of distinguishing ground motion destructiveness is improved by 35-150%. These results confirm that DLPGA has significant accuracy and timeliness advantages over artificially defined feature parameters in predicting PGA, which can greatly improve the effect of on-site EEW in judging the destructiveness of ground motion.

Probabilistic seismic hazard assessment of southern part of Ghana

This paper presents a seismic hazard map for the southern part of Ghana prepared using the probabilistic approach, and seismic hazard assessment results for six cities. The seismic hazard map was prepared for 10% probability of exceedance for peak ground acceleration in 50 years. The input parameters used for the computations of hazard were obtained using data from a catalogue that was compiled and homogenised to moment magnitude (Mw). The catalogue covered a period of over a century (1615-2009). The hazard assessment is based on the Poisson model for earthquake occurrence, and hence, dependent events were identified and removed from the catalogue. The following attenuation relations were adopted and used in this study-Allen (for south and eastern Australia), Silva et al. (for Central and eastern North America), Campbell and Bozorgnia (for worldwide active-shallow-crust regions) and Chiou and Youngs (for worldwide active-shallow-crust regions). Logic-tree formalism was used to account for possible uncertainties associated with the attenuation relationships. OpenQuake software package was used for the hazard calculation. The highest level of seismic hazard is found in the Accra and Tema seismic zones, with estimated peak ground acceleration close to 0.2 g. The level of the seismic hazard in the southern part of Ghana diminishes with distance away from the Accra/Tema region to a value of 0.05 g at a distance of about 140 km.

Estimating the upper limit of prehistoric peak ground acceleration using an in situ, intact and vulnerable stalagmite from Plavecká priepast cave (Detrekői-zsomboly), Little Carpathians, Slovakia-first results

Earthquakes hit urban centres in Europe infrequently, but occasionally with disastrous effects. Obtaining an unbiased view of seismic hazard (and risk) is therefore very important. In principle, the best way to test probabilistic seismic hazard assessments (PSHAs) is to compare them with observations that are entirely independent of the procedure used to produce PSHA models. Arguably, the most valuable information in this context should be information on long-term hazard, namely maximum intensities (or magnitudes) occurring over time intervals that are at least as long as a seismic cycle. The new observations can provide information of maximum intensity (or magnitude) for long timescale as an input data for PSHA studies as well. Long-term information can be gained from intact stalagmites in natural caves. These formations survived all earthquakes that have occurred over thousands of years, depending on the age of the stalagmite. Their 'survival' requires that the horizontal ground acceleration (HGA) has never exceeded a certain critical value within that time period. Here, we present such a stalagmite-based case study from the Little Carpathians of Slovakia. A specially shaped, intact and vulnerable stalagmite in the Plavecká priepast cave was examined in 2013. This stalagmite is suitable for estimating the upper limit of horizontal peak ground acceleration generated by prehistoric earthquakes. The critical HGA values as a function of time going back into the past determined from the stalagmite that we investigated are presented. For example, at the time of Jókő event (1906), the critical HGA value cannot have been higher than 1 and 1.3 m/s2 at the time of the assumed Carnuntum event (∼340 AD), and 3000 years ago, it must have been lower than 1.7 m/s2. We claimed that the effect of Jókő earthquake (1906) on the location of the Plavecká priepast cave is consistent with the critical HGA value provided by the stalagmite we investigated. The approach used in this study yields significant new constraints on the seismic hazard, as tectonic structures close to Plavecká priepast cave did not generate strong earthquakes in the last few thousand years. The results of this study are highly relevant given that the two capitals, Vienna and Bratislava, are located within 40 and 70 km of the cave, respectively.