Scholarlink

<p>Landslides are a major natural hazard in mountainous regions, often resulting in thousands of deaths and billions of dollars in property damage. Climate change is increasing the frequency and severity of these events. Seismic network monitoring has made it possible to detect landslides in real time. However, distinguishing seismic signals caused by landslides from those generated by earthquakes and background noise remains a key challenge. This study investigates the reliability of power spectral density (PSD) trends in seismic waveforms for identifying landslide-generated signals. By analysing seismic data from multiple stations within a 150-km radius of the event, we find a consistent PSD decay pattern across different landslides, regardless of their size, duration, or distance from the stations. We use the slope of the seismic waveform PSD in the frequency band 0.01–5&#xa0;Hz and skewness of the spectral power distribution as scalable entities for landslide detection<b>.</b> This confirms that landslides have unique spectral features, enabling them to be distinguished from other seismic sources. Our analysis suggests that the landslides show PSD slope values between −&#xa0;3 and −&#xa0;9. We have also noticed steeper slopes that match the slope of seismic waveforms of background noise when the station distances are above 200&#xa0;km. Although this limits landslide detection using distance stations, this can be ruled out when using the local networks for landslide monitoring. The study demonstrates that PSD analysis of seismic waveforms offers a stable and innovative method for real-time landslide detection in continuous seismic data. Utilizing these spectral signatures could greatly enhance landslide monitoring and early warning systems in high-risk areas.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Persistent PSD trends: a tool for seismic landslide detection

  • 期刊论文

  • 发表日期：

    2025年10月16日

查看全文
  • Mahesh Yezarla,
  • Rajesh Rekapalli,
  • Sandeep Gupta

摘要

Landslides are a major natural hazard in mountainous regions, often resulting in thousands of deaths and billions of dollars in property damage. Climate change is increasing the frequency and severity of these events. Seismic network monitoring has made it possible to detect landslides in real time. However, distinguishing seismic signals caused by landslides from those generated by earthquakes and background noise remains a key challenge. This study investigates the reliability of power spectral density (PSD) trends in seismic waveforms for identifying landslide-generated signals. By analysing seismic data from multiple stations within a 150-km radius of the event, we find a consistent PSD decay pattern across different landslides, regardless of their size, duration, or distance from the stations. We use the slope of the seismic waveform PSD in the frequency band 0.01–5 Hz and skewness of the spectral power distribution as scalable entities for landslide detection. This confirms that landslides have unique spectral features, enabling them to be distinguished from other seismic sources. Our analysis suggests that the landslides show PSD slope values between − 3 and − 9. We have also noticed steeper slopes that match the slope of seismic waveforms of background noise when the station distances are above 200 km. Although this limits landslide detection using distance stations, this can be ruled out when using the local networks for landslide monitoring. The study demonstrates that PSD analysis of seismic waveforms offers a stable and innovative method for real-time landslide detection in continuous seismic data. Utilizing these spectral signatures could greatly enhance landslide monitoring and early warning systems in high-risk areas.