home / skills / benchflow-ai / skillsbench / seismic-picker-selection

seismic-picker-selection skill

safe

/tasks/seismic-phase-picking/environment/skills/seismic-picker-selection

This skill helps you choose seismic event detection and phase picking methods by balancing generalizability, sensitivity, and speed for reliable catalogs.

npx playbooks add skill benchflow-ai/skillsbench --skill seismic-picker-selection

Review the files below or copy the command above to add this skill to your agents.

Files (1)

SKILL.md

4.0 KB

---
name: seismic-picker-selection
description: This is a summary the advantages and disadvantages of earthquake event detection and phase picking methods, shared by leading seismology researchers at the 2025 Earthquake Catalog Workshop. Use it when you have a seismic phase picking task at hand.
---


# Seismic Event Detection & Phase Picking Method Selection Guide

## Overview: Method Tradeoffs

When choosing an event detection and phase picking method, consider these key tradeoffs:

| Method | Generalizability | Sensitivity | Speed, Ease-of-Use | False Positives |
|--------|------------------|-------------|-------------------|-----------------|
| STA/LTA | High | Low | Fast, Easy | Many |
| Manual | High | High | Slow, Difficult | Few |
| Deep Learning | High | High | Fast, Easy | Medium |
| Template Matching | Low | High | Slow, Difficult | Few |

- **Generalizability**: Ability to find arbitrary earthquake signals
- **Sensitivity**: Ability to find small earthquakes


**Key insight:** Each method has strengths and weaknesses. Purpose and resources should guide your choice.

## STA/LTA (Short-Term Average / Long-Term Average)

### Advantages
- Runs very fast: Automatically operates in real-time
- Easy to understand & implement: Can optimize for different window lengths and ratios
- No prior knowledge needed: Does not require information about earthquake sources or waveforms
- Amplitude-based detector: Reliably detects large earthquake signals

### Limitations
- High rate of false detections during active sequences
- Automatic picks not as precise
- Requires manual review and refinement of picks for a quality catalog

---

## Template Matching

### Advantages
- Optimally sensitive detector (more sensitive than deep-learning): Can find smallest earthquakes buried in noise, if similar enough to template waveform
- Excellent for improving temporal resolution of earthquake sequences
- False detections are not as concerning when using high detection threshold

### Limitations
- Requires prior knowledge about earthquake sources: Need **template waveforms** with good picks from a preexisting catalog
- Does not improve spatial resolution: Unknown earthquake sources that are not similar enough to templates cannot be found
- Setup effort required: Must extract template waveforms and configure processing
- Computationally intensive

---

## Deep Learning Pickers

### When to Use
- Adds most value when existing seismic networks are sparse or nonexistent
- Automatically and rapidly create more complete catalog during active sequences
- Requires continuous seismic data
- Best on broadband stations, but also produces usable picks on accelerometers, nodals, and Raspberry Shakes
- **Use case**: Temporary deployment of broadband or nodal stations where you want an automatically generated local earthquake catalog

### Advantages
- No prior knowledge needed about earthquake sources or waveforms
- Finds lots of small local earthquakes (lower magnitude of completeness, Mc) with fewer false detections than STA/LTA
- Relatively easy to set up and run: Reasonable runtime with parallel processing. SeisBench provides easy-to-use model APIs and pretrained models.

### Limitations
- Out-of-distribution data issues: For datasets not represented in training data, expect larger automated pick errors (0.1-0.5 s) and missed picks
- Cannot pick phases completely buried in noise - Not quite as sensitive as template-matching
- Sometimes misses picks from larger earthquakes that are obvious to humans, for unexplained
reason

## References
- This skill is a derivative of Beauce, Eric and Tepp, Gabrielle and Yoon, Clara and Yu, Ellen and Zhu, Weiqiang. _Building a High Resolution Earthquake Catalog from Raw Waveforms: A Step-by-Step Guide_ Seismological Society of America (SSA) Annual Meeting, 2025. https://ai4eps.github.io/Earthquake_Catalog_Workshop/
- Allen (1978) - STA/LTA method
- Perol et al. (2018) - Deep learning for seismic detection
- Huang & Beroza (2015) - Template matching methods
- Yoon and Shelly (2024), TSR - Deep learning vs template matching comparison

Overview

This skill summarizes advantages and disadvantages of common seismic event detection and phase picking methods to help choose the right approach for a given task. It condenses guidance from leading seismology researchers at the 2025 Earthquake Catalog Workshop into practical tradeoffs and use recommendations. Use it to match detection goals, available data, and compute resources to the most appropriate method.

How this skill works

The skill compares four broad approaches—STA/LTA, manual picking, deep-learning pickers, and template matching—across metrics like generalizability, sensitivity, speed, and false-positive rates. For each method it lists concrete strengths, limitations, and typical deployment contexts so you can weigh outcomes (completeness, precision, setup cost) against constraints (data availability, compute, and expertise). It emphasizes when methods complement each other and when one is clearly preferable.

When to use it

Quick, real-time monitoring where ease and speed matter: STA/LTA.
High-precision, low false-positive catalogs when you have templates: Template matching.
Automated, rapid cataloging from continuous data in sparse networks or temporary deployments: Deep learning pickers.
Human-in-the-loop quality control or final catalog validation: Manual picking and review.
Hybrid pipelines combining deep learning for bulk picks and template matching for the smallest events.

Best practices

Match method to objective: prioritize sensitivity for completeness, precision for catalog quality, and speed for real-time alerts.
Run a small-scale validation: test pick accuracy and false-positive rate on representative data before full deployment.
Use deep-learning models with pretrained weights and augment with local retraining when possible to reduce out-of-distribution errors.
Combine methods: use deep-learning or STA/LTA for initial triggers, then apply template matching or manual review for critical events.
Maintain compute and review workflows: template matching requires more compute and manual curation; STA/LTA requires downstream review to reduce false positives.

Example use cases

Rapid earthquake catalog generation after an induced seismicity sequence using deep-learning pickers.
High-resolution aftershock sequence analysis in a well-instrumented region using template matching to recover small events.
Operational real-time alerting with STA/LTA at regional networks where speed and simplicity are priorities.
Targeted manual picking to finalize origin times and phases for a published catalog after automated detection.

FAQ

Which method finds the smallest earthquakes?

Template matching is the most sensitive for finding very small, repeating events similar to templates, but it requires representative template waveforms.

When should I prefer deep learning over STA/LTA?

Choose deep learning when you have continuous data and want more complete catalogs with fewer false positives than STA/LTA, especially in sparse or temporary deployments.

How much manual review is necessary?

All automated methods benefit from manual review for high-quality catalogs; STA/LTA typically needs the most manual refinement, while template matching and deep learning reduce but do not eliminate review.