NoisePy: Ambient Field Seismology in Python 


Marine Denolle, Assistant Professor, UW Earth and Space Sciences 

Yiyu Ni, PhD Student

Kuan-Fu Feng, Postdoctoral Scholar

SSEC Engineers

Carlos Garcia Jurado Suarez, Senior Principal Software Engineer

NoisePy is a software package used to image the Earth’s subsurface and monitor its changes through time. It performs large-scale cross-correlation of time series data recorded by networks of geophysical sensors. The collection of functions is designed to compute these cross-correlations between pairs of channels and sensors at multiple scales in space and time.  

The Earth’s ambient field contains tremendous information about the Earth’s structure. Earth changes occur at a vast range of temporal and spatial scales. Seismology can record these thanks to the long term deployment of seismic networks (~ 40 years) and the rapid expansion of sensing technologies. Seismological data volumes are exponentially growing, along with the number of sensors. The growth of sensors and data volume is a huge challenge for the seismological community: it needs user-friendly, scalable software that can run on modern cyberinfrastructure 

NoisePy hopes to aid in discerning between anthropogenic activities, ocean noise, and various other sources of seismic waves. These insights enable researchers to identify changes in subsurface features such as earthquake damage, volcanic unrest, and groundwater level changes.  

At present, the software exists across two Python repositories with differing implementations. Given the large amount of data present in such analyses, a robust deployment infrastructure is essential. To bring these tools to the greater seismology community, SSEC is collaborating with the Denolle Quake Lab at UW (University of Washington) to build a Python package which maintains performance while scaling capacity within an open-source framework. The package aims to calculate cross correlations on data at high volumes, handling eventual intake of up to 2TB a day. Upon completion, NoisePy and SSEC hope to enable one hundred scientists to process up to 100 TBs of seismic data. 

The renewed software engineering effort is already beginning to engage the seismology community. In May 2023, PI Marine Denolle deployed a beta version of the package via a hands-on workshop where attendees were able to successfully run the updated package.