Contaminated High-Frequency Data in Borehole Geophones from Induced Seismicity in the UK

Abstract:

Geophones deployed in monitoring wells close to induced seismicity are designed specifically for higher frequencies (>200 Hz) to properly capture microseismicity. Downhole arrays can produce catalogues of thousands of micro-earthquakes, covering events down to and even below local magnitudes of -2. The datasets offer new opportunities for understanding the mechanisms and characteristics of induced seismicity, for optimising the associated geo-energy activities and for mitigating seismic risks. However, signals retrieved from these arrays can sometimes be contaminated by high-frequency artefacts caused by geophone clamping issues, spurious frequencies, clipping, tube waves, etc., making the high-frequency content less reliable. Whilst these contaminations do not necessarily hinder earthquake detection and location, they can distort the observed earthquake spectrum in the higher frequencies and lead to miscalculations in source parameters such as magnitude, corner frequency, and rupture radius, as well as propagation parameters such as Q.

We examine seismic borehole data contaminated by high-frequency artefacts from the hydraulic fracturing site Preston New Road 1z in the UK. 15 Hz geophones with Nyquist frequencies of 1000 Hz were deployed by the operators at depths between 1500-2000 meters, producing a catalogue of 36,000+ events spanning local magnitudes between -4 and 1. We identified obvious high-frequency artefacts between 150 to 440 Hz, distorting the microearthquakes’ high-frequency spectra. With the goal of estimating source parameters using the spectral ratio method, a procedure which relies heavily on the shape of the spectrum, we first investigate to what extent the earthquake spectra can be recovered and how to avoid the noise contamination. We use seismic interferometry to examine if there are any clamping issues and tube waves present in the bore hole array. Furthermore, we analyse the particle motion polarization using hodograms to assess how much of the P- and S-waves are affected by the noise. By understanding how and to what extent the high-frequency artefacts affect the earthquake spectra, we seek to work around the issues and still recover earthquake information to some degree.