Strainmeters installed in deep (> 150m) boreholes in the Parkfield region show pronounced seasonal fluctuations, including a strong annual component. We suggest that a significant part of the annual variations in these strain records is caused by strain in the elastic part of the earth's crust induced by variations in atmospheric temperature at the surface. We test this hypothesis by computing the thermoelastic strain from the atmospheric temperature record observed at the weather station at Coalinga, and compare it to the strain recorded by dilatometers installed in boreholes near Parkfield. We use a thermoelastic strain model which computes the thermoelastic strain for an elastic half space with a decoupled unconsolidated (soil or gravel) upper layer. The source of the strain in the elastic solid is a spatially varying temperature field that travels through the unconsolidated surface layer, and whose local horizontal length scale is related to topography and lateral material heterogeneities. The strain in the underlying half-space, generated by temperature variations at the base of the decoupled surface layer, is calculated with an analytical solution for thermoelastic strain in a homogeneous half-space. Our simple model provides a good first-order fit to the annual signals recorded at the borehole-installed strainmeters, including one located as deep as 320 m. The two parameters yielded by the model, the thickness of the unconsolidated upper layer (~1 meter) and the wavelength of the source field (2.75 km) are sufficiently plausible to support the physical validity of the thermoelastic strain model. Removal of the thermoelastic strain signal from strainmeter records would provide better understanding of the remaining noise sources and allow clearer focus on tectonic signals.
AGU Fall Meeting Abstracts
- Pub Date:
- December 2007
- 1211 Non-tectonic deformation;
- 1294 Instruments and techniques;
- 1299 General or miscellaneous (1709)