Seasonal thermoelastic strain and postseismic effects in Parkfield borehole dilatometers

- We analyze seasonal variations of strain in 176-323 m deep Parkfield boreholes.
- The variations are modeled using atmospheric-induced thermoelastic strain.
- The effective source is the temperature field at the bottom of a thin surface layer.
- Only two periods have different persistent trends of observed and model signals.
- These periods follow the only two M>3.5 events in the study space-time domain.

Strainmeter records in three 176-323 m deep boreholes near Parkfield, CA, are dominated by seasonal fluctuations. We show that a significant part of the seasonal data may result from thermoelastic strain induced by atmospheric temperature variations. We test this hypothesis by computing thermoelastic strain in an elastic half-space covered by a thin unconsolidated layer from atmospheric temperature and comparing the results to the borehole strain records. The strain at depth is produced by the temperature field at the bottom of the unconsolidated layer. The model provides reasonable fits to the amplitudes and phases of the seasonal borehole signals. The two key model parameters, thickness of the unconsolidated layer (∼0.3-1.2 m at the used sites) and wavelength of the temperature field (3 km), are sufficiently plausible to support the physical validity of the model. Two instances with persistent deviations between the trends of the predicted thermoelastic strain and observed records may reflect shallow postseismic effects of M⩾4 nearby earthquakes.