Five creepmeters operate on the Hayward Fault. Typical creep rates are 5 mm/year, but because creep can be distributed over a wide zone, and the creepmeters are 30 m long and cross the fault obliquely (at 30 degrees), they do not always record the entire creep signal. The highest observed creep rate is recorded in Fremont (7.8-8.5 mm/year), and the lowest observed creep rate in Oakland Zoo. Processing details are included in the web page describing the Temescal Park creepmeter. Least-squares fits to 1993-1997 data are shown in the table below.
| site | creep rate, mm/yr | annual thermoelastic,mm | noise RMS mm |
| Point Pinole | 5.76+-0.05 | +-0.9 | 0.2 |
| Temescal Park | 4.06+-0.1 (1995-1998) | +-0.15 | 0.18 |
| Palisades, Hayward | 3.23+-0.05 | +-0.25 | 0.23 |
| Osgood, Fremont | 8.5+-1 | +-0.9 | 0.31 |
Hayward Creep Processes
The Hayward fault has slipped at 5 mm/year over the last several hundred thousand years. The fault between 5 and 12 km depth is believed to slip entirely in earthquakes, but the surface fault and deepest part of the fault also slips by a process of aseismic creep. In the past several decades, creep rates of 5 mm/year have severed buried utility pipes and cracked roads for 70 km between Fremont and Point Pinole. Creep has occurred at least since the last major earthquake on the fault in 1868. The 1906 earthquake on the San Andreas fault on the opposite side of the San Francisco Bay destroyed the Gallegos Winery (constructed in 1886 across the Hayward fault in Fremont), and its foundations have been offset more than 90 cm. This foundation was constructed after the last Hayward earthquake, but several other features offset in the 1896 earthquake have also continued to be offset by creep, indicating that creep incompletely releases strain energy applied to the fault.
The figure left illustrates how creep is confined to the top and bottom of an active fault, and how an oblique creepmeter measures surface creep. Earthquake ruptures occur at intervening depths, and presumably leak into the creeping zones with lesser slip. The displacement far from a fault can be measured with GPS methods to an accuracy of approximately 1 mm, but the displacement on the fault can be measured 100 times more precisel;y with a creepmeter. A creepmeter consists of a rigid bar attached to a stable monument on one side of the fault. Movement of the free end of the bar relative to a second monument on the other side of the fault is measured with a displacement transducer to an accuracy of approximately 10 microns (1/1000 mm), every 10 minutes.
Data from the creepmeters reveal that except at Fremont, the creep signal is approximately linear. Deviations from linearity north of the the city of Hayward are less than 0.2 mm RMS.