Abstract
Ocean bottom seismometer records seismograms using a clock in a recorder. The recorder has a very precise clock although a few seconds of time shift occurs during a one-year seafloor observation. Usually, we measure time difference between a recorder's clock and a GPS clock before and after an observation and calibrate recordings. Recently, we could not apply this method because of some trouble in a recorder. In this report, we show that we can estimate a time drift of recorder's clock using observed seismograms and calibrate recordings. One easy method to estimate a time drift is to measure a temporal change of differential travel time residuals of first P-wave between two stations. Other method is to measure a temporal change of ambient noise cross correlation function of two stations. In our recent seafloor observation by “Normal Oceanic Mantle project”, we had two troubled recorders of broadband ocean bottom seismometers (NM02, NM20). We have applied these methods and estimated time drifts of these two recorder's clocks as－10.49 ms/day, and －1.9 ms/day, respectively. We have also found that clock stability of these recorder's clocks is very well during seafloor observation from recordings by temporal change of cross correlation function and that a precision of the recorder's clock is well less than 100ms.
