Accuracy in GPS/Acoustic positioning on a moored buoy moving around far from the optimal position

Abstract:

For detecting the seafloor crustal deformation and Tsunami associated with large earthquakes in real-time, it is necessary to monitor them just above the possible source region. For this purpose, we have been dedicated in developing a real-time continuous observation system using a multi-purpose moored buoy. Sea-trials of the system have been carried out near the Nanakai trough in 2013 and 2014 (Takahashi et al., 2014). We especially focused on the GPS/Acoustic measurement (GPS/A) in the system for horizontal crustal movement. The GPS/A on a moored buoy has a critical drawback compared to the traditional ones, in which the data can be stacked over ranging points fixed at an optimal position. Accuracy in positioning with a single ranging from an arbitrary point is the subject to be improved in this study. Here, we report the positioning results in the buoy system using data in the 2014 sea-trial and demonstrate the improvement of the result. We also address the potential resolving power in the positioning using synthetic tests.

The target GPS/A site consists of six seafloor transponders (PXPs) forming a small inner- and a large outer-triangles. The bottom of the moored cable is anchored nearly the center of the triangles. In the sea-trial, 11 times successive ranging was scheduled once a week, and we plotted positioning results from different buoy position. We confirmed that scatter in positioning using six PXPs simultaneously is ten times smaller than that using individual triangle separately. Next, we modified the definition of the PXP array geometry using data obtained in a campaign observation. Definition of an array geometry is insensitive as far as ranging is made in the same position, however, severely affects the positioning when ranging is made from various positions like the moored buoy. The modified PXP array is slightly smaller and 2m deeper than the original one. We found that the scatter of positioning results in the sea-trial is reduced from 4m to 1.7m with the modified geometry.

Finally we produced a synthetic data with an artificial error in the array geometry and evaluated its effect on the positioning as a function of ranging point. This is interpreted with potential resolving power formulated in Kido (2007). In the presentation, we will show the results of synthetic test for systematic variation of the error condition.