NEW MADRID SEISMIC ZONE SURFACE DEFORMATION: AN EXPERIMENTAL INVESTIGATION OF GPS MONUMENT STABILITY WITH IMPLICATIONS FOR REGIONAL SEISMIC HAZARD ASSESSMENT

The 1811-12 NMSZ seismic events and their relationship, if any, to the well-recognized earthquake cycle associated with plate boundaries remains enigmatic. Recently, Smalley et al. (2005) reported that strain rates based on GPS observations from the GAMA network are well above background and may be as high as in plate boundary zones. Calais et al. (2005) argued, however, that the GAMA GPS data remain too uncertain and thus provide only an upper bound on strain accumulation in the NMSZ. The implication is that the current strain rate may yet be nearly zero in the NMSZ.

Several years may be required to obtain accurate site velocity estimates from GPS time series in areas of small strain such as the NMSZ; accordingly, a variety of errors with different timescales, whose relative contributions change with time, potentially corrupt the data. In particular, GPS errors related to monument instability may mask any small tectonic signal. While spatial averaging may reduce the size of both white and random-walk noise, it does not mitigate their relative effects on strain accumulation models. Long-term correlations thus can have a large effect on estimating deformation rates.

Monument motion is likely significant in the NMSZ, an area dominated by the unconsolidated sediments of the Lower Mississippi Valley. We have installed two new continuous GPS sites collocated at GAMA sites in Portageville, MO and Troy, TN to measure experimentally monument stability and noise. Our new sites use SCIGN-type drilled, deep-braced monuments, precision levels, and radomes, but in every other way are identical to GAMA sites ~10 m away. 30 s data have now been acquired for over 1.5 years. Analysis of the two monument types at over the same period of operation demonstrates that the new SCIGN monument yields more precise GPS component velocity estimates than the existing GAMA I-beam monument. In addition, baseline evolution between two monuments of similar type and spanning the Realfoot thrust, shows that the SCIGN monuments yield higher strain with improved precision over the GAMA I-beam monuments. This suggests that improved GPS infrastructure and additional time may soon be able to distinguish between elastic strain accumulation versus post-seismic relaxation models for current NMSZ deformation.