Resolving millimeter-level storm surge loading deformations using multi-GNSS data over the subdaily timescales

<p>Storm surges often strike the southern coast of North Sea in Europe during the winter season. The largest event on December 5-6, 2013 since 1953 pushed the water level to rise by up to 4 m within a few hours, which caused a transient but appreciable loading on the crust under the southern North Sea. Consequently, GNSS stations around this oceanic area experienced considerable displacements, which were up to 30 mm in the vertical while 5 mm in the horizontal component as predicted by the Proudman Oceanographic Laboratory Storm Surge Model (POLSSM). We processed the GPS/GLONASS data at 18 coastal stations from Nov. 1 until Dec. 31, 2013. We computed station displacements using precise point positioning ambiguity resolution every three hours to track subdaily loading deformations, and compared them with POLSSM predictions. The second- and third-order delays were mitigated using IGS global ionosphere map derived corrections; orbital repeat time (ORT) filtering, which aimed at reducing multipath effects, were enabled for both GPS and GLONASS on the observation level. We found that GNSS derived displacements presented high correlations of up to 0.7 with POLSSM predictions in the vertical direction over the 61 days; higher-order ionosphere corrections reduced the north RMS between GNSS solutions and POLSSM predictions by 0.2-0.3 mm, whereas the ORT filtering decreased the RMS by more than 10% for all three components. Introducing GLONASS data to GPS-only solutions further reduced the RMS to 5.9, 2.2 and 2.7 mm in the vertical, east and north components, suggesting a 6-12% improvement. Despite this millimeter-level agreement, the peak-to-peak vertical displacement of about 10 mm over the 6–24-h timescales for the largest surge event on December 5 was only marked marginally in the subdaily wavelet power spectra. Thanks to the spatial coherence among the 18 stations, the principal component analysis could enhance dramatically the resolution capability of subdaily GNSS in discriminating the subdaily vertical loading signals of 5-10 mm amplitude over the 6–24-h wavelet timescales. We demonstrate that multi-GNSS data have the potential to improve significantly the detection of subdaily geophysical signals dwelling on the periods of tens of minutes to hours.</p>