Using robust correlation matching to estimate sand-wave migration in Monterey Submarine Canyon, California

Abstract

Sandwave migration rate is of critical importance for the hydrodynamic research as well as the coastal engineering application. To acquire this information, the correlation matching algorithm is applied in this paper. By searching for similar seafloor areas between multi-temporal digital terrain models, the migration rate can be estimated. To account for the impact of the morphological distortion as well as outliers in the survey data, the correlation matching is adjusted based on the robust estimation theory. By executing the correlation matching based on the least median square criterion, a robust initial estimate is obtained. Afterwards, the potential outliers are detected based on the initial matching result, and the corresponding weights are adjusted to zero. Finally, correlation matching is executed with the updated weights. With the above procedure, the impact of outliers can be accounted for. The developed method is applied on the bathymetry time-series data collected in Monterey Canyon, California. Comparison shows that the presented algorithm significantly improves the credibility of the estimation. The generated migration vector field not only reveals the main tendency of the migration rate in the region, but also highlights the unique local migration patterns, which is helpful for understanding the mechanism of the sandwave migration.