Zhang, J.-Y.; Zhang, J.; Ma, Y.; Chen, A.-N.; Cheng, J., and Wan, J.-X., 2019. Satellite-derived bathymetry model in the Arctic waters based on support vector regression. In: Jung, H.-S.; Lee, S.; Ryu, J.-H., and Cui, T. (eds.), Advances in Remote Sensing and Geoscience Information Systems of Coastal Environments. Journal of Coastal Research, Special Issue No. 90, pp. 294-301. Coconut Creek (Florida), ISSN 0749-0208.

Bathymetric data are essential for navigation and marine engineering. Satellite-derived bathymetry (SDB) is an effective supplement to the ship-based echo sounder, especially for bathymetric measurement in the area that is difficult to reach. Accurate SDB in the Arctic waters is more complex by weak light, cloud, mist, floating ice and so on than in middle or lower latitude clean waters. This paper performs SDB based on Support vector regression (SVR) from multi-spectral imagery of Sentinel-2 in shallow water (0-20 m) of Pomorskiy Proliv of Northern Russia, and corrects the estimated water depth with the interpolation of residuals. The results indicate that the highest accuracy of SDB based on SVR is achieved by employing radial basis function on the full feature inputs, with the mean absolute error (MAE) of 2.5 m and the mean relative error of 44.9 %. After the residual interpolation correction, the overall MAE reduces to 2.0 m, and the root mean square errors decrease by at least 0.3 m among different depth ranges. The approach provides technical support for obtaining a wide range and relatively high spatial resolution of bathymetric data in the Arctic region.