Applicability of ordinary Kriging modeling techniques for filling satellite data gaps in support of coastal management

Effective coastal zone management relies on accurate and sound data of environmental parameters. In many cases, however, lack of pertinent environmental variable datasets hinders the development of sustainable policies and practices. Marine remote sensing has been used extensively to provide such datasets, with limited success due to different spatial resolutions, non-overlapping grids, and different spatial extent, since, in most cases, they are derived from miscellaneous sources. Several methods have been proposed aiming to overcome this problem and enable remote-sensing techniques to attain the highest possible applicability in coastal ecosystem management studies. Most of these techniques use interpolation modeling techniques for satellite data gap filling near the shoreline. This paper presents the use of the Ordinary Kriging (OK) methodology, as an interpolator of satellite data, from miscellaneous and variant environmental datasets of the territorial waters of the Heraklion Prefecture, on the island of Crete. The applicability of OK in satellite data gap filling was examined by the use of cross-validation, taking into account resolution levels for optimum estimations. In addition, a number of diagnostic statistics were employed to evaluate the performance of the interpolation models used. The results indicate that the proposed methodology is valid and independent of remote-sensing data characteristics, thus proving that OK can be used to homogenize effectively and integrate fully diverse satellite datasets. This study concludes that the methodology described in predicting estimates of missing values of satellite data with statistical certainty can be used as an effective tool in coastal management.