An approach for prediction of shoreline with spatial uncertainty mapping (SLiP-SUM).

Graphical abstract Highlights • A snake algorithm is an effective technique for delineating shorelines. • SLiP-SUM considers pixel sizes and estimation errors for calculation of uncertainty. • SLiP-SUM provides spatial uncertainty mapping along future shorelines. Abstract This study puts forward a semi-automatic shoreline detection and future prediction with spatial uncertainty algorithm called the SLiP-SUM (Shore Line Prediction with Spatial Uncertainty Mapping), which has five main steps: (1) preprocessing of data sets (i.e. aerial photos and/or satellite images), (2) extraction or delineation of the existing shorelines with snake algorithm, 3) prediction of the future shorelines using linear regression, (4) preparation of the spatial uncertainty mapping using cokriging, and (5) producing possible shoreline with spatial uncertainty. The proposed approach was tested on the coast of Kumluca, a dynamic coastal area in Turkey, and future shoreline predictions were made for 2020, 2025, 2030 and 2035 based on remotely sensed data (aerial photos and WorldView-2 images) acquired between 1971 and 2014. A cokriging interpolation technique was adopted that takes into account both estimation errors and the spatial resolution of the source data sets for the mapping of spatial uncertainty. The results indicate that the maximum and minimum annual changes of shoreline were -2.29 m/yr (transgression) in the west-central part of the study area, and + 0.36 m/yr (regression) in the east part. The general trend of the future shoreline regime was transgressional, and the mean uncertainty values along the predicted shorelines for 2020, 2025, 2030 and 2035 were 6.78 m, 2.02 m, 6.76 m and 7.06 m, respectively. [ABSTRACT FROM AUTHOR]