Evaluating shoreline identification using optical satellite images

Abstract

Abstract A technique to extract the shoreline location from optical satellite images has been developed and evaluated for the case study site of Progreso, Yucatan, Mexico. A novel method to extract a satellite-derived shoreline (SDS) was developed ensuring the maximum contrast between sea and land. The area under investigation is an 8 km length of shoreline that faces north into the Gulf of Mexico. The SDS was validated using quasi-simultaneous in situ shoreline measurements, both adjusted to equal water levels. In situ shoreline measurements recorded the instantaneous shoreward extent of the wave run-up when walking along the beach. The validation of SDS revealed that the SDS is located consistently seawards of the in situ shoreline, explained by: a) the water depth that optical satellite image requires to identify a pixel either as sea or land, and b) the shoreward extent of the wave run-up. The overall distance between SDS and in situ shoreline is 5.6 m on average with a standard deviation of 1.37 m (in the horizontal) over 8 km of shoreline. Confidence bounds considering the shoreward extent of the wave run-up, inter-tidal beach slope variation and tidal uncertainty were computed to assess the accuracy of the SDS. The SDS has been shown to be capable of detecting shoreline changes of less than 10 m and abrupt changes due to storms. The success of our method suggests that it should be applicable to other locations, after adapting the confidence bounds to the beach conditions.