Validating predictability of a sea level rise at a tide gauge station

Abstract

ABSTRACT Predicting coastal sea level rise during the 21st century is essential for risk assessments. It is, therefore, a central theme for numerous studies in predicting sea level rise at coastal regions using historical tide gauge measurements. However, these investigations rarely investigated if such an extrapolation of sea level rise at a coastal region is viable without evidence. This study proposes a broadly framed three-tier statistical validation approach for kinematic models that use tide gauge measurements in sea level predictions. It consists of a concurrent (contemporaneous) validation, retrospective (hindcast) validation and prospective (forecast) validation. The modus operandi for each stage is exemplified using sea level measurements at Brest tide gauge station since 1800. The proposed three-tiered validation process provide statistical information about the predictions to be made at this station.