Scholarly record

SIGNALS AND NOISES OF REGRESSIONAL AND NON-REGRESSIONAL SHORELINE CHANGE RATE METHODOLOGIES

E. Doukakis

Abstract

Shoreline change rate is the most critical feature of the coastal area because it reveals future harmful impacts driven by long-term climate changes and episodic events. Since the coastal zone is one of a nation’s greatest environmental and economic assets, the best methodology, as far as accuracy is concerned, estimating the rate of change of a shoreline is an open question. The accuracy both of the geoinformation used and the computed change rate is of crucial importance. The purpose of the present paper is to use historical shorelines extracted from three different Greek coastal regions and estimate their shoreline change rate.(signal). The most common and traditional statistical methods are used, i.e. end of point rate, average of rates, linear regression and Jackknife. The first two are the non-regressional and the last two the regressional methodologies. The error in position (EIP) of each methodology indicates its accuracy (noise). Since the true shoreline change rate of each coastal region is not known, it is impossible to assess which method is the most accurate unless the EIP is compared to the its signal. It is proved that the Jackknife method is the most accurate to use for shoreline evolution estimations. To evaluate further the derived conclusion, we introduce computer-derived (“synthetic”) shorelines and compute the signal-to-noise ratio for this ideal and errorless case. The results strengthen the main conclusion regarding the Jackknife ability to compute shoreline change rate with the best accuracy.

Publication details

Back to publication list