Quantification de l'évolution des côtes sableuses et rocheuses par photogrammétrie et lasergrammétrie = Quantifying morphological changes of sandy coasts by Photogrammetry and cliff coasts by lasergrammetry
Résumé
In the framework of coastal research at BRGM (French Geological Survey) two methods are being tested to monitor coastal evolution. The first method involves photogrammetric processing of ground-based optical imagery to measure sandy beaches topography. The system is both developed for hardware and software. The second method makes use of terrestrial laser scanner surveys to assess rock fall hazard of coastal cliffs. Concerning sandy beaches monitoring, this paper addresses the state of the system under development and addresses some practical issues encountered. The system still needs further developments both on the hardware and software side. For the second method, we present here the first results concerning rock fall hazard assessment at the experimental site of Mesnil Val, Normandy. The 700-m-long chalk cliff site, made of “Upper Lewes nodular chalk” and “Seaford chalk”, has been surveyed with terrestrial laser scanning equipment every 3-6 months. Each survey was processed to build 5x5 cm grid-cell Digital Surface Models (DSM) of the entire cliff-face. Here, we present an analysis of the data from the first two campaigns. Differences between DSM qualify as rock fall event, with 1 chance over 1000 uncertainty, if the topographic difference is larger than 3.3 cm, for a patch of 15x15 cm. 2202 such rock fall events occurred in the 101 days time span, between 7 December 2005 and 23 March 2006. Rock fall volumes span 6 orders of magnitude (from 7.4*10-4 à 1.7*102 m3). To our knowledge, this is the most complete rock-fall catalogue published to date. From this catalogue, we built the following experimental power law : F = 69.431*V-0.55, which describes the frequency F (in number of event per year) of seeing an event of given volume V (in m3) or larger occurring in a year. The parameters of the power law are in the ball park of other published rock fall hazard estimates, with the notable difference that our uncertainty bracket is much smaller, and our catalogue does not appear to suffer from usual censoring effect. Further comparison of the existing data set, on a