ANALYSIS OF LIDAR DATA FOR FLUVIAL GEOMORPHIC CHANGE DETECTION AT A SMALL MARYLAND STREAM

Abstract

Numerous detailed topographic measurements, which must be periodically repeated, are required to characterize stream bank and channel geometry. Light Detection and Ranging (LiDAR) is becoming more widely used, but its accuracy for change detection in and around small streams is not well quantified. Two LiDAR and one ground-surveyed elevation data sets are compared for a thickly vegetated riparian area in the Maryland Piedmont. Interpolated surfaces (prediction maps) and estimates of their uncertainty (standard error maps) are created from the point data using kriging. The LiDAR 2006 elevations are compared to ground-survey to evaluate accuracy. LiDAR 2002 and 2006 elevations are compared to evaluate the potential for change detection. When the estimated LiDAR system error is included in hypothesis testing, no statistically significant elevation differences are found between 2002 and 2006. Conclusions about geomorphic change based on LiDAR scenes should account for error and uncertainty in the data collection and processing.