High resolution remote sensing for native vegetation assessment and monitoring: an impact assessment approach

Abstract

The last decade has seen major advances in remote sensing technology, particularly in high-resolution satellite imagery and airborne laser scanning (ALS). Fundamental differences in data capture mean that new assessment techniques are required, particularly for vegetation structure and multi-temporal analysis. Here, high-resolution remote sensing tools are developed using a longwall mine subsidence impact assessment framework, based primarily on a scrubby forest-woodland setting on the Woronora Plateau, NSW Australia. Linear regression and t-tests were used to compare vegetation structural metrics from field and ALS data, with ANOVA and post-hoc tests used to determine solar energy and moisture controls on vegetation variation at hillslope scale. Landscape stratification was based on insolation and topographic wetness surfaces derived from ALS-based digital elevation models (DEM). Image matching and linear regression was used to test 3D-method orthorectification accuracy for off-nadir QuickBird imagery using different-resolution DEM. High resolution ALS-derived digital elevation models (DEM) allow pixel-accurate orthorectification of off-nadir imagery, a necessary precursor to multi-temporal image analysis.