Fine-resolution mapping of microforms of a boreal bog using aerial images and waveform-recording LiDAR

Boreal bogs are important stores and sinks of atmospheric carbon whose surfaces are characterised by vegetation microforms. Efficient methods for monitoring their vegetation are needed because changes in vegetation composition lead to alteration in their function such as carbon gas exchange with the atmosphere. We investigated how airborne image and waveform-recording LiDAR data can be used for 3D mapping of microforms in an open bog which is a mosaic of pools, hummocks with a few stunted pines, hollows, intermediate surfaces and mud-bottom hollows. The proposed method operates on the bog surface, which is reconstructed using LiDAR. The vegetation was classified at 20 cm resolution. We hypothesised that LiDAR data describe surface topography, moisture and the presence and depth of field-layer vegetation and surface roughness; while multiple images capture the colours and texture of the vegetation, which are influenced by directional reflectance effects. We conclude that geometric LiDAR features are efficient predictors of microforms. LiDAR intensity and echo width were specific to moisture and surface roughness, respectively. Directional reflectance constituted 4–34 % of the variance in images and its form was linked to the presence of the field layer. Microform-specific directional reflectance patterns were deemed to be of marginal value in enhancing the classification, and RGB image features were inferior to LiDAR variables. Sensor fusion is an attractive option for fine-scale mapping of these habitats. We discuss the task and propose options for improving the methodology.