LiDAR data provides an excellent data source for describing plantations, offering a range of useful metrics and allowing for the delineation of accurate boundaries. However, due to the complexity and associated cost LiDAR is generally captured infrequently – often only one observation is available, particularly when the data is sourced from national monitoring programs such as the NZ LiDAR Mapping Project.
To this end, initial LiDAR extents can be updated using freely available Sentinel-2 multispectral satellite images. As the LiDAR boundaries are of high accuracy, the areas checked for harvesting or other loss can be effectively constrained, reducing the processing time and the chance of false-positive errors.
The following demo uses LiDAR captured in 2013 to map plantations and small woodlots in the Wairarapa region (for more information, see the original blog post on this data here). In this example, successive Sentinel-2 scenes captured near the end of each year (from 2016 to date) are used to track harvest activities. Such an approach provides insights into annual harvesting rates and changes to a resources age-class and woodflow profile.