Joint use of space-borne SAR, optical imagery and air-borne LiDAR for improved mapping of forest structural types in New Zealand

Sustainable forest conservation and management practices require highly resolved and accurate maps of forest types. Extrapolation of field data, however, cannot achieve the necessary level of detail. The joint use of space-based optical imagery and structural information from synthetic aperture radar (SAR) and canopy metrics derived from air-borne Light Detection and Ranging (LiDAR) facilitates detailed classification of forest types. We present a segmentation-based support vector machine (SVM) classification using data from ESA’s Sentinel-1 and 2 missions, ALOS PALSAR and airborne LiDAR to create a map of structural types within indigenous forest in Greater Wellington, New Zealand. The model is evaluated using k-fold cross-validation with up-scaled field data. The highest classification accuracy of 80.9% is achieved for bands 2, 3, 4, 5, 8, 11, and 12 from Sentinel-2, the ratio of bands VH and VV from Sentinel-1, HH from PALSAR, and mean canopy height and 97th percentile canopy height from LiDAR. The classification based on the optical bands alone is 73.1% accurate. Our high-resolution regional map of structural forest types is fit-for-purpose for conservation management and we show that the inclusion of structural information from SAR and LiDAR can improve forest classification by 7.8%.