Potential of an ultraviolet, medium-footprint lidar prototype for retrieving forest structure

The aim of the paper is to carry on methodological development for retrieving forest parameters from medium-footprint lidar signals and for assessing the performance of different sampling strategies.The 2.4 m footprint lidar prototype (a profiler instrument using an ultraviolet laser) was flown above two different maritime pine stands: a young plantation (10 years old) and a mature, semi-natural stand (55 years old), both in the Landes forest, France.The vertical distribution of lidar measurements was studied for retrieving forest height parameters (mean total height, mean crown height and top height). The processing algorithm was based on an aggregation of successive signals followed by the correction of the signal attenuation along the travel through the vegetation. The performance of different sampling strategies was assessed by comparing the results for the full dataset (several fight lines over the stands) and for only a data subset (one flight line).In addition, the horizontal distribution of height measurements was studied for identifying the planting pattern of the stands and assessing the tree spacing of the semi-natural parcel, using geostatistics.We obtained a sub-metric estimation error (lidar – reference) of 0.2 m on the mean total height in the young stand (−0.7 m in the mature stand), a bias of −0.3 m (−0.3 m) on the mean crown height measurement and of 0.6 m (−1.0 m) on the top height. The planting pattern was also successfully identified, and the distance between trees was assessed in agreement with ground measurements.Having demonstrated its ability to assess forest structure, even with a unique flight line, the lidar prototype seems to be a valuable sensor for performing fast forest inventory at regional scale. In addition, this sensor opens the way to the development of bi-functional lidar for both atmosphere and vegetation remote sensing.

► Data acquisition on forest using a profiler lidar prototype, with different sampling strategies.
► Signal processing and data analysis to extract forest structure parameters (tree height, planting pattern, tree spacing).
► Comparison of lidar derived parameters with ground-truth data.
► Accurate assessment of forest structure parameters, even with a unique flight line.