Simulation study for finding optimal lidar acquisition parameters for forest height retrieval

Accurate and cost-effective measurements of forest inventory parameters are an essential input to forest management practice. Of these parameters, height has been shown to be valuable in the estimation of stem volume and site quality. The accuracy of height estimates from lidar data depends critically on the operational characteristics of the airborne instrument and the resulting sampling density relative to the individual tree spacing within the stand. Using a simulated lidar dataset for a forest plantation with varying stem density it was found that mean nearest neighbour spacing of the lidar point samples is linearly related to retrieved predominant height. The results also indicated that the accuracy of height retrieval may be poorer at the edge of the lidar swath due to uneven spacing of the sample points. Given the significant cost of lidar data acquisition, it is crucial that data acquisition planning maximizes the benefit of the data retrieved. The information gained through simulations such as those detailed in this paper may assist attaining this maximum benefit by identifying lidar operational parameters that produce sufficiently accurate results while minimizing cost. (c) 2005 Elsevier B.V. All rights reserved.