Prediction of L-band signal attenuation in forests using 3D vegetation structure from airborne LiDAR

In this study, we propose a novel method to predict microwave attenuation in forested areas by using airborne Light Detection and Ranging (LiDAR). While propagating through a vegetative medium, microwave signals suffer from reflection, absorption, and scattering within vegetation, which cause signal attenuation and, consequently, deteriorate signal reception and information interpretation. A Fresnel zone enveloping the radio frequency line-of-sight is applied to segment vegetation structure occluding signal propagation. Return parameters and the spatial distribution of vegetation from the airborne LiDAR inside Fresnel zones are used to weight the laser points to estimate directional vegetation structure. A Directional Vegetation Density (DVD) model is developed through regression that links the vegetation structure to the signal attenuation at the L-band using GPS observations in a mixed forest in North Central Florida. The DVD model is compared with currently-used empirical models and obtained better R2 values of 0.54 than the slab-based models. Finally, the model is evaluated by comparing with GPS observations of signal attenuation. An overall root mean square error of 3.51 dB and a maximum absolute error of 9.38 dB are found. Sophisticated classification algorithms and full-waveform LiDAR systems may significantly improve the estimation of signal attenuation.