Inferring the Spatial Distribution of Vegetation Height and Density in a Mesotidal Salt Marsh From UAV LIDAR Data

Unmanned Aerial Vehicles (UAV) have become an affordable and cost-efficient tool to quickly map targeted salt marshes areas. Although light detection and ranging (LIDAR) is effective at measuring surface elevations, laser penetration is limited in dense salt marsh vegetation. The limited ability of the laser to penetrate dense vegetation hinders its usefulness for surveying tidal marsh platforms. Accurate mapping of both elevation and plant characteristics in salt marshes is important for its management and conservation. For coastal salt marshes, methods to correct elevation data from airborne LIDAR are available. For UAV-borne lidar, a reliable method to extract ground elevation, vegetation density and vegetation height from high-resolution point clouds is yet not available. The present study aims at investigating a new empirical method to determine the vegetation characteristics from UAV LIDAR data in tidal salt marshes.

A LIDAR point cloud was collected using a Velodyne 16 LIDAR sensor mounted on a DJI Matrice 600 UAV. The study field consists of a vegetated tidal marsh at Sapelo Island, Georgia, USA. The data was collected in February 2019, at low spring tide, when narrow (<1.5 m width) tidal creeks were mostly dry so that their bathymetry could be detected. Vegetation height and density, and ground GPS-RTK elevations were collected in twentyone 0.5 m x 0.5 m plots in the study area. The entire domain was discretized by means of a triangular grid. The average dimension of the triangle sides was 0.20 m. An empirical formulation was obtained by comparing the point clouds in each bin with the field data. The spatial distribution of vegetation and ground elevations was then determined in the entire domain by means of our empirical formulation. Our high-resolution marsh morphology will be used to create a hydrodynamic and morphodynamic numerical model of the area.