Development of a multi-temporal effective leaf area index model from multi-spectral lidar; Scaling to Sentinel-1 and Sentinel-2 data using machine learning****

点击翻译按钮获取中文摘要

Lidar is a remote sensing technology that is used to map the variations in the three-dimensional structural properties of vegetation, including vegetation height. Mapping of foliage cover and leaf area index (LAI) are required to model how ecosystems use atmospheric carbon dioxide for photosynthesis. While single pulse, discrete return lidar systems show promise for estimating canopy cover, conversion to leaf area index requires knowledge of species type and the proportion of woody vs. leafy material. Identification of these may be achieved using a multi-spectral (MS) lidar system. MS lidar is a relatively new technology that combines rapid emission of laser pulses and reception of these pulses in three different wavelengths. The combination of these may be used to estimate the proportions of different woody and leafy components, and species. The development of this technology has important benefits for validation satellite optical remote sensing estimates of foliage area, which is often used as an indicator for LAI over broad and remote regions. Hatfield Inc. have developed optical remote sensing-based method for deriving LAI, which they will scale up from airborne lidar data and field validation methods. The University of Lethbridge will partner with Hatfield Inc. to develop MS lidar methods validated using digital hemispherical photography (DHP) for forest and wetland sites proximal to Fort McMurray and Fort McKay. The development of this application for mapping of LAI will add to the range of capability of this innovative Canadian MS lidar technology as a scalable 'plot-based' method for quantifying a critical parameter used to estimate carbon uptake by Canada's forests and wetlands.