Estimation de la biomasse forestière par des empreintes moyennes à grandes de données LIDAR

[Departement_IRSTEA]Territoires [TR1_IRSTEA]SYNERGIEInternational audienceForest covers play a tremendous role in our climate and have to be monitored from regional to global scale. Biomass is a primary indication of carbon sequestration rate in forest ecosystems and has been identified as an Essential Climate Variable (ECV) needed to improve our knowledge on carbon cycle and define efficient strategies for climate migration. Lidar, based on the emission and reception of laser pulses, is one of the most promising technologies for assessing forest parameters. We used a medium-footprint, ultra-violet (UV, 355 nm) profiler lidar prototype (LAUVA) onboard an Ultra-Light Aircraft (ULA), which has been developed by Commissariat à l'Énergie Atomique (CEA). In order to assess the capacity of this system to measure forest structure and to open the way towards spaceborne lidar systems dedicated to forest studies, we analysed data at two different scales. Parcels of maritime pines in the Landes forest (South-western France) have been sampled by the lidar. Then, data has been processed to simulate spaceborne observations with a resolution similar to the one of Geoscience Laser Altimeter System (GLAS, IR, 1064 nm, NASA) onboard Ice, Cloud and land Elevation Satellite (ICESat, http://icesat.gsfc.nasa.gov/). We have developed approaches to assess mean tree height and tree density for estimating aboveground biomass from medium footprint measurements. Mean tree height has been estimated through the process of lidar signals, tree density has been derived from the spatial distribution of measurements and Diameter Breast Height (DBH) has been calculated with statistical relation established on height measurement. Using an allometric equation relating biomass, height and DBH we estimated an aboveground biomass of 112 t.ha-1. The obtained model is compared to biomass estimation calculated from reference measurements, to calculate an error of 9%. Existing method developed for GLAS data will be adapted to process the large footprint simulated waveforms for estimating biomass at a larger scale. In particular, we will discuss the impact of 1) on aboveground biomass estimation accuracy of a larger footprint of lidar spaceborne missions performing data aggregating, 2) the reduction of signal to noise ratio within the forest structure, and 3) the dissipation along the optical path of the UV signal