Abstract
To recover ecosystem services, provide climate change mitigation, and ensure food security and protection of water sources, several law projects and international pledges were established with audacious goals of restoring forests on a broad scale. The use of images from different remote sensors, in different spatial, spectral and temporal scales, is of paramount importance for the planning and monitoring of large areas in restoration. In this project, we propose to develop the broad-scale monitoring of new forests in the state of São Paulo using images and data from low-cost or even free active and passive sensors associated with field data, to answer: 1) Where are the restoration forests of São Paulo? 2) What forest typologies make up the forests in restoration? 3) What are the successional stages of the different coverage of native vegetation? 4) What is the biomass and carbon stored in the forests in restoration? 5) Which remote sensing methodologies are more efficient for the broad-scale forest restoration monitoring from the integration of lidar data, radar multi and hyperspectral images and inventory? The sample design contains 20 sampling landscapes, wherein five of these, data from drone-borne lidar and hyperspectral sensors have been collected. For spatialization, we will use radar images (Sentinel 1, band C), as well as optical images Landsat and Sentinel 2, and high-resolution images from Planet company. As results we will present the spatial configuration of the new forests, we will define forest classes and successional stages, relate the attributes of the radar images and high-resolution optical images data to the aboveground biomass stocks modeled by lidar data and field plots. Our results will assist in the characterization of ecosystem services of new forests and support the monitoring of vegetation with focus on restoration, supporting decision-making by the environmental agencies.
|