Introduction: Forest systems are part of an important land-use strategy aimed at the maintenance of biodiversity, at the commercialization of wood products and other products, at quality of life and at environmental services such as carbon fixation. The greenhouse gas emission flows and the carbon stocked in the forests are indispensable information, and their estimation may support public policies on development and sustainability. Studies focused the development of methods for the assessment of greenhouse gases and forest carbon stock are concise due to difficulties in assessing and analyzing local terrestrial characteristics, such as carbon fixation capacity of the plants and soil, and generating methods and processes of extrapolation to geographically wider areas. This is where the project Classification and Spatialization of Carbon Stock in Natural and Planted Forests Using Remote Sensing (GeoSaltus), which is part of the network for Dynamics of Greenhouse Gases and Carbon Stock in Brazilian Natural and Planted Forests coordinated by Embrapa Forestry, fits. The project's strategy focuses on the development of methods for the assessment of greenhouse gases and forest carbon stock. The use of satellite images and geographic information systems for forest carbon stock estimation is a more adequate strategy because of the significantly lower time necessary to perform a forest classification and of the possibility of extrapolating characteristics obtained in local inventories to wide areas.

Objectives: To digitally classify and estimate the carbon stock of reference natural and planted forest areas of the Cerrado, Amazônia, Cerrado/Amazônia transition, Pantanal, Mata Atlântica and Pampa biomes using remote sensing. Among the specific objectives we intend to systematize a geographic database of information on greenhouse gas emissions and carbon stock in natural and planted forests at the reference biome areas studied; to obtain a set of spectral variables for natural forests from orbital remote sensors; to apply different digital classification models to native and planted forests by means of correlating spectral variables and biophysical parameters; and to spatialize the biophysical parameters of native and planted forests using carbon maps of the studied areas.

Expected results: We expect to produce results with technical and scientific impacts, by generating processes and methods of forest carbon stock estimation by means of remote sensing; with economical and environmental impacts, by generating carbon mapping protocols which are less costly than the traditional protocols and which offer possibilities of extrapolating local environmental characteristics to wider geographic regions; and with governance impacts, by producing a spatially explicit database on greenhouse gas and carbon stock which will support governmental decision-making and public policies for sustainable development.

Leader: Édson Luis Bolfe – Researcher at Embrapa Satellite Monitoring

Institutions involved:

• Embrapa,
• Universidade Federal do Mato Grosso (UFMT),
• Universidade Estadual de Campinas (UNICAMP),
• Serviço Florestal Americano (USDA).