Estoque de carbono e quantificação da incerteza propagada combinando inventário florestal e sensoriamento remoto
REDD+ is an instrument developed at UNFCCC conferences to financially reward developing countries for efforts to reduce deforestation and forest degradation. In 2010, the IPCC task force in Yokohama evaluated and recommended linking existing field and remote sensing works to forest emissions esti...
| Autor principal: | Celes, Carlos Henrique Souza |
|---|---|
| Grau: | Tese |
| Idioma: | por |
| Publicado em: |
Instituto Nacional de Pesquisas da Amazônia – INPA
2020
|
| Assuntos: | |
| Acesso em linha: |
https://repositorio.inpa.gov.br/handle/1/4982 http://lattes.cnpq.br/5735970963851152 |
| Resumo: |
|---|
|
REDD+ is an instrument developed at UNFCCC conferences to financially reward
developing countries for efforts to reduce deforestation and forest degradation. In 2010, the
IPCC task force in Yokohama evaluated and recommended linking existing field and remote
sensing works to forest emissions estimates (IPCC, 2010) in a system that should be
measurable, reportable and verifiable (MRV) (UNFCCC, 2009). In developing countries these
estimates are more difficult due to the lack of a field data collection system and the
availability of both current and temporal images to generate their emission history. The
adoption of techniques associated with mathematical modeling and computer system is
necessary to reach the recommendations for REDD + projects and were applied in the Ducke
Reserve of INPA. Forest inventory with ALS LiDAR and SRTM airborne data, RapidEye and
Landsat 8 was used. Linear models were used to establish relationships and the Monte Carlo
technique was applied to quantify the propagated error. The error in diameter measurement
could be measured and controlled in forest inventories by adopting remeasurement techniques
with tape and photogrammetry. The development of scripts for the processing of LiDAR data
allowed us to quantify and control errors in land estimates. The georeferencing of the plots for
the combination with high resolution data requires procedures that guarantee their accuracy
without compromising the field activities. The Monte Carlo method was important for the
estimation of the error mainly of the georeferencing of the field data, since the formula of the
error propagation does not allow this type of approach. The scripts are in development and
available to any user, in order to make the method replicable in different places. The models
and their uncertainties demonstrated spatial variation with recognized cause and consequence
effects, necessary for the reliability of the models. The most endangered mature forest areas
are the plateau regions around the Ducke Reservation suitable for REDD+ project and
sustainable development. |