Estimativa de biomassa utilizando dados lidar em floresta tropical

• Ano de defesa: 2018
• Autor(a) principal: Badin, Tiago Luis
• Orientador(a): Pereira, Rudiney Soares
• Banca de defesa: Silva, Emanuel Araújo , Pegoraro, Antoninho João
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Santa Maria; Centro de Ciências Rurais
• Programa de Pós-Graduação: Programa de Pós-Graduação em Engenharia Florestal
• Departamento: Recursos Florestais e Engenharia Florestal
• País: Brasil
• Palavras-chave em Português: Amazônia; Laser aerotransportado; Modelagem; Regressão linear; AGB
• Palavras-chave em Inglês: Airborne laser; Modeling; Linear regression; AGB
• Área do conhecimento CNPq: CNPQ::CIENCIAS AGRARIAS::RECURSOS FLORESTAIS E ENGENHARIA FLORESTAL
• Link de acesso: http://repositorio.ufsm.br/handle/1/14069

The Brazilian rainforests, mainly the Amazon, store in their biomass a large part of the global carbon stock, as a result of deforestation and degradation there has already been a considerable commitment, catalyzing the release of greenhouse gases into the atmosphere, aggravating the effects of global warming. In this context, the objective of this work was to estimate the above - ground biomass from data from airborne laser in Amazon rainforest. We used inventory data to calculate the biomass above the soil, values calculated through the model adjusted by Chave et al. (2015) adapted for tropical regions. Subsequently, the variables from the FUSION 3.6 software, derived from the airborne laser survey, were pre-selected using the Stepwise method. In the modeling, six models were tested: Linear, multiplication, exponential, parabola, polynomial of degree three and polynomial of degree four, where the variables Elev.CV, Elev.P99, Elev.MAD.mode and Elev.L3 from the laser composed the final model The best model was the polynomial of degree four, without intercept, which obtained coefficient of determination (R²) 0.76, standard error of estimate (Syx) 26.99, coefficient of variation (CV) 36,29, efficiency (E) 0.99, and absolute trend index (BIAS) -0.00005, and was therefore selected by the statistical criteria, later validated by the student's t-test. Thus, modeling with the inventory data related to LiDAR metrics proved to be efficient in the characterization of the tropical forest, showing that it is possible to use this technology to obtain estimates of above-ground biomass in tropical forests.