JFJosé Augusto Spiazzi Favarin

José Augusto Spiazzi Favarin

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


José Augusto Spiazzi Favarin has a total of 35 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Bioeconomy in the Amazon: Lessons and gaps from thirty years of non-timber forest products research

    by Emmanoella Costa Guaraná Araujo, Thiago Cardoso Silva, Ernandes Macedo da Cunha Neto, José Augusto Spiazzi Favarin, Joana Keila da Silva Gomes, Kyvia Pontes Teixeira das Chagas, Elaine Cosma Fiorelli, Artur Falqueto Sonsin, Emanuel Maia


  • Open Access

    Identification and characterization of gaps and roads in the Amazon rainforest with LiDAR data

    by J.A. Spiazzi Favarin, M. Sabadi Schuh, J. Marchesan, E. Alba, R. Soares Pereira

    Gap formations in the forest canopy have natural causes, such as bad weather, and anthropic ones, such as sustainable selective extraction of trees and illegal logging, which can already be detected through orbital remote sensing. However, the Amazon region is under frequent cloud cover, which makes it challenging to detect gaps using passive sensors. This study aimed to identify and delimit gaps in the Amazon forest canopy through airborne LiDAR (Light Detection and Ranging) sensor application while testing six different return densities. LiDAR and forest inventory data were obtained over an Amazon rainforest region, defining the minimum area as a forest canopy gap. The point cloud was processed to obtain six return densities with the generation of their respective CHM (Canopy Height Model), which were applied for segmentation and subsequent identification of gap areas and roads. The minimum gap area found was 34 m2, and the Kruskal Wallis test showed no significant difference among the six densities in gap detection; however, road identification decreased as the return density decreased. We concluded that LiDAR data proved promising as point clouds with low return density can be used without impairing gap identification. However, reducing the return density for road identification is not recommended.


  • Open Access

    Espectrorradiometria na distinção de espécies florestais utilizando Random Forest (RF), Support Vector Machine (SVM) e Artificial Neural Network (ANN)

    by Igor da Silva Narvaes, Mateus Sabadi Schuh, Pábulo Diogo Souza, Matheus Morais Ziembowicz, José Augusto Spiazzi Favarin, Janderson de Oliveira Silva, Noé dos Santos Hofiço, Laura Camila de Godoy Goergen, Rudiney Soares Pereira

    A distinção de espécies florestais na arborização urbana é fundamental para mitigação dos efeitos locais do aquecimento global. Neste sentido, foram utilizados os algoritmos de aprendizado de máquina RF, SVM e de aprendizagem profunda ANN. Os elevados valores de acurácia encontrados (F-1 score = 0,989; 0,9434; 0,9346, Acurácia Global = 0,989; 0,9444; 0,9333 e de índice kappa = 0,988; 0,9383; 0,9259) para o algoritmo ANN, SVM e RF, respectivamente. Os erros de classificação para a predição de algumas espécies para os classificadores analisados se dão em geral pela semelhança nos valores de reflectância nas regiões do red edge (700 a 720 nm) relacionados ao conteúdo similar de clorofila e nos comprimentos de onda específicos na região do infravermelho de ondas curtas (1400 e 1420 nm) responsáveis pelas diferenças no conteúdo de água e concentração química na planta e de lignina, respectivamente. Dado a complexidade dos classificadores, em especial o algoritmo de aprendizagem profunda ANN, e também aos de aprendizagem de máquina SVM e RF, recomenda-se a alteração de seus hiperparâmetros para se evitar o sobreajuste dos resultados, ou seja, mesmo que o algoritmo esteja adaptado a uma determinada região se torne ineficaz para prever novos resultados.

  • Aboveground biomass stock and change estimation in Amazon rainforest using airborne light detection and ranging, multispectral data, and machine learning algorithms

    by Juliana Marchesan, Elisiane Alba, Mateus Sabadi Schuh, José Augusto Spiazzi Favarin, Roberta Aparecida Fantinel, Luciane Marchesan, Rudiney Soares Pereira

    Developing an efficient method to accurately estimate aboveground biomass in tropical forests is critical to monitoring the carbon stock and implementing policies to reduce emissions caused by deforestation. Thus, the objective of the present study was to estimate aboveground biomass in areas of the Amazonian Forest with selective logging, using the random forest (RF), support vector machine (SVM), and artificial neural network (ANN) machine learning algorithms, using light detection and ranging (LiDAR) data and these combined with OLI/Landsat 8 variables, as well as mapping the biomass for the years 2014 and 2017, allowing one to analyze its dynamics between the years of analysis. The RF and SVM algorithms obtained the lowest error values in all datasets. The association of the variables increased the RF performance. Analyzing the dynamics of biomass, it was observed that the oldest exploration units (2006, 2007, and 2008) have lower biomass stocks. The highest biomass losses in 2017 came from units operated between 2012 and 2013 (the most recent record). Thus, with the method used in this study, it was possible to infer that the machine learning algorithms were efficient in estimating the biomass, emphasizing the RF and the SVM.


  • Open Access

    ABOVEGROUND BIOMASS ESTIMATION IN A TROPICAL FOREST WITH SELECTIVE LOGGING USING RANDOM FOREST AND LIDAR DATA

    by Juliana Marchesan, Elisiane Alba, Mateus Sabadi Schuh, José Augusto Spiazzi Favarin, Rudiney Soares Pereira

    The tropical forest is characterized by expressive biomass and stores high amounts of carbon, which is an important variable for climate monitoring. Thus, studies aiming to analyze suitable methods to predict biomass are crucial, especially in the tropics, where dense vegetation makes modeling difficult. Thus, the objective of the present study was to estimate aboveground biomass (AGB) in a tropical forest area with selective logging in the Amazon forest using the Random Forest (RF) machine learning algorithm and LiDAR data. For this, 85 sample units were used at Fazenda Cauaxi, in the municipality of Paragominas, Pará State. LiDAR data were collected in 2014 and made available by the Sustainable Landscapes Project. The software R was used for data analysis. Among the LiDAR metrics, the average height was used as it had the greatest significance to compose the model. The model presented a pseudo R² of 0.69 (value obtained by the RF), Spearman's Correlation Coefficient of 0.80, RMSE of 47.05 Mg.ha-1 (19.84%), and Bias of 2.06 Mg.ha-1 (0.87%). With the results, it was possible to infer that the average height metric was enough to estimate AGB in a tropical forest with selective logging, in addition, the RF algorithm the biomass to be estimated, which can be used to assist in monitoring and action management in areas of selective logging and serve as a basis for climate change mitigation policies.


  • Machine learning and generalized linear model techniques to predict aboveground biomass in Amazon rainforest using LiDAR data

    by Mateus Schuh, José Augusto Spiazzi Favarin, Juliana Marchesan, Elisiane Alba, Elias Fernando Berra, Rudiney Soares Pereira

    LiDAR remote sensing data combined with machine learning (ML) techniques have presented great potential for large-scale modeling of tropical forest attributes. However, the large amount of information that can be derived from an aerial LiDAR survey, summed with the intrinsic heterogeneity of tropical environments (e.g., the Amazon), makes it a challenge to accurately estimate forest biophysical variables. The aim of our work is to investigate the potential and accuracy of different ML techniques and a generalized linear model (GLM) to learn the relationships between LiDAR-derived metrics and forest inventory data for aboveground biomass (AGB) prediction in Amazon forest sites under selective logging regimes. The predictive performance of three ML techniques, namely random forest (RF), support vector machine (SVM), and artificial neural network (ANN), was compared against result from the GLM technique, across 85 sample plots. Interestingly, the GLM retrieved the most accurate estimations of forest AGB (rho Spearman’s coefficient = 0.87), compared with the ML techniques (RF = 0.77, SVM = 0.67, and ANN = 0.50). A number of possible factors affecting such results are listed and discussed in the text, including sample size and number of predictor variables. Continued research is necessary to improve the confidence of AGB estimation, especially over complex forest structures.


  • Open Access

    FUNCTIONAL ATTRIBUTES AS ECOLOGICAL PREDICTORS DURING SECONDARY SUCCESSION IN A SEASONAL DECIDUOUS FOREST IN SOUTHERN BRAZIL

    by Francieli de Fátima Missio, Solon Jonas Longhi, Matheus Degrandi Gazzola, Marina Scheuer, Rodrigo Pinto da Silva, José Augusto Spiazzi Favarin

    Through the variation of the functional attributes, it is possible to verify the functioning of an ecosystem, both by the species organization as by their responses to the environmental variations. In this sense, the objective of this study was to analyze the ecological strategies of the main tree species by analyzing their functional attributes. PCoA was used to verify which species are acquisitive and conservative; clustering was used to verify the functional attributes by groups; the Pearson’s correlation between the functional attributes and a CWM-RDA analysis was used to verify the environmental variables. Most species were classified as light-demanding climax and having zoochoric dispersion. The species presented phenotypic plasticity as an important ecological strategy in the composition of their functional attributes, especially when related to leaf area and specific leaf area. Most of the species belonged to the acquisitive group. The acquisitive and conservative groups indicate the resilience potential of the tree community and the change processes in ecological succession. There was a strong negative correlation between the leaf attributes, and a positive one with diameter and height, both correlations related to plant growth and development. Of the environmental variables only pH, K, and average elevation, were related to the attributes, indicating that environmental conditions are important for the establishment of the sampled species. Also, it was found that the species composition is linked to several conditions of ecological strategies associated with changes in the environment.