Avaliação de técnicas de análise espacial por interpolação IDW e krigagem ordinária na predição de variáveis dendrométricas para produção de Pinus taeda L.
Resumo
Resumo: Para o planejamento da produção de um povoamento florestal, é essencial conhecer o estoque volumétrico existente, o qual pode ser mensurado através do inventário florestal ou por metodologias de interpolação geoespaciais que estimam variáveis e são capazes de representar seu comportamento, extrapolando os valores para a população. Este trabalho teve como objetivo avaliar o potencial das ferramentas de análises espaciais e geoestatística para geração de estimativas em inventários florestais. Para isso, foram testados os métodos de interpolação espacial pela ponderação do inverso da distância (Inverse Distance Weighed - IDW) e a geoestatística com a aplicação de krigagem ordinária para a predição de quatro variáveis dendrométricas: volume, diâmetro, altura e área basal em plantios de Pinus taeda L. pertencentes a empresa Florestal Gateados Ltda., localizada no município de Campo Belo do Sul em Santa Catarina. Foram utilizados dois projetos, sendo o projeto denominado Porteira Queimada apresenta área total de 137,81 hectares, com 15 parcelas instaladas de 500 m² cada. O projeto denominado Prefeito tem área total de 153,71 hectares, com 32 parcelas de 500 m². Para a aplicação das técnicas de interpolação foi utilizado o software ArcGIS e os resultados foram comparados aos valores mensurados e processados pelo inventário florestal pré corte e ao volume real colhido em campo. Em função dos limites inferiores e superiores do intervalo de confiança apresentados pelo método de inventário pré-corte tradicional, a avalição das predições das variáveis pelos métodos espaciais está contemplada dentro dos intervalos de confiança. Portanto, é possível concluir que ao comparar os valores mensurados pelo Inventário Florestal Pré-Corte - IPC, preditos nas duas metodologias testadas juntos aos volumes colhidos em campo para os dois projetos, as predições não são diferentes sob a probabilidade de 95%, logo conclui-se pela maior facilidade da aplicação do IDW em relação a técnica de geoestatística para o caso em tela, sugere-se sua adoção em caso de necessidade de avaliação espacial de distribuição das variáveis aqui estudadas. No entanto, destaca-se que o inventário florestal tradicional resultou em valores estimados mais próximos aos valores de produção volumétrica colhido em campo, demonstrando, portanto, a consistência das técnicas tradicionais utilizadas nos inventários florestais. Abstract: For planning the production of a forest stand, it is essential to know the existing volumetric stock, which can be measured through the forest inventory or by geospatial interpolation methodologies that estimate variables and are able to represent their behavior, extrapolating the values to the population. This work aimed to evaluate the potential of spatial and geostatistical analysis tools to generate estimates in forest inventories. For this, the methods of spatial interpolation by weighting the inverse of the distance (Inverse Distance Weighed - IDW) and geostatistics with the application of ordinary kriging were tested for the prediction of four dendrometric variables: volume, diameter, height and basal area. in plantations of Pinus taeda L. belonging to the company Florestal Gateados Ltda., located in the municipality of Campo Belo do Sul in Santa Catarina. Two projects were used, the project called Porteira Queimada has a total area of 137.81 hectares, with 15 installed plots of 500 m² each. The project called Mayor has a total area of 153.71 hectares, with 32 plots of 500 m². For the application of interpolation techniques, ArcGIS software was used and the results were compared to the values measured and processed by the pre-cut forest inventory and the actual volume collected in the field. Due to the lower and upper limits of the confidence interval presented by the traditional pre-cutting inventory method, the evaluation of the variables' predictions by the spatial methods is contemplated within the confidence intervals. Therefore, it is possible to conclude that when comparing the values measured by the Pre-Cutting Forest Inventory - IPC, predicted in the two methodologies studied together with the volumes harvested in the field for the two projects, the predictions are not different under a probability of 95%, thus concluding - if for the greater ease of application of the IDW in relation to the geostatistics technique for the case in question, its adoption is suggested in case of need for spatial evaluation of the distribution of the variables studied here. However, it is noteworthy that the traditional forest inventory resulted in estimated values closer to the volumetric production values collected in the field, thus demonstrating the consistency of the traditional techniques used in forest inventories.
Collections
- Engenharia Florestal [253]