Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Virgínia Vieira Aires has a total of 3 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Using Trees as a Natural Weather Station for Wind Pattern Forecasting Applied to Forest Firefighting

    by Virgínia Vieira Aires, Cleonilson Protasio Souza, Orlando Rocha Baiocchi, Ane Polline Lacerda Protasio

    With the increasing frequency and intensity of extreme weather events, there is a growing need to develop innovative and accessible methods for environmental monitoring. This work presents a solution based on natural trees equipped with a low-cost embedded system. The innovative idea is that measuring only internal temperature of a tree around its trunk, it is possible to evaluate some weather parameters, such as wind speed and direction. To evaluate this relationship, a multiscale decomposition technique (Discrete Wavelet Transform) and machine learning models (Random Forest, Gradient Boosting, SVM, and Linear Regression) were applied. Among the models tested, Random Forest achieved the best results, demonstrating high accuracy with an error of 6.60% over a continuous monitoring period of 47 days. As an important outcome, the result shows that the proposed tree-based solution is viable for weather monitoring in hard-to-reach places, particularly in the context of forest fire prevention.

  • Open Access

    Using Trees as a Natural Weather Station for Wind Pattern Forecasting Applied to Forest Firefighting

    by Virgínia Vieira Aires, Cleonilson Protásio de Souza, Orlando Baiocchi, Ane Polline Lacerda Protasio

    With the increasing frequency and intensity of extreme weather events, there is a growing need to develop innovative and accessible methods for environmental monitoring. This work presents a solution based on natural trees equipped with a low-cost embedded system. The innovative idea is that measuring only internal temperature of a tree around its trunk, it is possible to evaluate some weather parameters, such as wind speed and direction. To evaluate this relationship, a multiscale decomposition technique (Discrete Wavelet Transform) and machine learning models (Random Forest, Gradient Boosting, SVM, and Linear Regression) were applied. Among the models tested, Random Forest achieved the best results, demonstrating high accuracy with an error of 6.60%. As an important outcome, the result shows that the proposed tree-based solution is viable for weather monitoring in hard-to-reach places, particularly in the context of forest fire prevention.