Models for forecasting deforestation alerts in the Legal Amazon
DOI:
https://doi.org/10.54766/rberu.v17i4.979Keywords:
Arima, Structural breaks, SeasonalityAbstract
This study sought to select a forecast model for deforestation alerts in the Legal Amazon
from data generated by satellite monitoring of DETER-B between August 2015 and December 2021. The time series of deforestation alerts was analyzed and then, forecasts of deforestation alerts were made, using seasonal ARIMA class forecasting models. The presence of structural break in May 2019 and stochastic seasonality was identified. Dynamic forecasts were made for six periods forward of the last value of the sample, to compare with the values outside the sample and verify the quality of the forecasts. The specifications were accurate in predicting the alerts six months forward, indicating that public policymakers can create reasonable expectations about the alerts, mainly by allowing the adoption of preventive measures for deforestation.
Downloads
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2024 ABER
This work is licensed under a Creative Commons Attribution 4.0 International License.
The submission of papers to the Journal implies the assignment of the copyright to the Brazilian Regional Science Association.
The content published by the 'Revista Brasileira de Estudos Regionais e Urbanos' (Brazilian Review of Regional and Urban Studies) is licensed under a Creative Commons Atribuição 4.0 Internacional license.