Spatio-temporal variability of fire occurrence and recurrence in the Caatinga Biome using MODIS sensor data

The indiscriminate use of fire, every year has been causing an imbalance in nature, which can be perceived globally. The remote sensing represents the main technological alternative in detecting, dimensioning and understanding the dynamics of fire. Thus, the objective of this study was to analyze the spatio-temporal distribution of the burned areas of the Caatinga Biome using the MODIS MCD64A1 product, from 2001 to 2018. For this, the monthly subsets of the Burned Area MCD64A1 product were used for the period of study. The Canadian Forest Service classification was also adopted, which defines the burned areas in five different classes: I (0-0.09 ha); II (0.1-4.0 ha); III (4.1-40.0 ha); IV (40.1-200.0 ha); V (>200.0 ha). The results achieved in this study reveal that the state of Piaui has a statistically higher average of fire occurrences and burned area in the times series. The months that had the largest burned areas in the biome were September, August and October and the greatest recurrences of burned areas from May to December. The burned area size classes that presented the highest occurrences were III, IV e V. The biome undergoes systematic growth of degradation, which enhances its fragility in the face of fire.