Fabiano Morelli

An Algorithm for Burned Area Detection in the Brazilian Cerrado Using 4 µm MODIS Imagery

The Brazilian Cerrado is significantly affected by anthropic fires every year, which makes the region an important source of pyrogenic emissions. This study aims at generating improved 1 km monthly burned area maps for Cerrado based on remote-sensed information. The algorithm relies on a burn-sensitive vegetation index based on MODIS daily values of near and middle infrared reflectance and makes use of active fire detection from multiple sensors. Validation is performed using reference burned area (BA) maps derived from Landsat imagery. Results are also compared with MODIS standard BA products. A monthly BA database for the Brazilian Cerrado is generated covering the period 2005–2014. Estimated value of BA is 1.3 times larger than the value derived from reference data, making the product suitable for applications in fire emission studies and ecosystem management. As expected the intra and inter-annual variability of estimated BA over the Brazilian Cerrado is in agreement with the regime of precipitation. This work represents the first step towards setting up a regional database of BA for Brazil to be developed in the framework of BrFLAS, an R and D project in the areas of fire emissions and ecosystem management planning.

Burned Area Mapping in the Brazilian Savanna Using a One-Class Support Vector Machine Trained by Active Fires

We used the Visible Infrared Imaging Radiometer Suite (VIIRS) active fire data (375 m spatial resolution) to automatically extract multispectral samples and train a One-Class Support Vector Machine for burned area mapping, and applied the resulting classification algorithm to 300-m spatial resolution imagery from the Project for On-Board Autonomy-Vegetation (PROBA-V). The active fire data were screened to prevent extraction of unrepresentative burned area samples and combined with surface reflectance bi-weekly composites to produce burned area maps. The procedure was applied over the Brazilian Cerrado savanna, validated with reference maps obtained from Landsat images and compared with the Collection 6 Moderate Resolution Imaging Spectrometer (MODIS) Burned Area product (MCD64A1) Results show that the algorithm developed improved the detection of small-sized scars and displayed results more similar to the reference data than MCD64A1. Unlike active fire-based region growing algorithms, the proposed approach allows for the detection and mapping of burn scars without active fires, thus eliminating a potential source of omission error. The burned area mapping approach presented here should facilitate the development of operational-automated burned area algorithms, and is very straightforward for implementation with other sensors.

Frequência espaço-temporal dos focos ativos em Minas Gerais durante o período de 1999 a 2009

The aim of this study was to rescue the history of hotspots in the state of Minas Gerais during the years of 1999 to 2009 for an exploratory data analysis. Through temporal analysis, we proceeded an analysis of the months and years where foci have been more frequent and of the critical period of occurrence of hotspots. The analysis considered the spatial frequency and the density of hotspots in the State Forestry Institute (IEF) nuclei. In the case of conservation units, it was also analyzed the annual recurrence. In total, 67,334 hotspots were detected in MG during the period from 1999 to 2009. Temporal analysis revealed that the year 2003 had the highest frequency, totaling 10,929 hotspots; the year 2009, in turn, had only 2,378 spots. October was the month with the largest frequency, with 24,149 of hotspots recorded. Nuclei with higher frequency, higher density and higher recurrence of hotspots in the case of conservation units, present the worst scenario, being considered as the most critical hotspots. However, low frequency of hotspots and high density are also indicatives of critical areas. In areas of high frequency of hotspots and low density, the size of the area directly influenced the high incidence of active hotspots. In these areas it is suggested an analysis of spatial distribution to check if hotspots are concentrated or sparse. The best scenario presented here is low frequency and low density of hotspots.