@InProceedings{LibonatiCamaSetz:2016:SpTePa,
author = "Libonati, Renata and Camara, Carlos C. da and Setzer, Alberto
Waingort",
affiliation = "{Universidade Federal do Rio de Janeiro (UFRJ)} and {Universidade
de Lisboa (UL)} and {Instituto Nacional de Pesquisas Espaciais
(INPE)}",
title = "Spatial and temporal patterns of burned area over Brazilian
Cerrado from 2005 to 2015 using remote sensing data",
booktitle = "Abstracts...",
year = "2016",
organization = "European Geophysical Society General Assembly",
abstract = "Although Cerrado is a fire-dependent biome, current agriculture
practices have significantly modified the native fire regime.
Moreover, over the last decades, climate conditions, such as
intensive droughts, have contributed to enhance the effects of
anthropogenic activities, and consequently fire, over the region.
For instance, during the 2010 extreme drought there was an
increase of 100% in the number of fire pixels detected by just one
polar orbiting satellite (information online at
http://www.cptec.inpe.br/queimadas). A better characterization of
spatial and temporal fire patterns over Cerrado is therefore
crucial to uncover both climate and anthropogenic influences in
this ecosystem. Additionally, information about the extent,
location and time of burned areas (BA) over Cerrado is especially
useful to a wide range of users, from government agencies,
research groups and ecologists, to fire managers and NGOs.
Instruments on-board satellites are the only available operational
means to collect BA data at appropriated spatial and temporal
scales and in a cost-effective way. Several global BA products
derived from remote sensed information have been developed over
the last years using a variety of techniques based on different
spatial, spectral and temporal resolutions. Although presenting
similar inter-annual variability, there are marked differences
among the products both in magnitude and location of the area
burnt. The development of regional algorithms which take into
account local characteristics such as vegetation type, soil and
climate is therefore an added value to the existing information.
We present a monthly BA product (AQM) for Brazil based on
information from MODIS 1km. The algorithm was specifically
designed for ecosystems in Brazil and the procedure represents the
first initiative of an automated method for BA monitoring using
remote sensing information in the country. The product relies on
an algorithm that takes advantage of the ability of MIR
reflectances to discriminate BA. Validation over Cerrado biome
indicates that the product is in accordance with BA maps from
reference data, making the product suitable for applications in
fire emission studies and ecosystem management. The AQM regional
database covers the 11-year period 2005-2015 over Cerrado and
allows analyzing the overall temporal and spatial distribution
patterns of BA for the last decade. The highest monthly mean
amount is observed in September, followed by October, and March
presents the lowest amount. The most severe year is 2007, followed
by 2005 and 2010; 2006 and 2009 are the years with less area
burned, followed by 2008. The spatial pattern of BA shows that the
north region of Cerrado presents the highest frequency of
occurrence. The intra and inter-annual variability of BA over
Cerrado are closely related to variability of precipitation but it
is worth emphasizing that, despite the major role played by
climate conditions, the human factor has also a prominent role on
fire dynamics in this region and cannot be disregarded.",
conference-location = "Vienna, Austria",
conference-year = "17-22 Apr.",
targetfile = "libonati_spatial.pdf",
urlaccessdate = "09 maio 2024"
}