**3. Analysis of the active fires in the state of São Paulo**

The fire detection method employed by INPE (National Institute for Space Research) is a digital non supervised clustering algorithm which selects pixels as burning if the AVHRR radiometric temperature exceeds 46°C [28], using the images of several satellites, which operate in the band between 3.7 and 4.1 μm, selecting the pixels (resolution elements) with higher temperature. Data from the morning overpass are used to identify fire pixels. Fire counts identified by INPE are provided in a weekly base in grid format. They consist of a 7 day sum at 0.5-degree increments from 17°N to 40°S and 85°W to 34.5°W [29]. Data of active fires obtained through the NOAA-12 satellite were used, for the period between January, 1999 and July, 2007 at the studied region (pixels located over the state of São Paulo).

Through the spatial distribution of fires (Figure 2) obtained by the NOAA-12 satellite, it is possible to notice that the regions with the highest number of outbreaks of fires during the selected period coincides with sugarcane crops, especially during the winter and spring, periods with the largest number of outbreaks. The central-northern, central, central-eastern and central-western regions of the São Paulo State are highlighted as the areas where there is a high rate of outbreaks due to the presence of the sugarcane culture.

The results of Figure 2 and Table 1 indicate that the winter was the season with the highest number of fire outbreaks during the period, when approximately 15,000 outbreaks were recorded, followed closely by spring with approximately 12,000 outbreaks, while in summer only 2000 spots were registered. During the harvest period, which occurs from May to November, the plantation areas are burnt a few hours before the manual cutting, and this period coincides with the dry season in Southeastern Brazil. From December to April the wet pattern prevails and there are only few activities for burning around the state of São Paulo. The number of fire outbreaks tends to increase from the month of March, achieving its maximum in the trimester July-August-September. The peak of the burning season (August) coincides with the least amount of precipitation in the region, while the reverse is also found, burning minimum in the months with maximum precipitation (December and January).

254 Atmospheric Aerosols – Regional Characteristics – Chemistry and Physics

contribution based in the Bucholtz's approach [27].

**3. Analysis of the active fires in the state of São Paulo** 

However, it is known that the LR depends on several physical-chemical parameters inherent to the aerosols being inspected, such as the aerosol refractive index, size and shape distribution of the aerosol particles [25]. To derive the appropriate "correct" values of the vertical profile of aerosol backscatter coefficient in the lower troposphere an iterative inversion approach was used (by "tuning" the LR values), based on the intercomparison of the AOD values derived by the Lidar and a collocated AERONET sunphotometer [26], assuming the absence of stratospheric aerosols and that the PBL is homogeneously mixed between ground and the altitude of 300 m, where the Lidar overlap factor is close to 1. Once the correct values of the vertical profile of the aerosol backscatter coefficient were derived (when the difference between the AODs derived by sunphotometer and Lidar was less than 10%), the Klett's inversion technique was reapplied, using the appropriate LR values, to retrieve the final values of the vertical profiles of the backscatter and extinction coefficient at 532 nm. The vertical profiles of pressure and temperature measured by radiosoundings launched twice a day, at 12 UTC and 00 UTC in a distance about 10 km from the place where the MSP-Lidar system is located, are used in order to obtain the molecular

The fire detection method employed by INPE (National Institute for Space Research) is a digital non supervised clustering algorithm which selects pixels as burning if the AVHRR radiometric temperature exceeds 46°C [28], using the images of several satellites, which operate in the band between 3.7 and 4.1 μm, selecting the pixels (resolution elements) with higher temperature. Data from the morning overpass are used to identify fire pixels. Fire counts identified by INPE are provided in a weekly base in grid format. They consist of a 7 day sum at 0.5-degree increments from 17°N to 40°S and 85°W to 34.5°W [29]. Data of active fires obtained through the NOAA-12 satellite were used, for the period between January,

1999 and July, 2007 at the studied region (pixels located over the state of São Paulo).

is a high rate of outbreaks due to the presence of the sugarcane culture.

Through the spatial distribution of fires (Figure 2) obtained by the NOAA-12 satellite, it is possible to notice that the regions with the highest number of outbreaks of fires during the selected period coincides with sugarcane crops, especially during the winter and spring, periods with the largest number of outbreaks. The central-northern, central, central-eastern and central-western regions of the São Paulo State are highlighted as the areas where there

The results of Figure 2 and Table 1 indicate that the winter was the season with the highest number of fire outbreaks during the period, when approximately 15,000 outbreaks were recorded, followed closely by spring with approximately 12,000 outbreaks, while in summer only 2000 spots were registered. During the harvest period, which occurs from May to November, the plantation areas are burnt a few hours before the manual cutting, and this period coincides with the dry season in Southeastern Brazil. From December to April the wet pattern prevails and there are only few activities for burning around the

**Figure 2.** Outbreaks of fire during the period from January, 1999 to July, 2007: (a) summer; (b) autumn; (c) winter and (d) spring.

The great quantity of active fires during the dry season, combined with low humidity in the region, can cause several health problems in the people living near the plantations and burnings of sugarcane. The association between internal and external exposure to the biomass smoke and its effects on health has been reported in some areas of Asia and India [30, 31]. Carbon deposition in the lungs occurred consistently in patients exposed to biomass burning [32]. Unlike most regions where external biomass burning is sporadic, the biomass burning in the region of São Paulo is a common and scheduled activity, due to the areas with sugarcane crops.

The seasonal and interannual variation of the active fires in the State of São Paulo can be analyzed in Table 1. Through this table it can be noticed that the highest value of active fires occurred in 1999 followed by a decrease until 2001. From this year on, a slightly constant tendency of decrease in numbers of forest fires was noticed, except for the two last measured years (2005 and 2006). It is important to highlight that the measurements and monitoring in 2007 only goes until July, before the burning maximum period – August, September and October (as seen on Figure 2), probably due to technical problems with the satellite acquisitions. Besides the annual active total fires, the seasonal variation of the

number of fires for each year can also be observed in the Table 1, where it can be clearly seen that the maximum active fires seasons are winter and spring, followed by autumn and summer in almost every analyzed year, except for the years of 2000 and 2004, where the total active fires detected were greater in spring and summer than in winter and autumn, respectively. For the year of 2004, it was observed that the rain rates for these two periods were normal according to the climatology, without an obvious reason for that variation. It can also be seen the difference between the number of fires during the maximum fires season (winter) and the minimum (summer), reaching a value 3 times higher at some periods.

Impacts of Biomass Burning in the Atmosphere

of the Southeastern Region of Brazil Using Remote Sensing Systems 257

maximum concentration of aerosols tends to be found near the surface, hindering the use

**Figure 3.** MODIS Maximum Aerosol Optical Depth over southeast of Brazil (a) 2003-2006 and (b) 2007- 2010. Years: (a) 2003 (orange); 2004 (blue); 2005 (green); 2006 (red) and (b) 2007 (orange); 2008 (blue);

Both the Amazon and Midwestern Brazil are great sources of aerosols into the atmosphere, and the maximum daily value of AOD found on the Southeastern region can be compared to these areas. In all of these regions the largest occurrences of elevated AOD are during the winter and spring in the Southern Hemisphere. The occurrence of higher AOD levels during this period for the Northern and Central region is basically due to the highest amount of biomass burning in the region (dry season). In Brazil, biomass burning in the Amazon region occurs with greater intensity during the dry season (July - October) and primarily affects the ecosystems of forest, pasture and Cerrado [34]. The small amount of precipitation and the increased atmospheric stability at this period contributes to the persistence of aerosols in the atmosphere, as well as a higher transport of biomass burning aerosols from

Analyzing the 146 days with maximum AOD higher than the 95th percentile (1.143) found between 2003-2010 it is observed that approximately 82% of the cases occurred during the spring in the Southeastern hemisphere (Figure 4). When considering the period corresponding for spring and winter, the same has 144 days with values above the 95th percentile (99% approximately). The emissions from these fires have significant impacts on the concentration of gases and aerosols. The results presented here emphasize the importance of monitoring the aerosols for the period between June and October. The values recorded during summer and fall may be related, among other reasons, with burning events in the Southeastern region and transport of biomass burning aerosols from distant regions, such as the Amazon and MidWestern region of Brazil regions, Bolivia, the Northern portion of Argentina and the North region of South America. Another possible explanation for the

values found are the urban aerosols emitted into the studied area.

of remote sensing for this task.

2009 (green); 2010 (red)

the Amazon region (Section 6).


**Table 1.** Active fires measured by the NOAA-12 satellite for the State of São Paulo – Brazil
