**4.1. Agro-industrial rural regions**

Seasonal variability in the major soluble ion composition of atmospheric particulate matter in the principal sugar cane growing region of central São Paulo State indicates that preharvest burning of sugar cane plants is an important influence on the regional-scale aerosol chemistry [34]. The size-distributed composition of ambient aerosols is used to explore seasonal differences in particle chemistry, and to show that dry deposition fluxes of soluble species, including important plant nutrients, increase during periods of biomass (sugar cane trash) burning [6, 10].

Concentrations of trace gases and aerosols were determined at six measurement sites of a regional network in São Paulo State (blue circles in Figure 6), installed in rural areas including the State´s central agricultural zone and the eastern coast [11] as part of an experimental research project to determine the anthropogenic component of nutrient deposition. The measurements were made over 12 months during 2008/2009 (one week of continuous sampling per month). Aerosols were collected onto 47 mm diameter Teflon filters using active samplers, and trace gases (NO2, NH3, HNO3 and SO2) were sampled using diffusion-based devices. The soluble ions NO3- , NH4+, PO43-, SO42-, Cl- , K+, Na+, Mg2+ and Ca2+ were analyzed in aqueous extracts of the aerosol filters, using ion chromatography. NO2, HNO3 and SO2 were similarly determined as NO2- , NO3 and SO42-, following aqueous extraction of the collection media. NH3 was determined using a colorimetric technique. Identification and quantification of nutrient sources was achieved using principal component analysis (PCA) followed by multiple linear regression analysis (MLRA) applied to the chemical data. Dry deposition fluxes were estimated using the measured atmospheric concentrations together with dry deposition velocities of gases and aerosols to different surface types, including tropical forest, savannah, sugar cane, pine, eucalyptus, orange, coffee, pasture and water. The annual cycle in deposition, to a sugar cane surface, of reactive nitrogen and sulphur in the gaseous, aerosol and dissolved phases is illustrated in Figure 7.

280 Atmospheric Aerosols – Regional Characteristics – Chemistry and Physics

measurements are available are marked with x.

**4.1. Agro-industrial rural regions** 

trash) burning [6, 10].

**4. Chemical composition of aerosols** 

using diffusion-based devices. The soluble ions NO3-

NO2, HNO3 and SO2 were similarly determined as NO2-

**Figure 6.** Aerosol monitoring sites in the State of São Paulo (except CETESB network) and 240 km ranges of IPMet's radars in Presidente Prudente (PPR) and Bauru (BRU). Sites from where lidar

Seasonal variability in the major soluble ion composition of atmospheric particulate matter in the principal sugar cane growing region of central São Paulo State indicates that preharvest burning of sugar cane plants is an important influence on the regional-scale aerosol chemistry [34]. The size-distributed composition of ambient aerosols is used to explore seasonal differences in particle chemistry, and to show that dry deposition fluxes of soluble species, including important plant nutrients, increase during periods of biomass (sugar cane

Concentrations of trace gases and aerosols were determined at six measurement sites of a regional network in São Paulo State (blue circles in Figure 6), installed in rural areas including the State´s central agricultural zone and the eastern coast [11] as part of an experimental research project to determine the anthropogenic component of nutrient deposition. The measurements were made over 12 months during 2008/2009 (one week of continuous sampling per month). Aerosols were collected onto 47 mm diameter Teflon filters using active samplers, and trace gases (NO2, NH3, HNO3 and SO2) were sampled

and Ca2+ were analyzed in aqueous extracts of the aerosol filters, using ion chromatography.

extraction of the collection media. NH3 was determined using a colorimetric technique. Identification and quantification of nutrient sources was achieved using principal component analysis (PCA) followed by multiple linear regression analysis (MLRA) applied

, NH4+, PO43-, SO42-, Cl-

, NO3-

, K+, Na+, Mg2+

and SO42-, following aqueous

**Figure 7.** Annual cycle in deposition fluxes to a sugar cane surface of: **(a)** sulphur in gaseous (S(g)), aerosol (S(a)) and rainwater (S(p)) phases; **(b)** nitrogen in gaseous (N(g)), aerosol (N(a)) and rainwater (N(p)) phases. Primary y-axes: gas and rainwater; secondary y-axes: aerosol. Data for Araraquara.

The sugar cane industry has a major impact on air quality and the characteristics of the atmospheric aerosol. During the dry season (May to October), the burning of the cane, a prerequisite of manual harvesting, has for many years resulted in very large emissions of pollutants, including high carbon content aerosols. These particles contain water-soluble organic carbon (WSOC), anions (sulphates, nitrates and chlorides), cations (potassium, ammonium, calcium, magnesium, sodium), black carbon (BC), insoluble organic carbon and trace metals. Carbonaceous material comprises the bulk of the aerosol mass, especially in fine particles [5-7, 35-39]. In 2004, the annual emission of nitrogen oxides (NOx) from sugar cane burning in Sao Paulo State was in excess of 45 Gg.N [40]. This is not only indicative of the scale of the emissions, but also of their potential for formation of secondary aerosols (containing nitrates, amongst other components).

In 2011, annual mean PM10 concentrations measured at automatic monitoring stations in the agro-industrial interior of São Paulo State were in the range 23-91 g.m-3, with the highest

values at locations affected by primary emissions from ceramics industries (Figure 5). At sites in the sugar cane production areas, annual mean PM10 concentrations were in the range of 32-41 g.m-3 [9]. The data revealed no obvious trends in PM10 concentrations during the period 2002-2011 (Figure 5).

Review of Aerosol Observations by Lidar and Chemical Analysis in the State of São Paulo, Brazil 283

, HCOO-

, CH3COO-

and Cl-

, but

in twelve size fractions, and used calculation of ion equivalent balances to show that during burning periods, the smaller particles (Aitken and accumulation modes) were more acidic,

insufficient NH4+ and K+ to achieve neutrality. Larger particles showed an anion deficit due to the presence of unmeasured ions, and comprised re-suspended dusts modified by accumulation of nitrate, chloride and organic anions. Increases of re-suspended particles during the burning season were attributed to release of earlier deposits from the surfaces of burning vegetation, as well as increased vehicle movement on unsealed roads. During the winter months, the relative contribution of combined emissions from road transport and industry diminished due to increased emissions from biomass combustion and other

**Figure 8.** Comparison of aerosol composition in the Araraquara region during winter (biomass

In separate work, biomass-burning aerosols were found to contribute around 60 and 25% of the mass of fine and coarse aerosols, respectively, in the Piracicaba sugar cane growing region [7]. A high proportion of the elements K, S, Cl, Br, Fe and Si in aerosols has been linked to biomass burning [45], indicative of both a combustion component (emissions of K,

burning) and summer (non-burning) periods: **(a)** Coarse particles; **(b)** fine particles.

S, Cl and Br) and a suspended soil dust component (emissions of Fe and Si).

containing higher concentrations of SO42-, C2O42-, NO3-

activities specifically associated with the harvest period.

A proportion of the primary material emitted during sugar cane fires is in the form of the large ash fragments notorious for causing domestic soiling problems in the region. During 1995-1996, CETESB investigated deposition rates, and measured the concentrations of PAHs, PCBs, dioxins and furans. The sedimented material was collected during the harvest period using plastic funnels lined with polyurethane foam, positioned near to plantations and in the urban area of the city of Araraquara. Samples were also collected in parallel using a high volume filter-based sampler. Levels of PCBs were in the range 4-12 ng.m-3, and showed no association with levels of carbonaceous material derived from the fires. Deposition fluxes of the dioxins and furans were in the range 1-17 pg.m-2.day-1, and were higher in greater proximity to plantations, indicating that sugar cane burning was a source of these compounds. The PAHs were found in two distinct groups. Naphthalene, fluorene, phenanthrene, anthracene, fluoranthene, and pyrene were present at concentrations exceeding 30 ng.m-3, while acenaphthene, chrysene, benzo(a)fluoranthene, benzo(k)fluoranthene, benzo(a)pyrene, dibenzo(a,h)anthracene, benzo(g,h,i)perylene and indeno(1,2,3,c,d)pyrene were found at up to 21 ng.m-3. Concentrations were always higher during the harvest period [41].

The presence of PAHs in ash from sugar cane fires was also reported by Zamperlini *et al.* [42, 43]. In the PM10 fraction, it was found that the most abundant polycyclic aromatic hydrocarbons were phenanthrene and fluoranthrene, and the least abundant was anthracene [44]. Cluster analysis of the total PAH concentrations for each day of sampling, and the corresponding meteorological data, suggested that concentrations of PAHs were independent of climatological conditions or season of the year. Vehicular sources were identified during both dry and wet seasons, although sugar cane burning emissions were the dominant source during the dry season.

Sugar cane burning is a major source of acidic gases that contribute to the formation of secondary aerosols. In Araraquara, Da Rocha *et al.* [36] reported concentrations of 9,0 ppb (HCOOH), 1,3 ppb (CH3COOH), 4,9 ppb (SO2), 0,3 ppb (HCl) and 0,5 ppb (HNO3). Extremely high concentrations of these gases were measured in the plumes downwind of sugar cane fires: 1160-4230 ppb (HCOOH); 360-1750 ppb (CH3COOH); 10-630 ppb (SO2); 4- 210 ppb (HCl); and 14-90 ppb (HNO3). Highest levels of SO2, HCl and HNO3 in Araraquara were measured during the harvest period, with peak concentrations in the evening (the time of the fires).

The distribution of soluble ionic material between fine (<3,5 μm) and coarse (>3,5 μm) aerosol fractions was determined by Allen *et al.* [5], who measured the ions HCOO- , CH3COO- , C2O42-, SO42-, NO3- , Cl- , Na+, K+, NH4+, Mg2+ and Ca2+. The fine and coarse particles showed acidic and basic properties, respectively, and concentrations of all major ions increased significantly during the dry season (Figure 8). Da Rocha *et al.* [6] collected aerosols in twelve size fractions, and used calculation of ion equivalent balances to show that during burning periods, the smaller particles (Aitken and accumulation modes) were more acidic, containing higher concentrations of SO42-, C2O42-, NO3- , HCOO- , CH3COO and Cl- , but insufficient NH4+ and K+ to achieve neutrality. Larger particles showed an anion deficit due to the presence of unmeasured ions, and comprised re-suspended dusts modified by accumulation of nitrate, chloride and organic anions. Increases of re-suspended particles during the burning season were attributed to release of earlier deposits from the surfaces of burning vegetation, as well as increased vehicle movement on unsealed roads. During the winter months, the relative contribution of combined emissions from road transport and industry diminished due to increased emissions from biomass combustion and other activities specifically associated with the harvest period.

282 Atmospheric Aerosols – Regional Characteristics – Chemistry and Physics

period 2002-2011 (Figure 5).

during the harvest period [41].

of the fires).

CH3COO-

the dominant source during the dry season.

, C2O42-, SO42-, NO3-

, Cl-

values at locations affected by primary emissions from ceramics industries (Figure 5). At sites in the sugar cane production areas, annual mean PM10 concentrations were in the range of 32-41 g.m-3 [9]. The data revealed no obvious trends in PM10 concentrations during the

A proportion of the primary material emitted during sugar cane fires is in the form of the large ash fragments notorious for causing domestic soiling problems in the region. During 1995-1996, CETESB investigated deposition rates, and measured the concentrations of PAHs, PCBs, dioxins and furans. The sedimented material was collected during the harvest period using plastic funnels lined with polyurethane foam, positioned near to plantations and in the urban area of the city of Araraquara. Samples were also collected in parallel using a high volume filter-based sampler. Levels of PCBs were in the range 4-12 ng.m-3, and showed no association with levels of carbonaceous material derived from the fires. Deposition fluxes of the dioxins and furans were in the range 1-17 pg.m-2.day-1, and were higher in greater proximity to plantations, indicating that sugar cane burning was a source of these compounds. The PAHs were found in two distinct groups. Naphthalene, fluorene, phenanthrene, anthracene, fluoranthene, and pyrene were present at concentrations exceeding 30 ng.m-3, while acenaphthene, chrysene, benzo(a)fluoranthene, benzo(k)fluoranthene, benzo(a)pyrene, dibenzo(a,h)anthracene, benzo(g,h,i)perylene and indeno(1,2,3,c,d)pyrene were found at up to 21 ng.m-3. Concentrations were always higher

The presence of PAHs in ash from sugar cane fires was also reported by Zamperlini *et al.* [42, 43]. In the PM10 fraction, it was found that the most abundant polycyclic aromatic hydrocarbons were phenanthrene and fluoranthrene, and the least abundant was anthracene [44]. Cluster analysis of the total PAH concentrations for each day of sampling, and the corresponding meteorological data, suggested that concentrations of PAHs were independent of climatological conditions or season of the year. Vehicular sources were identified during both dry and wet seasons, although sugar cane burning emissions were

Sugar cane burning is a major source of acidic gases that contribute to the formation of secondary aerosols. In Araraquara, Da Rocha *et al.* [36] reported concentrations of 9,0 ppb (HCOOH), 1,3 ppb (CH3COOH), 4,9 ppb (SO2), 0,3 ppb (HCl) and 0,5 ppb (HNO3). Extremely high concentrations of these gases were measured in the plumes downwind of sugar cane fires: 1160-4230 ppb (HCOOH); 360-1750 ppb (CH3COOH); 10-630 ppb (SO2); 4- 210 ppb (HCl); and 14-90 ppb (HNO3). Highest levels of SO2, HCl and HNO3 in Araraquara were measured during the harvest period, with peak concentrations in the evening (the time

The distribution of soluble ionic material between fine (<3,5 μm) and coarse (>3,5 μm) aerosol fractions was determined by Allen *et al.* [5], who measured the ions HCOO-

showed acidic and basic properties, respectively, and concentrations of all major ions increased significantly during the dry season (Figure 8). Da Rocha *et al.* [6] collected aerosols

, Na+, K+, NH4+, Mg2+ and Ca2+. The fine and coarse particles

,

**Figure 8.** Comparison of aerosol composition in the Araraquara region during winter (biomass burning) and summer (non-burning) periods: **(a)** Coarse particles; **(b)** fine particles.

In separate work, biomass-burning aerosols were found to contribute around 60 and 25% of the mass of fine and coarse aerosols, respectively, in the Piracicaba sugar cane growing region [7]. A high proportion of the elements K, S, Cl, Br, Fe and Si in aerosols has been linked to biomass burning [45], indicative of both a combustion component (emissions of K, S, Cl and Br) and a suspended soil dust component (emissions of Fe and Si).

In a study reported in [38], elemental analysis of individual and bulk aerosols collected in rural areas was followed by evaluation of the data using statistical hierarchical clustering, which revealed the contributions of two different types of carbonaceous material (biogenic and carbon-rich) and two aluminosilicate fractions (pure or mixed with carbon). These findings contrasted with the findings of similar work in the atmosphere of São Paulo city, where hierarchical clustering analysis revealed the presence of metal compounds, siliconrich particles, sulphates, carbonates, chlorides, organics and biogenic particles [46]. This reflects the very different characteristics of the aerosols found in the two regions.

Review of Aerosol Observations by Lidar and Chemical Analysis in the State of São Paulo, Brazil 285


) increased by between two and

,

Emissions of reactive nitrogen compounds are of concern due to their influence on both atmospheric acidity (production of HNO3 from reactions involving NO2) and the formation of photochemical oxidants such as ozone and peroxyacetyl nitrate (PAN). Reactions of acidic species with ammonia generate ammonium sulphates and nitrates, mainly in the long-lived accumulation mode size fraction. Deposition of reactive nitrogen can cause eutrophication of water bodies, as well as the release of trace metals in soils. Machado *et al.* [47] found that emissions of reactive nitrogen during sugar cane burning, in the forms of NH3, NOx and particulate nitrate and ammonium, were equivalent to 35% of the annual fertilizer-N application. The concentrations of nitrogen oxides showed a positive association with the number of fires, reflecting the importance of biomass burning as a major emission source, and mean concentrations of NOx in the dry season were twice those in the wet season. During the dry season, biomass burning was the main source of NH3, with other sources (wastes, soil, biogenic) predominant during the wet

cane burning in a planted area of about 2,2×106 ha were 11,0, 1,1, 0,2 and 1,2 Gg.N.yr-1,

The sources, atmospheric transport and reactions of the main inorganic reactive nitrogen (Nr) species (NO2, NH3, HNO3 and aerosol nitrate and ammonium) were investigated in a study conducted over a period of one year at six sites distributed across an area of about 130,000 km2 in São Paulo State [11]. Oxidized forms of nitrogen were estimated to account for about 90% of dry deposited Nr, due to the emissions of nitrogen oxides from biomass burning and road transport. NO2-N was important closer to urban areas; however, HNO3-N was the largest individual component of dry deposited Nr. A simple mathematical model was developed to enable determination of total Nr dry deposition from knowledge of NO2 concentrations. The model, whose error ranged from <1% to 29%, provided a new tool for

The sugar cane burning emissions radically alter the chemistry of precipitation water. Coelho *et al.* [122] found that concentrations of soluble ions (K+, Na+, NH4+, Ca2+, Mg2+, Cl-

six-fold during the harvest period. Principal component analysis revealed three main sources of the material in rainwater: biomass burning and soil dust re-suspension (52% of the total variance), secondary aerosols (26%) and vehicular emissions (10%). The biomass burning component diminished in the summer (non-burning period), when there was a relative increase in the importance of road transport/industrial emissions. The volumeweighted mean concentrations of ammonium (23,4 mol.L-1) and nitrate (17,5 mol.L-1) in rainwater samples collected during the harvest period were similar to those found in rainwater from São Paulo city, which emphasized the importance of including rural agroindustrial emissions in regional-scale atmospheric chemistry and transport models. There was evidence of a biomass-burning source throughout the year, which suggests that vegetation fires may continue to emit aerosols and their precursor gases, even after sugar

, C2O42- and HCO3-

, HCOO-

season. The estimated emission fluxes of NO2-N, NH3-N, NO3-

the mapping of reactive nitrogen deposition.

, PO43-, CH3COO-

respectively.

NO3-

, SO42-, F-

cane burning is phased out.

Da Rocha *et al.* [6] showed that dry deposition fluxes of important plant nutrients increased during the sugar cane burning season. During this period, the fine fraction aerosol was more acidic and contained elevated concentrations of SO42-, C2O42-, NO3- , HCOO- , CH3COO and Cl- , but insufficient NH4+ and K+ to achieve neutrality. Larger particles consisted of re-suspended dust, modified by inclusion of nitrate, chloride and organic anions. The increases in annual particulate dry deposition fluxes due to higher fluxes during the sugar cane harvest were 44,3% (NH4+), 42,1% (K+), 31,8% (Mg2+), 30,4% (HCOO- ), 12,8% (Cl- ), 6,6% (CH3COO- ), 5,2% (Ca2+), 3,8% (SO42-) and 2,3% (NO3- ). The contributions of dry deposition to total deposition (including precipitation scavenging, excluding gaseous dry deposition) were 31% (Na+), 8% (NH4+), 26% (K+), 63% (Mg2+), 66% (Ca2+), 32% (Cl- ), 33% (NO3- ) and 36% (SO42-).

Deposition rates of aerosol nutrient species to a range of natural and agricultural surfaces were reported in [10], using a size-segregated particle dry deposition model. Fluxes greatly exceeded those expected under pristine conditions, with deposition to tropical forest found to have increased by factors of 12,2 (NO3- ), 6,2 (PO43-) and 2,6 (K+) (Figure 9). Source apportionment using principal component analysis (PCA) and multiple linear regression analysis (MLRA) revealed that in central São Paulo State, biomass burning, products of secondary reactions and soil dust re-suspension contributed 43%, 31% and 21% of PM2.5 mass, respectively. Re-suspension and biomass burning contributed 22% and 19%, respectively, to PM10 mass, and re-suspension accounted for approximately half the mass of coarse particles. At least 40% of NO3- -N, 20% of phosphorus and 55% of potassium deposited originated from agriculture-related emissions.

**Figure 9.** Graph showing the present-day increase in aerosol dry deposition rates to a tropical forest surface, compared to deposition rates estimated for pristine conditions.

Emissions of reactive nitrogen compounds are of concern due to their influence on both atmospheric acidity (production of HNO3 from reactions involving NO2) and the formation of photochemical oxidants such as ozone and peroxyacetyl nitrate (PAN). Reactions of acidic species with ammonia generate ammonium sulphates and nitrates, mainly in the long-lived accumulation mode size fraction. Deposition of reactive nitrogen can cause eutrophication of water bodies, as well as the release of trace metals in soils. Machado *et al.* [47] found that emissions of reactive nitrogen during sugar cane burning, in the forms of NH3, NOx and particulate nitrate and ammonium, were equivalent to 35% of the annual fertilizer-N application. The concentrations of nitrogen oxides showed a positive association with the number of fires, reflecting the importance of biomass burning as a major emission source, and mean concentrations of NOx in the dry season were twice those in the wet season. During the dry season, biomass burning was the main source of NH3, with other sources (wastes, soil, biogenic) predominant during the wet season. The estimated emission fluxes of NO2-N, NH3-N, NO3- -N and NH4+-N from sugar cane burning in a planted area of about 2,2×106 ha were 11,0, 1,1, 0,2 and 1,2 Gg.N.yr-1, respectively.

284 Atmospheric Aerosols – Regional Characteristics – Chemistry and Physics

In a study reported in [38], elemental analysis of individual and bulk aerosols collected in rural areas was followed by evaluation of the data using statistical hierarchical clustering, which revealed the contributions of two different types of carbonaceous material (biogenic and carbon-rich) and two aluminosilicate fractions (pure or mixed with carbon). These findings contrasted with the findings of similar work in the atmosphere of São Paulo city, where hierarchical clustering analysis revealed the presence of metal compounds, siliconrich particles, sulphates, carbonates, chlorides, organics and biogenic particles [46]. This

Da Rocha *et al.* [6] showed that dry deposition fluxes of important plant nutrients increased during the sugar cane burning season. During this period, the fine fraction aerosol was more

but insufficient NH4+ and K+ to achieve neutrality. Larger particles consisted of re-suspended dust, modified by inclusion of nitrate, chloride and organic anions. The increases in annual particulate dry deposition fluxes due to higher fluxes during the sugar cane harvest were

(including precipitation scavenging, excluding gaseous dry deposition) were 31% (Na+), 8%

Deposition rates of aerosol nutrient species to a range of natural and agricultural surfaces were reported in [10], using a size-segregated particle dry deposition model. Fluxes greatly exceeded those expected under pristine conditions, with deposition to tropical forest found

apportionment using principal component analysis (PCA) and multiple linear regression analysis (MLRA) revealed that in central São Paulo State, biomass burning, products of secondary reactions and soil dust re-suspension contributed 43%, 31% and 21% of PM2.5 mass, respectively. Re-suspension and biomass burning contributed 22% and 19%, respectively, to PM10 mass, and re-suspension accounted for approximately half the mass of

**Figure 9.** Graph showing the present-day increase in aerosol dry deposition rates to a tropical forest

), 33% (NO3-

, HCOO-

) and 36% (SO42-).

), 6,2 (PO43-) and 2,6 (K+) (Figure 9). Source


), 12,8% (Cl-

). The contributions of dry deposition to total deposition

, CH3COO-

), 6,6% (CH3COO-

 and Cl- ,

), 5,2%

reflects the very different characteristics of the aerosols found in the two regions.

acidic and contained elevated concentrations of SO42-, C2O42-, NO3-

44,3% (NH4+), 42,1% (K+), 31,8% (Mg2+), 30,4% (HCOO-

(NH4+), 26% (K+), 63% (Mg2+), 66% (Ca2+), 32% (Cl-

to have increased by factors of 12,2 (NO3-

coarse particles. At least 40% of NO3-

deposited originated from agriculture-related emissions.

surface, compared to deposition rates estimated for pristine conditions.

(Ca2+), 3,8% (SO42-) and 2,3% (NO3-

The sources, atmospheric transport and reactions of the main inorganic reactive nitrogen (Nr) species (NO2, NH3, HNO3 and aerosol nitrate and ammonium) were investigated in a study conducted over a period of one year at six sites distributed across an area of about 130,000 km2 in São Paulo State [11]. Oxidized forms of nitrogen were estimated to account for about 90% of dry deposited Nr, due to the emissions of nitrogen oxides from biomass burning and road transport. NO2-N was important closer to urban areas; however, HNO3-N was the largest individual component of dry deposited Nr. A simple mathematical model was developed to enable determination of total Nr dry deposition from knowledge of NO2 concentrations. The model, whose error ranged from <1% to 29%, provided a new tool for the mapping of reactive nitrogen deposition.

The sugar cane burning emissions radically alter the chemistry of precipitation water. Coelho *et al.* [122] found that concentrations of soluble ions (K+, Na+, NH4+, Ca2+, Mg2+, Cl- , NO3- , SO42-, F- , PO43-, CH3COO- , HCOO- , C2O42- and HCO3- ) increased by between two and six-fold during the harvest period. Principal component analysis revealed three main sources of the material in rainwater: biomass burning and soil dust re-suspension (52% of the total variance), secondary aerosols (26%) and vehicular emissions (10%). The biomass burning component diminished in the summer (non-burning period), when there was a relative increase in the importance of road transport/industrial emissions. The volumeweighted mean concentrations of ammonium (23,4 mol.L-1) and nitrate (17,5 mol.L-1) in rainwater samples collected during the harvest period were similar to those found in rainwater from São Paulo city, which emphasized the importance of including rural agroindustrial emissions in regional-scale atmospheric chemistry and transport models. There was evidence of a biomass-burning source throughout the year, which suggests that vegetation fires may continue to emit aerosols and their precursor gases, even after sugar cane burning is phased out.
