1. INTRODUCTION AND OBJECTIVES

The Brazilian Cerrado represents the largest area of savanna in America, spanning approximately 2 million km² in Central Brazil (^{Castro et al., 2017}). Cerrado is characterized by different types of vegetation - pastures, open tree canopies, and dense forests - where spatial distribution is determined by many factors such as soil type and topography, and fire frequency and intensity (^{Batista et al., 2018}; ^{Meira et al., 2017}).

Soil is a basic component of the forest ecosystem and it is subject to changes by fire. However, its effects - which may alter edaphic factors - are still scarcely studied. According to ^{Resende et al. (2017}), although the Cerrado ecosystem is adapted to fire, it can lead to loss of nutrients, compaction, and erosion - a problem that affects vast areas of land. Conversely, when correctly used, controlled burning is known to be a useful tool in preventing high-magnitude fire events, since dry biomass accumulation contributes to increase the occurrence of fire (^{Batista et al., 2013}; ^{Camargos et al., 2015}).

Effects on the chemical composition of the soil chiefly results from burn severity, which consists of peak temperatures and duration of the fire (^{Abraham et al., 2018}; ^{Certini, 2005}). Thus, soil transformation is directly proportional to the intensity of the fires in the area (^{Lorenzon et al., 2014}).

Following the fire, essential nutrients, such as phosphorus, magnesium, calcium, and potassium, can be released by the ashes (^{Alcañiz et al., 2018}). Therefore, burning may favor an increase in fertility - although ephemeral, it is crucial for plant regeneration. However, according to ^{Knicker (2007}), the benefits of nutrient mineralization catalyzed by burning can be depleted in the medium term when burning is carried out during the dry season. This is because rain may cause the leaching of nutrients, which results in lower concentrations that may even be inferior to the ones observed in unburned soils.

Although it is a common practice in Brazil, the need for studies whose authors investigate the effects of burning on soil properties is evident, in such a way to improve the management of areas reserved for livestock activities aiming at reducing potential soil damages and contributing to its conservation.

The state of Tocantins represents an important agricultural border area in the country. Few studies have been carried out in Cerrado areas of municipalities in Tocantins measuring the impacts of burning on the chemical properties of the soil. This practice has also been widely used by most producers and pastoralists in the region.

Taking this into consideration, we aimed to determine the immediate effects of fire intensity on the superficial layer of the soil in a Cerrado sensu stricto area in the state of Tocantins, Brazil.

2. MATERIALS AND METHODS

The study was conducted in Fazenda Verdes Mares, in the municipality of Sucupira - state of Tocantins - located at a latitude of 11° 59’ 36” South and longitude of 48° 58’ 15” West, at 257 meters above sea level. According to the Köppen’s classification, it is characterized by a tropical wet-dry climate (Aw) (tropical with wet summer and dry winter), and the average annual precipitation is 1,500 mm. The soil is heterogeneous, with variations among Red-Yellow Latosol, Plinthosol, and Cambisol. Nevertheless, the soil of the experimental area has been described as Red-Yellow Latosol. It is worth noting that the study farm holds soy crops and pastures for cattle; however, the parcels of land composing the experimental area were unchanged, i.e., composed of natural vegetation. The study area is classified according to ^{Ribeiro & Walter’s (1998}) category of Cerrado sensu stricto. This savanna-like formation is chiefly characterized by an herb layer, predominantly covered by grasses, and a layer of trees and bushes, with irregular and twisted branches and coverage between 10 and 60% (^{Eiten, 1994}).

Prescribed burning was carried out by the team of Centro de Monitoramento Ambiental e Manejo do Fogo (CeMAF) in October 2015 between 11:05 a.m. and 6:05 p.m. A total of 14 parcels of land with 200 m² in area - 10-m wide and 20-m long - was delimited. The area was marked at its length in 2-meter intervals for measurements of the fire propagation speed (m.s^-1) and height. Before and after burning, the combustible material was randomly collected from each parcel in order to determine the available amount of fuel, humidity, and combustible material consumed by the fire.

Prior to burning, the average air temperature (°C), relative humidity (%), and average wind speed (m.s^-1) were measured by a portable meteorological station (Kestrel® 4000) located near the experimental area.

Fire behavior parameters, expressed by the fire line propagation speed (m.s^-1), flame height (m), and fire intensity (kcal.m^-1s^-1), have also been determined. Intensity was obtained from the ^{Byram’s equation (1959}), considering the calorific value of 3,705 kcal.kg^-1 determined by ^{Pivello et al. (2010}) for savanna areas.

Soil chemical properties were determined before and after burning (24 hours). For this purpose, three sampling points were used in each one of the 14 parcels of land for collecting samples composed of the first 5 cm of the superficial layer. Samples were sent to a private laboratory, where the following chemical properties of the soil would be determined: pH (CaCl²): potential of hydrogen; H + Al: potential acidity (cmol_c.kg^-1); P: available phosphorus (mg.kg^-1); S SO₄ ^-2: sulfur (sulphates) (mg.kg^-1); K⁺: available potassium (mg.kg^-1); Ca²⁺: exchangeable calcium (cmol_c.kg^-1); Mg²⁺: exchangeable magnesium (cmol_c.kg^-1); Al³⁺: exchangeable aluminum or acidity (cmol_c.kg^-1); OM: organic matter (g.kg^-1); OC: organic carbon (g.kg^-1); B⁺: boron (mg.kg^-1); Cu²⁺: copper (mg.kg^-1); Fe²⁺: iron (mg.kg^-1); Mn²⁺: manganese (mg.kg^-1); and Zn²⁺: zinc (mg.kg^-1).

Afterwards, we estimated the SB: sum of bases (cmol_c.kg^-1); CECt: effective cation exchange capacity (cmol_c.kg^-1); V: percent base saturation (%); m: percent aluminum saturation (%); Ca⁺/T: calcium saturation in cation exchange capacity (CEC) (%); Mg⁺/T: magnesium saturation in CEC (%); K⁺/T: Potassium saturation in CEC (%); Ca⁺/Mg⁺: calcium magnesium ratios; Ca⁺/K⁺: calcium potassium ratios; Mg⁺/K⁺: magnesium potassium ratios. Lastly, values were classified as very low, low, medium, and high, according to the methodology described by ^{Sousa & Lobato (2004}).

All variables were processed in Microsoft Excel (2010) spreadsheets and analyzed by the XLSTAT software, version 19.01 (2017). T-tests (p > 0.05) and the Pearson’s correlation were used for data analysis and comparison.

3. RESULTS AND DISCUSSION

On the day of the burning, the average air temperature was 41.82 °C; relative humidity, 15.74%; and wind speed, 0.40 m.s^-1 (Table 1). These meteorological variables changed throughout the day, which impacted variables of the fire behavior. According to ^{Soares & Batista (2007}), air temperature varies both in time and in space, whereas the maximum temperature is observed after midday.

According to our results, and based on ^{McArthur & Cheney’s (1966}) classification, the burning intensity was very low, accounting for a mean of 84.42 kcal.m^-1s^-1 with a standard deviation of 72.429 (Table 2). Nevertheless, it was not significantly related to the chemical variables.

However, the percentage of combustible material consumed by the fire (CMC) (Table 3), 55.70%, was the only burning-related variable presenting a significant correlation (p < 0.05) with soil chemical properties. These variables consist in the content of Ca²⁺, Fe²⁺, CO, SB, and the Ca⁺/T ratio (Table 4).

^{Oliveira et al. (2005}) reported that the action of fire with temperatures above 400 °C may cause the loss of P and N in the Cerrado environment, which affects the quality of the soil. Some authors have pointed out that when the fire consumes the initial biomass, its ashes return a significant amount of nutrients to the soil when compared with burned areas without biomass. This indicates to which extent the vegetation cover is important for the soil (^{Sampaio et al., 2003}; ^{Simon et al. 2016}).

Before burning, the chemical variables (Table 5) found in the soil of the experimental parcels of land presented very high acidity, with pH below 4.5, and potential acidity (H + Al) and exchangeable acidity (Al³⁺) on the medium scale. However, aluminum saturation (m) was high, whereas CECt and base saturation (V) presented low levels.

The soil presented an average of 13.5 g.kg^-1 of organic matter and 11.14 g.kg^-1 of organic content. Considering the macronutrient parameters presented by ^{Sousa & Lobato (2004}), contents of S, P, Ca²⁺, and Mg²⁺ were very low, although K⁺ was at a medium level. The micronutrient variables were found between the very low (Zn²⁺), low (B⁺ and Cu²⁺), medium (Mn²⁺), and high levels (Fe²⁺). ^{Pivello et al. (2010}) observed that, during the dry season, after burning in a Cerrado campo sujo area, the pH was below 4.5 and the potential acidity, above 70%. On the other hand, during the rainy season, the availability of nutrients was higher, including the organic matter content.

In our study, after burning, the organic matter (OM) content increased by 45.7%, followed by the organic carbon (OC), which increased by 20%. ^{Pomianoski et al. (2006}) observed the effects of fire on the soil OM in an agroforestry system, and concluded that, after prescribed burning, the availability of organic matter increased by 37% on the first layer (0-5 cm). According to ^{Oyedeji et al. (2016}), fire quickly accelerates the mineralization of soil organic matter.

Variables that determine soil acidity, such as pH and H + Al, changed, although not significantly. However, aluminum saturation (m) significantly decreased; nevertheless, Al³⁺ reduced by 26% (Table 5). ^{Faria et al. (2011}) found a slight increase in the pH and exchangeable acidity variables (Al³⁺) after burning, whereas the potential soil acidity (H + Al) presented significantly lower values. Conversely, ^{Batista & Soares (1995}) did not find a significant difference in characteristics related to soil acidity in a Pinus taeda plantation after burning.

We observed an increase in the saturation of micronutrients K⁺, Ca²⁺, Mg²⁺, S, and P (Table 5), although only the available phosphorus had no significant difference in the analysis. According to ^{Lorenzon et al. (2014}), who aimed to determine the consequent effects of fire on Red-Yellow Latosol, there was an increase in the content of P, from 1.61 to 4.19 mg.dm^-3; Ca²⁺, from 3.08 to 6.23 cmol_c.dm^-3; K, from 42.40 to 44.40 mg.dm^-3; and Mg²⁺, from 0.65 to 1.24 cmol_c.dm^-3. ^{Rheinheimer et al. (2003}) also observed an increase in the content of P after burning. However, they have also observed a sharp decrease of this nutrient on the superficial layer up to 60 days after burning. Yet, the soil presented higher K⁺ values in the burned parcels of land than in the unburned ones. ^{Simon et al. (2016}), upon studying the effects of burning in the Cerrado soil, verified that, in the 0.0-0.5 m depth layer, the Ca⁺/Mg⁺ ratio presented higher availability of these nutrients after burning.

Regarding changes in the content of micronutrients, we observed a significant increase in availability of Zn²⁺, Mn²⁺, and B⁺ contents (Table 5). ^{Couto et al. (2006}) studied the impact of fire on the availability of soil nutrients in the Pantanal region, and according to their results, there were no changes in Mg²⁺, B⁺, and Fe²⁺ values. Additionally, they found a slight increase in Cu²⁺ values.

According to the statistical analyses, we found a significant increase in the CECt, SB, and Ca⁺/T ratios (Table 5). However, the K⁺/T and Mg⁺/T ratios significantly decreased (Table 5). ^{Simon et al. (2016}) verified that, after burning, the superficial layer (0-5 cm) presented higher CEC values. This result was justified by the increase in availability of bases in the soil, besides the organic matter mineralization after burning. Thus, CEC increased due to the increase in bases and negative loads in the soil, which allowed these elements to be retained. Similar results have also been observed by other authors (e.g., ^{Dick et al., 2008}; ^{Rheinheimer et al., 2003}). According to these authors, the increase in cation concentration (Ca²⁺ and Mg²⁺) may be related to the release of oxides from the ashes.

4. CONCLUSIONS

Burning releases nutrients over a period of 24 hours and causes a number of chemical changes to the soil. These changes may be beneficial or harmful to chemical properties, and the degree of variation of their benefit or harm depends on several factors such as the use and type of burned soil, combustible material, time of day, duration, intensity, and frequency.

Low-to-moderate fire intensity may promote an increase in the availability of organic matter and other nutrients, such as Ca²⁺, Mg²⁺, K⁺, Mn²⁺, Zn²⁺, B⁺, and S, as well as in the ratios characterizing the soil, CECt, SB, Ca⁺/T, Ca⁺/Mg⁺, V, and Ca⁺/K⁺.