RELATIONSHIPS BETWEEN MICROBIAL ACTIVITY AND SOIL PHYSICAL AND CHEMICAL PROPERTIES IN NATIVE AND REFORESTED Araucaria angustifolia FORESTS IN THE STATE OF SÃO PAULO, BRAZIL

de Morais Pereira, Jamil; Baretta, Dilmar; Bini, Daniel; Vasconcellos, Rafael L. de F.; Nogueira Cardoso, Elke Jurandy Bran RELATIONSHIPS BETWEEN MICROBIAL ACTIVITY AND SOIL PHYSICAL AND CHEMICAL PROPERTIES IN NATIVE AND REFORESTED Araucaria angustifolia FORESTS IN THE STATE OF SÃO PAULO, BRAZIL Revista Brasileira de Ciência do Solo, vol. 37, núm. 3, 2013, pp. 572-586 Sociedade Brasileira de Ciência do Solo Viçosa, Brasil


INTRODUCTION
Araucaria angustifolia (Bert).O. Kuntze is a typical arboreal species of the Mixed Ombrophilous forest in Brazil (Veloso et al., 1991).The area of occurrence of Araucaria forests was formerly huge in the south and southeast of the country, but inadequate exploration led to a significant shrinkage of the original area, which once covered approximately 253,000 km 2 .Currently, there are only 32,021 km 2 (12.6 %) left, of which 981 km 2 are permanently preserved, representing only 0.39 % of the original forest area (Ribeiro et al., 2009).
For the preservation of this forest, the presence of a high diversity of arbuscular mycorrhizal fungi (Moreira et al., 2009) and diazotrophic bacteria (Lammel et al., 2007), heterotrophic bacteria and archaebacteria (Maluche- Baretta, 2007;Bertini, 2010), and soil invertebrates (Merlim, 2005;Baretta et al., 2010) is fundamental.Conservation is therefore an urgent issue that must be addressed without delay, since the soil quality is fundamental, not only for food, timber, fiber, and fuel production, but also for the maintenance of biodiversity and environmental quality (Doran & Zeiss, 2000;Bastida et al., 2006;Kaschuk et al., 2010).
The quality of a soil is related to its physical, chemical and biological properties, which are affected by management and land use type and can be measured by indicators that are sensitive to alterations (Doran & Parkin, 1994;Baretta et al., 2010).However, aside from the traditional physical and chemical properties, the indicators should include microbiological variables as well, which together can reflect the processes that influence soil quality (Tótola & Chaer, 2002;Bastida et al., 2008;Baretta et al., 2010).Among the microbiological indicators of soil quality, microbial biomass carbon (MBC) is one of the most promising and most commonly used, due to its immediate response to reflect environmental changes, i.e., more susceptible to variations than the physical and chemical properties, including soil organic carbon (Powlson & Jenkinson, 1981;Baretta et al., 2005;Nogueira et al., 2006;Baretta et al., 2010;Kaschuk et al., 2010).
The amount of mineralized carbon in the form of CO 2 , due to the decomposition of organic matter, indicates whether climatic conditions, soil management or the presence of pollutants affect microbial activity and is used as soil quality indicator.For example, disturbed ecosystems tend to become sources of CO 2, as a result of the increased decomposition rate and reduction or interruption of organic residue input into the soil (Baretta et al., 2008).From the relationship between soil basal respiration and MBC, the metabolic quotient (qCO 2 ) can be calculated, which is an important index of the metabolic condition of the microbial community, reflecting the degree of environmental stress (Anderson & Domsch, 1989).
Soil enzymes are used as soil quality indicators due to their high sensitivity and fast response to soil management and use, as well as the ease of determination (Tabatabai, 1994).The enzyme dehydrogenase is sensitive to changes in the soil microbial community and is active in living cells, responsible for electron transfer reactions in the respiratory chain (Bastida et al., 2006), thus reflecting the oxidative activity of the soil microbial biomass.
The soil microbial community is fundamental in any ecosystem because it takes part in organic matter decomposition and nutrient cycling, influencing the chemical and physical properties of the soil and consequently, primary productivity.On the other hand, it is also affected by the physical and/or chemical parameters.For example, the pH affects both microbial activity as well as nutrient availability for microorganisms and plants.Organic carbon is a crucial component for soil fertility, be it as substrate for microorganisms or as a source of energy, nitrogen, sulfur, and other mineral nutrients (Idowu et al., 2008).
The most widely used physical indicators of soil quality, correlated with the chemical and microbiological variables, were bulk density, porosity, aggregate stability, aeration, infiltration, and water retention capacity, among others (Schoenholtz et al., 2000).To deepen the understanding of the soil quality in Araucaria forests, relating several indicators, the purpose of this study was to assess which physical and chemical properties are most closely related to dehydrogenase enzyme activity, basal respiration and microbial biomass under native and reforested forests with A. angustifolia, in two contrasting seasons (winter and summer), and which of these properties are the main discriminating factors between the areas.

Description of the study areas
The study was conducted in areas of the Atlantic Forest biome, with Mixed Ombrophilous forest, in the state of São Paulo.Areas with native (NF) and reforested (RF) forests with Araucaria angustifolia were selected in three regions of the State (Figure 1).In each region, one 0.5 ha plot, considered a true replication, was marked in both forest types.
The first sampling area is part of the Ecological Station Bananal in the Serra da Bocaina, between the States of São Paulo and Rio de Janeiro.At a distance of 25 km away from the town of Bananal, the climate is humid mesothermal, with no dry season and mild summers (Cfb, according to the Köppen classification).
At the first sampling (August 2009), average temperature and rainfall were 17 o C and 40 mm, respectively, while at the second (February 2010), the average temperature rose to 22 °C and more than 300 mm rainfall was recorded in January.
The RF area (Area 2; at 1,126 m) had been replanted about 38 years before and the forest is still in development, together with other species such as Dicksonia sellowiana Hook.(tree fern), Euterpe edulis Mart.(Juçara palm), Myrcia rostrata DC., among others, herbaceous and shrub flora and gramineae at sites of minor vegetation density.The soil in NF, area 1 and RF area 2 was classified as Dystric Haplic Cambisol (WRB, 2006).
In the same region, the RF area (Area 4, 740 m) is part of the Experimental station of Itapeva, in the municipality of Itapeva, São Paulo, and had been planted 45 years before (Figure 1).Other plant species were observed, such as Gochnatia polymorpha (Less) Cabr.(Candeia), Chusquea ramosissima Lindm.(a bamboo type), as well as other herbaceous species in the understory, natural regeneration of Araucaria seedlings and scarce tree species, except for the The third forest replication was an area in the Parque Estadual Turístico of Alto do Ribeira (PETAR), in the municipality of Iporanga, São Paulo, and in the region surrounding the park, in Barra do Chapéu, São Paulo (Figure 1).The climate is mild and temperate without droughts, Cfb according to the Köppen classification (Souza, 2008).In the winter of sampling, April was the driest month (<50 mm) and June the coldest (on average 17 °C), while in the summer of evaluation, the wettest month was January (over 400 mm), with an average temperature of 24 o C. The soil was classified as Dystric Haplic Cambisol (WRB, 2006).The NF (Area 5,785 m) was covered by Mixed Ombrophilous forest (Figure 1).Souza (2008).
The corresponding RF (Area 6, 932 m), was represented by the "Núcleo do Areado", PETAR (Figure 1).Regenerating tree species, shrubs and herbs grow in this area, and grass on patches with lower plant density.The soil was classified as Dystric Haplic Cambisol (WRB, 2006).The physical-chemical properties of all sampled soils are listed in table 1.

Soil and litter sampling
In plots of 0.5 ha, established in each sampling area, 15 trees of A. angustifolia were randomly selected, spaced at least 20 m apart.From around the trees (2 m away from the trunk) 15 soil and litter samples were randomly collected, to evaluate the microbiological, physical and chemical soil and the chemical litter properties, as well as dry litter in both seasons (winter -August 2009 andsummer -February 2010).The soil was sampled five times per sampling point with a Dutch auger (0-20 cm layer), to form a composite sample.The litter of each tree was collected using a 25 x 25 cm frame.The samples were packed in polyethylene bags, refrigerated and transported to the laboratory in ice-cooled boxes.Thereafter the soil samples were sieved (2 mm), ground and stored in a refrigerator (4 o C) until analysis.Litter samples were oven-dried at 55 o C for 72 h and then ground in a mill.

Microbiological, physical and chemical evaluations of soil and chemical analysis of litter
The microbial biomass carbon (MBC) was determined by the fumigation extraction method (Vance et al., 1987), using 10 g of soil (moisture was adjusted to 60 % of water holding capacity -WHC).The samples were fumigated with ethanolfree chloroform for 24 h, maintaining non-fumigated controls.Then, MBC was extracted from samples and controls with K 2 SO 4 (0.5 mol L -1 ) under stirring at 180 rpm for 30 min, oxidation with K 2 Cr 2 O 7 (66 mmol L -1 ) and titration with ferrous ammonium sulfate ([(NH 4 ) 2 Fe(SO 4 ) 2 .6H 2 O] (33 mmol L -1 ) in the presence of barium diphenylamine sulfonate.The correction factor Kc used for the calculations was 0.33.
Microbial activity was estimated by basal respiration (CO 2 ) released from 100 g incubated soil samples (soil moisture was adjusted to 60 % of WHC).The soil was incubated in a glass flask containing NaOH (0.5 mol L -1 ), for 10 days in the dark, heated to 28 o C. The carbon released as CO 2 was measured every 24 h to determine the remaining NaOH by titration with a standardized HCl (0.5 mol L -1 ) solution, using the indicators phenolphthalein and early carbonate precipitation by the addition of BaCl 2 (4 mol L -1 ) (Alef & Nannipieri, 1995).The resulting CO 2 -C and MBC values were used to calculate the metabolic quotient (qCO 2 ), which represents the CO 2 -C release rate per unit of MBC, as described by Anderson & Domsch (1993).

Determination of dry matter and C, N and S in litter
The litter dry mass was determined after drying at 55 o C to constant weight.Then the litter was ground, sieved (100 mesh) and the C, N, and S contents were determined by dry combustion in an elemental analyzer for C, N and S.

Determination of soil physical and chemical properties
The soil pH was determined by potentiometry in 0.01 mol L -1 CaCl 2 solution at a ratio of 1:2.5;P, Ca, Mg were extracted with ion exchange resin and K by Mehlich-1.Phosphorus was determined spectrophotometrically by the molybdenum blue complex, K by flame emission spectrometry, Ca and Mg by atomic absorption spectrometry, H+Al by potentiometry in SMP solution at pH 7.0; Al was extracted with KCl solution (1 mol L -1 ) and determined by titration with 0.025 mol L -1 NaOH, and organic carbon (org-C) was determined by colorimetry, by oxidation of organic matter with Na 2 Cr 2 O 7. 2H 2 O and H 2 SO 4 , according to Raij et al. (2001).
The moisture (moist) percentage was calculated by the difference between 10 g of fresh soil and of soil dried at 105 o C for 48 h.For particle size analysis, the sand fraction was obtained by sieving (0.053 mm), the clay fraction by the hydrometer method (Gee & Or, 2002), and silt from the difference between the quantities of soil and sand plus clay in the sample.In the winter, at the same points sampled for chemical analysis, undisturbed soil samples were taken (0-5 cm layer), using stainless steel rings (diameter 5 cm).The samples were wrapped in plastic film and taken to a laboratory, where they were refrigerated until analysis.In the laboratory, soil samples were saturated by capillary action, by gradually increasing the water depth up to two thirds the height of the rings.Then, they were subjected to a matric potential of -1 and -6 kPa using a tension table (Embrapa, 1997).When in hydraulic equilibrium, the samples were weighed and subsequently dried at 105 °C for 48 h to determine water content (Embrapa, 1997) and bulk density (Bd) (Blake & Hartge, 1986).Macroporosity (Ma) was computed as the difference between the water content of saturated soil and the water content after application of 6 kPa (Embrapa, 1997).Microporosity was estimated as the water content retained at a tension of 6 kPa.Total porosity (Pt) was calculated as the sum of macroporosity and microporosity.Particle density was determined by a helium pycnometer (Danielson & Sutherland, 1986).

Statistical analysis
The microbiological properties (MBC, CO 2 -C, qCO 2 , and DHA) were subjected to analysis of variance (Twoway ANOVA) and the means were compared by the LSD (Least significant difference) test (p<0.05).These properties were also subjected to Canonical Discriminant Analysis (CDA) to identify which results were more relevant to discriminate the studied areas (Cruz-Castillo et al., 1994;Baretta et al., 2005).In case of significant differences between areas, the values of the standardized canonical coefficients (SCCs) were subjected to the mean comparison test LSD (p<0.05) as described by Cruz-Castillo et al., (1994).Additionally, the microbiological and chemical properties of soil [(pH CaCl 2 , org-C, P, K, Mg, Al, H+Al] and litter (litter C, litter N, litter S), as well as soil physical properties [(Bd, Ma, Pt and Moist) and litter dry mass (LDM) were subjected to Canonical Correlation Analysis (CCA).Univariate statistical analyses were performed using SAS version 8.2 (SAS, 2002).

RESULTS AND DISCUSSION
Significant differences were found for some of the physical, chemical and microbiological properties in both sampling seasons (Table 2).Regardless of the sampling time, the levels of potassium (K), phosphorus (P) and macroporosity in the soil of native Araucaria forests were higher than in the reforested areas, whereas in soil of reforested Araucaria areas, total porosity and moisture were higher (Table 2).
In summer, there was a reduction in the available K and P levels, coinciding with higher soil moisture, which may have accelerated waste decomposition because of greater use of these elements by the biological community of the soil.Aside from the differences between forests, significant differences of some of the properties between seasons were also observed (Table 2).The levels of C and N in litter in winter were higher for NF than RF.
Litter dry mass was also highest in NF soil, regardless of the sampling time.Most likely, this reflects the conditions of the more complete soil cover and plant diversification and the greater input of plant residues in NF (Table 2).The contribution of crop residues to soil is usually related to the soil C content, but in this case it did not differ significantly between areas, reflecting the recovery of vegetation in RF areas and the low fluctuations of this variable in a short period (season).However, the C content can increase at certain times of the year, in response to the increased organic matter and nutrient cycling, triggering a cycle of stimulation of the soil biological activity (Maluche- Baretta et al., 2007;Correia & Andrade, 2008).
In general, the soils are acidic under both forest types and, together with the other chemical properties, are within the range of values reported by Bertini (2010) and Carvalho et al. (2012) for soils under native and replanted Araucaria forest, in the State of São Paulo.Macroporosity was higher in the preserved native than the reforested area and the values found in the two ecosystems ( 0.10 m 3 m -3 ) (Table 2) were suitable for adequate oxygen diffusion and water infiltration into the soil.This shows a good structural quality, favorable for the successful development of the biological community (Grable & Siemer, 1968;Alves et al., 2007).Total porosity was higher under RF than NF, probably due to the higher microporosity and clay content in the RF soil (Table 1).
The levels of microbial biomass carbon (MBC) only differed between areas (p<0.05) in the summer (Figure 2).In this period, the highest and lowest MBC values were detected (893 g C g -1 ) in NF, which was approximately 21 % higher than in RF (706 g C g -1 ).This indicates that the maintenance of native vegetation ensures the best conditions (macroporosity, litter dry mass and K and P levels), with a positive influence on the development and establishment of the soil microbiota (Wardle, 1992) (Table 2).In addition, seasonal changes in the microbial activity may be due to the increase in temperature and rainfall, particularly in tropical regions, accelerating metabolic processes and promoting the activity of the soil microbial community (Joergensen et al., 1990;Wardle, 1992;Bastida et al., 2006).
In winter, the MBC values of NF and RF areas did not differ (818 and 723 g C g -1 soil), respectively (Figure 2).The MBC indicates changes in the soil ecosystem balance (Baretta et al., 2005(Baretta et al., , 2008;;Silva et al., 2009;Baretta et al., 2010;Kaschuk et al., 2010).The MBC values found in this study were almost double those reported by Baretta (2007) and Bertini (2010) and are within the range of values found by Carvalho et al. (2012) in soils under native and reforested Araucaria forests in the State of São Paulo, and within the range found for ecosystems of Atlantic forest (683 -1520 mg C kg -1 soil) (Kaschuk et al., 2010).
Unlike MBC, microbial basal respiration (CO 2 -C) did not differ between areas and sampling seasons (Figure 3).In the winter, NF and RF values were 64 and 76 g CO 2 -C g -1 soil d -1 , respectively, and in summer, 64 and 68 g CO 2 -C g -1 soil d -1 , respectively.The similarity of the CO 2 -C values in the different areas suggested similar microbial activity.The CO 2 -C values were within the same range as those found  by Bertini (2010) and Bini et al. (2013a), in soil under native and reforested Araucaria forest in the State of São Paulo and Paraná, respectively.
The respiration rates can increase when the soil microbiota uses more readily degradable carbon substrate or due to the rapid oxidation of this carbon (Islam & Weil, 2000).Other authors found no significant differences either, in relation to soil basal respiration between native and reforested Araucaria forest, under similar experimental conditions (Baretta, 2007;Bertini, 2010;Carvalho et al., 2012).
The metabolic quotient (qCO 2 ) indicates the changes in microbial activity between natural and disturbed ecosystems more clearly (Islam & Weil 2000;Baretta et al., 2005;Bastida et al., 2008).The only microbiological variable that differed (p<0.05) between sampling sites in both seasons was qCO 2 (Figure 4).The higher qCO 2 values in RF may indicate a higher consumption of carbon readily assimilated by the microbial community, requiring more energy for their support, associated with a condition of disorder, with greater CO 2 losses, a characteristic of ecosystems in development (Anderson & Domsch, 1993).Higher qCO 2 values in cropped soils indicate greater stress of the microbial community compared to soils with more stable systems, such as mature forests (Islam & Weil, 2000), similarly to healthy Araucaria forests harmed by fire in comparison with native Araucaria and planted Araucaria forests (Baretta, 2007;Bertini, 2010), or replanted Araucaria forests in the rainy and dry season (Carvalho et al., 2012), or agricultural cultivation with annual crops compared with Mixed Ombrophilous forest, Bini et al. (2013a).This tendency to higher qCO 2 values may indicate an environment of ecological stress and degradation or a high level of productivity in this area, which is in line with the history of the RF area, more exposed to human intervention and to heavier impacts on the soilclimatic factors.
In contrast, qCO 2 values in NF soil were lower, indicating more stable ecosystems, with greater efficiency of microorganisms to convert organic waste into microbial biomass and with greater sustainability, which is consistent with the history of the NF area and corroborates results published elsewhere (Tótola & Chaer, 2002;Gil-Sotres et al., 2005;Baretta, 2007).
The use of qCO 2 is based on the theory of ecological succession proposed by Odum (1969), according to which the qCO 2 value is reduced during the succession or regeneration, caused by some disturbance.In the climax stage however, as in NF, the soil microbial community would tend to become more efficient in conserving soil organic carbon.
The activity of the enzyme dehydrogenase (DHA) differed significantly (p<0.05) between the NF and RF areas in winter only (Figure 5).In this season, the DHA in NF and RF were 10.9 and 10.1 g TTF g -1 soil d -1 , respectively, and in summer 10.3 and 9.8 g TTF g -1 soil d -1 .The higher activity in RF than NF was related to the metabolism of viable soil microorganisms, indicating the oxidative activity of the microbial community which, together with MBC, CO 2 -C and qCO 2 , has been used as general biochemical parameter to measure the influence of management on soil quality (Gil-Sotres et al., 2005;Bastida et al., 2006;Carvalho et al., 2012;Bini et al., 2013b).
In winter, the thinner canopy and greater soil exposure in RF than NF may have resulted in greater temperature and soil moisture fluctuation, as well as of other factors that influence the increase of microbial  activity, represented by the higher DHA, in line with the qCO 2 (Fig. 5).
Bini ( 2009) and Bini et al. (2013a) found higher DHA in soils of an agricultural area (16.46 g TTF g -1 soil d -1 ), than of Araucaria forest (11.28 g TTF g -1 soil d -1 ) and in an area reforested with Araucaria (10.65 g TTF g -1 soil d -1 ) in the state of Paraná.The possibility of increased metabolic stress in RF, indicated by higher qCO 2 values, was confirmed by the higher DHA, especially in winter (Figures 4 and 5).

Canonical Correlation Analysis (CCA)
Four canonical correlations of microbiological properties (MBC, qCO 2 , basal respiration, and DHA) were included in the CCA and 12 environmental variables [pH (CaCl 2 ), H+Al, K, Al, org-C, available P, Mg, and soil moisture], and litter dry mass, and litter C, N and S levels.The microbiological variables were called biological (BIO1) and environmental (physical and chemical) were denominated chemical variables (CHEM1).The soil physical properties bulk density, macroporosity and total porosity were not used in the CCA because they were only evaluated in winter and may decrease the reliability of the correlations.Only the first canonical correlation of data was evaluated, with a higher value and significance level (p<0.0001)than of the other canonical correlations.
In winter, the first canonical correlation between the physical -chemical (CHEM1) and biological properties (BIO1) was 0.91 (p<0.0001), with 81 % of the variance of scores of the first canonical variable of the BIO1 properties explained by the scores of the first canonical variable of the CHEM1 properties (Figure 6a).The high value of canonical correlation (CC) between microbiological and physical-chemical properties was due to the higher CC between microbiological properties, especially of DHA (0.7410) and chemical properties such as pH (0.5936), Mg (0.3466), org-C (0.1970) and litter N (0.1306), aside from physical properties, e.g.: soil moisture (0.2652) and litter dry mass (0.1756) (Table 3).
The first biological canonical variable had higher standardized canonical coefficients (SCC) for the microbiological property dehydrogenase (1.0683), followed by basal soil respiration (0.4683) (Table 3), since the CC was highest for DHA (0.7470).The positive signs of DHA and CO 2 -C (SCC and CC) indicated high values of DHA and CO 2 emission and high sensitivity to the related physical-chemical properties (Table 3).Several physical and chemical properties were correlated with DHA and soil microbial basal respiration (CO 2 -C) in winter.
In winter, the SCC values of pH (1.2746) and exchangeable acidity (H+Al) (0.6320) were noteworthy.In summer, SCCs were highest for pH (0.5452), moisture (0.4561), Mg (0.2727) and organic carbon (org-C) (0.2220) and litter carbon (0, 1131), which were associated with increased DHA (0.7940) and CO 2 -C (0.3149) (Table 3).In this sense, it was suggested that these physical-chemical parameters are seasondependent and the metabolic behavior of microorganisms and the system functionality can be modified according to the time of the year.The qCO 2 value for SCC was negative in winter and positive but low in summer, with few positive correlations with the soil chemical properties.The SCC values of microbial biomass carbon (MBC) were negative, both in winter (-0.6796) and summer (-0.4769).This attribute was considered a suppressor property with little influence on the chemical properties of soil and litter (Table 3).However, both MBC and qCO 2 have been reported as important microbiological properties, reflecting soil quality in different ecosystems (Baretta et al., 2010;Kaschuk et al., 2010).The differences found for the physical-chemical properties indicate that these, after more than 25 years since the implementation of RF, still differ from NF soil (Figure 7).
Considering the first physical-chemical canonical variable, the properties pH and H+Al showed high levels of SCC, indicating that these are directly related to the microbiological properties, with a positive relationship, in particular with CO 2 -C and DHA (Table 3).Negative values of SCC and CC, both in the first physical-chemical canonical variable as in first biological canonical variable indicate that these properties were similar in the studied areas and were therefore considered suppressor properties, with low discrimination capacity of areas.
In the summer, the first canonical correlation between physical and microbiological properties was 0.86 (p<0.0001),whereas the microbiological variation was mostly explained by the first canonical variable of physical and chemical properties (0.73) (Figure 6b).This high value of canonical correlation (0.86) between microbiological and physical-chemical properties is a result of the higher CC values of the microbiological (DHA, CO 2 -C and qCO 2 ) and the chemical properties (pH, H+Al, Al, org-C, and Mg) (Table 3).
The first biological canonical variable in summer was similar to that observed in winter, since the standardized canonical coefficients (SCC) were higher for the microbiological attribute DHA, followed by soil respiration CO 2 -C (Table 3).In addition, the canonical correlation (CC) was highest for DHA (0.60).The positive signs of DHA, CO 2 -C and qCO 2 indicate a high DHA and microbial respiration in the areas, most marked in summer.In the first physical-chemical canonical variable, SCC was highest for the soil properties pH, organic-C, Mg, moisture and litter C content, indicating a direct relation with the microbiological properties (Table 3).

Canonical Discriminant Analysis (CDA)
The individual contribution of each property to the area differentiation was expressed by the parallel discrimination rate coefficient (PDRC), which is the product of the standardized canonical coefficients (SCC) by the respective canonical correlation (r) (Baretta et al., 2010).Thus, the coefficient of PDRC is considered the most suitable parameter to evaluate the effect of discrimination by soil properties analyzed in the areas.Positive PDRC values indicate effects of separation between the areas, that is, differences in the analyzed property, while negative values indicate similarities (Cruz-Castillo et al., 1994;Baretta et al., 2010).The r values reflect univariate information and show the isolated contribution of each property analyzed in the separation of the areas studied (Table 4 and Figure 7).
In general, the PDRC value of each property was different for each sampling time (Table 4).In this case, only variables with PDRC > 0.1 were discussed, which can be considered good quality indicators to distinguish areas (Baretta et al., 2010).
Regardless of the sampling time (winter and summer), the properties available P (0.162 and 0.138), moisture (0.304 and 0.237), S litter (0.116 and 0.178) and total porosity (0.152 and 0.236) were mainly responsible for the discrimination between areas in both seasons, due to their higher PDRC (parallel discrimination rate coefficient) values (Table 4).The microbial basal respiration also contributed in winter (0.111) and potential acidity (H+Al) in summer (0.121) (Table 4).
The first canonical discriminating function (DCF1) clearly separated the NF from the RF, with higher SCC values.The polarized distribution of the SCC values confirmed the high dissimilarity in the microbiological and physical-chemical properties between the areas (Figure 7a,b).
Soil moisture was the most discrepant property (highest PDRC) in the areas of discrimination, regardless of the season, since this property is greatly influenced by environmental conditions (temperature and rainfall) and vegetation cover, with little variation in less impacted areas (Zornoza et al., 2007).In addition, changes in vegetation cover induce changes in the residues that constitute the litter, as observed in winter in NF, where higher levels of nutrients such as N and P were related to a better litter quality and quantity produced in NF (Table 2).The level of available P in the soil was also relevant in the discrimination of the areas (Tables 2 and 4).Phosphorus, together with N, is considered one of the most limiting elements for biological activity, especially in tropical soils, where P and N availability are naturally low.In ecosystems where P fixation in the solid phase is possible, the conditions given by the chemical-physical characteristics of soil colloids, the pH, cations associated with phosphate (Ca, Al and Fe), and a low organic matter content are often a reason for difficulties in this nutrient cycling, as a result of changes in the type of management or land use, as in the case of reforestation with Araucaria (Table 2).
Total porosity was an important property in the discrimination of the areas (Table 4).Normally, porosity is related to the oxygenation capacity and moisture retention of the soil, determining the conditions for the development of organisms and the root system as well as for the functioning of the soil chemical and biological processes (Dexter, 1988).Although relevant in the discrimination of the areas, in this study total porosity was not correlated with other physical-chemical and microbiological properties, but its importance for soil microbiota has already been emphasized elsewhere (Schoenholtz et al., 2000;Idowu et al., 2008).
An analysis of the biological variables showed that they contributed little to the discrimination of the areas in the two seasons, with the exception of basal respiration in winter (Table 4).Thus, although the CDA demonstrated the difference between the two areas, mainly by means of the physical-chemical variables discussed above, little could be shown by the microbiological variables, given their limited influence to discriminate the areas in both seasons.The microbiological activity of the reforested area was shown to be similar to that in the native Araucaria forest, probably resulting from a metabolic redundancy of micro-organisms as well as the similar organic matter content in the two ecosystems.However, we emphasize the importance of an integrated assessment of physical-chemical and microbiological soil properties, for a more accurate assessment of soil quality, in view of the close interaction between them (Sparling et al., 2004;Baretta et al., 2008).
The small differences in microbiological activity observed between RF and NF suggest that not only the time in the replanted Araucaria area, but also the presence of other plant species had contributed to a recovery of the soil microbial community in RF, since the location of the planted Araucaria forest, inserted in Atlantic forest, favored the development of other plant species, thus contributing to a recovery of the soil microbial community in RF, in spite of the discrepant physical-chemical conditions of the soil.Another important aspect is the contribution of organic carbon content of soil and leaf litter, stimulating soil microbiota, aside from other factors such as soil pH.These are key parameters determining the soil biota, rather than the variables with lower canonical correlation, since these parameters contribute to more accurate interpretations from the microbiological point of view of soil health and recovery processes.
The above discussion reinforces the need to examine the relationships between physical-chemical and microbiological properties for a better understanding of a distinction of the two systems studied.For example, the activity of some enzymes such as dehydrogenase, evaluated together with soil basal respiration, microbial biomass and some physical-chemical parameters, may contribute to the interpretation of the functional status of different ecosystems (Matsuoka et al., 2003;Baretta et al., 2008) Since to date there are no publications in Brazil addressing the physical-chemical and microbiological properties in the same sampling area and using true replications, our results can be important for the formulation of new hypotheses about the qualityrelated soil properties that best assess this quality when used together, due to relations between physicochemical and microbiological properties, as demonstrated in this study.

CONCLUSIONS
1.Among the microbiological properties analyzed individually, the metabolic quotient was the most sensitive indicator to assess the state of conservation of forests, regardless of season, although microbial biomass carbon and dehydrogenase activity in summer and winter, respectively, were also clearly discriminating.
2. There were significant correlations between the soil chemical and microbiological properties, particularly with regard to dehydrogenase activity and basal respiration, for the microbial properties, and for pH and H+Al with regard to the chemical properties, in winter, whereas organic carbon, soil moisture, soil Mg, pH and litter carbon were most closely correlated with dehydrogenase activity and respiration, in summer.
3. The canonical correlation analysis (CCA) identified the properties available P, total porosity, litter S content and soil moisture as the most sensitive to discriminate the native Araucaria forest (NF) from the areas of Araucaria reforestation (RF), of which soil moisture was the most relevant, regardless of the sampling season.
4. Evaluating the results of CCA and canonical discriminant analysis together, similar microbiological activity was observed in NF and RF, while the differences in physical and chemical properties indicated a recovery in RF, mainly due to the formation of a soil cover and increasing plant diversity in this area.

Figure 1 .
Figure 1.Map of the State of São Paulo with the georeferenced points of the study locations, São Paulo Brazil, 2013.

Figure 2 .
Figure 2. Levels of microbial biomass carbon (MBC) in soil discriminating native Araucaria forest (NF) and reforested areas (RF) in August 2009 (winter) and February 2010 (summer).Lowercase and uppercase letters compare between forests in the same season.LSD test at 5 %.Vertical bars represent the standard deviation (n=45).

Figure 5 .
Figure 5. Dehydrogenase enzyme activity in soil native Araucaria forest (NF) and reforested areas (RF) in August 2009 (winter) and February 2010 (summer).Lowercase and uppercase letters compare forests in the same season.LSD test at 5 %.Vertical bars represent the standard deviation (n=45).

Figure
Figure 6.Canonical Correlation Analysis between the soil physical and chemical properties + chemical attributes of litter (CHEM1) and soil microbiological properties (BIO1).Soil microbiological (CO 2 -C, MBC, and DHA qCO 2 ) and chemical [pH(CaCl 2 ), org-C (H+Al), K, Al, Mg and P] properties, chemical litter (C, N and S) properties, and soil moisture + litter dry weight sampled in winter (a) and summer (b) in native Araucaria forest (NF) and reforested areas (RF), (n=45).

Figure
Figure 7. Standardized canonical coefficients (SCC) of canonical discriminating function 1 and 2 (DCF1and DCF2), discriminating native Araucaria forest (NF) and reforested area (RF), considering the microbiological and physicalchemical soil properties and chemical litter properties of samples collected in winter (August 2009) (a) and summer (February 2010) (b).São Paulo, Brazil.The full symbol represents the average value of SCC (centroid) for each area (n=45).