Araucaria angustifolia is an endangered tree species in Brazil and little is known about the soil attributes of these pine forests. This study was carried out to identify differences between natural and reforested Araucaria areas, in terms of soil microbiological and chemical attributes, with multivariate canonical discriminant analysis (CDA) and canonical correlation analysis (CCA). The studied areas included: 1. native forest with Araucaria (NF); 2. Araucaria reforestation (R); 3. Araucaria reforestation burnt by an accidental fire (RF); and 4. native grass pasture with native Araucaria and burnt by an intense accidental fire (NPF). In each area, 15 pine trees were selected and three soil samples collected from under each tree crown, in three different seasons. The CDA was applied to the soil microbiological attributes: carbon of microbial biomass (CMB), basal respiration (C-CO2) and metabolic quotient (qCO2), while CCA was performed with the soil microbiological and soil chemical attributes [pH (CaCl2), total organic carbon (TOC), and the contents of P, K, Ca, Mg, and H+Al]. The canonical correlation between soil microbiological and chemical attributes was highly significant, with a higher contribution of CMB and C-CO2 to the microbiological attributes, and of TOC and P to the chemical attributes. The CCA indicated that the variation of the chemical attributes can explain 52 % of the total data variability of the microbiological attributes, and the soil microbiological attributes of the areas 36 % of the total data variability of the chemical attributes. The CDA suggested that the contribution of each microbiological attribute to the discrimination of the areas was season-dependent, and discriminated CMB as the most important microbiological indicator, followed by C-CO2. The tools CCA and CDA proved to be essential in the study of soil quality indicators.
canonical correlation analysis; canonical discriminant analysis; soil quality bioindicator; carbon of the microbial biomass
