Quality reference values for trace metals in Podzols, Ferralsols, and Acrisols of Brazilian Atlantic rainforest

• Published in: Catena (Giessen) 2022-03, Vol.210, p.105879, Article 105879
• Main Authors: Barreto Mascarenhas, Renata, Bomfim de Jesus, Taíse, Gloaguen, Thomas Vincent, Del'Arco Vinhas Costa, Oldair, Wyzykowski, Jair
• Format: Article
• Language: English
• Subjects: Multivariate regression model; Pseudo-total concentrations; Tropical soils
• ISSN: 0341-8162; 1872-6887
• DOI: 10.1016/j.catena.2021.105879

•Podzols are enriched in Pb, Ferralsols in Ni and Zn, and Acrisols in Cr and Cu.•Reference quality thresholds must consider soil formation processes and horizons.•Multivariate regression is an excellent predictor of concentrations of trace metals. In tropical coastal forest ecosystems, human activities spread rapidly. It is, therefore, necessary to define soil quality. For this, the determination of the natural content of trace metals is an essential step. This study aims to understand the distribution of trace metals in the main soils of tropical coastal rainforest ecosystems: Ferralsols, Acrisols, and Podzols, in Bahia state, Brazil. After morphological, chemical, and physical characterization and classification of the soils, the concentration of trace metals was investigated both in soil profiles and in surface composite samples (national guideline recommendation), in five locations for each soil class. The soils were classified as Geric Ferralsol, Haplic Acrisol, Chromic Acrisol, and Ortsteinic Carbic Albic Endogleyic Podzol. Concentrations of metals differed significantly between soil classes. Even under tropical rainforest climate during soil formation, the parent rock material still influences geochemistry. Al, Fe, Mo, Ni, V, and Zn accumulate in Ferralsols, while Ba, Co, Cr, Cu, and Mn accumulate in Acrisols. In the Podzols, metals highly accumulate in B spodic horizon. The data highlights that composite sampling strictly based on soil layers and not horizons (as the national guideline recommends) can cause significant misinterpretations. Instead, the multivariate regression models provided consistent data, accurately predicting the pseudo-total concentrations of trace metals in individual soil samples. The regression method is fast and low cost and provides a reliable geochemical baseline for the definition of reference values.