AUTHOR=de Oliveira Edmar Almeida , Feldpausch Ted R. , Marimon Beatriz Schwantes , Morandi Paulo Sérgio , Phillips Oliver L. , Bird Michael , Murakami Alejandro Araujo , Arroyo Luzmila , Quesada Carlos Alberta , Marimon-Junior Ben Hur TITLE=Soil pyrogenic carbon in southern Amazonia: Interaction between soil, climate, and above-ground biomass JOURNAL=Frontiers in Forests and Global Change VOLUME=5 YEAR=2022 URL=https://www.frontiersin.org/journals/forests-and-global-change/articles/10.3389/ffgc.2022.880963 DOI=10.3389/ffgc.2022.880963 ISSN=2624-893X ABSTRACT=

The Amazon forest represents one of the world’s largest terrestrial carbon reservoirs. Here, we evaluated the role of soil texture, climate, vegetation, and distance to savanna on the distribution and stocks of soil pyrogenic carbon (PyC) in intact forests with no history of recent fire spanning the southern Amazonia forest-Cerrado Zone of Transition (ZOT). In 19 one hectare forest plots, including three Amazonian Dark Earth (ADE, terra preta) sites with high soil PyC, we measured all trees and lianas with diameter ≥ 10 cm and analyzed soil physicochemical properties, including texture and PyC stocks. We quantified PyC stocks as a proportion of total organic carbon using hydrogen pyrolysis. We used multiple linear regression and variance partitioning to determine which variables best explain soil PyC variation. For all forests combined, soil PyC stocks ranged between 0.9 and 6.8 Mg/ha to 30 cm depth (mean 2.3 ± 1.5 Mg/ha) and PyC, on average, represented 4.3% of the total soil organic carbon (SOC). The most parsimonious model (based on AICc) included soil clay content and above-ground biomass (AGB) as the main predictors, explaining 71% of soil PyC variation. After removal of the ADE plots, PyC stocks ranged between 0.9 and 3.8 Mg/ha (mean 1.9 ± 0.8 Mg/ha–1) and PyC continued to represent ∼4% of the total SOC. The most parsimonious models without ADE included AGB and sand as the best predictors, with sand and PyC having an inverse relationship, and sand explaining 65% of the soil PyC variation. Partial regression analysis did not identify any of the components (climatic, environmental, and edaphic), pure or shared, as important in explaining soil PyC variation with or without ADE plots. We observed a substantial amount of soil PyC, even excluding ADE forests; however, contrary to expectations, soil PyC stocks were not higher nearer to the fire-dependent Cerrado than more humid regions of Amazonia. Our findings that soil texture and AGB explain the distribution and amount of soil PyC in ZOT forests will help to improve model estimates of SOC change with further climatic warming.