Storage and depth distribution of organic carbon in volcanic soils as affected by environmental and pedological factors

It has been recognised that volcanic soils, particularly Andosols, can store large amounts of soil organic carbon (SOC). This study investigates the factors controlling the regional and vertical distributions of SOC in soils of volcanic origin. To this aim, we investigated the vertical distribution of SOC in a total of 212 soil profiles representing all combinations of soil-forming factors in a volcanic area with a high diversity of ecosystems and soil types. We analysed the SOC contents in relation to intrinsic (soil type and relevant soil properties: texture and pH) and extrinsic factors (climate, parent material, relief), and we studied the patterns of SOC distribution with depth by fitting the SOC contents to different curve models. Furthermore, we selected ten soil profiles for a more detailed study to assess the effect of vegetation by examining the relationships of the SOC storage and depth distribution to the amount and allocation of plant roots and litterfall. SOC storage was controlled by the interaction of climatic (rainfall), time (substrate age), topographic (slope) and biotic (plant-mediated) factors. Our results indicate that under humid conditions, large organic inputs and the inhibition of microbial degradation due to low pH, AI-toxicity and persistent anaerobiosis within soil microaggregates largely contributed to SOC accumulation. Soil type was a poor predictor of SOC storage, most likely due to the co-occurrence of young and evolved Andosols and a certain andic character in many soils that did not qualify as Andosols. The distributions of root carbon and SOC appeared to be closely interrelated, suggesting a major role of roots in the supply of organic matter and the lack of significant bioturbation. The depth distribution of SOC was best fitted by the quadratic, cubic and power models, the latter being a feasible alternative that should be used to this aim in volcanic soils rather than the widely used exponential model. (C) 2014 Elsevier BM. All rights reserved.