Clay and Climate Are Poor Predictors of Regional-Scale Soil Carbon Storage in the US Caribbean

Soils of the tropics play an important role in the global carbon (C) cycle as substrates for the most productive terrestrial ecosystems and as large C reservoirs. The amount of C stored in a soil is affected by soil properties such as texture, pH, and mineralogy, as well as climate, vegetation and land management, yet, uncertainty remains regarding the relative importance of these variables at different spatial scales. The goal of this study was to explore the factors influencing regional patterns in soil C storage. Soils were collected from 36 sites under four land cover types (forests, pastures, croplands and wetlands) and on 30 soil series in seven soil orders, representing the diversity in soil forming factors and major non-urban landscapes in Puerto Rico and the US Virgin Islands. Soils were analyzed at three depths, 0-30 cm, 30-100 cm, and 0-100 cm, for total carbon (TC), soil organic carbon (SOC), and total nitrogen (N). Traditional predictors of soil C storage such as clay content and climate were not sufficient to accurately capture regional soil C trends. The combination of clay plus fine silt was positively correlated with SOC stocks but explained a relatively small amount of the variation. Site temperature and mean annual precipitation were not correlated with SOC stocks. Soil order and land cover were weak predictors of SOC for a subset of data where forests and pastures were found on three of the same soil orders, allowing for a two-way ANOVA. Soil N showed similar trends as SOC. Patterns observed depended on the soil depth studied, reflecting relative differences in the strength of top-down and bottom-up factors influencing SOC and N storage. These findings suggest that mechanisms not described by the distribution of soil particle sizes need to be considered for better predictions of SOC dynamics. In a geologically diverse region, soil properties were better predictors of SOC than climate or land cover. Tropical regions represent a diversity of soil types yet parent material and soil physical properties other than texture are not currently included in most biogeochemical models attempting to predict the response of tropical soil C pools to environmental change. Recognition of the heterogeneity of geologic substrates and weathering gradients in tropical regions and incorporation of a more mechanistic understanding can improve soil C modeling and land management decisions.