Species, rotation, and life-form diversity effects on soil carbon in experimental tropical ecosystems

Extensive areas of species-rich forests in the tropics have been replaced by tree monocultures over the last two decades, and the impact on biogeochemical cycles is unclear. We characterized effects on soil carbon dynamics of species identity and rotation frequency in experimental plantations containing three native, non-N-fixing tree species, Hyeronima alchoreoides, Cedrela odorata, and Cordia alliodora, grown in monocultures and in polycultures with two monocot species, Euterpe oleracea and Heliconia imbricata. Over all treatments, change in total soil organic carbon (TSOC, 0-15 cm) after 10 years ranged from a loss of 24% (0.9 mg/ha in 1-yr rotation of Cedrela) to an increase of 14% (0.6 mg/ha under Hyeronima polycultures). Species differed in their effects on quantities of TSOC (P = 0.038), but differences were more pronounced in light particulate organic matter (LPOM; P = 0.001), a biologically active, sand-size soil fraction that constituted 6% of TSOC. Effects of rotation frequency were strong; in Cedrela and Cordia, the 4-yr rotations had higher soil C stocks than did long-term monocultures, where soil C stocks had declined under 10-yr-old trees. Under Cedrela and Cordia, polycultures had significantly higher stocks of soil C than monocultures, whereas soil C stocks were high under Hyeronima in both cultures. In polycultures, Hyeronima dominated detrital inputs, contributing 88% of litterfall and fine-root growth, whereas Cedrela and Cordia contributed <34%. Root C:N ratio and fine-root growth accounted for most of the variability in changes in soil C stocks after 10 years in long-term rotations (partial R-2 = 0.70 and 0.14, respectively). These data suggested that roots drove soil C accrual in long-term rotations, and that mechanisms involving root chemistry, and not quantity of detrital inputs, best explained effects of species on soil C sequestration.