Soil properties in tropical montane cloud forests influence estimates of soil CO2 efflux

Soils in tropical montane cloud forests contain large amounts of organic carbon but few data are available on the flux of carbon dioxide (CO2) from them. Routine methods used to measure soil CO2 efflux, however, may be susceptible to measurement artifact in these ecosystems due to the high porosity and low bulk density of their soils. Here, we measured soil CO2 efflux in two tropical montane cloud forests, one lowland tropical forest and a garden lawn in the Peruvian Andes-Amazon and used ‘swap tests’ (repeated alternating measurements) to explore sources of measurement artifact when using two different closed-dynamic chamber systems (made by LI-COR Biosciences, NE, USA): a ‘multiplexed system’, consisting of 104-long term chambers, 8150-multiplexer and 8100-infra-red-gas analyzer (IRGA); and a ‘survey system’, consisting of a 20 cm-survey chamber and 8100-IRGA. We found that, in the cloud forests, soil CO2 efflux was significantly higher (by 54 ± 2%) when measured using the multiplexed system compared to when using the survey system. In contrast, there was no significant difference in soil CO2 efflux measured by the two systems in the lowland tropical forest, on the garden lawn, and in the cloud forests when the soil collars were inserted in the mineral soil. A series of diagnostic tests indicated that the multiplexed system overestimated soil CO2 efflux in the cloud forests by promoting mass-flow of CO2 from the organic soil of low bulk density. The measurement artifact was significantly reduced by reducing flow rates and diverting chamber air in-flow away from the soil surface, suggesting that the artifact was due to turbulence within the chamber (‘Venturi effects’). Future studies of soil CO2 efflux in tropical montane cloud forests should be undertaken with consideration of these artifacts, and gas flux monitoring systems should be tested in conditions to represent such environments.