Soil fluxes of carbonyl sulfide (COS) across four distinct ecosystems

Soils are additional but poorly resolved sinks of carbonyl sulfide (COS) in terrestrial ecosystems. COS has been proposed as a tracer for quantifying gross photosynthesis based on the coupled stomatal uptake of COS and CO2. But applying this tracer requires the soil COS flux to be subtracted from the ecosystem flux to obtain the actual plant flux. To simulate soil COS fluxes, we have built a 1-D diffusion-reaction model accounting for vertical transport in the soil, microbial sinks and sources, and a litter layer. Uptake and production of COS in the soil column are linked with soil temperature and moisture through empirical functions adapted from enzyme kinetics and lab incubations. We have measured soil COS fluxes and the related soil variables in four distinct ecosystems: a wheat field (Southern Great Plains, OK, USA), an oak woodland (Santa Monica Mountains, CA, USA), a tropical rainforest (La Selva Biological Station, Costa Rica) and a boreal pine forest (Hyytiälä, Finland). Across all sites, a lower soil temperature and a humid climate are generally favorable to soil COS uptake. Strong COS emissions were observed in the wheat field at high soil temperatures after harvesting but were absent in other ecosystems, indicating that COS exchange may behave differently in agricultural soils. We simulated the soil fluxes in all ecosystems using the diffusion-reaction model, and optimized the source/sink strength parameters with field data. The optimized model provides insights that are not attainable from data analysis alone: For example, the wheat field soil must have continued uptake activity even when it showed net emissions, and leaf litter contributed dominantly to the COS sink after rain in the oak woodland. We expect the new model to be useful for simulating global soil COS fluxes as field data on soil fluxes from a broader range of ecosystems become available.