Winter is coming – ecosystem-scale COS exchange during senescence of a deciduous forest

The gross uptake of CO₂ on ecosystem level (GPP) can’t be measured directly, but has to be inferred from models or proxies. One of the newly emerged constrains on GPP is the trace gas carbonyl sulfide (COS). COS enters the plant leaf through the stomata and diffuses through the intercellular space, the cell wall, the plasma membrane and the cytosol like CO₂. Within the cytosol, it is then catalyzed by the enzyme carbonic anhydrase (CA) in a one-way reaction to H₂S and CO₂. Basically, this one way flux would make COS a very promising tracer for GPP on ecosystem level, but there is growing evidence that plants are also capable of emitting COS. Mosses and even vascular plants that are under high stress like drought and fungal infection, have been reported to emit COS. Furthermore, a winter wheat field, that showed a good correlation between the CO₂ and COS ecosystem fluxes during the peak growing phase turned into a source for COS after going into senescence. This indicates that yet unknown COS emission processes likely related to plant degradation, could complicate the use of COS as a tracer for GPP.

Since the majority of studies have focused on measuring COS ecosystem fluxes during peak growing times or on evergreen forests, we seek to quantify the relationship between the ecosystem-scale exchange of CO₂ and COS of an ecosystem going into senescence.

Between September and November 2019 we deployed our quantum cascade laser (Aerodyne Research Inc., MA, USA) at a beech forest in Leinefelde, Germany to conduct eddy covariance measurements for COS, CO₂ and H₂O. Our observations started when the beech forest was still green and in full leaf and ended when most of the trees had already shed their leaves. The ecosystem fluxes of COS and CO₂ concurrently decreased over the course of our campaign up to the point when we could not observe a net uptake of CO₂ anymore. We will further compare the GPP estimates resulting from classical flux partitioning and flux partitioning with the additional use of COS to determine if the model differences increase towards the end of the season.