Resistance modelling of ammonia exchange over oilseed rape

Ammonia (NH3) surface/atmosphere exchange is bi-directional and as such resistance models must include canopy concentrations. An existing single layer model that describes the exchange in terms of adsorption to leaf cuticles and bi-directional transport through leaf stomata, which is governed by a stomatal compensation point (chi (s)), is applied here to NH3 exchange over oilseed rape and compared with measured fluxes. For the first time the model is tested using values of chi (s) based on the apoplastic ratio [NH4+]/pH (Gamma (s)) measured directly in the field. Strong NH3 emission from decomposing leaf litter at the ground and the likelihood of high [NH4+] in the siliques complicate the exchange pattern with oilseed rape and Limit the application of the original model. This is therefore extended by: (a) the inclusion of a litter layer (2-layer model), with an emission potential (Gamma (1)), (b) additionally dividing the plant canopy into a foliage- and a silique-layer (3-layer model) and (c) considering the relative humidity (h) dependency of Gamma (1). The 2-layer model is able to predict night-time emission, but daytime emission is estimated to originate from the litter layer, which is in contradiction to the NH3 sources and sinks derived for this canopy. The 3-layer model using a constant value of Gamma (1) requires an emission potential for the siliques of about 1300, which is consistent with bioassay estimates. Together with a parameterization of Gamma (1) that increases with h this model indicates that during daytime emission originates from the siliques, in agreement with the source/sink analysis. It is concluded that the leaf stomata were an effective NH3 sink, whereas the leaf litter dominates night-time emissions and the silique-layer (probably) daytime emissions. Although the 2-layer model reproduces the net exchange, the 3-layer model appears to be the mechanistically more accurate description. (C) 2000 Elsevier Science B.V. All rights reserved.