Ammonia (NH3) surface/atmosphere exchange is bi-directional and as such res
istance 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 s
tomatal compensation point (chi (s)), is applied here to NH3 exchange over
oilseed rape and compared with measured fluxes. For the first time the mode
l 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 decom
posing leaf litter at the ground and the likelihood of high [NH4+] in the s
iliques complicate the exchange pattern with oilseed rape and Limit the app
lication of the original model. This is therefore extended by: (a) the incl
usion of a litter layer (2-layer model), with an emission potential (Gamma
(1)), (b) additionally dividing the plant canopy into a foliage- and a sili
que-layer (3-layer model) and (c) considering the relative humidity (h) dep
endency of Gamma (1). The 2-layer model is able to predict night-time emiss
ion, but daytime emission is estimated to originate from the litter layer,
which is in contradiction to the NH3 sources and sinks derived for this can
opy. The 3-layer model using a constant value of Gamma (1) requires an emis
sion potential for the siliques of about 1300, which is consistent with bio
assay estimates. Together with a parameterization of Gamma (1) that increas
es 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 d
ominates night-time emissions and the silique-layer (probably) daytime emis
sions. 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.