Stomatal conductance of rose whole plants in greenhouse conditions: Analysis and modelling

This paper deals with the modelling of the equivalent ''whole plant'' conductance (g(s)) of roses in greenhouse conditions. The purpose was to provide more information about g(s) as a key control variable of gas exchange and its variation with respect to relevant environmental or plant variables. Using previously published data on seasonal gas exchange of whole rose plants in the glasshouse and assuming the whole plant behaves like a ''big leaf'', two schemes of modelling the stomatal conductance were adopted: (i) the multiplicative model of Jarvis (J model) assuming that g(s) is driven by microclimatic factors, and (ii), the conductance-assimilation model of Ball, Woodrow and Berry (B model) that describes g(s) with respect to net CO2 assimilation (A(n)) and environmental variables at the leaf surface. Identification of the model parameters were performed with a data set corresponding to the spring season and showed that the performances of both models depend upon the characterization of humidity (air vapour pressure deficit, D-a, leaf to air deficit, D-1, leaf surface deficit, D-s, surface relative humidity, h(s)) The best predictive performances of the models were obtained with hyperbolic functions where D = D-1 (J model) and D = D-s (B model). The values of the parameters of the best performing version of the B model (B-L model, amended by Leuning) compare well with published values for leaves of woody species. When the two models were tested for their predictive ability (using the data set of the autumn and winter periods), the J model failed to give a good estimation of g(s) under conditions of low PPFD and D, leading to a significant underestimation of g(s). Prediction of g(s) with the B-L model provides a better description of the experimental data.