Specific leaf area (SLA, leaf area/leaf dry weight) is an important variable in crop growth models, as it relates dry matter production to leaf area expansion and, consequently, to light interception and photosynthesis.
Several hypothesis have been used for its modelling.
SLA may be considered as constant all along the growth cycle and within the canopy (e.g. van Keulen et al, 1982), or changing with the development stage of the crop (e.g. Wilkerson et al, 1983). In both cases, leaf area expansion is calculated by multiplying the increase in leaf dry weight by SLA. Therefore, daily changes in leaf area and dry weight are considered as strictly correlated.
However there is some experimental evidence that this correlation does not exist on a short term basis: SLA may fluctuate by some 50% within some days (Longuenesse, 1984). Jones et al (1989) observed that SLA was correlated, negatively with irradiance and CO2 concentration and positively with temperature.
They introduced this relation in a model of tomato (Lycopersicon esculentum Mill.) growth and yield, TOMGRO. A more mechanistic approach was suggested by Warren-Wilson (1972) for Impatiens parviflora and Thornley and Hurd (1974) for tomato plants.
They observed that SLA (or leaf area ratio i.e. leaf area/shoot dry weight) was negatively correlated with net assimilation rate.
Variations in SLA would only be due to changes in the size of a leaf assimilate pool, located in leaf blades.
The structural SLA i.e. the ratio of leaf area to leaf structural dry weight (total dry weight minus carbohydrates) would remain constant.
Thornley and Hurd (1974) validated their model with plants grown in constant climatic conditions and Warren-Wilson (1972) showed it simulated correctly the consequences of an increase of light level on growth rate and leaf area ratio.
The purpose of the present work was to compare the ability of Jones' and Warren-Wilson's models to predict daily variations in SLA observed on tomato plants under changing climatic conditions.
Both were implemented in TOMGRO. If the original version of TOMGRO (Jones et al, 1989) provided good simulations of long-term growth and mean SLA, a compartment of leaf assimilates had to be added to improve predictions of daily changes in SLA, which confirmed the validity of Warren-Wilson's approach.
If the stability of structural SLA can be discussed, it seems valuable to model the time-course of the leaf carbohydrate pool.