Recent studies by MOREL (1978) and PLATT, SATHYENDRANATH, CAVERHILL and LEWIS (1988) have demonstrated the relative stability of the relationship between available photosynthetic energy at the ocean surface and energy stored by algal photosynthesis, once it has been normalized with respect to the column integrated chlorophyll biomass. Therefore the cross section vis a vis photosynthesis and per unit of areal chlorophyll, phi*, in m2 (g Chl)-1, should be relatively stable in spite of the various environmental and trophic situations possibly encountered in the open ocean. Such an ecological or biogeochemical "constant" is of importance when trying to transform biomass maps (obtained from remotely sensed ocean colour data) into primary production maps. Its approximate constancy has to be understood and deserves analysis. This analysis rests on the use of the local and instantaneous growth rate equation (KIEFER and MITCHELL, 1983) which has to be triply integrated with respect to wavelength, depth and time. Such an integration is, in effect, the core of a light production model which schematically includes the following steps: (i) to compute as a function of the sun elevation for various days, latitudes and atmospheric conditions (aerosols, water vapour, etc.) the photosynthetic energy impinging at the ocean surface (direct and diffuse components); (ii) to account for the loss by reflection at the air-sea interface; (iii) to propagate this radiant energy (in terms of spectral scalar irradiance) throughout the water column and according to given pigment vertical profiles; (iv) to evaluate what part of energy is absorbed by algae within the productive column; (v) and finally to compute that part of absorbed energy which is stored in the form of organic carbon added to the initial biomass; this last step implies that the yield for growth be modelled as a function of the irradiance level and temperature. Sensitivity tests have been effected with respect not only to the physical parameters which can be accurately modelled, but also with respect to the physiological factors for which values and parameterisation are more uncertain. Non-linear and interactive influences cause the phi* values obtained by running the model, to vary within a rather restricted range (within a factor 2, for most of the trophic and environmental conditions), which are similar to those resulting from field studies. The variability of the biomass-normalized primary production can be explained and the seasonal or zonal trends illustrated. The effect of cloudiness is also analyzed. This spectral light - photosynthesis model (a sub model in the more general study of the biomass evolution) can be used either to reproduce primary production experiments and also as a predictive tool in the oceanic carbon fixation problem. The global scale made accessible by satellite techniques requires that a climatological field of the phi* parameter be produced. This can be done by operating the present model, provided that the physiological factors which intervene are sufficiently ascertained and adequately parameterised, and also provided that the vertical distribution of the algal biomass can be inferred from the partial information (restricted to the upper layer) delivered by ocean colour sensors.
