ESTIMATION OF THE PHOTOSYNTHETIC ACTION SPECTRUM - IMPLICATION FOR PRIMARY PRODUCTION MODELS

A simple method for estimating the photosynthetic action spectrum is d eveloped. The method uses the shape of the absorption spectrum of phyt oplankton pigments, scaled to the magnitude of the initial slope of th e photosynthesis-light curve as established for broad-band illuminatio n. The method was tested by comparing the estimated action spectra wit h those measured during a cruise in the North Atlantic, in the fall of 1992. The agreement between the constructed and the measured spectra was good. Both the measured and constructed action spectra were then u sed to compute daily water-column primary production (P-Z,P-T) using a spectrally resolved model. The results showed that, at most of the st ations, the P-Z,P-T computed using the constructed action spectrum was not significantly different from P-Z,P-T calculated using the measure d spectrum. Daily water-column primary production was also computed at each station using the average shape of the measured action spectra ( spectra averaged over all stations), scaled to the magnitude of broad- band initial slope at that station. The results were similar to the P- Z,P-T values computed using the action spectra constructed for individ ual stations. The errors that may affect the constructed action spectr um are assessed through a sensitivity analysis. The analysis suggests that, for our data, the presence of photosynthetically inactive pigmen ts causes negligible errors in the computed P-Z,P-T. An assessment of the effects of random errors in the action spectrum showed that the er ror in the computed primary production was on average 1.5% (under the conditions chosen for the computation), when random errors of up to +/ -20% were introduced into the action spectrum. However, given similar conditions, systematic errors of similar magnitude in the action spect rum cause an average error of about 6% in the computed water-column pr imary production.