Action spectrum and maximum quantum yield of carbon fixation in natural phytoplankton populations: implications for primary production estimates in the ocean

Spectral and non-spectral measurements of the maximum quantum yield of carbon fixation for natural phytoplankton assemblages were compared in order to evaluate their effect on bio-optical models of primary production. Field samples were collected from two different coastal regions of NW Spain in spring, summer and autumn and in a polar environment (Gerlache Strait, Antarctica) during the austral summer. Concurrent determinations were made of spectral phytoplankton absorption coefficient [a(ph)(lambda)], white-light-limited slope of the photosynthesis-irradiance relationships (alpha (B)), carbon uptake action spectra [alpha (B)(lambda)], broad-band maximum quantum yields (phi(m)), and spectral maximum quantum yields [phi(m)(lambda)]. Carbon uptake action spectra roughly followed the shape of the corresponding phytoplankton absorption spectra but with a slight displacement in the blue-green region that could be attributed to imbalance between the two photosystems PS I and PS II. Results also confirmed previous observations of wavelength dependency of maximum quantum yield. The broad-band maximum quantum yield (phi(m)) calculated considering the measured spectral phytoplankton absorption coefficient and the spectrum of the light source of the incubators was not significantly different form the averaged spectral maximum quantum yield [phi(max)(lambda)] (t-test for paired samples, P = 0.34). These results suggest that maximum quantum yield can be estimated with enough accuracy from white-light P-E curves and measured phytoplankton absorption spectra. Primary production at light limiting regimes was compared using four different models with a varying degree of spectral complexity. No significant differences (t-test for paired samples, P = 0.91) were found between a spectral model based on the carbon uptake action spectra [alpha(B)(lambda) - model a] and a model which uses the broad-band phi(m) and measured a(ph)(lambda) (model b). In addition, primary production derived from constructed action spectra [a(c)(B)(lambda) from a(ph)(lambda) and alpha(B) (model c) was also not significantly different from that derived from total spectral model a (t-test for paired samples, P = 0.60). It was found, however, that primary production at low light regimes can be strongly overestimated (44%) when a(ph)(lambda) is derived from chlorophyll concentrations. A white-light model based on broad-band alpha(B) (model d), which does not consider phytoplankton light absorption, yields values 17% lower than those of model a. It is concluded that primary production at light-limited conditions can be computed accurately from broad-band maximum quantum yield estimates or from constructed action spectra provided that a(ph)(lambda) is measured. However, given that phytoplankton absorption coefficients are necessary for both approaches and as computations based on phi(m) showed less variability, we suggest that the maximum quantum yield proxy should be used. (C) 2000 Elsevier Science B.V. All rights reserved.