Regionalization and Dynamic Parameterization of Quantum Yield of Photosynthesis to Improve the Ocean Primary Production Estimates From Remote Sensing

Quantum yield of photosynthesis (phi) expresses the efficiency of phytoplankton carbon fixation given certain amount of absorbed light. This photophysiological parameter is key to obtaining reliable estimates of primary production (PPsat) in the ocean based on remote sensing information. Several works have shown that phi changes temporally, vertically, and horizontally in the ocean. One of the primary factors ruling its variability is light intensity and thereby, it can be modeled as a function of Photosynthetically Available Radiation (PAR). We estimated phi utilizing long time-series collected in the North Subtropical Oligotrophic Gyres, at HOT and BATS stations (Pacific and Atlantic oceans, respectively). Subsequently the maximum quantum yield (phi(m)) and K-phi (PAR value at half phi(m)) were calculated. Median phi(m) values were similar to 0.040 and 0.063 mol C mol photons(-1) at HOT and BATS, respectively, with higher values in winter. K-phi values were similar to 8.0 and 10.8 mol photons m(-2) d(-1) for HOT and BATS, respectively. Seasonal variability in K-phi showed its peak in summer. Dynamical parameterizations for both regions are indicated by their temporal behaviors, where phi(m), is related to temperature at BATS while K-phi to PAR, in both stations. At HOT, phi(m) a was weakly related to temperature and its median annual value was used for the whole data series. Differences in the study areas, even though both belong to Subtropical Gyres, reinforced the demand for regional parameterizations in PPsat models. Such parameterizations were finally included in a PPsat model based on phytoplankton absorption (PPsat-aphy-based), where results showed that the PPsat-aphy-based model coupled with dynamical parameterization improved PPsat estimates. Compared with PPsat estimates from the widely used VGPM, a model based on chlorophyll concentration (PPsat-chl-based). PPsat-aphy-based reduced model-measurement differences from similar to 62.8 to similar to 8.3% at HOT, along with well-matched seasonal cycle of PP (R-2 = 0.76). There is not significant reduction in model-measurement differences between PPsat-chl-based and PPsat-aphy-based PP at BATS though (37.8 vs. 36.4%), but much better agreement in seasonal cycles with PPsat-aphy-based (R-2 increased from 0.34 to 0.71). Our results point to improved estimation of PPsat by parameterized quantum yield along with phytoplankton absorption coefficient at the core.