Evaluation of three contrasting models in estimating primary production from ocean color remote sensing using long-term time-series data at oceanic and coastal sites

Abstract

Accurate estimates of depth-integrated Net Primary Production (NPP, mg C m−2 d−1) and the creation of a robust climate data record of NPP for the global oceans are essential goals of the ocean color remote sensing community. Here, we take advantage of in situ NPP measurements from three long-term time-series sites, the HOT (Hawaii Ocean Time-series), BATS (Bermuda Atlantic Time-series Study) and CARIACO (Ocean Time-Series Program from the Cariaco basin), spanning over 30 years to evaluate three contrasting models in estimating NPP from ocean color remote sensing. These models for NPP estimation include the Absorption-based Model (AbPM), which relies on phytoplankton absorption coefficient, the Vertically Generalized Production Model (VGPM), which centers on chlorophyll-a concentration, and the Carbon-based Productivity Model (CbPM) centering on phytoplankton carbon. In addition to the accuracy of NPP estimation from these models, we laid great emphasis on evaluating their skills in capturing the monthly to seasonal variations and interannual trends in NPP at the three sites. Comparison with in situ NPP at all three long-term sites (∼20 years) showed that AbPM yielded the highest coefficient of determination (R2 = 0.67) and the lowest uncertainties (Bias = 0.03 and unbiased root mean square difference = 0.17). Seasonal and interannual variations apparent in the in situ NPP time-series records were best captured by AbPM. These results showcase the robust capabilities of AbPM and its superiority for global carbon cycling and climate change studies, largely because it takes into account optical and photosynthetic parameters of local phytoplankton.