Assessment and propagation of uncertainties in input terms through an ocean-color-based model of primary productivity

Abstract

Net primary productivity (NPP) fields, derived from satellite observations of ocean color, are commonly published without relevant information on uncertainties. In this study, we assessed the uncertainty in NPP estimates of the Vertically Generalized Productivity Model using a Monte Carlo approach. We did not consider the uncertainty stemming from the basic model formulation, but restricted the uncertainty analysis to input terms, which were generated by, or related to, remote sensing. The study was based on global monthly remote sensing data from 2005. We found that the typical distribution of uncertainty around the model output could be approximated by a lognormal probability density function. On average, NPP value in a grid cell was overestimated by 6%, relative to the mean of the corresponding uncertainty distribution. The random component of uncertainty in NPP, expressed as the coefficient of variation, amounted to an average of 108%. The systematic positive errors in individual grid cells built up to an overestimate of 2.5 Pg C in the annual global NPP of 46.1 Pg C. The largest individual contributor to the random uncertainty in NPP was the input term that describes the physiological state of phytoplankton. However, the biggest contribution to the systematic uncertainty in the model output came from the parameter that represents changes in the rate of chlorophyll-normalized photosynthesis with depth. Therefore, improvements in the accuracy of these two terms would have the largest potential to decrease the input-related uncertainty in the model NPP estimates.