Specifying Algorithm Uncertainties In Satellite-Derived Par Products

Satellite ocean-color project offices routinely generate Level 2 and Level 3 daily Photo-synthetically Available Radiation (PAR) products. Accuracy is currently evaluated against in-situ measurements from buoys and fixed platfouns at a few locations, but specifying algorithm (and other) uncertainties on a pixel-by-pixel basis is needed to assess product quality. Expressing uncertainties requires modeling the measurement, identifying all possible error sources (e.g., noise in the input variables, imperfect/incomplete mathematical model), and deteunining the combined uncertainty. In the present study, algorithm uncertainties associated with PAR products are considered, i.e., those due to model approximations and parameter errors (e.g., decoupling effects of clouds and clear atmosphere, neglecting diurnal variability of clouds, using aerosol climatology) assuming that the input variables (TOA reflectance at wavelengths in the PAR spectral range) are known perfectly. A procedure is provided to estimate and provide, for each pixel of a product, this uncertainty component of the total uncertainty budget, which is expected to dominate. The bias and standard deviation of the daily PAR estimates are calculated as a function of clear sky PAR and cloud factor (i.e., the effect of clouds on daily PAR). The uncertainty characterization is accomplished using an extended simulation dataset covering the 2003-2012 time period using hourly MERRA-2 input data. The large number of data points allows one to sample well atmospheric variability and in particular many variations of daytime cloudiness, for all latitudes. Selected maps of global daily and monthly PAR and associated uncertainties (bias, standard deviation), obtained from MERIS data, are analyzed. Comparisons with match-up data at the COVE calibration/evaluation site reveal that experimental uncertainties are similar to the theoretical uncertainties obtained from simulated data.