Systematic Underestimation of MODIS Global Chlorophyll-a Concentration Estimation Algorithm Associating With Scale Effect

A framework for characterizing the scale effects of ocean chlorophyll 3 algorithm of moderate resolution imaging spectroradiometer (MODIS) (OC3M) in predicting chlorophyll-a concentration (chla) from global ocean is presented. In order to reveal the systematic underestimation of OC3M algorithm caused by scale problem in MODIS global chla product, an eight neighborhood algorithm is used to coarsely estimate the distribution variance of chla concentration, D(R_3M), from MODIS imageries. Based on the analysis of the global data series of monthly composites of MODIS Aqua satellite data from 2008 to 2010 at 4-km resolution, the chla concentration estimated by OC3M algorithm would be at least 1.29% smaller than the field measurements in theory, even if the atmospheric effects has been accurately removed owing to the scale effects. If the scale effect error is taken into account, the underestimation of chla in the Pacific, Indian, and Southern oceans, and overestimation at Atlantic oceans associated with packaged cell are much larger than these conclusions obtained by previous researchers. The D(R_3M) may be greatly underestimated in this paper, which would lead to the underestimation of scale effect errors of OC3M algorithm. Accurately estimating D(R_3M) may be a possible area for scale effect study in the future.