Creating Unified Ocean Color Data Records with Uncertainties
The generation of unified satellite data records through the merging of ocean color data from multiple sensors has proven beneficial to the science users community at various levels. First, merged products offer improved coverage of the ocean at daily to monthly time scales, which reduces the uncertainties in estimates derived from those products for both local and global studies. Second, merged data products often have lower uncertainties than the same product from any single sensor. Last, data merging has also proven a powerful tool to identify inconsistencies among the different data sources or issues with the sensors radiometry. In all, data merging benefits both the ocean color and biogeochemistry science that uses its data and the inter-sensors calibration/validation activities.
Here, we propose to continue the development of unified and coherent ocean color time series through the merging of data from multiple sensors. We will continue the development of merged ocean color products from the GSM semi-analytical model. This model merges data at the Remote sensing reflectance level and derives several biogeochemically relevant data products simultaneously along with uncertainty estimates at each pixel. In addition, we will also generate merged products from higher level data (e.g. chlorophyll-a concentration) as such products are no longer available to the science community. We will also develop new merged ocean color products. In particular, we will develop a merged remote sensing reflectance product that will allow users to work with a data set with improved spectral resolution and lower uncertainties. Last, uncertainty estimates for all merged products will be generated on a pixel-by-pixel basis. All products and uncertainty estimates will be validated through matchup analyses. The merged records will cover the time span over which multiple ocean color sensors are or will be available (SeaWiFS, MODIS, MERIS, VIIRS, OLCI,&). Both global (9-4 km resolution from level-3 data) and regional (1-4 km resolution from level-2 data) merged products will be developed.
Stephane Maritorena - PI, University of California at Santa Barbara
Page Last Updated: May 2, 2019 at 12:02 PM EDT