Satellite-based phytoplankton diagnostic pigment maps for the northern Gulf of Mexico. 2003-2018

Published: 24 June 2022| Version 1 | DOI: 10.17632/nvxy6bd4hm.1
Andy Stock, Ajit Subramaniam


Satellite-derived daily maps, multi-day composites and climatologies of four diagnostic pigments that serve as biomarkers for different phytoplankton types (fucoxanthin, zeaxanthin, 19'-butanoyloxyfucoxanthin and 19'-hexanoyloxyfucoxanthin). The data contain proportions of total diagnostic pigments (according to Vidussi et al., 2001) for the northern Gulf of Mexico 2003-2018 at 4km spatial resolution, and were created using random forests with various satellite data products as inputs. Linear correlations between predicted and observed values obtained by spatial leave-one-out cross-validation, when averaged over separation distances between training and test observations from 0.1km to 200km, were: 0.85 (fuco), 0.72 (zea), 0.74 (but-fuco), and 0.69 (hex-fuco). Mean absolute errors estimated in the same way were 0.13 (fuco), 0.11 (zea), 0.02 (but.fuco), and 0.07 (hex.fuco). All data are provided in GeoTIF format. Dates in the filename for 8-day composites refer to the first of the 8 covered days. The folder hex_pan includes geotifs for 19'-hexanoyloxyfucoxanthin created with an alternative, polynomial algorithm broadly following Pan et al. (2010, 2013) for comparison purposes. Details are provided in a research article under review that will be linked to once published.



Lamont-Doherty Earth Observatory, The University of British Columbia, Earth Institute at Columbia University


Oceanography, Remote Sensing, Marine Biology