This repository is a set of matlab functions used to partition the global open ocean based on the presence patterns of 536 phytoplankton species. To this end, we use a SOM-PCA-HAC approach. First, we assess the similarity of observations using the presence/absence patterns in a self-organizing map (SOM). Second, we perform a dimensionality reduction of the 536 attributes (species) using a principal component analysis (PCA). Finally, we group the reduced data into clusters using hierarchical agglomerative clustering (HAC).
The manuscript detailing the resulting biomes has been published in Progress in Oceanography, and can be cited as:
Hofmann Elizondo U., Righetti D., Benedetti F., and Vogt M. (2021) Biome partitioning of the global ocean based on phytoplankton biogeography. Progress in Oceanography, Volume 194, ISSN 0079-6611, https://doi.org/10.1016/j.pocean.2021.102530
