A multivariate analysis of phytoplankton assemblages in the western North Atlantic

• Published in: Journal of Plankton Research
• Main Authors: matta, James F., Marshall, Harold G.
• Format: Text
• Language: English
• Published: Oxford University Press 1984
• Subjects: Articles; North Atlantic
• Online Access: http://plankt.oxfordjournals.org/cgi/content/short/6/4/663; https://doi.org/10.1093/plankt/6.4.663

Principal component analysis (PCA) of a large data set of phytoplankton records, 1111 stations collected in an uneven sampling pattern over a 9-year period from the US north-eastern shelf waters is discussed. Year-to-year variation was found to be the largest component of variation in the data and was more important than seasonal variation and all but the grossest of geographic variation. Strong seasonal and geographic patterns were detected when long-term temporal variation was removed and it was possible to identify groups of critical species influencing geographic patterns. In this study PCA techniques were most successful when used on data collected from a small geographic area or data collected from a narrow spectrum of time (e.g., a single year or a single month) or both. Interpretation of patterns is difficult if several major sources of variation are present. The major sources of variation in phytoplankton populations were identified as annual changes, seasonal changes and geographic patterns. Distinct trends were identified in the geographic pattern with the southern estuaries, the Mid-Atlantic Shelf and offshore versus coastal zones identified as having distinctive fauna] assemblages.