RECOGNIZING PATCHY PALEOCOMMUNITY GRADIENTS WITH DIFFERENT ECOLOGICAL ORDINATION TECHNIQUES (PCA AND FSPA): AN EXAMPLE FROM THE YORKTOWN FORMATION (PLIOCENE) OF VIRGINIA

Different ordination techniques reveal different aspects of paleocommunity structures. PCA (Principal Components Analysis) ordinates paleoecologic samples based primarily on the relative abundance of the 2-3 most abundant species. FSPA (Flexible Shortest Path Adjustment of Ecologic Distances) is an ordination technique similar to PCA or DCA, but it uncovers patchy ecologic distributions more effectively as well as removing the “arch effect” (a common problem in PCA).

FSPA and PCA were performed on species abundance and paleoecological guild data from the middle sequence of the Pliocene Yorktown Formation (Rushmere and Morgart's Beach Members). This sequence is unconformity-bounded and records a paleoenvironmental transition from shallow marine silty sand to muddy sand bottom conditions. Samples of these members were collected at 30 cm intervals at 5 localities in Virginia.

The two ordination techniques brought out different aspects of the paleocommunity changes through the paleoenvironmental transition. Based primarily on the relative abundance of two very abundant species, PCA revealed two strongly defined end-member paleocommunities from the top and bottom of the sequence, but did not offer much information about the nature of the transition between these end-members. FSPA, on the other hand, exposed several paleoecological trajectories for the transition between end-members at different locations, highlighting the variability in paleocommunity responses to the paleoenvironmental transition. The ordination uncovered the importance of local controls on the structure of paleocommunities and the resultant patchiness in paleocommunity distributions; the same lithologic transition from silty sand to muddy substrate produced recognizably different faunal patterns in FSPA ordination space. On the other hand, FSPA ordinations of guild data revealed the powerful control of substrate (silty sand versus muddy) on paleoecologic composition. These results highlight the importance of using multiple ordination techniques when exploring paleoecological patterns.