BIOLOGICALLY-BASED ESTUARINE SALINITY ZONES DERIVED FROM A MULTIVARIATE-ANALYSIS

Principal Component Analysis (PCA) was used to derive estuarine salini ty zones based on field data on the salinity ranges of 316 species/lif e stages in the mid-Atlantic region (chiefly Chesapeake Bay and Delawa re Bay species). Application of PCA to the data matrix showed that the structure underlying a diversity of salinity distributions could be r epresented by only five Principal Components corresponding to five ove rlapping salinity zones: freshwater to 4 parts per thousand, 2-14 part s per thousand, 11-18 parts per thousand 16-27 parts per thousand, and 24 parts per thousand to marine. The derived salinity zonation showed both differences and similarities to the Venice System of estuarine z onation. However, unlike the static and essentially descriptive Venice System, the new method will allow researchers to establish biological ly-relevant local salinity zones, and then develop hypotheses about th e processes that give rise to the resulting patterns. Examples of this procedure are given for the mid-Atlantic region. The method used here may also be useful for studying distributions across other environmen tal gradients, such as temperature, pH, substrate, turbidity, vegetati on, or latitude.