APPLICATION OF A SELF-ORGANIZING MAP AND CANONICAL CORRESPONDENCE ANALYSIS IN MODERN BENTHIC FORAMINIFERAL COMMUNITIES: A CASE STUDY FROM THE PEARL RIVER ESTUARY, CHINA

Abstract

Benthic foraminiferal assemblages were studied in 327 surface sediments collected at water depths ranging from 2 to 100 m in the Pearl River Estuary and its adjacent continental shelf. A self-organizing map (SOM) was created to display the benthic assemblages based on species abundance, with six faunal assemblages identified. The distribution pattern of the assemblages on the SOM matched that of the sampling sites. The relationship between the benthic fauna and natural environmental parameters was investigated using a canonical correspondence analysis (CCA). The CCA Axis 1 (37% of the faunal variance) was primarily related to depth (r = 0.88), while CCA Axis 2 (11% of the faunal variance) was related to sediment texture (r = 0.72). We hypothesize that the salinity gradient derived from depth, along with substrate, are the most important factors controlling the distributions of the foraminiferal assemblages. This study verifies that the SOM is efficient in classifying the foraminiferal community and that benthic foraminifera are valuable bioindicators reflecting ecological changes over long time scales.