Abstract:As primary producers, phytoplankton play an important role in ensuring the laws of material recycling and energy flow in aquatic ecosystems. Phytoplankton abundance and community structure are indicators of the marine en- vironment. Thus, they influence fisheries resources through a bottom-up effect. The Qiansan Islets have been rich in fisheries resources for a long time, but over-fishing and environmental degradation have put tremendous pressure on this marine ecosystem in recent years. Thus, the local government conducted an artificial reef construction project beginning in 2005 to improve the marine environment and enhance fisheries resources. In this study, we used multivariate statistical analysis to compare phytoplankton community structure at different times (four seasons) and locations (artificial reef and control areas) to assess primary production in this region after constructing an artificial reef. Four cruise survey samplings were carried out at the artificial reef island and the adjacent waters around the Qiansan Islets in Haizhou Bay from September 2012 to August 2013. Eighty-five species in 35 genera and three phyla were identified, including 67 species of Bacillariophycophyta, 17 species of Pyrrophycophyta, and one species of Chryso- phycophta. Phytoplankton abundance values in spring and autumn (16.5 105 and 13.31 105 cells/m3, respectively) were significantly higher than those in summer and winter (6 105 and 3.8 105 cells/m3, respectively). The dominant species included Chaetoceros densus, Chaetoceros constrictus, Chaetoceros pseudocurvisetus, Coscinodiscus subtilis var. subtilis, Coscinodiscus debilis, Chaetoceros castracanei, Chaetoceros curvisetus, and Chaetoceros debilis. C. den- sus was the most dominant species in the spring. Phytoplankton community parameters (species number, abundance, chlorophyll a concentration, Margalef’s richness, Shannon-Wiener diversity, and the Pielou evenness index) were not significantly different (P>0.05) between the reef and control areas, except the Shannon–Wiener diversity and the Pielou evenness indices in spring, but significant differences were found among seasons (P<0.01). Environmental factors were also significantly different seasonally, but no differences were found between the reef and control areas. Multidimen- sional scaling and cluster analyses showed that phytoplankton community composition was very similar between areas but significantly different among seasons. The similar phytoplankton community structure was probably due to the close proximity of the reef to the control area, the small scale and deep water in the sampling area, the sampling sites’ close proximity to the island, and the short settling time of the artificial reef. A redundancy analysis showed that tem- perature, SiO3-Si, dissolved oxygen, PO4-P, BOD5, dissolved inorganic nitrogen, chemical oxygen demand, and trans- parency were the main factors influencing the phytoplankton community.