Time-series analyses of fish abundance from an artificial reef and a reference area in the central-Adriatic Sea

This study evaluates the variation in population abundance overtime of twenty commercial fish species associated with an artificial reef and a reference site. Mean yearly catch rates were computed from data collected monthly from 1988 to 2012 using trammel nets. The log-transformed ratios of catches obtained at the reef and the reference site were calculated for each species. Statistical methods employed to study the changes in abundance of such fish species were multidimensional scaling (MDS), Min/Max Auto-correlation Factor Analysis (MAFA), Dynamic Factor Analysis (DFA) and chronological clustering. The time-series analyses were performed on three groups of species showing similar patterns of temporal cross-correlation. These time-series analytical techniques were utilized to identify common trends, the influence of some environmental variables, and changes in group trends. The analyses indicated a decreasing trend in the catch ratio for two species groups (mostly reef-dwelling species) while the third species group indicated an inverse pattern. Changes in the trends of abundance in some species were likely related to the general deterioration of the artificial reef modules.