Investigating possible causal relations among physical, chemical and biological variables across regions in the Gulf of Maine

Abstract

We examine potential causal relations between ecosystem variables in four regions of the Gulf of Maine under two major assumptions: (i) a causal cyclic variable will precede, or lead, its effect variable; e.g., a peak (through) in the causal variable will come before a peak (through) in the effect variable. (ii) If physical variables determine regional ecosystem properties, then independent clusters of observations of physical, biological and interaction variables from the same stations will show similar patterns. We use the leading–lagging-strength method to establish leading strength and potential causality, and we use principal component analysis, to establish if regions differ in their ecological characteristics. We found that several relationships for physical and chemical variables were significant, and consistent with ‘‘common knowledge’’ of causal relations. In contrast, relationships that included biological variables differed among regions. In spite of these findings, we found that physical and chemical characteristics of near shore and pelagic regions of the Gulf of Maine translate into unique biological assemblages and unique physical–biologi- cal interactions