THE relation between productivity and species diversity in temperate–arctic marine ecosystems

Energy variables, such as evapotranspiration, temperature, and productivity explain significant variation in the diversity of many groups of terrestrial plants and animals at local to global scales. Although the ocean represents the largest continuous habitat on earth with a vast spectrum of primary productivity and species richness, little is known about how productivity influences species diversity in marine systems. To search for general relationships between productivity and species richness in the ocean, we analyzed data from three different benthic marine ecosystems (epifaunal communities on subtidal rock walls, on navigation buoys in the Gulf of St. Lawrence, and Canadian Arctic macrobenthos) across local to continental spatial scales (<20 to >1000 km) using a standardized proxy for productivity, satellite‐derived chlorophyll α Theoretically, the form of the function between productivity and species richness is either monotonically increasing or decreasing, or curvilinear (hump‐ or U‐shaped). We found three negative linear and three hump‐shaped relationships between chlorophyll α and species richness out of 10 independent comparisons. Scale dependence was suggested by more prevalent diversity–productivity relationships at smaller (local, landscape) than larger (regional, continental) spatial scales. Differences in the form of the functions were more closely allied with community type than with scale, as negative linear functions were restricted to sessile epifauna while hump‐shaped functions occurred in Arctic macrobenthos (mixed epifauna, infauna). In two of the data sets, (St. Lawrence epifauna and Arctic macrobenthos) significant effects of chlorophyll α co‐varied with the effects of salinity, suggesting that environmental stress as well as productivity influences diversity in these marine systems. The co‐varying effect of salinity may commonly arise in broad‐scale studies of productivity and diversity in marine ecosystems when attempting to sample the largest range of productivity, often encompassing a coastal–oceanic gradient.