Long-term primary production trends in the Laurentian Great Lakes: a comparison of geochemical methods

Sediment cores from 12 locations throughout the Laurentian Great Lakes basin were analyzed for geochemical indicators of primary production. Sediment analytes included organic and inorganic contents, carbonates, sediment accumulation rates, total organic carbon and nitrogen concentrations, carbon and nitrogen isotope composition, and trends in spectroscopically inferred chlorophyll a (and its main diagenetic products). When multiple indicators were considered, production records related to recent cultural eutrophication and catchment activities were clear in the paleorecords. Indicators derived from loss-on-ignition (organic and inorganic content) were strongly associated with periods of human settlement in the catchments that increased overall sediment loads to the lakes. The ratio of carbon to nitrogen tracked catchment inputs of carbon, especially during periods of watershed development in western Lake Superior. Sediment records such as chlorophyll a and δ¹³C_org appear to be reliable indicators of trends in past algal abundance, particularly in Lake Erie, which has a well-known history of higher production and cultural eutrophication. These analytes also correlated well with past measured water quality surrogates for lake primary productivity and stressor data such as human populations in adjacent watersheds. A comparison among indicators revealed that several show redundancy as good proxies of production or productivity, though context was important. For instance, heavier isotopes of carbon and nitrogen are often prescribed as sedimentary indicators of lake productivity, but these two analytes were negatively correlated in Lake Superior, possibly due to long-term increases in cyanobacteria or changes in the nitrogen source. An increase in sediment carbonates can indicate summer blooms of cyanobacteria (a phenomenon that was clearly apparent in Lake Ontario) or catchment erosion. To make convincing geochemical inferences of primary production in the Great Lakes, it is recommended that a weight of evidence be built through the use of multiple indicators.