David M. Warner

Summary: Lakes Michigan and Huron, which are undergoing oligotrophication after reduction of phosphorus loading, invasion by dreissenid mussels and variation in climate, provide an opportunity to conduct large‐scale evaluation of the relative importance of these changes for lake productivity. We used remote sensing, field data and an information‐theoretic approach to identify factors that showed statistical relationships with observed changes in chlorophyll a (chla) and primary production (PP). Spring phosphorus (TP), annual mean chla and PP have all declined significantly in both lakes since the late 1990s. Additionally, monthly mean values of chla have decreased in many but not all months, indicating altered seasonal patterns. The most striking change has been the decrease in chla concentration during the spring bloom. Mean chlorophyll a concentration was 17% higher in Lake Michigan than in Lake Huron, and total production for 2008 in Lake Michigan (9.5 tg year⁻¹) was 10% greater than in Lake Huron (7.8 tg year⁻¹), even though Lake Michigan is slightly smaller (by 3%) than Lake Huron. Differences between the lakes in the early 1970s evidently persisted to 2008. Invasive mussels influenced temporal trends in spring chla and annual primary production. However, TP had a greater effect on chla and primary production than did the mussels, and TP varied independently from them. Two climatic variables (precipitation and air temperature in the basins) influenced annual chla and annual PP, while the extent of ice cover influenced TP but not chla or primary production. Our results demonstrate that observed temporal patterns in chla and PP are the result of complex interactions of P, climate and invasive mussels.

Acoustic surveys were conducted in late summer/early fall during the years 1992-1996 and 2001-2013 to estimate pelagic prey fish biomass in Lake Michigan. Midwater trawling during the surveys as well as target strength provided a measure of species and size composition of the fish community for use in scaling acoustic data and providing species-specific abundance estimates. The 2013 survey consisted of 27 acoustic transects (546 km total) and 31 midwater trawl tows. Mean prey fish biomass was 6.1 kg/ha (relative standard error, RSE = 11%) or 29.6 kilotonnes (kt = 1,000 metric tons), which was similar to the estimate in 2012 (31.1 kt) and 23.5% of the long-term (18 years) mean. The numeric density of the 2013 alewife year class was 6% of the time series average and this year-class contributed 4% of total alewife biomass (5.2 kg/ha, RSE = 12%). Alewife ≥age-1 comprised 96% of alewife biomass. In 2013, alewife comprised 86% of total prey fish biomass, while rainbow smelt and bloater were 4 and 10% of total biomass, respectively. Rainbow smelt biomass in 2013 (0.24 kg/ha, RSE = 17%) was essentially identical to the rainbow smelt biomass in 2012 and was 6% of the long term mean. Bloater biomass in 2013 was 0.6 kg/ha, only half the 2012 biomass, and 6% of the long term mean. Mean density of small bloater in 2013 (29 fish/ha, RSE = 29%) was lower than peak values observed in 2007-2009 and was 23% of the time series mean. In 2013, pelagic prey fish biomass in Lake Michigan was similar to Lake Huron, but pelagic community composition differs in the two lakes, with Lake Huron dominated by bloater.