Mapping the spatial distribution of the biomass and filter‐feeding effect of invasive dreissenid mussels on the winter‐spring phytoplankton bloom in Lake Michigan

Abstract

Summary The effects of the invasive bivalves Dreissena polymorpha (zebra mussel) and Dreissena rostriformis bugensis (quagga mussel) on aquatic ecosystems, including Lake Michigan, are a topic of current interest to scientists and resource managers. We hypothesised that the winter–spring phytoplankton bloom in Lake Michigan is reduced at locations where the fraction of the water column cleared per day by Dreissena filter feeding approached the net growth rate of phytoplankton, when the water column was not stratified. To test this hypothesis, we compared the spatial distribution of Dreissena filter‐feeding intensity (determined from geostatistical modelling) to the spatial distribution of chlorophyll (determined from satellite remote sensing). To map the spatial distribution of Dreissena biomass and filter‐feeding intensity, we developed a geostatistical model based on point observations of mussel biomass measured in Lake Michigan in 1994/1995, 2000, 2005 and 2010. The model provided fine‐scale estimates of the spatial distribution of biomass for the survey years and provided estimates, with their uncertainty, of total biomass lakewide and within subregions. The approach outlined could be applied more generally to map the distribution of benthic biota in lakes from point observations. Total biomass of Dreissena in Lake Michigan, estimated from the geostatistical model, increased significantly over each five‐year period. The total biomass in units of 106 kg ash‐free dry mass (AFDM) (with 90% confidence interval) was 6 (4–8) in 1994/1995, 18 (14–23) in 2000, 408 (338–485) in 2005 and 610 (547–680) in 2010. From 1994/1995 to 2005, increases were observed in all regions of the lake (northern, central and southern) and in all depth zones (<30, 30–50, 50–90 and >90). However, from 2005 to 2010, for depths of <50 m, biomass declined in the northern region, remained constant in the central region and increased in the southern region; biomass continued to increase in all three lake regions for depths >50 m. The filter‐feeding intensity of Dreissena exceeded the benchmark spring phytoplankton growth rate of 0.06 day−1 in 2005 for depths <50 m (lakewide). In 2010, the filter‐feeding impact exceeded 0.06 day−1 within depths <90 m (lakewide), which greatly increased the spatial area affected relative to 2005. A regression analysis indicated a significant relationship between the reduction in satellite‐derived chlorophyll concentration (pre‐D. r. bugensis period to post‐D. r. bugensis period) and spatially co‐located filter‐feeding intensity (fraction of water column cleared per day) during periods when the water column was not stratified (December to April).

Food depletion regulates the demography of invasive dreissenid mussels in a stratified lake

“…Habitat differences in mussel productivity implied by our results have key implications for other Great Lakes (e.g., Michigan, Ontario, and Huron), where both overall dreissenid biomass and its recent increase have occurred primarily in the extensive profundal habitats (Rowe et al b ; Nalepa and Baldridge ), as well as for lakes with long periods of summer stratification more broadly. First, observed declines in mussel growth rates with depth and in profundal vs. littoral benthic habitats reflect a strong gradient in overall dreissenid impacts on pelagic resources because mussels in profundal habitats have access to (and consequently impact) epilimnetic primary production only during brief mixing periods.…”

“…Our analysis of long‐term trends in seston concentrations (turbidity) 1 m above the bottom and of our biophysical modeling results indicate that this effect is driven by depletion of a limited food supply in profundal habitats during long periods of summer stratification when phytoplankton production in the euphotic zone is not efficiently transported to the bottom by vertical mixing. At the same time, slower growth and higher longevity also explain the slower rates of biomass increase over time exhibited by profundal dreissenid populations in the Great Lakes (e.g., Rowe et al b ). Finally, limited recruitment in the profundal habitats owing to competition over limited food resources may ultimately limit mussel abundance and, if sufficiently strong, may produce long‐term fluctuations in quagga mussel populations and their ecosystem impacts.…”

“…Dreissenids are known to reduce food resources (e.g., seston and chlorophyll concentrations) at both lakewide scales (reviewed in Karatayev et al ; Higgins and Vander Zanden ; Pothoven and Fahnenstiel ; Rowe et al b ) and more locally within lake basins during thermal stratification (MacIsaac et al ; Ackerman et al ; Edwards et al ). For example, Rowe et al ( b ) showed reduced satellite derived chlorophyll a (Chl a ) concentration in Lake Michigan during periods of deep mixing, spatially associated with areas of high mussel biomass around the 30‐m to 50‐m depth contours. Pothoven and Fahnenstiel () found that Chl a in the hypolimnion (bottom 20 m) of Lake Michigan declined 63% and 54% between the pre‐ and post‐quagga mussel invasion periods in early and late summer, respectively.…”

“…Although similar dynamics have been observed in the nearshore zones (0–30 m) of lakes Erie, Michigan, and Ontario, offshore densities and biomass stabilized only in Lake Erie, which has the smallest profundal zone and was invaded by quagga mussels earlier than other Great Lakes (Karatayev et al b ). In deeper (> 90 m) profundal habitats of lakes Michigan and Ontario quagga mussels density continue to slowly increase (Nalepa et al ; Rowe et al, a ; Nalepa and Baldridge ), suggesting that mussels have not yet exhausted their food resources and reached their carrying capacity in the deepest areas of these lakes.…”

How are coastal benthos fed?

“…Blue line shows 50 m isobath. Mussel biomass redrawn from data by Rowe et al (), where mussel ash free dry weight (AFDW) converted to dry weight (DW) using ratio of 0.88 : 1.…”