Back to Results
	A multi agency project funded by US EPA's STAR Program

Results: Diatom-based Inference Models

We aimed to devise and test a new diatom-based model for water quality variables in the coastal Great Lakes. Diatoms, the siliceous yellow-brown algae, have been used extensively as indicators of water quality conditions. They are well-accepted as indicators by the U.S. EPA and other agencies, and have been used to evaluate aquatic stressors worldwide, including Europe and several areas of North America (including the Great Lakes). Diatom species respond with great fidelity to stressors associated with major “pressure” indicators in the Great Lakes (e.g., nutrient and salinity loading, siltation, and factors affecting water transparency, including exotic species). Diatom remains preserve well in sediment, so they also provide opportunities to reconstruct historical information at a site (paleolimnology). A number of other features make diatoms robust environmental indicators:

1.  ubiquitous; occur in virtually any aquatic environment
2.  diverse; can provide a fine-grained assessment of environmental conditions
3.  versatile; sensitive to a variety of stressors, particularly water chemistry
4.  short turnover rate; respond rapidly to changing conditions
5.  more time-integrative than “snapshot” environmental measurements
6.  narrow tolerances and specific optima to environmental conditions

Sampling Design

Diatom and chemical samples were collected from 237 coastal sites. At each site, a suite of important water quality parameters was collected, including chemical (e.g., nutrients) and physical (e.g., turbidity) variables. Diatom assemblages from each site were counted and identified to the most specific taxonomic level possible using light microscopy. Diatom data were compiled for comparison to corresponding environmental data and development of indicator models.

Indicator Application

We developed a diatom-based model to infer total phosphorus concentration (a water quality parameter typically related to augmented nutrient load and cultural eutrophication). Over 300 taxa were identified as comprising the majority of the coastal diatom population in the Great Lakes, and their environmental optima and tolerances were identified. Model testing (weighted average regression and calibration) has confirmed that the environmental characteristics of the diatoms provide a robust means to infer phosphorus concentrations for a study site (left). In the future, the evaluation of diatoms in coastal ecosystems will provide cost-effective management advantages over traditional chemical measurements; the diatom assemblages are a direct measure of the health of the primary producer community, and diatom-inferred data integrate conditions than can be overlooked with instantaneous sampling of chemical parameters. For instance, the diatom assemblage in a periphytic sample provides a time-integrated assessment at a site that is subject to erratic pulses of nutrient contamination. Indeed, it was clearly recognized that diatom-based reconstructions better predicted water quality than snapshot measurements from Great Lakes coastlines!