Field Ecology

Assessing Aquatic Health in the Frenchman's Bay Watershed

Benthic macroinvertebrates and road-salt runoff in an urbanized Lake Ontario lagoon

2024 · ENV316 Ecology · University of Toronto

GIS map of the Frenchman's Bay watershed in Pickering with three georeferenced TRCA macroinvertebrate sampling stations marked above the lagoon.
Key result

Benthic macroinvertebrate sampling returned a Hilsenhoff Biotic Index of 5.73, placing the Frenchman’s Bay watershed in the "fair" water-quality band despite sitting in one of Canada’s most densely urbanized watersheds.

Objective

The question was simple: how healthy is the water in Frenchman’s Bay, a 27 km² watershed in Pickering that drains some of Canada’s oldest and most densely urbanized land before it reaches western Lake Ontario? Rather than rely on a single chemical reading, the assessment used the animals living in the sediment as the primary signal, then checked road-salt runoff as a second line of evidence.

Approach

Benthic macroinvertebrates were collected and sorted following the Ontario Stream Assessment Protocol (OSAP), then identified to order under a stereo microscope (specimen #10, for example, keyed out to Megaloptera). Counts from the whole class were pooled into a single community dataset. From there I calculated taxa richness, the Shannon-Wiener diversity index, Jaccard similarity, and Bray-Curtis dissimilarity to compare groups, and the Hilsenhoff Biotic Index to translate pollution tolerance into a water-quality score.

TRCA invertebrate records from a south-shore site (2015 to 2016) gave a short time series. Sodium was chosen as the water-quality parameter because road salt is the dominant urban contaminant here, and a t-test in R compared sodium concentrations between years. Sampling stations (4854517, 4853833, and 4854328) were georeferenced and mapped in GIS so the biology stayed tied to real locations.

Findings

The pooled Hilsenhoff Biotic Index came out at 5.73, which falls in the “fair” band. Trichoptera (caddisflies) were the most abundant order, and class taxa richness reached 15. Across the 2015-to-2016 TRCA record, abundance rose from 2,808 to 2,918 and Shannon-Wiener evenness climbed from 0.86 to 1.00 while richness held at 10. The sodium t-test returned p = 0.088, above the 0.05 threshold, so the year-over-year change wasn’t statistically significant, and reported sodium stayed under the 50 mg/L guideline.

Why it matters

For a watershed under heavy urban pressure, “fair” plus the presence of sensitive taxa points to limited pollution rather than collapse. The honest caveat is the small sample: with only two years of sodium data (df = 1), the t-test can’t rule out a real trend, and it shouldn’t. The clearest takeaway is methodological. Pairing biotic indices with water chemistry and GIS gives a fuller, more defensible read on watershed health than any one measure alone, and it flags exactly where longer monitoring is needed.

Figures

Diversity and similarity metrics from TRCA invertebrate data for a south-shore site, 2015 to 2016. Abundance and Shannon-Wiener evenness both rose while taxa richness held at 10.
Summary table of six ecological metrics for a Frenchman’s Bay site across 2015 and 2016: abundance 2,808 then 2,918, richness 10 and 10, Shannon-Wiener 0.86 then 1.00, HBI 1.85 then 2.09, Jaccard 0.2, Bray-Curtis 0.074.

Diversity and similarity metrics from TRCA invertebrate data for a south-shore site, 2015 to 2016. Abundance and Shannon-Wiener evenness both rose while taxa richness held at 10.

Specimen #10 identified under the microscope as Megaloptera (alderfly/dobsonfly larva). Order-level ID of each picked specimen fed the taxa-richness counts and Hilsenhoff scoring.
Stereo-microscope photo of a curled brown benthic macroinvertebrate larva with paired hooked anal prolegs, labeled specimen #10 Megaloptera.

Specimen #10 identified under the microscope as Megaloptera (alderfly/dobsonfly larva). Order-level ID of each picked specimen fed the taxa-richness counts and Hilsenhoff scoring.

Raw R output from the sodium t-test. At p = 0.088 the year-to-year change in sodium was not statistically significant, so the data could not confirm a rising road-salt trend on this record.
R console output for a t-test on sodium: t = 7.2018, df = 1, p-value = 0.08784, with a 95 percent confidence interval and mean of x 4289.4.

Raw R output from the sodium t-test. At p = 0.088 the year-to-year change in sodium was not statistically significant, so the data could not confirm a rising road-salt trend on this record.