The movement of ballast water is a major pathway for introducing aquatic invasive species. Ballast water management systems (BWMS) aim to reduce species introductions, but their performance varies under challenging water quality conditions. This study assessed BWMS performance in a freshwater port with impaired water quality by comparing harbour conditions to challenge water criteria for type approval testing defined by the International Maritime Organization and U.S. Environmental Protection Agency, and by analyzing paired harbour uptake and ballast discharge samples. Of 18 harbour samples, 67% approached or exceeded challenge water criteria for ≥50 μm organism abundance, while 22% and 11% exceeded criteria for dissolved organic carbon and total suspended solids. All nine discharge samples met limits for small organisms (≥10 to <50 μm), but eight exceeded limits for larger organisms (≥50 μm) despite >99% reduction from uptake levels. Abundance of larger organisms in discharge was affected by uptake water quality (ultraviolet transmittance and organism abundance), and BWMS treatment type/filter size. The findings suggest that raising challenge test criteria may improve BWMS reliability across environmental conditions.

Monitoring of environmental water quality in Hamilton Harbour, Lake Ontario, Canada, was conducted by long-term deployment of a YSI EXO multiparameter water quality sonde (YSI Incorporated, Yellow Springs, Ohio, USA) and semi-monthly collection of water samples for more comprehensive laboratory analysis. The sonde was installed at a depth of 5 meters inside a perforated PVC housing attached to the pier wall at Pier 25 (Agrico Canada L.P., Hamilton, Ontario, Canada, 43.272881, -79.786398) from April to December 2022 to take in situ measurements of temperature, pH, dissolved oxygen, salinity, turbidity, chlorophyll a (chl a), and fluorescent dissolved organic matter (fDOM) at hourly intervals.

Discrete samples of harbour water were collected semi-monthly (18 sample collection events) for assessment of live organism abundances and water chemistry. Briefly, a 4-liter Van Dorn bottle (Halltech Environmental and Aquatic Research Inc., Guelph, Ontario, Canada) was repeatedly deployed to a depth of 5 meters to collect a total of 12-15 liters of water into a 20 L carboy. The carboy was gently inverted five times and split into sample bottles (1 L for live counts of organisms in the ≥10 to <50 μm size class, 250 mL for particulate organic carbon (POC) and particulate organic nitrogen (PON), 125 mL for dissolved inorganic carbon (DIC) and dissolved organic carbon (DOC), 500 mL for total suspended solids (TSS) and volatile suspended solids (VSS), and 3 × 30 mL for measurement of Ultraviolet Transmittance (UVT)).

For collection of organisms ≥50 µm, a 30-L Schindler-Patalas trap (Science First/Wildco, Yulee, FL, USA) was deployed to a depth of 5 meters and the collected organisms were concentrated in the 35 µm mesh cod-end and rinsed into a 1-L Nalgene sample bottle using ambient water that had been pre-filtered through a 10 µm Nitex mesh sieve (Sefar Inc., Depew, New York, USA). The sample was topped up to 500 mL with 10 µm filtered ambient water. All samples were stored in the dark in a cooler at the ambient temperature of the harbour water (7–26 °C ± 2 °C) until laboratory analysis.

For the ≥10 to <50 μm size class, samples were analyzed using epifluorescence microscopy with fluorescein diacetate (FDA, Sigma-Aldrich Canada, Oakville, Ontario, Canada) as a vital marker, following the counting protocol outlined by Adams et al. (2014). Briefly, a 5 mL subsample was collected from a well-mixed 1 L sample of unconcentrated harbour water, stained with FDA for 10 minutes, with 1 mL subsequently transferred to a Sedgewick-Rafter counting chamber. The sample was examined for 20 minutes using a Zeiss Axio Vert.A1 microscope (Carl Zeiss Canada, Ltd, Toronto, Ontario, Canada); this process was repeated six times, and the mean abundance was calculated from the six replicate subsamples. All live organism counts were conducted within six hours from the time of collection.

For the ≥50 µm size class, samples were mixed by gentle inversion prior to splitting in half by weight. One half sample was concentrated to 40-60 mL using a 35 µm mesh sieve and 10 µm filtered harbour water for analysis. Aliquots of 0.5-2.5 mL, depending on the number of organisms, were examined in a modified Bogorov chamber using a Nikon SMz800N Zoom stereoscope at 30 – 80x magnification, with 50 μm fluorescent microspheres added as a size reference. Aliquot size was determined by the analyst to target a coefficient of variation (CV) of 10% while spending a maximum of 1-1.5 hours per sample (completed within 6 hours from the start time of collection). Analysts typically counted 17% of the half sample volume. For each aliquot, first, the number of dead ≥50 µm organisms were counted as per the ETV Protocol. Organisms were considered dead if they showed no movement or response to tactile stimuli. Secondly, the total number of organisms were counted. The number of live organisms were determined by subtracting the number of dead from the total (NSF International, 2010).

Water chemistry parameters, including particulate organic carbon (POC), particulate organic nitrogen (PON), dissolved inorganic carbon (DIC), dissolved organic carbon (DOC), and total suspended solids (TSS) were measured using standard analytical methods (National Laboratory for Environmental Testing in Burlington, Ontario, Canada), while UVT was measured using a RealUV254 Field Meter (Real Tech Inc., Whitby, ON, Canada).

Ships were sampled opportunistically at terminals within Hamilton Harbour, Ontario, Canada between April and December 2023. Ambient harbour water was sampled as close as safely practicable to each ship’s sea chest intake during ballast water uptake and, once ballast uptake operations were completed and the specified BWMS hold time for treatment was met, a treated ballast water sample was collected during discharge.

In situ water quality parameters (temperature, pH, dissolved oxygen, conductivity, salinity, turbidity, total algae, and depth) were measured using a YSI EXO multiparameter sonde installed prior to ballast water uptake, just outside each ship’s mooring lines, at a distance 31–70 m astern (mean distance: 51 m) and a depth of 5 m. The sonde was housed within a stainless steel protective cage (Nexsens Technology, Inc., Fairborn, Ohio, USA), secured to the pier wall, and fitted with anti-fouling measures. Data were recorded at 5-minute intervals throughout ballast uptake operations , with data retrieval conducted after the completion of all ballast operations at berth. Grab samples were taken for live counts of organisms in the ≥10 to <50 μm and ≥50 μm size classes, as well as for laboratory-based water chemistry analyses; these were collected at the start of ballast uptake operations and processed as described previously for harbour monitoring.

Treated ballast water samples were collected during ballast discharge using the ship-supplied sample port installed on the main ballast pipe in the ships’ engine room, using in-line isokinetic continuous sampling methods, a large volume collection device and a plankton net collection system consistent with prior studies (Bailey et al., 2022). Ballast water sample flow rates and volumes were measured using a flow meter with digital display (Burkert S030 paddlewheel body and SE35 digital display, Burkert Fluid Control Systems, Germany). Three 1 m³ continuous samples of ballast water were collected and concentrated using a 35 μm mesh net (30 cm diameter; Wildco-Science First, USA), kept partially submerged in a 75 L sampling bin outfitted with two valved discharge lines to direct disposal of filtrate in accordance with ship crew instructions. Following sampling, each net was rinsed with 10 μm filtered ballast water, and the concentrate was transferred to a 1 L Nalgene bottle. All samples were stored in dark, insulated coolers with ice packs to maintain temperatures close to ambient ballast conditions.

A continuous subsample of whole (unconcentrated) treated ballast water (total 20L) was collected for the duration of the ≥50 µm sampling for analysis of organisms ≥10 to <50 μm using a small sample probe (Moser et al., 2018; Casas-Monroy and Bailey, 2021) installed within the large volume collection device. Samples were stored at or just below ambient water temperature in dark, insulated coolers with ice packs until processing. A 5 L subsample was removed following mixing by gentle inversion and concentrated to 125 mL for live counts of the ≥10 to <50 µm size class using epifluorescence microscopy with FDA as described previously, though samples were concentrated to improve precision (Yardley et al., 2024).

Live counts of the ≥50 μm size class in ballast discharge samples were analyzed following the methods described in Bailey et al. (2022). Briefly, each 1 L sample was concentrated to 20-25 mL using a 35 µm sieve and 10 µm filtered sample water. Aliquots of 0.5 to 2.5 mL were examined using a modified Bogorov chamber and a Nikon SMZ800N Zoom stereoscope under 30 – 80x magnification, aiming for a target of 25–50 live organisms per aliquot, with 50 μm fluorescent microspheres added as a size reference. In contrast to the harbour monitoring counting method described previously, only ≥50 µm organisms showing movement or response to stimuli (NSF International, 2010) were counted since few living organisms were expected in samples post treatment. Samples were concentrated to maximize the volume assessed within 60-90 minutes (within six hours from the time of collection) to minimize mortality due to handling/containment. The abundance of live organisms was calculated cumulatively across the three net samples considering the volume counted and the concentration factor.