Can natural variability trigger effects on fish and fish habitat as defined in Environment Canada's metal mining environmental effects monitoring program?

Abstract

The Metal Mining Effluent Regulations (MMER) took effect in 2002 and require most metal mining operations in Canada to complete environmental effects monitoring (EEM) programs. An ''effect'' under the MMER EEM program is considered any positive or negative statistically significant difference in fish population, fish usability, or benthic invertebrate community EEMdefined endpoints. Two consecutive studies with the same statistically significant differences trigger more intensive monitoring, including the characterization of extent and magnitude and investigation of cause. Standard EEM study designs do not require multiple reference areas or preexposure sampling, thus results and conclusions about mine effects are highly contingent on the selection of a near perfect reference area and are at risk of falsely labeling natural variation as mine related ''effects.'' A case study was completed to characterize the natural variability in EEM-defined endpoints during preexposure or baseline conditions. This involved completing a typical EEM study in future reference and exposure lakes surrounding a proposed uranium (U) mine in northern Saskatchewan, Canada. Moon Lake was sampled as the future exposure area as it is currently proposed to receive effluent from the U mine. Two reference areas were used: Slush Lake for both the fish population and benthic invertebrate community surveys and Lake C as a second reference area for the benthic invertebrate community survey. Moon Lake, Slush Lake, and Lake C are located in the same drainage basin in close proximity to one another. All 3 lakes contained similar water quality, fish communities, aquatic habitat, and a sediment composition largely comprised of fine-textured particles. The fish population survey consisted of a nonlethal northern pike (Esox lucius) and a lethal yellow perch (Perca flavescens) survey. A comparison of the 5 benthic invertebrate community effect endpoints, 4 nonlethal northern pike population effect endpoints, and 10 lethal yellow perch effect endpoints resulted in the observation of several statistically significant differences at the future exposure area relative to the reference area and/or areas. When the data from 2 reference areas assessed for the benthic invertebrate community survey were pooled, no significant differences in effect endpoints were observed. These results demonstrate weaknesses in the definition of an ''effect'' used by the MMER EEM program and in the use of a single reference area. Determination of the ecological significance of statistical differences identified as part of EEM programs conducted during the operational period should consider preexisting (background) natural variability between reference and exposure areas. Integr Environ Assess Manag 2013;9:155-163. ß 2012 SETAC

Development of an adaptive monitoring framework for long‐term programs: An example using indicators of fish health

“…In some studies, long-term collections have been done, but few sites were sampled (Mazor et al 2009). Other studies have shown differences larger than a default CES, but used few reference sites (Mackey et al 2013). Studies with interpretive limitations would benefit from the addition of more information collected and comparable over multiple scales (Scarsbrook et al 2000;Ballantine and Davies-Colley 2014), but this is clearly easier said than done.…”

Evaluation of Lethal Fish Sampling and the Nonlethal Alternative Under the Canadian Metal and Diamond Mining Environmental Effects Monitoring Program

“…"Guidance" refers to the sample size recommendations provided by EC (2012a). This statistical disadvantage has been exemplified in the data analysis performed by Mackey et al (2013), who conducted a mock EEM study in a baseline environment sampled prior to treated effluent release. CAN = Canada; EC = Environment Canada; EEM = environmental effects monitoring; SK = Saskatchewan.…”

“…"In Practice" data were totaled at the mine and provincial levels, whereas at the national level data were extrapolated using Saskatchewan data, which led to the inclusion of standard deviations. When CES were applied to lethal effect endpoints that statistically differed between reference and exposure areas, Mackey et al (2013) found fewer differences that would be considered effects requiring focused monitoring at an operational mine site under the revised regulations. population alternative under the EEM program.…”

“…From a mine proponent perspective, the lack of nonlethal CES in the MDMER is a statistical disadvantage because it does not provide a regulatory threshold that distinguishes between natural variability and ecologically relevant changes. This statistical disadvantage has been exemplified in the data analysis performed by Mackey et al (2013), who conducted a mock EEM study in a baseline environment sampled prior to treated effluent release. When CES were applied to lethal effect endpoints that statistically differed between reference and exposure areas, Mackey et al (2013) found fewer differences that would be considered effects requiring focused monitoring at an operational mine site under the revised regulations.…”

An Adaptive Environmental Effects Monitoring Framework for Assessing the Influences of Liquid Effluents on Benthos, Water, and Sediments in Aquatic Receiving Environments

“…The minimum acceptable design includes 1 reference and 1 exposure area, despite decades of peer-reviewed literature urging the use of more complex designs that more robustly describe reference-area variability (Green 1979;Stewart-Oaten et al 1986;Underwood 1991Underwood , 1994Underwood and Chapman 2003). Huebert et al (2011) and Mackey et al (2012) identified that the minimum design has a good probability of resulting in the detection of differences that are unrelated to exposure to effluent and repeated the need for better design and the use and interpretation of data from more comprehensive study designs. Huebert et al (2011) and Mackey et al (2012) identified that the minimum design has a good probability of resulting in the detection of differences that are unrelated to exposure to effluent and repeated the need for better design and the use and interpretation of data from more comprehensive study designs.…”