Warning This Web page has been archived on the Web.

Archived Content

Information identified as archived on the Web is for reference, research or recordkeeping purposes. It has not been altered or updated after the date of archiving. Web pages that are archived on the Web are not subject to the Government of Canada Web Standards. As per the Communications Policy of the Government of Canada, you can request alternate formats on the Contact Us page.

Help the Government of Canada organize its website!

Complete an anonymous 5-minute questionnaire. Start now.

Skip booklet index and go to page content

National Assessment of Pulp and Paper Environmental Effects Monitoring Data: Findings from Cycles 1 through 3

7.2 Summary of Effect Sizes

The Cycle 3 distribution and range in measured exposure versus reference area percent differences for abundance, taxon richness, Simpson’s evenness and the Bray-Curtis index of dissimilarity were similar to those in Cycle 2 (Fig. 7; Lowell et al. 2003). As was found for fish, this underscores the reproducibility of measured differences at a national scale in Cycle 3 relative to Cycle 2 and likely reflects the consistency of response patterns from one cycle to the next (see also section 6.4). The measured differences were calculated as the exposure area mean minus the reference area mean, expressed as a percentage of the reference area mean. Figure 7 provides a complete summary of the differences calculated for control/impact mills (i.e., includes both statistically significant and non-significant differences). This figure does not include mills that conducted a gradient design, since percent differences (as computed here)annot be calculated for this type of study design; note, however, that standardized effect sizes can be calculated for gradient designs for the purposes of meta-analysis (see section 3 and section 6.4).

Similar to Cycle 2, the abundance endpoint showed the most extreme range, varying from a decrease of 80% to an increase of over 5500%. The percent differences distribution for taxon richness ranged from a decrease of 67% to an increase of 147%. Ranges for the Bray-Curtis index and the Simpson’s evenness index were -33% to 471% and -58% to 227%, respectively. Note that most of the Bray-Curtis values were positive, due to the method of calculation. The few negative values were due to unusual data distributions.

Distribution of measured percent differences between exposure and reference areas for the benthic invertebrate survey (control/impact designs only) for a) abundance, b) taxon richness, c) Bray-Curtis and d) Simpson’s evenness

Figure 7: Distribution of measured percent differences between exposure and reference areas for the benthic invertebrate survey (control/impact designs only) for a) abundance, b) taxon richness, c) Bray-Curtis and d) Simpson’s evenness

Figure 8 was developed using all of the Cycle 3 results (control/impact and gradient designs) and shows the number of mills that had no significant effect versus a significant effect for the four endpoints. The figure also further divides the mills that had a significant effect into those where the effect was less than the CES of ± 2 standard deviations (± 2SD) and those where the effect exceeded the CES of ± 2SD. For both Cycle 2 and Cycle 3, the highest percentages of the mills were non-significant for taxon richness, abundance and Simpson’s evenness. The Bray-Curtis index was the most sensitive endpoint in both cycles, and this is shown by a high number of mills having a significant effect that exceeded ± 2SD.

Number of mills showing no significant difference, a significant difference less than the CES of ± 2SD and a significant difference greater than the CES of ± 2SD

Figure 8: Number of mills showing no significant difference, a significant difference less than the CES of ± 2SD and a significant difference greater than the CES of ± 2SD. Includes both control/impact designs (n = 67) and gradient designs (n = 20). Note that CES have not yet been developed for gradient designs; therefore, mills that used a gradient design and found a statistically significant effect are included under “significant and > CES.” See text for additional information. Note also that for the gradient designs, the Bray-Curtis index was not calculated because it requires highly site-specific information that was not available at a national scale (Lowell et al. 2003).

Although Figure 8 shows that many mills did not show a significant effect when considering each endpoint individually (except Bray-Curtis), approximately 54% of the mills that conducted benthic invertebrate studies found at least one significant effect in at least one of the benthic invertebrate endpoints in Cycle 3. Approximately half of these mills exceeded the CES of ± 2SD for taxon richness or abundance in Cycle 3.

For control/impact designs, a mill will conduct focused monitoring (including magnitude and geographic extent) for Cycle 4 if the CES for abundance or taxon richness was exceeded (with statistical significance) in the same direction for two consecutive cycles (Cycles 2 and 3). For gradient designs, if a mill measured a statistically significant effect (significant correlation coefficient) for abundance or taxon richness in the same direction for both Cycles 2 and 3, the mill would conduct more focused monitoring in Cycle 4. Given the inherent lower power for detecting effects when using the gradient design relative to the control/impact design (Lowell et al. 2003), it is likely that only gradient mills having particularly large effects would conduct more extensive monitoring. Similarly, only control/impact mills having particularly large effects would conduct more extensive monitoring, due to the dual requirement for control/impact designs of statistical significance and exceedance of the CES.

By examining the number of exceedances of CES in both Cycles 2 and 3 as an indicator of sites where more serious impacts are occurring, it is estimated that more extensive monitoring effort would be undertaken at approximately 20% of the mills conducting a benthic invertebrate community survey. Approximately half of these mills exceeded the CES for abundance, about one-third exceeded the CES for taxon richness, and the remaining mills exceeded the CES for both abundance and taxon richness. The majority of these mills exhibited significantly higher endpoint values in the exposure area; however, most mills where the CES was exceeded solely for taxon richness showed significantly lower endpoint values in the exposure area. This was also observed in Cycle 2. The significant increases observed in abundance were usually an indication of overall eutrophication. The decreases in taxon richness, in turn, were likely a reflection of more pronounced eutrophication or toxicity/smothering (see following section).