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

Three main response patterns were observed in the Cycle 2 EEM fish survey (Lowell et al. 2003). These responses, and the changes in the EEM endpoints associated with them, have been widely described in the literature (see Munkittrick et al. 1991, 1994, 2000 for reviews). It should be noted, however, that other response patterns may also occur at some mills. The first of the three main patterns, nutrient enrichment, was generally associated with increases in gonad and liver weight, as well as increases in condition and often growth rate (weight at age). The second main pattern, nutrient limitation together with chemical toxicity or other inhibitory effects, was associated with decreases in these endpoints. Nutrient limitation is defined broadly here to include some combination of limited availability of food, appetite suppression and/or internal alteration of food absorption (leading to decreases in several endpoints). Factors other than nutrient availability (such as chemical toxicity) may also contribute to decreases in several of the endpoints. Chemical toxicity may occasionally lead to increased liver size (as part of the detoxification mechanism), together with decreases in the other endpoints.

The third, and most prominent, response pattern seen in Cycle 2 was associated with increases in condition and liver weight, and decreases in gonad weight. This was the national average response pattern (Fig. 4) and is generally indicative of nutrient enrichment coupled with metabolic disruption (Munkittrick et al. 2000). Thus, the national averages (grand means) in Figure 4 show that fish in effluent-exposed areas were significantly faster growing (increased weight at age), were significantly fatter (increased condition), and had significantly larger livers than fish in reference areas, but exposed fish also had significantly smaller gonads. This response pattern can occur when effluent exposure disrupts normal allocation of resources to gonadal development and may include some element of endocrine disruption associated with difficulties in producing sufficient sex steroid hormones (Munkittrick et al. 1991, Van Der Kraak et al. 1992, Damstra et al. 2002).

Figure 4: Grand means for five key fish endpoints for Cycles 2 (C2) and 3 (C3). Error bars represent 95% confidence intervals. Number of comparisons for: Age (C2 = 133; C3 = 138), Condition (C2 = 123; C3 = 123), Gonad (C2 = 126; C3 = 124), Liver (C2 = 128; C3 = 129), Weight at Age (C2 = 100; C3 = 105).

While Cycle 3 showed a similar diversity of response patterns, the national average response was quite similar to Cycle 2, with exposed fish exhibiting significant increases in growth rate, condition and liver size, but a significant decrease in gonad size (Fig. 4). Thus, in Cycle 3, the national average pattern was again consistent with one of nutrient enrichment coupled with metabolic disruption. Although the depression in gonad size remained unchanged since Cycle 2, the increase in condition and growth rate was lessened, perhaps signalling a reduction in nutrient enrichment effects in Cycle 3. An additional change that was observed between cycles was a switch from increased average age for exposed fish in Cycle 2 to decreased average age in Cycle 3. Note that the age and weight at age endpoints are included here for informational purposes, although they should be interpreted with care due to difficulties in aging fish at some mills (for this reason, these two endpoints are not being used to direct mills to more extensive monitoring in Cycle 4). Future cycles of data collection will provide more information on this issue.

Figure 5 shows the results of meta-analyses carried out to look at the influence of gender on fish responses. Greater differences were observed between genders in Cycle 3 compared with Cycle 2, where gender differences were observed only for the gonad endpoint. When breaking the results down by gender for Cycle 3, it is apparent that the lessening of nutrient enrichment effects was most pronounced for females (liver and condition endpoints). The change in the age response was also greater for females. In contrast, the female versus male gonad response in Cycle 3 was almost identical to that in Cycle 2, with the gonad size reduction being significantly greater for females in both cycles.

Figure 5: Five key fish endpoints, by gender, for Cycles 2 (C2) and 3 (C3). M = male, F = female. Error bars represent 95% confidence intervals. Number of comparisons for Age (C2: F = 67, M = 66); C3: F: 73, M = 65); Condition (C2: F = 61, M = 62; C3: F = 65, M = 58); Gonad (C2: F = 66, F = 60; C3: F = 65, M = 59); Liver (C2: F = 64, M = 64; C3: F = 70, M = 59); Weight at Age (C2: F: 50; M = 50; C3: F = 56, M = 49).

The national responses for Cycle 3 can also be subdivided by species (Fig. 6). The white sucker national average response pattern was the same as the overall national average pattern, and was consistent with nutrient enrichment in combination with metabolic disruption (characterized by increased liver size, condition and growth rate and decreased gonad size). This was likely a reflection of the large proportion of surveys that used white suckers as sentinel species (see Table 4). None of the other species was used at a large numbers of mills in Cycle 3. Nonetheless, some of the species that were less commonly used are included in Figure 6 for comparative purposes. Interpretations of the average responses shown by these other species should be done with care, given their low frequency of use. Of note, however, is the response pattern observed for mummichogs. This was the most commonly used marine species, and it exhibited a significant decrease in all the endpoints, generally indicative of a nutrient limitation or toxicity response pattern. This response is consistent with the overall inhibitory response pattern observed for benthic invertebrate communities in marine habitats, which will be discussed in section 6. The fish data were not subdivided by habitat due to a lesser diversity of sampled habitats relative to the invertebrates (leading to low sample sizes for some habitat types) as well as their higher mobility relative to benthic invertebrates (so more difficult to assign to either erosional or depositional habitat types).

Figure 6: Five key endpoints, by species, for Cycle 3. Error bars represent 95% confidence intervals. LS = long-nose sucker (n = 10), WS = white sucker (n = 34), MU = mummichog (n = 7), YP = yellow perch (n = 6). Note that the number of comparisons (n) for each species is an average of all five endpoints. There were cases where certain measurements were not completed on some fish; therefore, the number of comparisons would be slightly lower for those endpoints.

Overall, the response patterns in Cycle 3 were quite similar to those observed in Cycle 2 (Fig. 4). The reduction in gonad size was unchanged between cycles. The Cycle 3 responses did, however, suggest a lessening in nutrient enrichment effects relative to Cycle 2. That is, while there were still significant increases in condition and growth rate in exposed areas in Cycle 3, they were not as great as observed in Cycle 2. This slight shift in fish response may have been due to a variety of possible causes, as discussed in section 8.