Features Equipment & Systems
Tracking the Benefits of Mill Environmental Investments Designed for Fish Protection

Worldwide, the pulp and paper industry undertook a major environmental modernization in the 1980s and 1990s. In Canada, this modernization helped the industry meet new effluent regulations and generally improved effluent quality. While these achievements are impressive, the main objective of pollution abatement programs is to protect the ecological integrity of waterways, particularly fish populations and communities, and their habitats. Hence, there is a need to gauge the success of industry efforts by studying fish in the wild.

For a variety of reasons, including natural variability and complex biotic-abiotic interactions in the receiving waters, the interpretation of such studies is difficult. In particular, establishing causality when differences between fish from reference areas and areas receiving mill effluent are observed. The methods for impact assessment work must be scientifically sound and capable of clearly establishing degraded environmental conditions. Equally, they must be able to document improvements or absence of harm.

There are many different ways to assess the biological status of wild fish, but they can be broadly classified into “sentinel species” and “community-based” approaches. The sentinel species approach is where one or two species are considered to be indicators of the condition of all the fish inhabiting a particular waterway. When sentinel species are used, the biological status of fish can be evaluated at the biochemical, tissue, or whole organism level. A further assumption is that inferences on potential effects at the higher level of biological organization (i.e., population, community) can be made from measurements at the lower levels of biological organization (e.g. biochemical).

On the other hand, community-based approaches provide a direct assessment of fish community structure such as the proportion of pollution-intolerant species.

Most of the field studies over the last decade have used sentinel species to assess the potential influence of mill discharges on fish. In fact, Canadian regulations require mills to conduct environmental effects monitoring (EEM) studies on two sentinel species in the receiving waters every three years. With sentinel species, distinct patterns were described by various research groups of what was considered typical for fish exposed to pulp and paper mill effluent. These included reduced plasma steroids and increased hepatic ethoxyresorufin-O-deethylase (EROD) activity at the biochemical level, increased liver size, decreased gonad size and fecundity from EEM-type studies, as well as increased frequency of health anomalies (e.g., gill erosion). Studies that directly assess the potential effects of mill effluents on the entire fish community are rare as they are considered to be more laborious.

Both the sentinel species approach and the community-based approach are thought to have unique as well as common advantages and disadvantages. Consequently, the simultaneous use of these approaches to assess the condition of fish in relation to mill effluents would be expected to provide valuable leads as to which methods provide the most promise for gauging the outcome of industry strategies designed for environmental protection.

At Paprican, scientists have have researched both sentinel species and community-based field approaches to assess the biological status of wild fish. The St. Franois River in Quebec, FIG. 1a and b, provided an ideal opportunity for such a study. Three mills discharge effluent to this river. Two of the three mills installed secondary treatment facilities in 1995. This resulted in their satisfying new BOD regulations, as well as those for suspended solids and acute lethal toxicity. There was also a general improvement in effluent quality determined by laboratory sub-lethal tests.

The third mill was already in compliance as it had operated a biological treatment system since the mill start-up in 1987. Further improvements were made in 1995 through process changes and fine-tuning of the treatment system. These changes at the mills made it possible to study the fish in the river on the basis of both temporal (i.e., before and after process/treatment changes at the mills) and spatial (i.e., upstream-downstream from the discharges) comparisons.

The temporal comparison is particularly attractive as it overcomes concerns about the virtually impossible task of selecting ‘true’ reference sites for studies based only on spatial (upstream/downstream) comparisons. A true reference site is one that is exactly like the exposure site except for the absence of effluent. The existence of dams that restrict fish movement between sites upstream and downstream from the mills on the St. Franois River minimized concerns about fish mobility and life history. These concerns plague all studies done in the field.

Different sentinel species approaches

The sentinel species selected for this study was the white sucker (Catostomus commersoni). The sampling occurred at six locations, Fig. 1b, during September and October in 1994, 1995 and 1998.

The fish survey component of the EEM, done at all six sites in those three years, requires information on sex, length, total body weight, liver weight, gonad weight, egg size/fecundity for female fish and age. These variables are used to derive the principal fish parameters related to size-at-age, condition, gonad weight, fecundity and liver weight which are estimated as regressions of one variable on another (typically length and weight) using analysis of covariance (ANCOVA).

The health assessment index (HAI) approach dealt with 15 variables including blood parameters, tissue and organ anomalies. Each variable is assigned a numeric value where the normal condition is rated as zero. Any abnormalities in the spleen, kidney, liver, eyes, gills and pseudobranchs are given values of 30. For the other variables, such as the thymus, abnormalities are rated as 10, 20 or 30 depending on the severity of the anomaly. For each fish, the variable rankings are summed and an HAI for the population is calculated by summing the individual HAIs and dividing by the number of fish in the population. The HAI was completed at six sites in 1998 and at three sites (1, 3 and 4) in 1995.

In the biochemical measurements approach, the liver and blood of individual white suckers were used to measure hepatic EROD activity and plasma steroid levels, respectively. Blood samples taken from females were analyzed for estradiol and testosterone while samples taken from males were analyzed for testosterone only. In 1995, fish from the sampling site (no. 2) downstream from mill A at East Angus were not tested for biochemical parameters. Fish from all sites were analyzed in 1998.

Fish community assessment

Between August 5 and 19, 1998, sampling for assessing the status of fish communities was done. The sampling locations, Fig. 1b, corresponded to those selected by the Quebec Ministry of the Environment for a 1991 study on the same stretch of the river. Each of the 13 sites was sampled along 500 m of shoreline on both sides of the river by using an electric shocker. All fish were identified to the species level, measured for length and weighed. A maximum of 20 fish of each species was examined for external anomalies such as skeletal deformation, fin erosion, lesions, tumors and the presence of parasites. For each sampling site, an index of biotic integrity (IBI) was calculated. According to this method, the lowest IBI score is 12, representing a very poor fish community. The highest score is 60, representing an excellent fish assemblage in terms of diversity, lack of anomalies as well as presence of pollution-intolerant species.

Comparison of methods for studying fish in the wild

A comparison of the results of the four approaches (three sentinel species and one community-based) used to study the condition of fish in the St. Franois River is summarized in Table I. At no time were the typical responses attributed to mill effluents seen when using the regulatory EEM or the biochemical (EROD/steroid) appro
aches. While differences between sites were found, there was no apparent interpretable pattern. Rather, the differences seemed to be more the result of natural variability. Thus, based on these methods, the fish were not considered to be deleteriously affected by mill effluents either before or after the installation of secondary treatment.

It appears that compounds such as phytosterols and wood extractives, previously shown or hypothesized to be capable of causing such responses, were not present in the river or were at levels below those able to cause effects. Based on the limited study using the HAI, some increases in abnormalities were found downstream from two of the three mills, but these could not be exclusively linked to the effluent discharges. Also, the type of abnormalities that contributed to the increased HAI, such as eye and liver anomalies, were more of the kind that can be found in fish from virtually all locations, as opposed to the types of anomalies previously reported for fish from sites affected by industry. In addition, the health of two other species of fish (darters and smallmouth bass) was found to be the same or even better downstream from the mills than upstream.

In contrast to the sentinel species approaches, the Quebec Ministry of the Environment concluded from its fish community assessment in 1991 that the fish assemblages were impaired (rated poor) below the two mills (A and B) discharging non-biologically treated effluents. This was not the case below the mill (C) discharging secondary-treated effluent where the fish community was found to be average. Three years after the installation of secondary treatment at the two mills, however, the fish communities downstream of the effluent discharges improved. They went from a poor classification to a level considered good and average, respectively, for Mills A and B such that there was little, if any, difference relative to the upstream stations. Downstream from mill C, the fish community improved to a good classification, probably reflecting better water quality throughout the river (also confirmed by local fishermen) as well as process modifications and fine-tuning of the treatment system at this mill. An attractive feature of the IBI approach is that it studies the community itself, the actual level of biological organization of interest, without the need for prediction from lower levels. Furthermore, it provides direction as to what is acceptable (e.g., average, good, excellent) and what is not (poor, very poor). Finally, in contrast to previous concerns regarding the level of effort required to conduct fish community studies, the time required to complete the IBI study was actually less than what was required for sentinel species.

Based on the results of this work, it is clear that the selection of the appropriate method for impact assessment work is crucial. Even when the appropriate method is chosen, the study design would likely influence interpretability. Issues related to reference site selection and fish mobility were mentioned earlier. Simple upstream/downstream (control/impact) designs make it difficult to draw clear conclusions about causality. The before-after-control-impact (BACI) study design used in this study is one way to overcome the interpretation problems related to simple control-impact (upstream/downstream) study designs. The opportunity for such a comparison strengthened the interpretability of the results of this study.

A further important conclusion from this work is that the interpretation of field study results should be made with careful consideration of the limitations, including study design, associated with the approach selected for a particular study. This is particularly important as the environmental performance of mill operations is based increasingly on the results of field studies.

Pierre Martel, Tibor Kovacs, Ron Voss, Paprican, Pointe-Claite, QC.

TABLE I. Summary results for field work conducted between 1991 and 1998 on the St. Franois River. The shaded area represents sampling done before the installation of secondary treatment at mills A and B. The narrative statements describe the condition of the downstream station for IBI and the conditions relative to the upstream station for EEM, EROD/steroids and HAI.

1991

1994

No typical response

1995

1998

ND: Not done