Northern Watershed Project (Report 3). Cumulative Effects of Watershed Disturbances on Stream Fish Communities in the Kakwa and Simonette River Basins, Alberta


Author(s)

Garry Scrimgeour, Paul Hvenegaard, John Tchir, Sharon Kendall and Alan Wildeman

Summary

Study Focus

Developing management strategies that minimize the cumulative effects of human-induced disturbances on ecological systems is arguably the single largest challenge to sustainable resource management. In Alberta and elsewhere, rapid expansion of the forestry and oil and gas sectors, combined with conversion of forested lands to agriculture has raised concerns about the ecological sustainability of the boreal forest. The current lack of understanding of the cumulative effects of watershed disturbances on stream fish within Alberta’s boreal forest is a central challenge to the management of stream fish assemblages.

We quantified the cumulative effects of watershed disturbances arising from industrial activities on stream fish communities in the Kakwa and Simonette River basins, Alberta, Canada using data collected between 1994 and 2001. Data on fish abundance, community composition and watershed attributes, including descriptors of disturbance, to address the following three focal questions:

1) Is the presence of fish, game fish and individual species affected by watershed disturbances?
2) Are species density and biomass affected by watershed disturbances?
3) Is fish community structure affected by watershed disturbances?

Summary of Study Findings

Overall patterns in fish communities

Analyses of fish communities from a total of 528 stream reaches revealed marked differences in fish communities between the Kakwa and Simonette River basins. The Kakwa watershed supports 9 species of fish comprising representatives from five familial groups whereas the Simonette River Basin supports 20 species of fish from nine Families. I

n the Kakwa River Basin, bull trout and sculpin were the most frequently occurring species followed by Arctic grayling, mountain whitefish and rainbow trout. Longnose sucker, longnose dace, burbot and white sucker occurred at relatively few sites.

In the Simonette River Basin, sculpin, lake chub, bull trout, and white sucker were the most frequently occurring species followed by Arctic grayling, mountain whitefish, longnose sucker, northern redbelly dace, longnose dace, pearl dace, brook stickleback and redside shiner. Troutperch, burbot, emerald shiner, finescale dace, flathead chub, northern pike and largescale sucker occurred at relatively few sites.

Based on density estimates, fish communities in the Kakwa were numerically dominated by sculpin, rainbow trout, bull trout and Arctic grayling. When combined, these four species accounted for the vast majority of all fish encountered. In contrast, northern redbelly dace, sculpin, lake chub, white sucker, brook stickleback and pearl dace were the predominant species in the Simonette River Basin. Overall density of fish in Kakwa watershed was about four-fold lower than that in the Simonette watershed.

Focal question 1: Is the presence of fish, game fish and individual species affected by watershed disturbances?

Logistic regression indicated that the presence of fish, game fish and individual species were moderately to highly predictable based on watershed area, stream width, elevation and to a lesser extent reach slope, size composition of the substratum and stream bank width.

At the stream reach scale, the occurrence of fish in both the Kakwa and Simonette River basins was strongly affected by stream size (i.e., order) and to a lesser extent watershed type.

With only two exceptions, the presence of fish, game fish and individual species were unrelated to watershed disturbances arising from the cumulative effects of industrial activities.

The two exceptions reflect the negative relations between bull trout presence and the: i) cumulative percent watershed disturbance in the Kakwa River Basin and ii) cumulative density of stream crossings in the Simonette River Basin.

Focal question 2: Is fish density and biomass affected by watershed disturbances?

Regression analyses generally showed that total fish density and density of the predominant species in the Kakwa watershed were primarily related to stream wetted width (i.e., the width of the water surface), elevation and reach slope. Fish density was generally highest in small streams or those at high elevation and decreased with increasing width or lower elevation. In general regression models explained relatively little of the overall variance in total density and density of the most abundant species and non-linear models did not typically explain appreciably more variance than linear models.

With some exceptions, our analyses generally showed that total fish density and biomass and density and biomass of the numerically dominant species and species groups were unrelated or poorly related to watershed disturbances attributes including harvesting, stream crossing attributes and their underlying attributes (e.g., percent of the watershed disturbed by roads, harvest blocks, seismic lines, pipelines, and stream crossings by roads, seismic lines, power lines and pipe lines).

In the Kakwa River Basin, the notable exceptions to these findings were the positive relationships between: i) total fish density and percent watershed disturbance and, ii) density of sculpin and percent watershed disturbance.

In the Simonette watershed, the notable exceptions were a): the positive relationship between: i) total fish density and stream crossing density, ii) density of dace and stream crossing density, iii) total biomass and percent watershed disturbance, iv) biomass of sculpin and seismic line density, and iv) biomass of shiner and crossing density.

While we report some statistically significant relations between fish density and biomass and watershed disturbance attributes, the majority of these relations did not explain substantial amounts of variance in fish density or biomass.

Focal question 3: Is fish community structure affected by watershed disturbances?

We quantified the cumulative effects of watershed disturbances on fish community structure in the Kakwa and Simonette River basins using a reference condition approach. This approach evaluates the extent to which potentially impacted sites contain fish assemblages predicted by relationships between fish community structure and habitat variables derived from reference (i.e., least-impacted) sites. If the empirical model (discriminant function model) derived from the least-impacted sites also explains similar amounts of variance in fish assemblage membership in the potentially impacted sites then it is assumed that impacts are not detectable.

Kakwa River Basin

In the Kakwa River Basin, cluster analyses of 62 reference sites (i.e., least disturbed sites) using percent composition data identified three discrete fish assemblages. Assemblage 1 consisted primarily of bull trout, assemblage 2 was dominated by sculpin whereas mountain whitefish, and to a lesser extent rainbow trout, bull trout, Arctic grayling, and sculpin dominated assemblage 3.

Discriminant function analyses were used to determine linkages between the three fish assemblage types and habitat variables. Results of these analyses showed that stream wetted width, stream reach slope, site elevation and percent small gravel were moderately powerful discriminators among the three assemblage types and had an overall classification success of 71.0%.

Stream reaches supporting fish assemblage 1 were typically located at high elevations, were relatively narrow, with high reach slope and stream beds with low amounts of small gravel. Reaches supporting fish assemblage 2 were typically located at lower elevations, with broader stream channels, and higher amounts of small gravels.

Stream reaches supporting fish assemblage 3 were typically located at low elevations, were relatively broad, with moderate reach slope and stream beds that contained low amounts of small gravel. We evaluated larger patterns in fish communities initially by completing cluster analyses of least-impacted sites and potentially impacted sites. In addition to identifying the three fish assemblages from the leastimpacted sites, clustering also identified two more assemblage types. The first new assemblage consisted primarily of sites dominated by Arctic grayling, and to lesser extent bull trout mountain whitefish and rainbow trout. The second new assemblage type was numerically dominated by primarily by rainbow trout which likely originated from fish stocked into the adjacent Musreau Lake.

We used discriminant function analyses to quantify how well fish communities at the potentially impacted sites (i.e., test sites where watershed disturbance ranged from 10 to 43%) were predictable based on the four watershed variables of stream wetted width, reach slope, elevation and percent small gravel within the substratum.

Our results showed that the four habitat variables of stream wetted width, reach slope, elevation and percent small gravel were poor predictors of fish assemblage type of the potentially impacted test sites and overall had a classification success of only 50%.

These results indicate that the discriminant function model developed from the least-impacted reference sites was a poor predictor of fish assemblage structure in the potentially impacted sites and that fish communities in the Kakwa River Basin are affected by the cumulative impacts of current levels of industrial activities.

Simonette River Basin

In the Simonette River Basin, cluster analyses of 106 reference (i.e., least impacted) sites identified five relatively discrete fish assemblages. Assemblage 1 consisted primarily of white sucker and to a lesser extent finescale dace, lake chub, trout-perch and redside shiner whereas assemblage 2 was dominated by lake chub and to a lesser extent white sucker, longnose sucker, sculpin and finescale dace. In contrast, northern redbelly dace, lake chub, pearl dace dominated assemblage 3. Fish communities comprising assemblages 4 and 5 were numerically dominated by mountain whitefish, pearl dace and Arctic grayling (assemblage 4) and sculpin (assemblage 5).

As for the Kakwa River Basin, discriminant function analyses were used to determine linkages between the five fish assemblages and habitat variables. Results of these analyses showed that stream wetted width, stream reach slope, site elevation and percent of fine materials in the substratum were moderately powerful discriminators among the five fish assemblages and had an overall classification success of 71.1%.

Stream reaches supporting fish assemblage 1 were located within moderately large, low elevation reaches within moderately large watersheds and characterized by low reach slope and with substrata dominated by fine sediments. Reaches supporting assemblage 2 were moderately small, low gradient systems located at intermediate elevations whereas assemblage 3, were typically located in small, low elevation reaches. Fish communities comprising assemblages 4 and 5 were located at high elevations with moderate reach slope but differed in stream size and the percent of fine materials within the stream bottom. Streams comprising assemblage 4 were typically large with a predominance of fine materials in the river bed compared with the small streams comprising assemblage 5 where fine sediments were relatively rare.

We used discriminant function analyses to quantify how well fish communities at the potentially impacted sites (i.e., test sites where watershed disturbance ranged from 21 to 61.8%) were predictable based on the four watershed variables of stream wetted width, reach slope, elevation and percent fine materials within the substratum.

Our results showed that the four habitat variables of wetted width, reach slope, elevation and percent fine materials within the substratum were moderately poor predictors of fish assemblage type and overall had a classification success of only 57.1%. T

hese results indicate that the discriminant function model developed from the least-impacted reference sites was a poor predictor of fish assemblage structure in the potentially impacted sites and that fish communities in the Simonette River Basin are affected by the cumulative impacts of current levels of industrial activities.

Management Implications and Recommendations

Towards an improved understanding of the cumulative effects of humaninduced activities of stream fish communities

Rationale

The expansion of Alberta’s forest industry since the mid-1980’s combined with conversion of forest lands to agriculture and increased oil and gas activities has raised concerns about the ability of ecological sustainability of stream fish communities in northern Alberta. These industrial activities have the potential to affect stream fish communities by influencing the quantity and quality of habitat for stream fishes. Our results showed that current levels of industrial activity have detectable cumulative effects on stream fish communities in the Kakwa and Simonette River basins. Such effects were linked with forest harvesting and linear disturbances that intercept streams.

Recommendation

If management of the boreal forest is to be based on ecological considerations in addition to economic, social, political and cultural factors values, a more detailed understanding of the cumulative effects of multiple industrial activities on fish communities is required. Our pursuit of a knowledge base sufficient to understand the cumulative impacts of industrial activities on stream fish communities is in its infancy and this lack of knowledge challenges our ability to manage resources in a sustainable fashion. As a result we recommend an increased commitment to understand the cumulative effects of industrial activities on stream fish and other biotic communities in forested regions of Alberta.

Development and implementation of stream fish monitoring program

Rationale In Alberta, fish communities and aquatic environments are protected under both provincial and federal legislation. For instance, consequences of watershed activities on aquatic environments and the biological diversity that they support are considered within Provincial Acts and Regulations (e.g., The Water Act, Timber harvest and planning operational Ground Rules [Anonymous 1994]). These provisions include operating ground rules for forest harvesting practices, and provincial codes of practice for: 1) watercourse crossings, 2) pipeline and telecommunication line crossings and 3) temporary diversion of water for hydrostatic testing. Operating ground rules define a set of forest harvesting practices including the protection of 30 and 60 m buffers adjacent to small and large permanent streams and practices that reduce inputs of organic matter into stream and avoidance of stream channels. Fish communities and aquatic habitats are also protected under habitat protection provisions (Section 35 [1]) of the Federal Fisheries Act which prohibits works or undertakings that result in the harmful alteration, disruption or destruction of fish habitat, while Section 35 [2], allows for authorization by the Minister, or under regulation, of harmful alteration, disruption or destruction of fish habitat. While these provisions are intended to protect stream fish populations, a scientifically rigorous program to monitor the consequences of industrial activities on stream fish communities does not exist.

Recommendation

The absence of effective monitoring programs compromises our ability to manage stream fish communities. Rigorous monitoring programs are required to: i) understand current trends in fish populations, ii) evaluate the ecological effects of anthropogenic and natural disturbances on stream fish communities and iii) evaluate the effectiveness of restoration measures, iv) to critically assess the effectiveness of current watershed management practices. As such, we recommend the allocation of resources to develop and implement a scientifically rigorous stream monitoring program in northern Alberta. This monitoring program should also include evaluations of focal watersheds, such as the Prairie Creek sub-basin, where ecological impacts have been detected and where information on recovery of fish communities is required.

The role of empirical modeling in stream fish management

Rationale

The use of empirical models to quantify the impacts of human activities on aquatic ecosystems has expanded rapidly over the last 15 years and biological assessments using multi-metric or multivariate approaches are now common place. These statistical methods represent important tools that can assist with the management of stream fish communities.

Recommendation

We suggest that the increased use of empirical modeling would assist with the management of stream fish communities by: i) providing techniques to understand large-scale patterns in fish communities, ii) gaining insights to potential cause-effect relationships, iii) evaluating environmental impacts, iv) quantifying temporal variance in fish communities, v) identifying fish community types and the environmental gradients that may drive them, and vi) monitoring the effectiveness of current management actions. As a result, we recommend the increased use of empirical and other modeling tools to assist with the management of stream fish communities.

Download PDF