In most Canadian Areas of Concern, fish and wildlife populations and their habitats (i.e. Beneficial Use Impairments 3 and 14) have been listed as impaired. While much work has addressed other Beneficial Use Impairments, there has often been a lack of specific data and methodologies for evaluating fish and wildlife populations and their habitats. This article presents a methodology for refining delisting criteria for wildlife and habitat Beneficial Use Impairments in the Bay of Quinte Area of Concern using indices of condition in a coastal wetland monitoring framework. Data have been collected to provide information on loss of fish and wildlife habitat (water quality and submerged aquatic vegetation and aquatic macroinvertebrate communities) and degradation of fish and wildlife populations (fish, amphibian and breeding bird communities). Three potential models for delisting are presented using the submerged aquatic vegetation community data as an example. For all coastal wetland attributes considered through the framework, Bay of Quinte coastal wetlands were generally in better condition than other Canadian sites along the Lake Ontario shoreline.
Over the last two centuries, Great Lakes coastal wetlands have been degraded and, in many extreme cases, lost entirely (Snell, 1987; Krieger et al., 1992; Schaefer, 1994; Environment Canada, 2002). These losses have impacted a wide range of fish and wildlife communities (Danz et al., 2005), as coastal wetlands provide breeding and migratory habitat for wildlife (Mitsch and Gosselink, 1993; Austen et al., 1994; Hecnar, 2004; Hanowski et al., 2007), and critical spawning and nursery areas for fish (Chubb and Liston, 1986; Klarer and Millie, 1992; Jude et al., 2005). Despite these critical functions and values, coastal wetlands continue to be degraded by anthropogenic stressors such as land use change, pollution, nutrient and sediment loading, fragmentation, invasive species, shoreline alteration and water level control (Rodriguez and Holmes, 2009).
The widespread occurrence of coastal wetland degradation emphasizes the need for restoration action and policy development for protection and restoration. Monitoring can identify current wetland status and trends and responses to disturbances, information that can subsequently be used to make sound decisions for the restoration and conservation of coastal wetlands.
Bay of Quinte Area of Concern
The Bay of Quinte Remedial Action Plan (BQ RAP) Coordinating Committee (1990) listed ten Beneficial Use Impairments (BUIs), including BUIs 3 and 14: degradation of fish and wildlife populations, and loss of fish and wildlife habitat (BQ RAP Restoration Council; herein BQ RAP RC, 2003) for the Bay of Quinte Area of Concern (AOC). Of the 14 delisting criteria for BUIs 3 and 14 for Bay of Quinte (BQ RAP RC, 2007a, b), five relate to coastal wetlands. While there has been much work completed in the limnetic and littoral zones of the bay to address fish-related aspects of BUIs 3 and 14 (see BQ RAP RC, 2007a and other papers in this issue), relatively little had been completed in the transition zones from aquatic to terrestrial habitats, represented by coastal wetlands.
Need to refine delisting criteria
The existing delisting criterion for BUI 3 is more conceptual than quantifiable: “demonstrate that key fish and wildlife species are present in numbers consistent with a stable, diverse and healthy aquatic ecosystem” (BQ RAP RC, 2003). Karr's (1996) definition of biotic integrity demonstrates clear linkages to the existing delisting criterion: the ability of a habitat “to support and maintain a balanced, integrated, adaptive biological system having the full range of elements expected in the natural habitat of a region.” Therefore, an Index of Biotic Integrity (IBI) can be used to report on elements of coastal wetland fish and wildlife population and habitat conditions. This represents discrete elements of the more conceptual delisting criterion for degradation of fish and wildlife populations.
Environment Canada – Canadian Wildlife Service (EC-CWS) implemented methodologies using IBIs developed through the Great Lakes Coastal Wetlands Consortium (GLCWC, 2007; Burton et al., 2008) and in the Durham Region coastal wetlands in Lake Ontario (Figure 1) through the Durham Region Coastal Wetland Monitoring Project (DRCWMP; EC and Central Lake Ontario Conservation Authority (CLOCA), 2004). In 2005, BQ RAP RC assessed the DRCWMP for its ability to report on fish and wildlife population and habitat impairments in the Bay of Quinte AOC.
For loss of fish and wildlife habitat, the BQ RAP RC (2003) recognized two sources: loss of habitat footprint (infilling) and loss of habitat quality. Fish, breeding bird, and amphibian community condition in Lake Ontario coastal wetlands decrease with reduced habitat quality, as measured by the water quality index (WQI) and submerged aquatic vegetation (SAV) community IBI (EC-CWS, Toronto, ON, Canada, unpublished data; also see Seilheimer and Chow-Fraser, 2006). This article presents a framework for collecting and presenting data for coastal wetland monitoring in Great Lakes AOCs, specifically the Bay of Quinte. It demonstrates how these data can be applied to report on biotic communities and habitat. Lastly, the potential applications of this monitoring program are presented for refining five of the 14 delisting criteria for BUIs 3 and 14.
The DRCWMP assesses the biotic community condition by monitoring SAV, breeding birds, amphibians, fish and aquatic macroinvertebrates. The DRCWMP also monitors water quality as a measure of disturbance in the wetland, as actions in the watershed and adjacent land use can affect wetland water quality (Chow-Fraser, 2006). In applying this framework to Bay of Quinte AOC delisting criteria, fish, bird, and amphibian communities report on BUI 3, while the WQI and SAV community IBI are indicators for BUI 14.
Data collection summaries – biotic communities
Fish were captured by electrofishing over 44-metre transects stratified by habitat type. Species were identified and weighed and their fork length measured (EC and CLOCA, 2007).
Breeding bird community
Data were collected using the Marsh Monitoring Program (2003) protocol with augmentations suggested by Meyer et al. (2006). Surveys used a 100-m radius point count method at established survey stations.
Data were collected using the Marsh Monitoring Program (2003) protocol.
Twenty 1-m by 1-m quadrats were placed in the open water basin. Total percent cover and individual percent coverage of species were recorded (EC and CLOCA, 2007).
Three replicate sub-samples of approximately 150 aquatic macroinvertebrates (⩾500 μm) were taken by sweep-netting through the water column in the emergent communities. Aquatic macroinvertebrates were identified to the lowest taxonomic group possible.
Calculating the IBI
IBIs were developed in EC and CLOCA (2004). Community attributes were examined for a response against a disturbance gradient. The overall disturbance in each wetland was derived statistically using the method described by Hughes et al. (1998). Community attributes suspected to respond to disturbance were identified through existing literature regarding similar habitats (fish – Minns et al., 1994; aquatic macroinvertebrate – Burton et al., 1999; SAV – Albert and Minc, 2004) except for the breeding bird and amphibian community IBIs, which were developed in EC and CLOCA (2004) and refined in Burton et al. (2008). For all IBIs, attributes that responded to disturbance (raw metrics; Table 1) were transformed into standardized metrics using a linear function with a minimum value of 0 and maximum value of 10, as outlined in Minns et al. (1994). The standardized metric values were added and normalized to create an IBI score ranging from 0 to 100, with 0 being the most degraded condition.
Data were collected at three replicate sampling locations within three metres of the emergent vegetation (generally Typha spp.). Temperature, pH, conductivity (μS cm−1), and turbidity (NTU) were measured on-site at each sampling station using standard probes (Hydrolab DS5 or Quanta).
Water Quality Index
Chow-Fraser (2006) developed a WQI for Great Lakes coastal wetlands using 12 water quality parameters. Additional WQI models were included that incorporated a subset of the 12-parameter suite. EC-CWS found value in using Equation 7 (Table 5.6 in Chow-Fraser, 2006), as it used parameters already collected through the DRCWMP (i.e. temperature, pH, conductivity and turbidity) and had a strong association with the 12-parameter model (r2= 0.898).
Bay of Quinte representative sites
Ten representative sites (sentinel sites) were selected within the Bay of Quinte AOC (Figure 2) because monitoring all coastal wetlands within this very large AOC is impractical. These sites represent a range of sizes (medium: 25 ha to 100 ha; large >100 ha), hydrogeomorphic types (open embayment and protected embayment) and disturbance factor (low and high), as well as geographic location within the Bay of Quinte (upper and middle bay; EC-CWS, 2007).
In general, Bay of Quinte coastal wetlands had much higher scores for all biotic communities than other Lake Ontario coastal wetlands outside the AOC. From 2006 to 2009, SAV IBIs for Bay of Quinte coastal wetlands ranged from 57.6 to 98.2 while for other Lake Ontario coastal wetlands, they ranged from 0 to 87.2 (Table 2). Macroinvertebrate IBIs ranged from 33.2 to 91.9 in Bay of Quinte and 5.1 to 87.9 in other Lake Ontario coastal wetlands (Table 2). Bird IBIs ranged from 28.2 to 100.0 in Bay of Quinte and 2.2 to 96.4 in other Lake Ontario coastal wetlands (Table 3). Fish IBIs ranged from 55.4 to 90.9 in Bay of Quinte and 6.3 to 73.3 in other Lake Ontario coastal wetlands (Table 3). Amphibian IBIs ranged from 66.3 to 90.9 in Bay of Quinte and 0 to 93.8 in other Lake Ontario coastal wetlands (Table 3). Water quality was also in better condition in Bay of Quinte than the non-AOC coastal wetlands in the other sampled Lake Ontario sites, ranging from -0.80 to 1.24 versus -2.74 to 0.74 (Table 4).
The implementation of the GLCWC-developed monitoring protocols into the DRCWMP and Bay of Quinte AOC demonstrate the applicability of this framework to coastal wetlands throughout Lake Ontario. The data collection and analysis methodologies can be applied to refine delisting criteria in several ways, a decision that ultimately rests with the respective RAP Restoration Council, or equivalent.
Potential applications of the methodologies are discussed below using the 2009 SAV community data as an example; similar methods would apply for other biotic communities and water quality.
Potential Models for Delisting
Comparison with non-AOC coastal wetlands in the rest of Lake Ontario
The Bay of Quinte AOC average can be compared with the other Lake Ontario coastal wetlands average using a t-test. When the Bay of Quinte AOC average is not significantly lower than the non-AOC average for that year, the delisting criterion has been met. Following this model, the SAV community condition delisting criterion has been met in 2009.
This concept of delisting is useful because it can account for lake-wide variability in the indicator, in contrast to a static delisting criterion (to be considered delisted, mean SAV IBI in the ten representative sites must be greater than 75). Furthering this example, if the SAV community condition decreased from a mean of 80 to 70 in a sampling year, the SAV community delisting criterion would be considered impaired. However, if there were non-anthropogenic, stochastic lake-wide environmental factors (e.g. extreme weather phenomena) that drove the decrease in lake-wide SAV community condition, it may not be reasonable to consider the Bay of Quinte SAV community condition as impaired in that particular year.
Additionally, the non-AOC wetlands compared here were not randomly selected; it would be more appropriate to randomly select a subset of coastal wetlands in Lake Ontario, rather than comparing Bay of Quinte to sites that are known to be degraded (i.e. Durham Region) due to extensive development in the surrounding land use. Such randomization is expected to be possible under projects funded through the Great Lakes Restoration Initiative (GLRI), as sites on Lake Ontario are being chosen using a probabilistic design.
A limitation of this approach in the Bay of Quinte AOC is that the biotic community conditions within its coastal wetlands are generally much higher than other non-AOC conditions in the rest of Lake Ontario. This is the case with the SAV community (Figure 3), WQI (Figure 4) and all surveyed biotic communities, although bird communities tend to show higher variability within regions (EC-CWS, Toronto, ON, Canada, unpublished data). As such, conditions within the Bay of Quinte can decrease far too much before they would be considered impaired. Using data to generate Figure 3, a power analysis (alpha = 0.05 and beta = 0.80) reveals that for the mean Bay of Quinte SAV IBI to be worse (statistically lower) than the rest of Lake Ontario, the mean would have to drop from 87.2 to 15.4, a difference of 71.8 IBI points. From a RAP and good environmental stewardship standpoint, this paradigm for delisting the Bay of Quinte fish and wildlife population and habitat impairments does not appear suitable.
Comparison within Bay of Quinte
Under this approach, annual mean indices for representative sites would be statistically compared to benchmark year data to detect differences. Using data to generate Figure 3, a power analysis (alpha = 0.05 and beta = 0.80), based on a paired t-test reveals that to be considered impaired the mean representative site SAV IBI would need to significantly decrease from the proposed benchmark based on 2009 data of 87.2 to 83.4. This would help the BQ RAP determine the status of delisting criteria as they become affected by anthropogenic impacts in the bay. However, this approach does not account for lake-wide variability, for example, variation driven by a new water level regulation plan for Lake Ontario. Following this model, the SAV community condition delisting criterion has been met for 2009.
Setting targets based on historical conditions
A common perception within the Bay of Quinte environmental community is that coastal wetlands are still degraded, despite having some of the best wildlife and habitat communities in Lake Ontario. Nonetheless, the presented framework can still be applied by setting a higher target based on expert opinion because relevant historic data are lacking. For example, the previous models propose benchmarks of 30.6 and 87.2 for SAV IBI in 2009 to be considered unimpaired. If historic conditions within the AOC were considered to be higher, different benchmarks could be set using expert opinion and examining the individual constituent metrics for the particular IBI. Setting new targets for individual metrics allows calculation of a new target IBI; if the Bay of Quinte conditions fall below, the delisting criteria will not have been met and those BUIs would continue to be listed as impaired.
The problem with this approach is that there is a lack of data regarding historical conditions of coastal wetlands in the Bay of Quinte. The BQ RAP Coordinating Committee (1993) stated that there was no evidence of wildlife population degradation; however, BQ RAP RC (2003) acknowledged that in the absence of wildlife population data, wildlife populations cannot be assumed to be unimpaired. Thus, targets would have to be set based on professional opinion and knowledge of the system and by more closely examining the individual metrics for each IBI and their responses to disturbance.
A paradigm within some RAP communities for delisting criteria is to demonstrate the BUI is not more impaired within the AOC than outside the AOC. Regardless of the model selected to define delisting criteria, this framework can be applied for wildlife communities and habitat in coastal wetlands. It is possible that a combination of two or more of these models be applied, though the exact statistical and methodological steps required to implement that would require exploration.
The application of a coastal wetland monitoring framework in the Bay of Quinte AOC for BUIs 3 and 14 represents considerable progress for fish and wildlife monitoring in coastal wetlands. The method developed for Durham Region and applied to the Bay of Quinte gave a quantifiable way to evaluate the integrity of these communities and their habitat. It clearly identified which variables need to be monitored by demonstrating linkages between those metrics and disturbance. This framework provides a method of evaluating fish and wildlife populations and their habitats for those AOCs that have listed BUIs 3 and 14 across the Great Lakes basin; however, it also represents significant progress in coastal wetland monitoring at other scales across the Great Lakes basin. Elements of the coastal wetland monitoring framework developed through the GLCWC have proven to be successful both in their implementation on a regional scale (i.e. Durham Region) and in their extension to other areas (i.e. Bay of Quinte). The DRCWMP demonstrates the development of a framework that can be feasibly implemented and applied to other regions. This standardized methodology allows for the comparisons among data sources, which will facilitate the identification of monitoring trends from site-level to basin-wide.
The authors thank Quinte Conservation, Lower Trent Conservation Authority and Durham Region Coastal Wetland Monitoring Project partners for data collection, as well as Lesley Dunn, Nancy Patterson and two anonymous reviewers for their valuable comments on the manuscript. We also thank Sarah Rice for providing support with mapping.