Bryant M. Serre
Figure 2: Ontario Benthic Biomonitoring Network Kick and Sweep Training Session, May 2019 (Photo Courtesy of Bryant Serre [per TRCA]).
Benthic macroinvertebrates are bottom dwelling organisms in lotic and lentic ecosystems that have been, and continue to be, surveyed as part of environmental assessment. Benthic organisms refer to the consortia of small aquatic animals and larval starts of insects (e.g., dragonflies or infraorder Anisoptera) that occupy the sediment for a significant duration of their life history (Brinkhurst 1975; Hutchinson 1961) . Benthic macroorganisms confine themselves to river and lake sediment for a series of reasons, such as predation avoidance behaviors and camouflage and providing optimal conditions for initial larval development. However, as they are inhabit aquatic sediment environments, they are in subject to the physical, chemical and biological conditions of streams and lakes. Indeed, as benthos are niche-constrained, populations are concomitantly subject to both short and long term stressors while being easily measured and informative indicators of aquatic health (Read et al. 1983; Reynoldson and Metcalfe-Smith 1992).
Within ecotoxicological assessment of the benthos, individual organism health and state is rarely measured (Fargasova 1999; Bettinetti et al. 2020). Instead, the relative abundance of benthic macroinvertebrates is used to elucidate the biological conditions of aquatic ecosystems. For example, the decreasing relative proportion of Ephemeroptera (mayflies), Plecoptera (stoneflies), and Trichoptera (caddisflies) (EPT) taxonomic groups, in relation to all other sampled taxa, is used to indicate increasing river or lake stress (United States Environmental Protection Agency (USEPA) 2021). To date, a suite of different metrics have been developed and continually used globally for aquatic monitoring (USEPA 2021).
Under anthropocentric human influence, drainage basins have been urbanized extensively, leading to potentially deleterious effects in aquatic systems. The series of effects on watercourses and waterbodies are diverse; for instance, surface runoff and stormwater collection can mobilize heavy metals (notably mercury, lead, and arsenic), asphaltenes (polycyclic and polynucleic aromatic hydrocarbons or PAHs), macronutrients (e.g., runoff from farm fields promoting excessive algal development), and plastic pollution to cause chronic (i.e., reproductive), lethal (i.e., mortality) and sublethal (i.e., behavioral) biological effects (Muller et al. 2020; Scholz 2015). Moreover, the heightened temperature of urban surfaces (alia the Urban Heat Island Effect, Oke 1982) is often transferred to surface water flows, significantly raises temperatures of receiving watercourses and waterbodies, and concomitantly shifting the range of tolerance for benthic macroinvertebrates and subsequent ecotoxicological impact (Jones et al. 2012). In the context of sustainable development and fighting climate change, the relevance to assessing and continually monitoring of benthic macroinvertebrates is axiomatic.
Currently, benthic health assessments have been limited to individual watersheds and jurisdictions. In Canada, Environment and Climate Change Canada’s (ECCC) Canadian Aquatic Biomonitoring Network (CABIN) has reported on benthic health in individual case areas (e.g., Southwestern Hudson Bay drainage area, Great Slave Lake drainage area): there has not been a comparative analysis of benthic indices across the entire data collection. Moreover, analyses of benthic indices to urbanization across watersheds will elucidate whether benthic macroinvertebrate indices are suitable indicators of aquatic ecosystem health (personal communication, C. Metcalfe (CABIN) 9 March 2022).
In North America, there has been use, across several levels of environmental governance, to analyze the relative abundance of benthic macroinvertebrates as a means to discern biological health of lotic (flowing water), lentic (stagnant water), coastal and transitional (e.g., river deltas, mangrove forests) ecosystems. Since 1972, the International Joint Commission (IJC) has used benthic macroinvertebrates, in conjunction with a suite of indicators for monitoring the status of the Laurentian Great Lakes. The use of benthic macroinvertebrate indices was suggested under the the Great Lakes Water Quality Agreement (GLWQA), and they are still used to measure North American's progress towards improving the water quality of watercourses and waterbodies (notably Lake Erie which exhibited severe eutrophication and poor biological quality). In the United States, the United States Environmental Protection Agency’s (USEPA) National Aquatic Resource Surveys (NARS) uses indices of relative benthic populations, in conjunction with measuring river chemical (e.g., conductivity, dissolved oxygen, soil chemistry, macronutrient concentrations), physical (e.g., human disturbance, clarity/turbidity, riparian vegetation condition) and recreational (e.g., presence of filamentous blue-green algae microcystin aeruginosa, enterococci, and faecal coliforms) indicators, to assess the quality and condition of watercourses across the contiguous United States (USEPA 2020). Similarly, Environment and Climate Change Canada’s (ECCC) Canadian Aquatic Biomonitoring Network (CABIN) measures changes in benthic biological communities to assess the state of Canadian lotic systems (ECCC 2018). Regionally, examples can be seen in many watersheds. In Ontario, Canada, the Credit Valley Conservation (CVC) Authority in uses benthic indicators as part of their approach to discern whole watershed health, beyond the confines of the lotic ecosystem itself (CVC 2017). Similarly, the largest Canadian conservation authority, Toronto and Region Conservation (TRCA), collects benthic macroinvertebrate samples frequently as part of their Regional Watershed Monitoring Program (TRCA 2020). While these instances exemplify how there are multiplicitous assessments of water quality using benthos abundance, and across varying continental, federal and provincial extents, benthic indices have not been evaluated wholly across all of Canada as to allow relative comparisons of water quality and potential aquatic ecosystem stress.
Given the foregoing findings, it is the goal of this project to evaluate the health of aquatic ecosystems across Canada, using indices of benthic macroinvertebrate populations. Specifically, it is the objective of this study to test the relationship of the Ephemeroptera, Plecoptera and Trichoptera (EPT) benthic macroinvertebrate index to upstream watershed land cover for each existing sample point. In turn, this will help understand the potential ecotoxicological effects of urbanization on aquatic ecosystem health. This study is guided by the following question: how do benthic population assemblages respond to changes in urbanization within the drainage basin? This objective is accomplished through computing the EPT index for each sample site and the land cover composition of upstream drainage basins for sample sites, and subsequently, testing the statistical relationships between basin land cover and the population EPT index.
Figure 3: Example of Benthic Macroinvertebrates from a Ponar Sample, Utah State University Extension (2020). * Chronomids, Hyallela Azteca (bright orange crustacea), Mayflies, Caddisflies and Stoneflies can be seen in the sample. Sorted sample on the right.
This analysis was carried out across all of Canada, using existing benthic macroinvertebrate survey data collected under the CABIN program. Under the CABIN program, eleven (11) major drainage areas were surveyed; including 1) the Maritime Provinces Drainage Area (DA), 2) the St. Lawrence DA, 3) the Northern Quebec and Labrador DA, 4) the Southwestern Hudson Bay DA, 5) the Nelson River DA, 6) the Western and Northern Hudson Bay DA, 7) the Great Slave Lake DA, 8) the Pacific DA, 9) the Yukon River DA, 10) the Arctic Drainage DA, and 11) the Mississippi River DA (ECCC 2018). Under the general confines of macroinvertebrate sampling, species assemblages are presumed to be in Panmixia (free mouvement, ability to be present) across all drainage areas (SERAS 2003: SOP 2054 0.0).
Figure 4: Watersheds included in the Environment Canada (EC) Canadian Aquatic Biomonitoring Network (CABIN) benthic sampling program.
The macroinvertebrate surveys under the CABIN program are collected on-site, by researchers across Canada using Kick Net (98.69% of samples) and Ponar (selectively, 1.3% of samples) methods. Coordinates are provided for each site survey, which allows this data to be utilized as a point (vector) dataset. Kick Net sampling occurs when a net is held downstream, and one uses their foot to dislodge sediment upstream, which is then incorporated into the water column, permitting benthos to flow downstream into a sampling net (ECCC 2018).
To delineate the boundary of upstream land area, the ArcGIS Watershed Analysis tool requires 1) a polyline layer for watercourses and 2) a digital elevation model. In this case, the National Hydrographic Network (NHN) file is most appropriate, given the resolution of detail to higher reaches of catchement (and respectively smaller streams), where samples were retrieved (i.e., 1st and 2nd Strahler orders) (Natural Resources Canada (NRCAN) 2019). Moreover, watershed delineation involves flow direction calculations, which are in turn used to make flow accumulation layers and to delineate watershed pour points (Parmenter and Melcher 2012). The digital elevation model used for this project is the Canadian Digital Elevation Model (CDEM) (NRCAN 2015). This is a Digital Surface Model (DSM) raster layer with 2 m resolution which will be used for z-coordinates (elevation) as to delineate basin boundaries.
The Agriculture and Agri-food Canada (AAFC) Annual Crop Inventory land cover dataset was used for land use classifications. This layer was selected for it’s yearly time intervals, corresponding with the different time periods CABIN benthic samples were retrieved (2009-2022). This raster layer has a final spatial resolution of 30 meters, and is considered sufficiently accurate for non-agriculture land cover estimation (AAFC 2015); the raster has a thematic resolution of 72 categories/land cover classes which is believed sufficient to parse out developed from undeveloped (or naturalized) land. As the AAFC cropping layer is specific to just Southern Canada, the classified and non-distributed land cover for all of Canada, across all years, was retrieved from members at Environment Canada (M. Luymes, personal communication, 14 March 2022)
This project required several steps to be able to compute the upstream basin cover for each sample point. The steps to achieve the project objective of testing the relationship between benthic macroinvertebrate indices (BMI) to the upstream land cover for each point are as follows.
In Microsoft Excel, non-viable sample points were removed. The EPT index, as defined per the USEPA Benthic Monitoring Framework (i.e., a list of all indices), was computed for all CABIN sampling points. Sampling coordinates were appended from CABIN reports to the individual sample site ID's. The resulting .xlsx file was converted to a .csv and imported into ArcGIS Pro 2.9.0. In ArcGIS, the .csv file was converted into a point vector file using the "Create Points from Table" (or Display XY Data in ArcMap) function. These were considered the pour points for the watersheds, in which the basins would be clipped to.
Figure 5: Sample points from the CABIN program (in red) with there respective upstream drainage basins (in blue).
The digital elevation file (Canadian DEM) was imported into ArcGIS, where the Watershed Geoprocessing toolbox was used. Firstly, a depressionless DEM was created by filling surface depressions in the digital elevation file. Secondly, a flow direction raster was computed using the "Flow Direction" tool; a flow direction raster takes into account the eight adjacent cells and assigns a reasonable trajectory of flow for each cell. Typically, the next step is to use the flow direction raster to create a flow accumulation raster, using the "Flow Accumulation" tool; however, as the pour points were already known (i.e., the sample points), this step was avoided. Subsequently, watersheds were delineating using the "Snap Watershed" tool where the flow direction layer and the pour points were used to delineate upstream drainage areas. This process was looped over all points in the dataset to create a total of 1244 watersheds.
Figure 6: Depressionless DEM for Southern Ontario, Canada, derived from the CDEM.
Figure 7: Flow direction raster (hillshade) for Southern Ontario, Canada, derived from the depressionless DEM.
Land cover rasters were imported for all sample years, using the full coverage of the AAFC data layer (courtesy of Environment Canada). In this step, the upstream land cover was delineated for the year corresponding to the benthic sample. To loop this over all sample points, watersheds were separated into different shapefiles (.shp) for each sample year. Then, individual loops were written so that watershed land cover could be summarized for the corresponding year of the sample using the "Summarize Raster Within" tool in ArcGIS Pro. Once analyzed, the land cover for each pour point (upstream basin of the sample points) was correlated to the EPT index scores for each sample point. moreover, as a final step, points in Lake sediment were removed due to the size of their drainage areas (nearly 100x the size of the average river sample watershed). This left N = 514 sites within the analysis.
Overall, a total of N = 514 sites were analyzed for their relative abundance of Ephemeroptera, Plecoptera and Trichoptera (EPT) taxonomic groups. When correlated to upstream basin cover, particularly urban cover and agricultural area (added to test the potential deleterious effects of soil macronutrient and pesticide leaching on watercourses), there was no relationship found. The results of this study are inconsistent with the findings in the literature where benthic indices were successfully related to upstream land cover, as a metonym for anthropogenic stress (Pedersen and Perkins 1986; Hutton et al. 2015).
Figure 8: Results of the correlation matrix between EPT benthic indices and upstream urban and agricultural land cover.
While the benthic metrics did not show a relationship with upstream land cover, the finding of insignificant results in itself is still interesting. Often, benthic indices are unchallenged indicators of aquatic toxicology; while their history of use is profound, and when conjoined with other water chemistry and biological assessment, they provide a comprehensive picture of watercourse health and state, their singly use may lead to inappropriate deductions on the state of watercourse health. This tends to be universal for many evaluations of environmental health, but it is by no means specific to benthic macroinvertebrates, but many studies who conclude upon the importance of more rigorous environmental science plainly.
Moreover, there are issues in interpreting metric scores due to random population effects. For instance, there are occasions where a species is not present, however, it may not be an indication of ecotoxicological stress. Differences in stream substrate, and the ability of the niche to support a population may dictate organism presence/or a lack thereof. Problematically, benthic population metrics assume populations are in panmixia; this process in genetics is often understood as the idealized theorem that if they can be there, they will be there. Subsequently, when benthic samples are interpreted without the ecological conditions of the site, false deductions on watercourse health are made. Or, causation between the environmental parameter measured (e.g., upstream land cover) is assigned to metric scores, perhaps falsely even if a strong correlation existed. As the results of this study go to affirm, there is a need for studies to continue to search for environmental relevance when conducting ecotoxicological assessment or environmental monitoring. It is central to the scientific method, but ever more paramount to reiterate in the context of fighting anthropogenic climate change.
The first limitation to this study was the inability to write an index and match function into the iterative process of measuring upstream basin cover. Initially, it was the goal that within the for loop for the "Summarize Raster Within" function, that i would be able to call the properties for each iterated list item, and then have the corresponding yearly land cover layer for that sample to be pulled. This proved to be difficult and after multiple attempts, I separated out the for loops, thereby computing individual sampling year with their corresponding land cover. In future, this would ideally be rectified should this type of analyses be expanding to other datasets (e.g., NARS in the United States). While in the initial proposal it was assumed that stream coverage would be an issue, the National Hydrographic Network (NHN) file was comprehensive and extensive-enough to cover the entirety of the sampling points in the CABIN dataset.
Overall, it is the goal of this project to assess the relationship between upstream land cover (as an indicator of the stress of human development) on downstream water quality, through multi-species benthic macroinvertebrate indices. While the results of this study did not indicate there is a relationship between human stressors and aquatic health, and therefore cannot provide insights into the comparative health of watercourses and waterbodies across Canada, the results nonetheless exemplify the importance of environmental relevance in assessing the state of aquatic ecosystems. When measured without understanding stream habitat, or without quantifying the ecological niche for each sample point and the subsequent carrying capacity of the landscape, solely using benthic indices to indicate aquatic health can lead to erroneous conclusions.
Agriculture and Agri-Food Canada (AAFC) (2015) Annual SpaceBased Crop Inventory for Canada. Agroclimate, Geomatics and Earth Observation Division, Science and Technology Branch. Retrieved from https://open.canada.ca/data/en/dataset/ba2645d5-4458-414d-b196-6303ac06c1c9
Barnes K, Cartwright L, Portiss R , Midwood J ,Boston C, Granados M, Sciscione T, Gibson C & Obembe O (2020) Evaluating the effectiveness of aquatic habitat restoration implemented using the Toronto Aquatic Habitat Restoration Strategy. Retrieved from https://trcaca.s3.ca-central 1.amazonaws.com/app/uploads/2021/01/07074413/TWAHRS-assessment-FINAL-technical-document-Nov-2020.pdf
Bettinetti, R., Zaupa, S., Fontaneto, D., & Boggero, A. (2020). Biological, chemical, and ecotoxicological assessments using benthos provide different and complementary measures of lake ecological status. Water, 12(4), 1140.
Brinkhurst, R. O. (1975). The benthos of lakes. Macmillan International Higher Education.
Credit Valley Conservation (CVC) (2017) Annual Report 2016 Credit Valley Conservation Integrated Watershed Monitoring Program (IWMP). Retrieved from https://www.creditvalleyca.ca/wp-content/uploads/2017/11/IWMP_2016_Web.pdf
Environment and Climate Change Canada (2018) National Science Forum 2017. Retrieved from https://publications.gc.ca/collections/collection_2018/eccc/En4-296-2017-eng.pdf
Fargašová, A. (1999). Ecotoxicology of metals related to freshwater benthos. General Physiology and Biophysics, 48-53.
Hutchinson, G. E. (1961). The paradox of the plankton. The American Naturalist, 95(882), 137-145.
Hutton, M., Venturini, N., García-Rodríguez, F., Brugnoli, E., & Muniz, P. (2015). Assessing the ecological quality status of a temperate urban estuary by means of benthic biotic indices. Marine pollution bulletin, 91(2), 441-453.
Jones, M. P., Hunt, W. F., & Winston, R. J. (2012). Effect of urban catchment composition on runoff temperature. Journal of Environmental Engineering, 138(12), 1231-1236.
Müller, A., Österlund, H., Marsalek, J., & Viklander, M. (2020). The pollution conveyed by urban runoff: A review of sources. Science of the Total Environment, 709, 136125.
National Resources Canada (NRCAN) (2015) Canadian Digital Elevation Model (CDEM) [data file]. Retrieved from https://open.canada.ca/data/en/dataset/7f245e4d-76c2-4caa-951a-45d1d2051333
National Resources Canada (NRCAN) (2019) National Hydrographic Network. Retriev from https://www.nrcan.gc.ca/science-and-data/science-and-research/earth-sciences/geography/topographic-information/geobase-surface-water-program-geeau/national-hydrographic-network/21361
Oke, T. R. (1982). The energetic basis of the urban heat island. Quarterly Journal of the Royal Meteorological Society, 108(455), 1-24.
Parmenter B, Melcher J (2012) Watershed and Drainage Delineation by Pour Point in ArcMap 10. Retrieved from http://sites.tufts.edu/gis/files/2013/11/Watershed-and-Drainage-Delineation-by-Pour-Point.pdf
Pedersen, E. R., & Perkins, M. A. (1986). The use of benthic invertebrate data for evaluating impacts of urban runoff. Hydrobiologia, 139(1), 13-22.
Read, P. A., Anderson, K. J., Matthews, J. E., Watson, P. G., Halliday, M. C., & Shiells, G. M. (1983). Effects of pollution on the benthos of the Firth of Forth. Marine Pollution Bulletin, 14(1), 12-16.
Reynoldson, T. B., & Metcalfe-Smith, J. L. (1992). An overview of the assessment of aquatic ecosystem health using benthic invertebrates. Journal of aquatic ecosystem health, 1(4), 295-308.
Scholz, M. (2015). Wetland systems to control urban runoff. Elsevier.
SERAS (2003). Standard Operating Procedures Benthic MacroInvertebrate Sampling. Retrieved from https://clu-in.org/download/ert/2054-r00.pdf
U.S. Environmental Protection Agency (USEPA) (2021) Indicators: Benthic Macroinvertebrates. Retrieved from https://www.epa.gov/national-aquatic-resource-surveys/indicators-benthic-macroinvertebrates
U.S. Environmental Protection Agency (USEPA) (2020) National Rivers and Streams Assessment 2013–2014: A Collaborative Survey. EPA 841-R-19-001. Retrieved from https://www.epa.gov/national-aquatic-resource-surveys/nr