Data sources and methods for Water quality in Canadian rivers
Data sources and methods
What are the data sources
Water quality data for 325 monitoring sites are gathered from federal, provincial and territorial monitoring programs from across Canada. The complete list of data sources from Federal and Provincial monitoring networks can be found in Annex A.
Water quality guidelines for the protection of aquatic life are used to calculate the indicator. They come from the Canadian Council of Ministers of the Environment, the United States Environmental Protection Agency, and provincial and territorial government sources. A complete list of water quality guidelines used by each jurisdiction can be found in Annex B.
Additional information from Statistics Canada, Natural Resources Canada, Agriculture and Agri-Food Canada, and Environment and Climate Change Canada are used to assess land use.
For the 2013 to 2015 period, water quality data from 178 sites were used to compile the national indicator. These data were drawn from monitoring sites in Canada's 16 southernmost drainage regions (Figure 1).
Figure 1. Geographic extent of the 16 drainage regions selected for the national water quality indicator
The map of Canada shows the drainage regions included in the national water quality indicator. Of Canada's 25 drainage regions, 16 are used to report on water quality nationally.
Data for 4 additional sites in Yukon, 4 sites in Alberta, 1 site in Saskatchewan, and 10 sites in the Northwest Territories were used to improve coverage of the Mackenzie River region and of Yukon in the Pacific Ocean region in the regional indicator.
Water quality is evaluated at an additional 147 monitoring sites across Canada. Water quality results for all 325 sites can be explored using the interactive water quality map.
Data used to calculate the indicator include concentrations for a total of 40 measured substances, physical parameters, and data, such as pH, temperature or hardness, required to calculate certain guidelines, for all 325 sites across Canada from 2002 to 2015. Sample timing and frequency is set by monitoring programs and vary among sites.
Each data record collected is tagged with the site name, the date the sample was collected, the name and chemical form of the parameter, and land use and ecological information. Water quality parameter data for the indicator, along with water quality indicator scores and site information from the monitoring programs, are stored in a central water quality indicator dictionary housed within a larger database at Environment and Climate Change Canada.
Land use characterization for all monitoring sites was completed in 2008. At that time, land use at each site was determined using:
- population density from Statistics Canada's 2006 Census of Population
- mine locations using Natural Resources Canada's 2006 Census of Mines
- point-source pollutant releases from industrial and commercial facilities using Environment Canada's 2007 National Pollutant Releases Inventory
- agricultural activity locations using Statistics Canada's 2006 Census of Agriculture
- land cover using Natural Resources Canada's land cover mapsFootnote ,Footnote 
Data quality assurance and quality control
Data quality assurance/quality control is performed within each monitoring program providing data for the water quality indicator. Each monitoring program follows standardized methods for sample collection in the field. Chemical analyses are performed in Canadian laboratories accredited by the Canadian Association for Laboratory Accreditation or the Standards Council of Canada.
Environment and Climate Change Canada performs further quality assurance/quality control processes to ensure datasets meet minimum data requirements for the analysis and that calculation standards are respected. This process verifies the number of samples, sample timing, location of monitoring sites, and calculations. It also leads to removal of parameters due to low sampling frequencies or to detection limits being higher than the guidelines used in the calculation. Unusually high or low values in the monitoring datasets are double-checked and confirmed through consultation with the data provider.
Minimum data requirements
Calculating the water quality status for most sites requires a minimum of 4 samples per year collected over 3 years. A minimum of 3 samples per year is permitted for northern and remote sites, as access during winter months can be difficult, dangerous and costly. A sensitivity analysis found that there was no significant difference in the water quality index score when mid-winter samples were excluded.Footnote 
Minimum sampling requirements for the 2013 to 2015 period were not met at 12 sites: 8 in Manitoba, 2 in Newfoundland and Labrador, 1 in Saskatchewan, and 1 in Ontario. The results for these sites were evaluated by local water quality experts who concluded the data could be included because they were consistent with previous years and were considered representative of local water quality.
For a parameter to be included in the calculation of the indicator, a sample value must be available for each year for at least 33% of the total number of samples.
The indicator was calculated using data from 2013 to 2015, the most recent data available from all monitoring programs. For 5 sites, data from late December 2012 or early January 2016 were used to meet requirements for minimum number of samples.
How is this indicator calculated
This indicator is calculated using the water quality index as endorsed by the Canadian Council of Ministers of the Environment.Footnote  For each site, 5 to 15 water quality parameters are compared to their guideline value using the index calculation. The index produces a score between 1 and 100. Sites are assigned a water quality category based on the score. The results are grouped into 5 regions for presentation in the Regional water quality in Canadian rivers indicator.
Trends in water quality are evaluated using a guideline deviation ratio calculated using data from the first year of data collected at the site to 2015. To calculate the guideline deviation ratio, the concentration of each water quality parameter result at a site was divided by its guideline and averaged annually to obtain the guideline deviation ratio at a site. A Mann-Kendall test was used to assess whether there was a statistically significant increasing (improving water quality) or decreasing (deteriorating water quality) trend in the annual guideline deviation ratios at a site.
Annex B contains a complete list of parameters and guidelines used in each jurisdiction. Information on water quality parameters and guidelines used at individual sites can be found in the interactive water quality map.
Federal, provincial and territorial water quality professionals select water quality chemical substances and physical properties to be assessed at each site based on their knowledge of local water quality stressors. Typically, at least one form of the following parameter groups is reported at each monitoring site: nutrients (for example, phosphorus, nitrate, nitrite, total nitrogen), metals (for example, zinc, copper, lead), and physico-chemical parameters (for example, pH, turbidity), as well as 2 to 4 regionally specific parameters (for example, chloride, ammonia, dissolved oxygen, pesticides).
The exceptions are British Columbia and Yukon. In these regions, a common set of parameters is assessed at all sites with site-specific parameters added as required. Dissolved oxygen, phosphorus, pH, nitrogen and water temperature are included at sites when data are available.
Water quality guideline selection
Water quality guidelines for the protection of aquatic life are recommended limits or statements for a variety of chemical substances and physical parameters, which, if exceeded, may impair aquatic life. These guidelines are based on existing knowledge of a substance's environmental fate, behaviour, and chronic or acute toxicity. The water quality indicator uses chronic water quality guidelines for the protection of aquatic life, except for Quebec, where acute water quality guidelines for metals are used.
Federal, provincial or territorial water quality experts select the guidelines to use in the calculation of the water quality indicator based on their local relevance. The Canadian Freshwater Quality Guidelines for the Protection of Aquatic Life are recommended if locally relevant. Annex B provides a complete list of guidelines used by provinces and territories and their source.
Background concentrations of naturally-occurring substances and other local river characteristics can impact the measured concentration and toxicity of some substances. In these cases, site-specific guidelines are developed using procedures based on background concentrationsFootnote  or a rapid assessment approach. The rapid assessment approach uses long-term monitoring data and adjusts for natural events, such as high flows, that may influence results.Footnote 
Selection of national core sites for the development of the national indicator
Among Canada's 25 drainage regions (Figure 1), 16 were selected based on highest human density and land use to create the water quality indicator core network for national water quality reporting. Within the 16 selected drainage regions, core sites were selected to ensure site drainage areas do not overlap and are independent of one another. The upstream drainage area of each monitoring site was delineated by Statistics Canada using the National Hydro Network.Footnote  Where the upstream drainage areas of monitoring sites overlapped, the site furthest downstream was retained for the core network, as the downstream site is impacted by the maximum area in the river basin and, to some degree, reflects the cumulative impact of all upstream stresses. For 14 large rivers, core sites were chosen in the upper, mid and lower portions of the main river and at the most downstream sites on each tributary, when available. Additional core sites were included on these rivers, because water travels thousands of kilometres from the source to the mouth of these rivers. Water quality changes along the way and cannot be summarized by a unique downstream monitoring site. The final selection of core sites ensures monitoring sites are well distributed among provinces and drainage regions.
The number of core sites changes from year to year because samples are missed or lost and, as a result, the site may not have the minimum data required to be reported.
Classification of sites
Land use was assessed in the drainage area of core sites and classified according to the criteria presented in Table 1 using the drainage area of each monitoring site. For this analysis:
- Agricultural land cover corresponds to land cover classes 26, 27, 28 and 29
- Undisturbed land cover corresponds to land cover classes 0, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 31, 32, 33 and 38Footnote 
Land use upstream of 16 core sites in Newfoundland and Labrador and Quebec was defined based on knowledge by local water quality professionals. Land use for 8 sites in Northwest Territories and Yukon was not classified because (1) they are trans-boundary sites and the United States portion of the sites is challenging to classify or (2) they are close to the ocean.
|Agriculture||> 20% of drainage area is agricultural land cover|
|Mining||Presence of at least one mine|
|Mixed pressures||Agriculture and mining OR|
Agriculture and population density > 25 persons/km2 OR
Mining and population density > 50 persons/km2
|Undeveloped||> 95% of drainage area is undisturbed land cover|
Calculating Water Quality Status
The water quality indicator is calculated using the water quality index, as endorsed by the Canadian Council of Ministers of the Environment. The water quality index calculation considers three factors to summarize water quality at a site: scope, frequency and amplitude (Equation 1). Scope (F1) is the percentage of parameters for which the water quality guidelines are not met. Frequency (F2) is the percentage of samples for which the water quality guidelines are not met. Amplitude (F3) refers to the amount by which the water quality guidelines are not met. The score is normalized to yield a score between 1 and 100.Footnote  The full set of equations for the water quality index is described in Canadian Council of Ministers of the Environment Water Quality Index 1.0 Technical Report (PDF; 1.40 MB).
The Canadian Council of Ministers of the Environment's water quality index score is calculated as 100 minus the square root of the ratio of the sum of the scope (F1) squared plus the frequency (F2) squared plus the amplitude (F3) squared divided by 3.
Water quality scores are grouped into 5 categories following the Canadian Council of Ministers of the Environment's water quality index (Table 2).
(95.0 to 100.0)
|Water quality measurements never or very rarely exceed water quality guidelines.|
(80.0 to 94.9)
|Water quality measurements rarely exceed water quality guidelines and, if they do, it is usually by a narrow margin.|
(65.0 to 79.9)
|Water quality measurements sometimes exceed water quality guidelines and may do so by a wide margin.|
(45.0 to 64.9)
|Water quality measurements often exceed water quality guidelines and/or exceed the guidelines by a considerable margin.|
(0 to 44.9)
|Water quality measurements usually exceed water quality guidelines and/or exceed the guidelines by a considerable margin.|
The 3-year roll-up is intended to dampen temporal variability in the results caused by annual fluctuations in weather and hydrology to make the water quality indicator more representative of how humans are impacting water quality in rivers.Footnote 
Calculation of trends in the water quality
To investigate if water quality at a site has changed through time, a separate set of calculations and metrics from the water quality index were carried out. The trend analysis allows for the detection of improving or deteriorating trends in water quality at a site, whether they occur above or below guideline values. The water quality index formulation can only detect change once parameter values exceed their guidelines, making it a metric that is much less sensitive to change over time.
For each year a guideline deviation ratio was calculated by dividing each parameter concentration (C) by its guideline value (G) for each sampling date. The logarithm of the ratios was calculated and averaged for each year to produce a mean annual value (Equation 2). The ratios were multiplied by -1 to invert the values so improving water quality has a positive slope to match how water quality is portrayed with the water quality index.
For each year:
i = parameters
j = samples
n = total number of samples
p = total number of parameters
C = measured concentration
G = guideline value
As the concentration of a parameter gets closer its guideline, the guideline deviation ratio gets closer to zero. A ratio below zero means the parameter is above its guideline. When parameters are well below the guideline, the ratio is close to 1.
The guideline deviation ratio is calculated as minus 1 times the sum of the logarithm of the annual average of the measured concentration (i) of parameter (j) divided by its guideline and divided by the total number of measurement and parameters over parameters (j) and measurements (i).
Three parameters were exceptions:
- Dissolved oxygen and total alkalinity have guidelines for which measurements must be above, rather than below like the majority of parameters. The guideline deviation ratio for dissolved oxygen was calculated by dividing the guideline by the concentration.
- pH measurements must lie within a range of generally 6.5 and 9. For this parameter, measurements within the guideline range were given a value of 1. The guideline deviation ratio for pH values less than 6.5 was calculated by dividing the guideline by the concentration. For pH values greater than 9, the guideline deviation ratio was calculated by dividing the concentration by the guideline.
- Where temperature was used as a parameter, the absolute value of the guideline deviation ratio was used if temperatures were below zero.
Current parameters and guidelines at each site were used through the entire record to avoid mistaking methodological changes in the water quality indicator for water quality change. When historical data were missing for a parameter, the parameter was dropped from the trend analysis. In one case, there was a change in the analytical form of a parameter. In 2012, Quebec began reporting unionized ammonia instead of dissolved ammonia. The ammonia data in the older data set were left as dissolved ammonia for this analysis because there is no way to convert between the two forms.
A Mann-Kendall test using the Kendall package of the statistical software R was used to detect the presence of statistically-significant trends in the guideline deviation ratios. A count of sites with increasing, declining and no trends in the water quality indicator was compiled for the indicator of change through time.
The year in which sampling started at each site varies from 2002 for 73 sites, 2003 for 54 sites, 2004 for 12 sites, 2005 for 7 sites, 2006 for 29 sites and 2007 for 3 sites.
What has recently changed
This indicator has undergone a major transformation since it was last published in June 2015.
- The Land use impacts of freshwater quality indicator has been integrated into the indicator results.
- Previously, trends in the water quality indicator were determined by comparing 95% confidence intervals around the water quality scores. When confidence intervals for the first and last index period did not overlap, water quality was considered to have changed. The new method uses the annual average of all ratios of the water quality measurements to their guidelines. The new method allows the use of the data underlying the indicator and can detect changes in water quality below the guideline.
- The regional analysis of water quality is no longer reported on the drainage region scale. Results are now reported based on how water flows through 5 large regions of Canada. Data for sites in northern Alberta, the Northwest Territories and Yukon have been added to the Mackenzie River and Pacific Ocean regions to fill in details about northern water quality.
- Complete data for all sites sampled under federal monitoring programs are now available online. Parameter data that underlie the calculation of the water quality index, the index scores and associated categories, and the yearly trend ratios can be found in the files.
What are the caveats and limitations
This indicator reflects the state of water quality in rivers in southern Canada. Northern Canada is under-represented.
An additional 19 non-core sites were included in the regional indicator to allow for coverage of the Mackenzie River region and on Yukon in the Pacific Ocean region, which are not included in the national water quality indicator.
The indicator only uses data for which guidelines exist. It does not cover all potential water quality issues in Canada.
The indicator is based on the impacts of a concentration of a number of chemicals at each site. These concentrations do not show the effect of spills or other transient events unless samples were collected right after the spill happened or their effect on water quality is long-lasting.
Water quality guidelines are derived from laboratory studies that do not consider, among other things, the impact of flow on sediment loads in a river. Although site-specific guidelines try to take into account the impact of elevated flows on parameter concentrations, elevated levels of naturally-occurring substances, such as minerals, nutrients, glacier deposits and soils, can lower water quality ratings.
The water quality indicator does not directly measure biological integrity; it measures whether physical and chemical characteristics of freshwaters are acceptable for aquatic life. Although physical and chemical measurements provide good proxies of biological integrity, only biological information provides a direct measurement of conditions for aquatic life.
The water quality indicator only assesses the quality of surface waters. Groundwater is not considered in this indicator.
The trends reported in this indicator are based on annual scores that aggregate parameter data. In the aggregation, negative and positive trends may cancel each other out. The trends may be different from analyses performed on a parameter by parameter basis.
It can be difficult to compare water quality index scores among sites due to flexibility in the selection of parameters and guidelines to reflect local and regional water quality concerns. The water quality categories assigned based on the scores, however, are comparable. A site classified as marginal has water quality guidelines that are being exceeded frequently and by a considerable margin, even if the parameters are not exactly the same.
Only parameters for which water quality guidelines exist can be included in the indicator. The absence of a water quality guideline for a parameter does not mean the parameter is unimportant.
The water quality indicator scores are sensitive to the number of parameters and samples used in their calculation. The number of parameters used in this indicator varies from 5 to 15 depending on the monitoring site, and between 9 and 60 samples can be used for a given parameter. In general, as the number of parameters, or samples, used to calculate the index increases, the score decreases because there is a greater chance of a guideline exceedance.Footnote 
Water quality varies naturally with weather and hydrological cycles. Although the water quality indicator uses a 3-year average to dampen the influence of specific rain fall and snow melt events on the water quality indicator score, care must be taken in comparing one period to another.
- Date modified: