Background
The Athabasca Oil Sands Region is an active area of open-pit and subsurface bitumen, oil, and natural gas extraction in northern Alberta and Saskatchewan, Canada. Open-pit mining is localized predominantly to a ~10,000 square-km region surrounding the town of Fort Mackay, Alberta, Canada. Expansion of mine operation since the 1980's (Korosi et al., 2016; Cooke et al., 2017) has intensified the need for long term and high temporal resolution water quality monitoring to observe and respond to potential impacts of open-pit mining on downstream water chemistry. These potential impacts include the mobilization of trace metals (e.g., Pb, V, Cu, Zn) to downstream systems where the metals, if bioavailable and in high concentrations, can be toxins and negatively impact downstream aquatic systems through bioaccummulation and biomagnification in food webs. As part of its mandate, Alberta Environment and Parks has collected twice-weekly routine and trace metal water chemistry samples and continuous sonde measurements of tributary rivers in the Alberta Oil Sands Region since ~2016 (data availability extends to pre-2000, but is site-dependent).
Sondes (Figure 2) are data loggers which collect continuous (high frequency ~15 min) data for environmental variables, including:
|
Correlations between sonde variables (such as turbidity or conductivity) and specific water quality parameters (such as particulate or dissolved trace elements) are poorly understood in the Alberta Oil Sands Region. Relations between turbidity or conductivity and water chemistry are highly dependent on bedrock and surficial geology, groundwater flow paths, and other geological and geomorphological catchment characteristics. Therefore, these relations should be investigated in any new region where water quality modeling using sonde data is to be applied. In this study I use data from nine monitoring sites along four tributary rivers in the Alberta Oil Sands Region to assess water quality, differences between monitoring sites, and the utility of sonde variables to represent trace metal chemistry variables in this specific region.
Significance
Environmental monitoring of open-pit mining the Alberta Oil Sands Region (Figure 3) is one of the primary activities of Alberta Environment and Parks operations. More than 10% of Alberta Government's ~57 M annual (2018-2019) Oil Sands Monitoring budget was spent on surface water quality monitoring, including technician salaries, helicopter contracts, and laboratory work for the purpose of sampling the Athabasca River and its Alberta Oil Sands Region tributaries on a regular, high-frequency basis. This water quality monitoring is essential to identifying and responding to potential contaminant concentration increases which could adversely impact downstream aquatic life, food webs, and peoples who live within and downstream of the Alberta Oil Sands Region .
In comparison to chemical data, sonde data is far less costly and easier to obtain. Investigating whether sonde data can approximate water quality parameters with high accuracy on a continuous basis for the tributary rivers in the Alberta Oil Sands Region could inform stakeholders about the potential to reduce sampling frequency at some or all sites. Secondly, if some sites are classified as lower priority, then sampling frequency could also be reduced for these specific sites.
The implementation of sondes as proxies for modeling water chemistry, and the reduction of high frequency monitoring sites, could benefit taxpayers whilst still maintaining a high standard of surface water quality monitoring.
In comparison to chemical data, sonde data is far less costly and easier to obtain. Investigating whether sonde data can approximate water quality parameters with high accuracy on a continuous basis for the tributary rivers in the Alberta Oil Sands Region could inform stakeholders about the potential to reduce sampling frequency at some or all sites. Secondly, if some sites are classified as lower priority, then sampling frequency could also be reduced for these specific sites.
The implementation of sondes as proxies for modeling water chemistry, and the reduction of high frequency monitoring sites, could benefit taxpayers whilst still maintaining a high standard of surface water quality monitoring.
Objectives: Questions and Hypotheses
Our study has three main objectives: (A) generally assess water quality across the Alberta Oil Sands Region , (B) assess differences between monitoring sites, and (C) assess the utility of sondes as proxies for water chemistry.
A) Describe water chemistry at the nine sites along four tributaries in the Alberta Oil Sands Region open-pit mining region:
B) Assess if contaminant concentrations at the 9 sites and 4 tributary rivers are different?
C) Assess if sonde sensors can serve as proxies for water chemistry parameters at all nine sites or at individual tributaries:
A) Describe water chemistry at the nine sites along four tributaries in the Alberta Oil Sands Region open-pit mining region:
- What is the range and distribution of chemical concentrations?
- Use summary statistics and box plots.
- What water chemistry parameters covary at the nine sites and four tributaries?
- Compare site species scores to loadings of principal component analysis: (a) at all 9 sites together, and (b) at the four tributaries and their 2 (or 3) sites simultaneously.
B) Assess if contaminant concentrations at the 9 sites and 4 tributary rivers are different?
- Are trace element concentrations higher downstream of mining activities?
- Compare upstream and downstream sites using Discriminant analysis on stream order and one tailed t-tests.
- HYPOTHESIS: trace metal (Cu, Pb, V, Zn, Al) and suspended sediment concentrations will be significantly higher downstream of mining activity.
- Compare upstream and downstream sites using Discriminant analysis on stream order and one tailed t-tests.
- Are any of the sites "high priority", or "lower priority", (have generally higher or lower contaminant concentrations)?
- Compare variable loadings of potential contaminants to position of sites in Canonical Discriminant Analysis ordination.
- Are differences between sites more strongly accounted for by surface geology, or upstream-downstream site location? What chemical variables explain these differences?
- Visual inspection of Canonical Discriminant Analysis ordination site scores and variable loadings.
- HYPOTHESIS: difference between upstream and downstream sites will be greater than difference between tributary systems.
- Visual inspection of Canonical Discriminant Analysis ordination site scores and variable loadings.
C) Assess if sonde sensors can serve as proxies for water chemistry parameters at all nine sites or at individual tributaries:
- Can sonde-turbidity be a proxy for total trace element concentrations, including potential contaminants (Cd, Cr, Cu, Pb, Ni, V), associated with sediment mobilization? For example, does an increase in turbidity correspond to an increase in total trace element concentrations?
- Compare covariance of PCA loadings between turbidity and total trace element concentrations.
- Calculate correlation between turbidity and total trace element concentrations.
- HYPOTHESIS: turbidity will covary (PCA) or correlate (regression) strongly with total trace elements with low mobility likely to exist in a particulate phase, such as Pb, Al, Cu, and Zn.
- Can sonde-conductivity be a proxy for dissolved trace element concentrations (including potential contaminants commonly found in dissolved phases: As, U) associated with ground water?
- Compare covariance of PCA loadings between conductivity and dissolved trace element concentrations.
- Calculate correlation between conductivity and dissolved trace element concentrations.
- HYPOTHESIS: conductivity will covary (PCA) or correlate (regression) strongly with dissolved trace elements typical of carbonate dissolution and groundwater flow paths, such as Ca, Sr, As, U.
Disclaimer: all data and writing within this webpage are for statistics-training purposes only, and were performed on test data.