Question 7 of 7: Design and Discussion of an Environmental Monitoring Program

Question 7: Design and Discussion of an Environmental Monitoring Program (20 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

The illustrative program described below is a groundwater monitoring program at a municipal fuel-storage/vehicle-fuelling facility in British Columbia, of the kind an EGBC-registered environmental consultant is routinely retained to design and run under provincial contaminated-sites regulation.

Objectives. The short-term objective, in the first year, was to establish site-specific groundwater background concentrations and confirm whether petroleum hydrocarbons (from historic fuel handling) had reached groundwater above the applicable BC Contaminated Sites Regulation (CSR) generic numerical standards. The long-term objective, once a release was confirmed, shifted to tracking the areal extent and temporal trend (stable, expanding or naturally attenuating) of the dissolved-phase plume, to support a remediation or risk-management decision and eventual site closure with regulatory sign-off.

What was sampled and how. Groundwater was sampled from a network of permanently installed 50 mm monitoring wells screened across the water table. Each well was purged of at least three casing volumes (or purged to stabilized field parameters — pH, conductivity, dissolved oxygen, turbidity) using low-flow sampling techniques before a grab sample was collected directly into laboratory-supplied, pre-preserved containers, following the field/preservation protocol described in Question 4 (VOC vials collected headspace-free, metals containers acid-preserved, all samples cooled to 4°C in an insulated cooler for same-day courier to the laboratory).

Sampling design. A combination of design types was used: judgmental (authoritative) placement for the initial well network (one up-gradient background well, several down-gradient wells clustered around the known source area — the former fuel dispenser island and underground storage tank pit), refined after the first round of results by adaptive placement of two additional wells to bound the observed plume edge more tightly. Because the same fixed wells were resampled on every round, the design is also a repeated-measures (paired-in-time) design at the network level, directly analogous to the paired before/after structure in Question 2.

Frequency, sample count and duration. Sampling was conducted quarterly for eight consecutive quarters (2 years) to capture seasonal water-table fluctuation, followed by semi-annual monitoring for a further 2 years once the plume was shown to be stable. Each round sampled 8 wells (1 background, 7 network wells) plus 1 field duplicate and 1 trip blank per cooler, for 10 samples per round and roughly 80 primary groundwater samples plus QA/QC samples over the full 4-year program.

Statistical hypotheses tested. Two hypotheses were tested at each monitoring round: (1) whether down-gradient well concentrations were significantly greater than the site background concentration, $H_0:\mu_{DG}\le\mu_{BG}$ vs. $H_1:\mu_{DG}>\mu_{BG}$ (a one-tailed comparison directly analogous to Question 2's paired-treatment test, here run as an independent two-sample or Mann–Whitney test between background and each down-gradient well's time series); and (2) whether the plume trend at each well was increasing, stable or decreasing over time, via the nonparametric Mann-Kendall trend test on each well's concentration-vs-time series (chosen because concentration data are typically non-normal and often partially censored, Question 1b/6b, which the Mann–Kendall test handles without requiring an assumed distribution).

Statistical analyses and laboratory work. Each round's dataset was summarized with the same descriptive/EDA tools used in Question 5 (boxplots by well, to flag anomalous rounds), before the two hypothesis tests above were run; between-well and between-round comparisons used ANOVA-style variance partitioning (Question 3) where the data supported it. Laboratory analyses were principally EPA/CCME-referenced methods for BTEX and F1–F4 petroleum hydrocarbon fractions by GC-FID/GC-MS (instrumental analysis, Question 6d), run by a laboratory accredited under CALA/ISO 17025, with every batch including method blanks, matrix spikes and surrogate recoveries as internal QC.

Relevant standards. Results were compared against the BC Contaminated Sites Regulation Schedule 3.1/3.2 numerical water standards (drinking-water and aquatic-life-protective, as applicable to the receiving environment), and, where drinking-water use was potentially affected, the Guidelines for Canadian Drinking Water Quality (Question 6f).

Other relevant issues. Seasonal water-table fluctuation was found to be a significant confounder in early rounds (dilution in wet-season high-water-table conditions masked a real, slowly increasing dry-season trend), which is why the program moved to a nonparametric, trend-based (Mann-Kendall) rather than a simple round-to-round comparison once this was recognized; well-specific factors (a partially fouled well screen giving anomalously low turbidity-affected results in one round) were also flagged and resolved through the field-duplicate QC program (Question 4) rather than being taken at face value.