Determining reasonable response actions following a fly ash spill in the headwaters of Banff National Park, Canada

It can be challenging for practitioners to determine reasonable response actions following an environmental spill because there are risks associated with the recovery process, acute constraints on time, and few case studies available from antecedent events. Here, we evaluate environmental risk using a screening level assessment (SLA) and describe risk management actions during the response phase of a train derailment that released 600 tonnes of fly ash into a headwater creek in Banff National Park, Canada. Trace metal concentrations, and physico-chemical parameters from downstream of the derailment site were compared to Canadian environmental quality guidelines and upstream reference values. There was a 1 to 2.2-fold exceedance of sediment quality guidelines (As, Cd and Se) as well as a 3.6 to 17.5-fold exceedance of water quality guidelines (Al, Cd, Fe and turbidity) downstream of the train derailment. Despite uncertainty about site-specific toxicity when using a SLA, we did require the removal of the settled fly ash from the creek based on the multiple exceedances of guidelines, regulatory context, wilderness setting and potential contribution to cumulative effects downstream. Case studies that evaluate risk and describe risk management actions help practitioners make consistent and efficient decisions during the response phase of a spill.

Methods
The methods of data collection are given in more detail within the accompanying peer-reviewed article. Abbreviated version here:  

At approximately 03:00 on December 26, 2014, a train derailment occurred on the transcontinental rail line at a stream crossing at the west end of the Town of Banff, in Banff National Park, AB, Canada. The derailment released 600 tonnes of fly ash and a carload of soybeans into Forty-Mile Creek, a fourth-order valley-bottom creek that meanders through montane forests. Forty-Mile Creek enters the Bow River 300 m downstream of the derailment site. The Bow River is a fifth-order river that originates from the glacier-fed Bow Lake at the continental divide. The discharge of the Bow River is approximately 20 times greater than the discharge in Forty-Mile Creek during baseflow conditions. Based on the product MSDS and fly ash literature, we identified a list of sediment and water trace metals in addition to water physico-chemical parameters as contaminants of potential concern (COPCs).

We sampled fly ash by scooping material directly from the rail car into a 500 mL polypropylene container. Twelve spatially replicated sediment samples were collected once, two weeks following the incident (January 1st, 2015), from a 200 m reach immediately downstream of the derailment site (S-DN140Mile to S-DN1240Mile). Twelve samples were also collected from an upstream reference reach, randomly spaced starting at 500 m upstream and ending at 1.5 km upstream of the derailment site (S-RF140Mile to S-RF1240Mile; Figure 1), but downstream of the landfill and dam. Samples were located by walking back and forth in the vicinity of each site over the frozen river and stopping randomly at sampling locations within the reach. The ice was cored using a 20-cm dia. hand ice auger (Nils; U.S.). Sediment was collected by reaching through 30–90 cm of water and scooping surface sediment (approx. top 10 cm layer) into a 500 mL polypropylene container. The contents were then immediately dumped into a polypropylene bag provided by the analytical laboratory (Maxxam Environmental Services, Calgary, AB). The following trace metals were analysed from the fly ash using inductively coupled plasma/mass spectrometry (ICP-MS; Agilent 7700 ICPMS): Ag, Al, As, soluble Boron (B), Ba, Be, Cd, Co, Cr, Cr6+, Cu, Fe, Hg, Mo, Ni, Pb, Sb, Se, Sn, Tl, U, V, Zn. Sediment concentrations of the same trace metals were also analysed using ICP-MS, but due to an error, Al and Fe were not included in the testing package. Samples were dried at 55°C, sieved and digested in a nitric-hydrochloric acid and de-ionized water mixture to solubilize the solid matter and remove organic material.   

We sampled and metered water daily on Forty-Mile Creek starting at 14:00 on December 27th, 2014, thirty-three hours following the train derailment, until January 2nd, 2015. We sampled daily again for one week, beginning on January 23rd, 2015, five weeks following the incident. We sampled daily for one week, again a third time, beginning on March 14th, 2015, fourteen weeks following the incident. Water was sampled from a site 150 m downstream from the derailment site (W-DN140Mile), 300 m downstream of the derailment site (W-DN240Mile) and a reference site 1.7 km upstream of the derailment, but downstream of the landfill and dam (W-RF40Mile; Figure 1). Water was also sampled and metered from the Bow River at the same schedule as Forty-Mile Creek, from sites 800 m downstream of the derailment site (W-DN1Bow), 17 km downstream of the derailment site (W-DN2Bow), and an upstream reference site 9.2 km upstream of the derailment site (W-RFBow; Figure 1). Selection of sampling sites included consideration of road accessibility because sampling occurred during winter. Therefore, we sampled water from all sites at least once/day for a total of 21 days. Water samples were collected by wading into the water while moving away from the shoreline and plunging 250 ml high-density polyethylene bottles, provided by the laboratory (Maxxam Environmental Services, Calgary, AB), into the water facing upstream. Bottles were rinsed with the flowing water three times before taking the samples. Water samples were analyzed for the following trace metals based on a general total metals package: Al, Ag, As, B, Ba, Be, Ca, Cd, Co, Cr, Cu, Fe, K, Li, Mg, Mn, Mo, Na, Ni, P, Pb, S, Sb, Se, Si, Sn, Sr, Tl, Ti, U, V, Zn using ICP-MS and –OES (Varian Vista Pro ICPOES and Agilent 7700 ICPMS) by Maxxam Analytics (Calgary, Alberta, Canada). However, not all these trace metals were evaluated (See Initial Testing Program below). We also recorded physico-chemical parameters (turbidity and pH) using a portable device (YSI ProDSS multiparameter meter; YSI Incorporated, Yellow Springs, OH, U.S.) at W-DN1, W-DN2 and W-RF in both Forty-Mile Creek and the Bow River at the time of water sampling.