Stormwater quality assessment and management for the town of Jasper in Alberta, Canada


 Stormwater pollutants can have deleterious impacts on the aquatic life of receiving water bodies. This paper presents the findings of a stormwater quality monitoring program performed in the town of Jasper in Alberta, Canada. This is one of the very few studies done on a small urban settlement to identify key pollutants of concern, characterize stormwater and identify unique pollutant sources in the town. A total of 14 stormwater quality parameters were found to be of high concern to aquatic life. The most prominent pollutants were total suspended solids (TSS), metals and phosphorus which compared well with studies conducted in large urban settlements. Tourist influx contributed to high metal and petroleum hydrocarbon loads during the peak season due to high vehicular activity. Elk were found to reside in the town during summer and their fecal droppings resulted in elevated fecal coliform concentrations. It was found that winter road salting was responsible for excessive chloride concentrations observed during the spring melt. TSS concentrations were statistically correlated with various metals as well as phosphorus using Spearman's rank correlation. It was found that the current street sweeping schedule in the town coincided with lower TSS and metal loads in stormwater.


INTRODUCTION
It has been well established that pollutants on a watershed (Canada) Ltd ), were found to focus on flood prevention, leaving limited resources to investigate quality issues in depth. Case studies that characterize stormwater in small urban settings are needed so that representative data can be collected, and the mechanisms responsible for runoff pollutants can be accurately identified so as to ultimately facilitate the implementation of suitable quality controls.
Meteorological events are the primary mechanism for pollutant wash off and they contribute to the temporal variability of pollutant concentrations in stormwater monitoring data (Kayhanian et al. ; Liu et al. ). Previous research has sampled rainfall (wet weather) events to assess the influence of storm intensity and duration as well as antecedent dry days on pollutant loading (Liu et al. ). Parallel to these efforts, but on a much smaller scale, the effects of urban snowmelt on runoff quality have also been addressed in colder regions. Furthermore, according to Janke et al. (), dry weather events, which are associated with background runoff, have been found to wash-off substantial contaminant loads in catchments and should be monitored. Galfi et al. () presented one of the few studies that characterized stormwater by sampling dry, wet and snowmelt events. More studies that monitor a range of weather events are needed to better understand how seasonal changes correlate with runoff quality variations in order to characterize and manage stormwater appropriately.
The focus of this research was to conduct a field-based investigation on stormwater quality in a small town in Jasper National Park, Alberta, Canada, called the town of Jasper. The aim was to monitor key water quality parameters (physical, chemical and biological) that may impose negative effects on sensitive ecosystems of receiving water bodies. Grab samples were collected at multiple sites and during various weather events to enable a complete characterization of the stormwater. A comprehensive approach was taken whereby the monitoring data were combined with historical reports as well as field observations and interviews, in order to understand how unique and seasonally variable activities in the townsuch as tourism and animal migrationmay have an impact on stormwater quality. This is a novel approach in the field of stormwater research as no other studies have investigated the effects of specific land-use activities, such as tourism, on stormwater quality. The small study site lends itself to conduct a detailed study with fewer variables at play compared to a large urban center enabling an in-depth understanding of the variables that may affect stormwater quality in the town.

Site description and historical reports
Jasper is a small remote town located in Jasper National Park, which is situated in the Rockies mountain range in Alberta, Canada. The park is at the headwaters of some of Canada's major river systems and has numerous pristine lakes, creeks and wetlands scattered throughout its region.
Parks Canada (PC) prioritizes the conservation of these water bodies, because they sustain abundant wildlife and attract millions of tourists every year. The town of Jasper is situated at the confluence of three valleys which is a vital wildlife movement corridor for a range of animals such as elk and bears. Stormwater outfalls in the town drain runoff into local creeks and lakes, which ultimately feed into the Athabasca River, a Canadian Heritage River.
Major outfalls in the town as well as the surrounding water bodies are indicated in Figure 1. These water bodies serve as habitats to sensitive aquatic ecosystems and provide drinking water for wild animals. Therefore, it is of great importance to ensure that stormwater from the town does not have a negative impact on the natural ecosystem of receiving water bodies. At the same time, it is important to preserve the aesthetic and recreational value of receiving water bodies for the residents and tourists in Jasper.
The town has a permanent resident population of only 4,700 people, but it attracts many tourists during the summer months, with a capacity to support 20,000 visitors daily during the peak season (July to October). The high influx of tourists may result in disproportionately high stormwater contamination, as studies have found that pollutant buildup is largely attributed to anthropogenic sources such as population density and vehicular traffic (Adams & Papa ). The town of Jasper experiences small but frequent storms in the summer which are associated with high pollutant wash off (Alberta Guidelines ). Furthermore, winter brings sub-zero temperatures, followed by spring melt, which may result in high pollutant loading from accumulated pollutants. The stormwater drainage system for the Municipality of Jasper is largely comprised of a conventional subsurface storm sewer collection system. This system is a complete curb-gutter system and was built for a 5-year design storm.
It was established in the early 1900s after the town had developed and has expanded together with the expansion of the town. There are three main catchments, namely North, Central and Cabin Creek, and they drain runoff to the NO1, CE1 and CA2 outfalls, respectively ( Figure 1).
The North catchment is the largest with an area of 72 ha and a total storm sewer length of 7.5 km. All catchments have gentle grades. The average slopes of the North, Central and Cabin Creek catchments are 3.3, 2.7 and 2.5%, respectively, and the flushing conditions were assumed to be the same for all catchments. The directly connected impervious area ranges from 20% in the residential areas to 70% in the commercial areas, whereby the North catchment is a mixture of both land-use types, the Central catchment is mostly commercial and Cabin Creek is residential only. Outfall CE1 conveys runoff into the North Twin Lake, whereas CA2 and NO1 drain stormwater to Cabin and Cottonwood creeks, respectively, which travel from high lying wetlands to the Athabasca River.
Stormwater management is a shared responsibility between PC and the Municipality of Jasper. There is little evidence of flooding or erosion damage caused by runoff because the stormwater system is well established and the soil is well drained. Street sweeping is the only form of quality control currently applied in the town. However, the stormwater quality is largely unknown, and there is currently no management plan in place to mitigate potential stormwater impacts. Previous monitoring reports, which were provided by PC, were completed by independent consulting firms, and there lacks consistency in monitoring methods to directly compare results between studies in order to characterize stormwater. It is therefore necessary to design comprehensive, long-term studies to investigate stormwater issues.  Table 1 shows that outfalls experienced no flow during the dry event (Event 1) on 5 October 2017, which was expected. The samples obtained during Event 1 are representative of background conditions (runoff that results from cross-connections, irrigation drainage and groundwater flow). Event 6 on 2 November 2018 was also reported as generating 'no flow' at outfalls (Table 1), because the storm was small. Therefore, it was assumed that surface runoff was too low to mobilize the first flush of contaminants and that observations represent background pollutant levels.
Events 2-4 in Table 1 were collected during typical rainfall and snowmelt runoff conditions in the town. The three outfalls CA2, CE1 and NO1 experienced low flows during these events. It is only during Event 5 on 2 July 2018 that  Analytical parameters that exceeded either PCLT or CWQG guidelines for two or more sampling events during the course of the test period were considered POIs. Parameters that exceeded guidelines on one occasion were isolated loading cases and were not considered POIs, unless they were a cause for concern in previous studies.

Analysis methods
The magnitude and frequency of POI exceedances were assessed to determine whether they are a cause for concern for the health of aquatic life. This was done by visually inspecting concentration graphs for each POI. Furthermore, POI concentration patterns were compared to identify potential correlations and relevant relationships (Burton & Pitt ). Apparent correlations between pollutants were subject to statistical hypothesis testing with a Spearman rank-order correlation, performed with the XLSTAT software. Finally, spatial and temporal concentration patterns of POIs were evaluated to characterize stormwater and determine potential POI sources.

Identifying and analyzing POIs
The results from the analytical testing of samples were evaluated, and a total of 14 parameters were identified as POIs. for each event (x-axis) at the outfalls CA2 (green), CE1 (blue) and NO1 (red). The horizontal dashed line in each graph delineates the applicable CWQG or PCLT threshold limit, and the concentration value of the threshold limit is marked in the upper right corner of the graph. POIs that exceeded guidelines significantly and frequently were judged to be of greater concern to the health of aquatic life than marginal and occasional exceedances. A qualitative evaluation of the analytical results in Figure 2 shows that POIs of highest concern are phosphorus, total suspended solids (TSS) and metals. These findings compare well with the literature which was conducted for large urban areas. The impact of fecal coliforms on aquatic environments is of high concern, because analysis of the samples showed PHC fraction 2 (PHC-F2) only exceeded guidelines during Event 5 at the NO1 outfall; however, it is included as a POI because it was identified as a cause for concern in previous studies at the same location. Exceedance of PHC-F2 was significant, reaching almost 12 times more than the guideline, and occurred during the event of highest flow. In addition, a hydrocarbon sheen and odor were occasionally observed at either CE1 or NO1 during Events 3-6.
Metal concentrations were found to have a statistically

Stormwater characterization
The spatial and temporal concentration patterns of POIs were evaluated to characterize stormwater and determine potential POI sources. Figure 3(a) is a land-use map of Jasper and Figure 3(b) is a map that indicates road features in Jasper which may be responsible for certain pollutant loading.
TSS concentrations were generally lower for low-flow and dry events and highest during Event 5 (the major storm event). Further inspection of the data reveals that TSS concentrations tend to be higher at the CE1 and NO1 outfalls compared to the CA2 outfall. However, TSS concentrations were unexpectedly lower (especially at CE1) during Event 5the same period when street sweeping was implemented in Jasper. In communication with the Municipality maintenance staff, it was determined that street sweeping was implemented frequently (on a weekly basis) during the summer season, predominantly in the central catchment (Figure 3(b)). This corresponds to the particularly low TSS at the CE1 outfall. It is thus likely that the sweeping was responsible for TSS reductions at the CE1 and NO1 out-  (Figure 3(b)). The data findings suggest that the vehicles on this road are a significant contributor to oil leaks in Jasper because the road provides ample parking and primary access to the commercial center, and has four gas stations along with it where vehicular oil leaks were frequently observed (Figure 3(b)). The road has especially high volumes of traffic during the tourist season, and this is reflected in the data which show that PHC-F2 concentration spiked during Event 5.
The POI concentration graphs in Figure 2 suggest that  which found that TSS, metals and TP were the most common pollutants found in stormwater.
It was found that the prominent source of metal and PHC loading was from vehicular traffic in commercial areas. Furthermore, these parameters were present in significantly higher concentrations during Event 5, which was attributed to the high influx of tourists in the summer months. Fecal coliforms were present in significant magnitude in the summer which was indicative of, amongst other things, the presence of fecal droppings from the resident elk population which is unique to the town of Jasper.
TSS and metal concentrations were statistically correlated whereby Fe showed the closest correlation to TSS, followed by Al, Pb, Zn, Co, Cu, Cr, Mn and Cd. Furthermore, it was found that street sweeping in the town coincided with reduced TSS and metal loads at the Central and North outfalls. The concentration patterns of TP and Cl were found to respond to meteorological events, with similar findings in the literature for large catchments. TSS concentrations were statistically correlated with TP concentrations, and thus, the source of TP was linked to soil features (organic debris). Chloride was detected during spring melt events, and the source was presumed to be road salting.
The overall findings from this study are based on limited data, and thus, it is recommended that future sampling programs focus on capturing additional sampling events to reduce uncertainty. Nevertheless, the findings from this study provide a platform for future stormwater research in the field of small urban studies.

FUNDING
Financial support was provided by PC Agency, the Royal Military College and Queen's University for the sampling of stormwater, as well as processing and analyzing of the analytical results.