Journal Pre-proof Severe western Canadian wildfire affects water quality even at large basin scales Craig A. Emmerton, Colin A. Cooke, Sarah Hustindss, Uldis Silins, Monica B. Emelko, Ted Lewis, Mary K. Kruk, Nadine Taube, Dongnan Zhu, Brian Jackson, Michael Stone, Jason G. Kerr, John F. Orwin PII: S0043-1354(20)30608-4 DOI: https://doi.org/10.1016/j.watres.2020.116071 Reference: WR 116071 To appear in: Water Research Received Date: 23 January 2020 Revised Date: 4 June 2020 Accepted Date: 15 June 2020 Please cite this article as: Emmerton, C.A., Cooke, C.A., Hustindss, S., Silins, U., Emelko, M.B., Lewis, T., Kruk, M.K., Taube, N., Zhu, D., Jackson, B., Stone, M., Kerr, J.G., Orwin, J.F., Severe western Canadian wildfire affects water quality even at large basin scales, Water Research (2020), doi: https:// doi.org/10.1016/j.watres.2020.116071. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2020 Published by Elsevier Ltd. Following a wildfire, shortlived increases in chemicals associated with ash occurred in the waters of very large, low-relief rivers 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Severe western Canadian wildfire affects water quality even at large basin scales Emmerton, Craig A.1,2*, Cooke, Colin A.1,3*, Hustins, Sarah1, Silins, Uldis4, Emelko, Monica B.5, Lewis, Ted6, Kruk, Mary K.1, Taube, Nadine1, Zhu, Dongnan1, Jackson, Brian 1, Stone, Michael 7, Kerr, Jason G.1, Orwin, John F.1 Affiliations: 1 Alberta Environment and Parks, Edmonton, Alberta, Canada Dept. of Biological Sciences, University of Alberta, Edmonton, Alberta, Canada 3 Dept. of Earth and Atmospheric Sciences, University of Alberta, Edmonton, Alberta, Canada 4 Dept. of Renewable Resources, University of Alberta, Edmonton, Alberta, Canada 5 Dept. of Civil and Environmental Engineering, University of Waterloo, Waterloo, Ontario, Canada 6 Hatfield Consultants, North Vancouver, British Columbia, Canada 7 Dept. of Geography and Environmental Management, University of Waterloo, Waterloo, Ontario, Canada 2 Corresponding Author (*) Emails: Emmerton: craig.emmerton@gov.ab.ca Cooke: colin.cooke@gov.ab.ca 20 21 Abstract 22 However, our understanding of these impacts is founded primarily from studies of small 23 watersheds with well-connected runoff regimes. Despite the predominance of large, low-relief 24 rivers across the fire-prone Boreal forest, it is unclear to what extent and duration wildfire- 25 related material (e.g., ash) can be observed within these systems that typically buffer upstream 26 disturbance signals. Following the devastating 2016 Fort McMurray wildfire in western Canada, 27 we initiated a multi-faceted water quality monitoring program that suggested brief (hours to 28 days) wildfire signatures could be detected in several large river systems, particularly following 29 rainfall events greater than 10 millimeters. Continuous monitoring of flow and water quality 30 showed distinct, precipitation-associated signatures of ash transport in rivers draining expansive 31 (800–100,000 km2) and partially-burned (<1–22 percent burned) watersheds, which were not 32 evident in nearby unburned regions. Yields of suspended sediment, nutrients (nitrogen, 33 phosphorus) and metals (lead, others) from impacted rivers were 1.2 to 10 times greater than 34 from those draining unburned regions. Post-fire suspended sediment concentrations in impacted 35 rivers were often larger than pre-fire 95% prediction intervals based on several years of water 36 sampling. These multiple lines of evidence indicate that low-relief landscapes can mobilize 37 wildfire-related material to rivers similarly, though less-intensively and over shorter durations, 38 than headwater regions. We propose that uneven mixing of heavily-impacted tributaries with 39 high-order rivers may partially explain detection of wildfire signals in these large systems that 40 may impact downstream water users. 41 Keywords: wildfire, river, water quality, suspended sediment, Boreal, continuous monitoring Wildfires can have severe and lasting impacts on the water quality of aquatic ecosystems. 42 43 1. Introduction Communities within the wildland-urban interface of forested regions are experiencing 44 more frequent and destructive wildfires (Hanes et al., 2018; Mell et al., 2010). This changing 45 wildfire regime is associated with a warming climate that is intensifying dangerous fire weather 46 and drought conditions (Diffenbaugh et al., 2015; Jolly et al., 2015), exposing these communities 47 to increased risk of catastrophic impacts (Radeloff et al., 2018). Besides the evident 48 consequences to lives, natural capital and infrastructure, wildfires often impact downstream 49 surface water resources resulting from changes in runoff and degraded, increasingly variable 50 water quality that can affect aquatic ecosystem functioning and drinking water treatability 51 (Bladon et al., 2014; Gresswell, 1999; Hohner et al., 2019; Smith et al., 2011; Writer et al., 52 2014). 53 Wildfires can enhance meteoric runoff from watersheds through chemical soil sealing 54 and reduction in evapotranspiration due to vegetation removal (Larsen et al., 2009; Neary et al., 55 2005). Depending on fire severity and the timing of storm events, delivery of ash and sediment 56 (Moody and Martin, 2001; Santín et al., 2015), organic matter and nutrients (Burd et al., 2018; 57 Earl and Blinn, 2003; Lane et al., 2008; Rhoades et al., 2011), major ions (Bayley et al., 1992; 58 Chanasyk et al., 2003; Mast et al., 2016), and trace contaminants (Kelly et al., 2006) into 59 receiving waters can occur and cause deteriorated water quality (Bladon et al., 2014; Smith et al., 60 2011). Ensuing biogeochemical processing of this mobilized wildfire-related material can induce 61 secondary ecological impacts on aquatic ecosystems resulting from eutrophication and 62 sedimentation (Earl and Blinn, 2003; Gresswell, 1999; Silins et al., 2014). However, our 63 understanding of fire-related impacts on surface waters has been overwhelmingly established 64 from investigations conducted at relatively small catchment scales (e.g., <10s–100s km2) and in 65 catchments with responsive hydrologic connectivities between landscapes and receiving waters 66 (Belillas and Rodà, 1993; Earl and Blinn, 2003; Gerla and Galloway, 1998; Inbar et al., 1998; 67 Lane et al., 2008; Minshall et al., 2001; Neary et al., 2005; Prepas et al., 2003; Scott et al., 1998; 68 Smith et al., 2011). 69 Landscape disturbances and related impacts on downstream river conditions are difficult 70 to detect as watershed and river sizes increase (Blöchl et al., 2007; Elhadi et al., 1984). High- 71 order river systems aggregate water and material from broad drainage networks and dampen 72 biogeochemical signals from upstream landscapes through lateral mixing, dilution, and other in- 73 stream processes (Temnerud and Bishop, 2005). Consequently, assessments of wildfire-related 74 impacts on water quality in rivers draining diverse watershed sizes—especially very large ones— 75 are scant (Smith et al., 2011). Watershed hydrologic connectivity also influences river response 76 to landscape disturbance (Bracken and Croke, 2007). For instance, steeper topographic settings 77 and well-drained soils can promote rapid hydro-geochemical connections between disturbed 78 landscapes and rivers, and contribute to sustaining downstream and temporal wildfire impacts on 79 water quality (e.g.,(Rust et al., 2018). Though several studies have described wildfire impacts on 80 water quality in larger (Burke et al., 2005; Emelko et al., 2016; Rhoades et al., 2011) or low- 81 relief (Townsend and Douglas, 2000) watersheds, these investigations have been conducted in 82 landscapes that are hydrologically well-connected. Critically, wildfire impacts on water are not 83 well understood in large river systems with relatively low hydrologic connectivity, in which 84 contaminants may be more efficiently retained on land surfaces conferring greater apparent 85 storage or watershed buffering capacity (Ebel and Mirus, 2014; Prepas et al., 2003). These 86 attributes are characteristic of many areas across the circumpolar Boreal forest, where low 87 topographic relief, heterogeneous glacial deposits and extensive peatland landscapes result in 88 relatively low landscape-water coupling (Devito et al., 2017). While wildfires have been 89 investigated at plot- and subwatershed-scales in these Boreal landscapes (e.g.,(Depante et al., 90 2018; Olefeldt et al., 2013; Prepas et al., 2003; Wilkinson et al., 2018), wildfire impacts on water 91 quality at large basin scales in these regions have not been previously reported. This a critical 92 knowledge gap considering the high future probability of more severe fires occurring across the 93 circumpolar Boreal forest. 94 The primary objective of this study is to assess the impacts of the May 2016 Fort 95 McMurray (Horse River) wildfire in northeastern Alberta, Canada on the water quality of large 96 (>1,000 km2 drainage areas) and low-relief (<5% drainage area slope) rivers in the region. Using 97 a multi-tiered monitoring approach to target signal detection in these large rivers, we first 98 compared suspended sediment, dissolved material (ion), nutrient, and metal concentrations and 99 yields in rivers draining burned areas with an unburned control station. Second, at each station, 100 we assessed if precipitation events and subsequent river flow mobilized more wildfire-related 101 material relative to non-precipitation periods. Finally, we evaluated suspended sediment, 102 dissolved material and flow relationships in rivers draining burned landscapes with pre-fire data 103 from the same station. We hypothesized that rivers draining burned watersheds would be 104 resistant to unbuffered overland flow and therefore show no detectable differences in water 105 quality relative to rivers draining unburned landscapes or extensive historical data. 106 107 108 2. Methods 2.1 Setting Our study region is contained within the Boreal Plains ecozone, which is characterized by 109 flat and rolling landscapes with thick glacial till soils, and extensive coverage of wetlands and 110 coniferous forests (Chanasyk et al., 2003; Devito et al., 2017). Due primarily to the Rocky 111 Mountains barrier ~600 km to the west, much of this ecozone is relatively dry (419 mm annual 112 precipitation at Fort McMurray; 1981–2010) with a cold, continental climate (Environment and 113 Climate Change Canada, 2018). Human disturbances in the region are mostly from local 114 urbanization in Fort McMurray and oil sands strip-mining and exploration-related impacts to 115 landscapes (Figure 1). 116 The Athabasca River drains approximately 140,000 km2 of western Canada and occupies 117 23% of the province of Alberta (Figure 1). This Strahler-order nine river is large and voluminous 118 (~500 m wide; >500 m3 s-1 mean annual flow) in its lower reaches and exhibits a spring 119 snowmelt-freshet seasonality in flow (Figure 2). In this portion of the river, concentrations of 120 suspended sediment and associated water quality parameters (e.g., organic carbon, nutrients, 121 metals) derived from upstream tributaries and channel re-suspension can be high relative to 122 national water quality guidelines during much of the year (Glozier et al., 2018). Several high- 123 order river systems empty into the lower Athabasca River including the Clearwater and 124 Hangingstone rivers (Figures 1, 2). 125 The 2016 Fort McMurray wildfire (Figure 1; May 01/16–August 02/17) was the most 126 costly natural disaster in Canadian history ($8.9 billion in direct costs;(Alam et al., 2018). By 127 May 03/16, the wildfire advanced quickly from the southwest of Fort McMurray and triggered 128 the dramatic evacuation of ~90,000 people, resulting in a loss of 3,000 structures and eventual 129 consumption of ~600,000 hectares of forest and peatland. Within the final fire perimeter (nearly 130 established by May 17/16), 77% of the area was impacted by fire, half of which burned at a high 131 severity (see supplementary material for methodology). 132 2.2 Monitoring and historic data 133 Post-fire river water quality monitoring in this study focused on three rivers: the 134 Athabasca, Clearwater and Hangingstone. Three burned (impacted) stations were established: (1) 135 on the Athabasca River upstream of Fort McMurray (left bank); (2) the Clearwater River near 136 the mouth; and (3) the Hangingstone River at Fort McMurray (Figure 1). These locations drain a 137 gradient of watershed size (Hangingstone<Clearwater<Athabasca) and reflected a contrasting 138 gradient of relative burned area and area burned at a high severity in their watersheds 139 (Athabasca<Clearwater< Hangingstone; Table 1). An unburned (control) station was also 140 initiated on the Athabasca River at Grand Rapids, upstream of the wildfire perimeter. Each of 141 these four water quality stations incorporated multiple measurement approaches and, where 142 possible, were co-located near meteorological and flow stations to better understand hydrological 143 influences on post-fire water quality. 144 2.2.1 Water quality measurements 145 A multiple-method approach for sampling river chemistry of each of the four water 146 quality stations was implemented after the fire to assess differences between unburned and 147 burned locations (Figure 1, Table S1). Nearly two weeks after the fire began, we deployed multi- 148 probe data sondes (Hydrolab DS5X, OTT Hydromet, USA) just offshore at the two Athabasca 149 stations (burned and unburned), and near mid-channel at the smaller Clearwater and 150 Hangingstone stations. Two additional sondes were deployed downstream of Fort McMurray just 151 offshore of each bank of the Athabasca River to assess river mixing conditions (Figure 1). Each 152 calibrated sonde recorded general water quality including water temperature, pH, specific 153 conductivity, turbidity and dissolved oxygen at 15-minute intervals. Sondes were retrieved and 154 replaced with calibrated instruments every week to limit impacts from fouling and sensor drift. 155 All sonde monitoring concluded in mid-October 2016. 156 In addition to the sondes, we collected daily water quality samples at the four main 157 stations (Figure 1, Table S1). Surface water autosamplers (Teldyne ISCO, USA) were deployed 158 on the river bank at each station (burned: May–August, 2016; control: June–August, 2016) with 159 water intakes placed mid-column just offshore at the two Athabasca locations, and mid-channel 160 in the Clearwater and Hangingstone rivers. Each autosampler collected equal aliquots of river 161 water every three hours into 1 L ProPak® plastic bags in holders for a time-integrated sample 162 each day. Weekly grab samples of surface waters were manually collected at each station from 163 May to August 2016 to supplement automated sampling. After wading waist-deep at each 164 location, water samples were collected at 30-cm depth directly into pre-cleaned or new plastic or 165 glass bottles, depending on chemical methods. At the three burned stations (Athabasca-Fort 166 McMurray, Clearwater-mouth, Hangingstone-Fort McMurray), grab sampling continued 167 monthly through 2016, 2017 and 2018 via other programs. 168 All water samples were analyzed at accredited laboratories for suspended sediment, total 169 ion (Na, K, Mg, Ca, Cl), total nutrient (total organic carbon, total Kjeldahl nitrogen, total 170 phosphorus) and total metal (29-metal scan) concentrations. Only total measurements were 171 considered for autosampler water samples to limit sample chemical transformation concerns 172 during storage. 173 2.2.2 Precipitation and river flow measurements 174 175 To identify periods when wildfire-related material may be mobilized in burned watersheds, we delineated notable local precipitation events and subsequent river flow response 176 periods at each water quality station. Total daily precipitation data from meteorological stations 177 nearest to sampling locations (Figure 1, Table S1) were used to identify noteworthy rainfalls 178 equalling or exceeding 10 mm (94th percentile of daily rain events at Fort McMurray A, 1999– 179 2015). Precipitation events at each station (see supplementary material) are defined in time here 180 as the day of the noteworthy rainfall plus five subsequent days to encompass flow and 181 biogeochemical responses of each river. Days not within a precipitation event period are defined 182 as non-event periods. 183 Hourly and daily flow data before and after the wildfire were retrieved from hydrometric 184 stations close to water quality locations, where possible (Figures 1, 2, Table S1). Discharge at the 185 Athabasca River upstream of Fort McMurray (left bank) station was calculated by difference in 186 measured daily flows between the nearby Athabasca River downstream of Fort McMurray and 187 Clearwater River at Draper stations. Flow at Athabasca River at Grand Rapids was calculated 188 using daily runoff yields from the upstream Athabasca River at Athabasca hydrometric station 189 scaled to the drainage area at the Grand Rapids location (Table 1, Figure 1). 190 2.2.3 Historic water quality data 191 Historic water quality data from long-term provincial and oil sands monitoring programs 192 (2001-2016) were used to quantify pre- and post-fire water quality conditions at the three 193 impacted stations (Table S1). Monthly surface water sampling has been ongoing in the 194 Athabasca River upstream of Fort McMurray for many decades, while the Clearwater and 195 Hangingstone rivers have been monitored for several years. Monthly grab sampling also 196 occurred in 2015 (pre-fire) and 2016 (post-fire) at burned (High Hills, Christina, Clearwater, 197 Hangingstone) and unburned (Calumet, Mackay, Ells, Firebag) river stations via regional 198 monitoring networks (Table S1, Figure S1a). Water collection and analytical methods from 199 historical programs were consistent with those used during the post-fire sampling program. 200 2.3 Quantitative analyses 201 2.3.1 Water quality assessment 202 River turbidity and conductivity are typically negatively correlated during high-water 203 events due to flow-related sediment concentration and ion dilution (Schlesinger, 2007). 204 However, after wildfires, ash-laden runoff pulses during storms can trigger concurrent increases 205 in turbidity and conductivity in receiving waters (Dahm et al., 2015; Earl and Blinn, 2003). To 206 identify existence of concentration-dilution and concentration-concentration periods in each 207 river, mean hourly sonde and flow data were included in an unconstrained principal components 208 analysis (PCA; Canoco v.5.03; Biometris, The Netherlands). Six-hour covariances (using a 60- 209 minute moving window) of turbidity and conductivity were quantified at each station (including 210 sonde-only locations) to directly compare the magnitude and frequency of suspected ash 211 mobilization periods (i.e., positive covariances) between stations. We also performed a 212 laboratory ash-mixing experiment to demonstrate how ash affected turbidity and conductivity 213 measurements in an aqueous solution. We collected loose grey and black ash from the soil 214 surface of burned forested plot (0.25 m2) in the Clearwater River watershed (N56.6849; W- 215 111.2560) on May 13/16, prior to post-fire rainfall events. Later in the laboratory, we 216 incrementally added ~500 mg of ash to 10 L of deionized water and continuously stirred the 217 solution while turbidity, conductivity and pH were measured using a calibrated sonde. 218 Descriptive statistics were calculated to quantify differences in yields of selected particle- 219 bound and dissolved chemicals between impacted and control river stations on the Athabasca 220 River. To link assessments of selected chemicals at each station to other measured parameters, 221 we correlated all parameter concentrations with concurrent measurements of flow, turbidity, 222 conductivity and pH to organize parameters as particle-bound type (positive correlation with 223 flow and turbidity) or dissolved type (positive correlation with conductivity and pH). 224 225 2.3.2 Precipitation and flow assessment All water quality data (i.e., sonde, samples) at each station were binned into precipitation 226 event (≥10 mm) and non-event (<10 mm) periods for comparisons. This approach was used on 227 sonde turbidity-conductivity covariance data and concentrations and yields of sampled 228 chemicals. Assessment of hydrologic responses of impacted rivers before and after the fire is 229 described in the supplementary material. 230 2.3.3 Historic water quality assessment 231 Post-fire changes in particle-bound and dissolved river water chemistry, relative to 232 historic records, were evaluated in the three rivers draining burned watersheds (Table S1) using 233 concentration-discharge (C-Q) relationships. Monthly, pre-fire water quality data (suspended 234 sediments as a proxy of particle-bound chemicals, total calcium as a proxy for specific 235 conductivity/dissolved chemicals; see Table S2) from each river station were log-transformed 236 and linearly regressed with log-transformed, pre-fire daily discharge data. Subsequently, 95% 237 prediction intervals were calculated (Sigmaplot v13, Systat, USA) to serve as a boundary of what 238 C-Q relationships would be expected in the river based on historic conditions. When post-fire C- 239 Q relationships (2016–18) were compared with historic relationships, data points that were 240 outside of the bounds of the prediction intervals were considered noteworthy. Other monthly 241 grab sample and flow data from four unburned and four burned river stations were used in a 242 repeated measures ANOVA (SPSS v.25; IBM, USA) to assess lower-resolution pre-fire (2015) 243 and post-fire (2016) impacts on water quality. 244 3. Results 245 3.1 Water quality 246 3.1.1 Sonde water quality 247 High-frequency water quality monitoring by sondes at the four main stations typically 248 showed flow-concentration for turbidity and flow-dilution for conductivity (Figure 3) and pH, as 249 well as contrasting seasonal changes for water temperature and dissolved oxygen. These trends 250 were evident in PC1 and PC2 scores of the PCA (Table S3). However, sharp peaks in turbidity 251 and conductivity together occurred during short periods (hours to days) often along the rising 252 limb of the hydrograph at burned stations only. The third principal component at each burned 253 station reflected this covariation between turbidity and conductivity in the absence of association 254 with flow. At the Athabasca River control station, PC3 scores also showed periods of concurrent 255 increases in turbidity and conductivity, but were positively associated with flow, perhaps 256 suggesting an influence of saline groundwater (Gibson et al., 2016). Mean six-hour covariance 257 scores of turbidity and conductivity more distinctly showed differences between stations than did 258 the PCA. Covariances from burned stations were 3–25 times higher than the control station, 259 though means were influenced by the highest values (Figure 4). Highest covariances at impacted 260 stations typically occurred during precipitation events (mean: 14-113) and were 2 to 22 times 261 higher than non-event periods. These observations were consistent with results of the ash mixing 262 experiment that demonstrated concurrent increases in turbidity, specific conductivity and pH as 263 ash was added to deionized water (Figure S2). Further, though pre-fire grab sampling in 264 impacted rivers showed consistent negative associations between turbidity concentrations and 265 specific conductivity, 11-38% of post-fire, precipitation event turbidity and conductivity samples 266 were greater than the upper historic 95% prediction interval, suggesting a more positive 267 association between the parameters during those periods (Figure S3). 268 Within the burned reach of the Athabasca River, turbidity and conductivity covariances 269 on the left bank decreased from a mean of 7.7 (maximum: 1,071) upstream to a mean of 7.0 270 (maximum: 574) downstream. However, the downstream right bank station observed much 271 higher covariance values (mean: 17.2; maximum: 3,782) than either station on the left bank, 272 which together suggests an unmixed chemical profile longitudinally and laterally in the river 273 (Figure 3,4). 274 3.1.2 Sample water quality 275 Daily automated and grab sample water quality results indicated only slight differences 276 between burned and control stations when aggregated to the entire sampling season (Figure 5). 277 For example, seasonal mean daily sediment yields (June 22–August 30/16) from the directly 278 comparable stations of the Athabasca River at Grand Rapids (unburned; GR) and the Athabasca 279 River upstream of Fort McMurray (burned; FMM) were within one standard error of each other 280 (GR: 0.68±0.10 kg ha-1 d-1; FMM: 0.75±0.09 kg ha-1 d-1). Results of other sediment-associated 281 yield comparisons including total lead (GR: 8.1±1.0; FMM: 10.3±1.8 mg ha-1), total Kjeldahl 282 nitrogen (GR: 3.8±0.3; FMM: 3.3±0.2 g ha-1 d-1), and total phosphorus (GR: 0.54±0.07; FMM: 283 0.45±0.04 g ha-1 d-1) were consistent with the suspended sediment results post-wildfire and likely 284 other correlates (Table S2). Similarly, monthly suspended sediment yields across a subset of 285 several rivers draining burned and unburned watersheds in the region before and after the fire 286 indicated no fire effect (Figure S1b). Total calcium, used as a proxy measure for dissolved 287 material (Table S2), also showed similar yields between the impacted (0.23±0.01 kg ha-1 d-1) and 288 the control station (0.21±0.01 kg ha-1 d-1) on the Athabasca River (Figure 5). Differences in 289 monthly calcium yields between a subset of rivers draining burned and unburned watersheds 290 were also statistically-similar (Figure S1b). Alternatively, total organic carbon (TOC), most of 291 which is in the dissolved form in the Athabasca River (Glozier et al., 2018), had mean yields 292 greater at the control (91±5 g ha-1 d-1) relative to the impacted (67±3 g ha-1 d-1) station. 293 Substantial differences occurred within and between stations when classifying data by 294 precipitation event and non-event periods. At impacted stations only, peaks of concentrations 295 (Figure 6) and yields during precipitation events were observed relative to non-event periods for 296 suspended sediment (concentrations: 1.7–6.6 times higher; yields: 1.8–10.3x), total phosphorus 297 (1.9–2.6x; 2.0–5.3x), total Kjeldahl nitrogen (1.3–1.4x; 1.2–2.4x), and total lead (2.1–3.3x; 2.0– 298 6.3x). In contrast, such ratios at the control station were lower than impacted stations and were 299 often near unity (concentrations: SS=0.9, TP=1.3, TKN=0.9, Pb=1.1x; yields: SS=1.0, TP: 1.4; 300 TKN: 0.9; Pb: 1.2x). Total calcium and TOC yields were largely unchanged between 301 precipitation and non-event periods (0.8–1.1x) at all stations except for the burned Hangingstone 302 River (Ca: 1.7x; TOC: 1.5x). 303 Despite observing measureable impacts to water quality at burned stations, most 304 monitored parameters rarely exceeded national surface water quality guidelines for the protection 305 of aquatic life (Government of Canada, 2019), or the relative number of exceedances were 306 comparable to pre-fire conditions or those at control stations (Table S4). 307 308 3.2 Historic water quality Elevated suspended sediment concentrations observed at burned stations did not result 309 from higher flows than sampled historically, nor wholesale changes in the hydrological 310 conditions of impacted river systems (see supplementary material; Figures S4-S6). Specifically, 311 the prediction intervals calculated using historic C-Q relationships for suspended sediment from 312 the burned Athabasca River upstream of Fort McMurray station were exceeded on four 313 occasions after the wildfire (Figure 7). Two of these exceedances occurred on June 09 and July 314 31/16, when the two largest recorded rainfalls in the Fort McMurray region occurred during that 315 summer. Sampling during subsequent years showed no similar notable concentrations. Five 316 exceedances of the upper historic prediction interval occurred in the Clearwater River during 317 precipitation events in 2016, but none during the following years. The Hangingstone River 318 experienced 17 exceedances relative to its historic concentration-discharge relationship. 319 However, these exceedances occurred during precipitation event and non-event periods in 2016, 320 as well as across multiple years after the fire, including during spring runoff in 2017. These C-Q 321 results were likely similar to other particle-bound measures (Table S2). Compared to suspended 322 sediment concentrations, only four exceedances of the upper historic prediction interval for total 323 calcium concentration were observed across all rivers, and likely similar for other dissolved-type 324 chemicals (Table S2). Collectively, suspended sediment concentration exceedances of historic 325 95% prediction intervals in impacted rivers became more frequent as watershed size decreased, 326 and the proportion of burned area and area burned at high severity across watersheds increased 327 (Table 1, Figure 7). 328 4. Discussion 329 330 4.1 Episodic detections of fire-related material in large river systems Despite (a) naturally turbid and organic-rich water qualities prior to disturbance (Figures 331 3,5,6;(Glozier et al., 2018), (b) the potential for considerable dilution including from substantial 332 subsurface contributions to regional river flow (Gibson et al., 2016), and (c) poor hydrologic 333 connectivity between the landscapes and low-relief rivers (Chanasyk et al., 2003; Devito et al., 334 2017; Gibson et al., 2016; Ireson et al., 2015), wildfire impacts on river water quality were 335 detectable at the very large basin scale following the 2016 Fort McMurray wildfire. Three 336 independent monitoring approaches reported brief, precipitation-related increases in wildfire- 337 related material (i.e., suspended and dissolved material, nutrients, metals;(Burton et al., 2016; 338 Earl and Blinn, 2003; Reale et al., 2015) in rivers draining burned watersheds relative to a 339 control station or historic, pre-fire data (Figures 3-4,6-7). These findings were also supported by 340 an ash mixing experiment (Figure S2) and other historic data. Together, this weight-of-evidence 341 suggests rainfall-induced surface erosion delivered ash and soil from local, burned hillslopes into 342 large rivers that briefly impacted water quality without shifting overall metrics of river water 343 quality relative to unburned regions (Figure 5). The role of precipitation events in mobilizing 344 particle-bound chemicals across burned landscapes is well-established in the literature 345 (e.g.,(Bladon et al., 2014; Burton et al., 2016; Langhans et al., 2016; Townsend and Douglas, 346 2000) and in more extensively-burned regions of the Boreal Plains (Burke et al., 2005). 347 However, detection of suspended chemicals in large, wetland-dominated, low-slope river 348 watersheds with low proportions of areas burned (Table 1) was a key outcome of this study. 349 Alternatively, total measurements mostly in dissolved form (e.g., total calcium, total organic 350 carbon; others; Table S2) showed inconsistent changes between impacted and control stations, as 351 well as between precipitation and non-event periods, perhaps reflecting uneven burning of peat at 352 depth where non-erosive flows dominate (Devito et al., 2012; Wilkinson et al., 2018). 353 354 4.2 Factors shaping wildfire signal detection in large river systems Following wildfire, loss of ground cover, decreased surface roughness, and soil crusting 355 and sealing (Larsen et al., 2009) can impede infiltration and expose soils to splash and fluvial 356 erosion (Shakesby and Doerr, 2006). These processes often allow storms to mobilize loads of 357 sediment, nutrients and contaminants downstream (Bayley et al., 1992; Betts and Jones, 2009; 358 Moody and Martin, 2001; Neary et al., 2005). However, the magnitude and duration of these 359 water quality impacts are shaped by the intensity of precipitation events and the erosive potential 360 of impacted landscapes (Langhans et al., 2016). Headwater systems, with typically small, 361 extensively-burned, steep and flashy landscapes, often show substantial sediment yields from 362 burned watersheds relative to those unburned (typically >100x;(Smith et al., 2011) with water 363 quality impacts often detected for months to years (e.g.,(Inbar et al., 1998; Owens et al., 2013). 364 In this study of lower-relief and wetland-dominated landscapes, catchments are large, burn less- 365 extensively and runoff storage is prevalent (Table 1;(Devito et al., 2012). These conditions 366 resulted in suspended sediment yield ratios (impacted:control) ≤1.3x with detectable water 367 quality impacts at only hourly and daily scales (Figures 3-4,6). However, fire-related impacts on 368 water quality in this study (Table 1, Figure 6), as well as those in headwater regions (Rhoades et 369 al., 2011) demonstrate that differing basin sizes, proportions of area burned, and area burned at 370 high severity, can impact water quality of river systems differently. This suggests that 371 mobilization of fire-related materials downstream is likely driven by similar processes across 372 differing landscapes, but are shaped by catchment- and fire-related characteristics. However, we 373 postulate that incomplete hydro-chemical mixing may also play an important role in wildfire 374 signal detection in these large rivers expected to buffer landscape chemical signals (Blöchl et al., 375 2007). 376 Large rivers can substantially modify upstream and locally-derived terrestrial material via 377 mixing, dilution and sedimentation processes (Temnerud and Bishop, 2005). However, we 378 suggest that the complexities of mixing river waters of different qualities and quantities may 379 have also been responsible for the detection and downstream persistence of wildfire chemical 380 signals observed in these large rivers, particularly in the Athabasca River. Other studies have 381 reported a persistence of plumes/mixing zones downstream of river confluences because of 382 differences in water densities, particle loads, or the chemical composition of sediments (Bouchez 383 et al., 2010; Herrero et al., 2018; Stone et al., 2011). Measurements assessing cross-sectional 384 changes in water quality of the Athabasca River suggest incomplete mixing occurs well 385 downstream of anthropogenic or tributary inputs (Glozier et al., 2018). Following the wildfire 386 and precipitation events, we observed that both low- and high-order tributaries draining severely 387 burned watersheds delivered plumes of ash to the Athabasca River that were visible for several 388 kilometers downstream of confluences (Figure 8). Empirical evidence from the three Athabasca 389 River sondes supported these visual observations. For example, we observed strong differences 390 in turbidity, conductivity, and turbidity-conductivity covariance values measured on the left and 391 right banks of the Athabasca River at the same longitudinal location (Figures 3,4). Higher 392 turbidity values on the right bank were coincident with turbidity and conductivity spikes in the 393 smaller but more extensively-burned Clearwater and Hangingstone rivers that empty upstream 394 from the right bank. The entry of these rivers established a distinct, fire-impacted plume on the 395 right bank of the Athabasca River through to where sonde measurements occurred. Though it 396 was unclear where the right bank plume mixed fully with the Athabasca River downstream, we 397 did observe likely mixing and dilution of fire-influenced plumes on the left bank of the 398 Athabasca. Near the upstream of Fort McMurray station, several severely-burned low-order 399 streams entered the Athabasca River on the left bank and passed though the sonde just 400 downstream, resulting in defined, precipitation-related turbidity and conductivity spikes up to 401 1000 NTU and 350 µS cm-1, respectively (Figure 3,8). Downstream of Fort McMurray on the 402 same bank, another sonde measured turbidity and conductivity spikes often 10-60% as large with 403 lower turbidity-conductivity covariances (Figure 4) compared to the upstream location, and 404 similar to the unimpacted Grand Rapids station. These results suggest efficient mixing of the 405 small, fire-impacted plumes on the left bank with larger volumes of unimpacted Athabasca River 406 water over just a few kilometers of this reach. Regardless, this uneven mixing of waters from 407 impacted tributaries with the large Athabasca River may simultaneously explain observed 408 downstream persistence of wildfire signals, but also overall signal attenuation due to dilution 409 (i.e., relative to headwater regions), both of which were defining observations of this study. 410 411 4.3 Human-related impacts of wildfires in large river systems The occurrence of and damage from wildland-urban interface fires has increased over the 412 previous decade in western North America (Schoennagel et al., 2017; Westerling et al., 2006). 413 As such an event, the Fort McMurray wildfire burned regions both within and surrounding the 414 remote urban center of Fort McMurray, Alberta. However, this wildfire presented a unique 415 situation where very large, low-relief watersheds were expected to buffer wildfire impacts on 416 local surface water resources. Interestingly, though we observed relatively muted water quality 417 changes in these very large rivers compared to headwater regions, these arguably inconsequential 418 basin-scale effects still resulted in impacts to post-fire drinking water treatment costs borne by 419 the local municipality, similar to headwater regions (Bladon et al., 2014; Emelko et al., 2011; 420 Hohner et al., 2019; Thurton, 2017). Preliminary analyses (unpublished) suggest unmixed 421 plumes in the Athabasca River from heavily-impacted, low-order tributaries discharging 422 upstream of the drinking water treatment plant intake on the river may have been responsible. 423 Thus, exclusive focus on the magnitude and persistence of the impacts of wildfire on surface 424 water quality may fail to uncover its true stresses on surface water resources. This consideration 425 may be particularly applicable to future wildland-urban interface fires as export of complex 426 chemicals (e.g., plastics) from residential, commercial, and industrial settings may present 427 unique water quality conditions in receiving waterbodies of human importance (Oliver et al., 428 2012). 429 430 5. Conclusions This study presents a multi-tiered approach to monitoring the flow and water quality of 431 large rivers in a low-relief region affected by an expansive and severe wildfire. While broader 432 wildfire impacts to water quality of regional rivers were markedly lower than commonly 433 reported in smaller watersheds of greater relief, water quality impacts were nonetheless observed 434 across a range of very large river basin scales. Novel use of continuous flow and general 435 chemical monitoring was key to identifying impacted periods in the rivers studied herein, similar 436 to few other investigations (Cooke et al., 2016; Dahm et al., 2015; Mast et al., 2016). In 437 particular, our study emphasizes the importance of considering the spatial frame of reference 438 (entire basin vs. local) when evaluating wildfire impacts on water quality in large river basins. 439 The proximity and scale of this and other recent wildfires to major urban municipalities further 440 highlights the importance of shifting North American wildfire regimes as increasing threats to 441 community surface water resources. 442 443 444 445 446 447 448 449 450 451 452 6. Acknowledgements Funding for this work was provided by Canada’s emergency disaster recovery fund and the Alberta Government. We would like to thank the tireless work during the emergency by Alberta Environment and Parks water quality technicians, particularly Chris Ware, Monica Polutranko, Trina Ball, Shelley Manchur, Brittany Kereliuk, Meghan House, Sarah Lamb, Tye Dubrule and Jessica Pope. We also thank government Data Management staff including Chris Rickard, Doreen LeClair, Lisa Reinbolt and Jenny Pham. Burn severity mapping was graciously provided by the Wildfire Operations sections of Alberta Agriculture and Forestry. We appreciate input by Dr. Bill Donahue, Dr. Paul Drevnick, Dr. Jennifer Graydon and Dr. Fred Wrona during its development. Martin Davies provided valuable interpretation of sonde water quality data. Finally, we thank two anonymous reviewers for taking the time to improve this manuscript. 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 7. References Alam, R., Islam, S., Mosely, E., Thomas, S., Dowdell, V., Doel, D., 2018. Rapid impact assessment of Fort McMurray Wildfire. Edmonton. Bayley, S.E., Schindler, D.W., Parker, B.R., Stainton, M.P., Beaty, K.G., 1992. Effects of forest fire and drought on acidity of a base-poor boreal forest stream: similarities between climatic warming and acidic precipitation. Biogeochemistry 17, 191–204. https://doi.org/10.1007/BF00004041 Belillas, C.M., Rodà, F., 1993. The effects of fire on water quality, dissolved nutrient losses and the export of particulate matter from dry heathland catchments. J. Hydrol. 150, 1–17. https://doi.org/10.1016/0022-1694(93)90153-Z Betts, E.F., Jones, J.B., 2009. Impact of Wildfire on Stream Nutrient Chemistry and Ecosystem Metabolism in Boreal Forest Catchments of Interior Alaska. Arctic, Antarct. Alp. Res. 41, 407–417. https://doi.org/10.1657/1938-4246-41.4.407 Bladon, K.D., Emelko, M.B., Silins, U., Stone, M., 2014. Wildfire and the future of water supply. Environ. Sci. Technol. 48, 8936–8943. https://doi.org/10.1021/es500130g Blöchl, G., Ardoin-Bardin, S., Bonell, M., Dorninger, M., Goodrich, D., Gutknecht, D., Matamoros, D., Merz, B., Shand, P., Szolgay, J., 2007. At what scales do climate variability and land cover change impact on flooding and low flows? Hydrol. Process. 21, 1241–1247. https://doi.org/10.1002/hyp.6669 Bouchez, J., Lajeunesse, E., Gaillardet, J., France-Lanord, C., Dutra-Maia, P., Maurice, L., 2010. Turbulent mixing in the Amazon River: The isotopic memory of confluences. Earth Planet. Sci. Lett. 290, 37–43. https://doi.org/10.1016/j.epsl.2009.11.054 Bracken, L.J., Croke, J., 2007. The concept of hydrological connectivity and its contribution to understanding runoff-dominated geomorphic systems. Hydrol. Process. 21, 1749–1763. https://doi.org/10.1002/hyp.6313 Burd, K., Tank, S.E., Dion, N., Quinton, W.L., Spence, C., Tanentzap, A.J., Olefeldt, D., 2018. Seasonal shifts in export of DOC and nutrients from burned and unburned peatland-rich catchments, Northwest Territories, Canada. Hydrol. Earth Syst. Sci. 22, 4455–4472. https://doi.org/10.5194/hess-22-4455-2018 Burke, J.M., Prepas, E.E., Pinder, S., 2005. Runoff and phosphorus export patterns in large forested watersheds on the western Canadian Boreal Plain before and for 4 years after wildfire. J. Environ. Eng. Sci. 4, 319–325. https://doi.org/10.1139/s04-072 Burton, C. a., Hoefen, T.M., Plumlee, G.S., Baumberger, K.L., Backlin, A.R., Gallegos, E., Fisher, R.N., 2016. Trace Elements in Stormflow, Ash, and Burned Soil following the 2009 Station Fire in Southern California. PLoS One 11, e0153372. https://doi.org/10.1371/journal.pone.0153372 Chanasyk, D.S., Whitson, I.R., Mapfumo, E., Burke, J.M., Prepas, E.E., 2003. The impacts of forest harvest and wildfire on soils and hydrology in temperate forests: A baseline to develop hypotheses for the Boreal Plain. J. Environ. Eng. Sci. 2, S51–S62. https://doi.org/10.1139/s03-034 Cooke, C.A., Schwindt, C., Davies, M., Donahue, W.F., Azim, E., 2016. Initial environmental impacts of the Obed Mountain coal mine process water spill into the Athabasca River (Alberta, Canada). Sci. Total Environ. 557–558, 502–509. https://doi.org/10.1016/j.scitotenv.2016.03.058 Dahm, C.N., Candelaria-Ley, R.I., Reale, C.S., Reale, J.K., Van Horn, D.J., 2015. Extreme water 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 quality degradation following a catastrophic forest fire. Freshw. Biol. https://doi.org/10.1111/fwb.12548 Depante, M., Petrone, R.M., Devito, K.J., Kettridge, N., Macrae, M.L., Mendoza, C., Waddington, J.M., 2018. Potential influence of nutrient availability along a hillslope: Peatland gradient on aspen recovery following fire. Ecohydrology 11, 1–13. https://doi.org/10.1002/eco.1955 Devito, K., Mendoza, C., Qualizza, C., 2012. Conceptualizing water movement in the Boreal Plains. Implications for watershed reconstruction. Synthesis report preapred for the Canadian Oil Sands Network for Research and Developement 164. Devito, K.J., Hokanson, K.J., Moore, P.A., Kettridge, N., Anderson, A.E., Chasmer, L., Hopkinson, C., Lukenbach, M.C., Mendoza, C.A., Morissette, J., Peters, D.L., Petrone, R.M., Silins, U., Smerdon, B., Waddington, J.M., 2017. Landscape controls on long-term runoff in subhumid heterogeneous Boreal Plains catchments. Hydrol. Process. 31, 2737– 2751. https://doi.org/10.1002/hyp.11213 Diffenbaugh, N.S., Swain, D.L., Touma, D., 2015. Anthropogenic warming has increased drought risk in California. Proc. Natl. Acad. Sci. 112, 3931–3936. https://doi.org/10.1073/pnas.1422385112 Earl, S.R., Blinn, D.W., 2003. Effects of wildfire ash on water chemistry and biota. Freshw. Biol. 1015–1030. Ebel, B.A., Mirus, B.B., 2014. Disturbance hydrology: challenges and opportunities. Hydrol. Process. 28, 5140–5148. https://doi.org/10.1002/hyp.10256 Elhadi, N., Harrington, A., Hill, I., Lau, Y.L., Krishnappan, B.G., 1984. River mixing—A stateof-the-art report. Can. J. Civ. Eng. 11, 585–609. https://doi.org/10.1139/l84-076 Emelko, M.B., Silins, U., Bladon, K.D., Stone, M., 2011. Implications of land disturbance on drinking water treatability in a changing climate: Demonstrating the need for “ source water supply and protection” strategies. Water Res. 45, 461–472. https://doi.org/10.1016/j.watres.2010.08.051 Emelko, M.B., Stone, M., Silins, U., Allin, D., Collins, A.L., Williams, C.H.S., Martens, A.M., Bladon, K.D., 2016. Sediment-phosphorus dynamics can shift aquatic ecology and cause downstream legacy effects after wildfire in large river systems. Glob. Chang. Biol. 22, 1168–1184. https://doi.org/10.1111/gcb.13073 Environment and Climate Change Canada, 2018. Canadian Climate Normals 1981-2010 [WWW Document]. URL http://climate.weather.gc.ca/climate_normals/ (accessed 2.14.18). Gerla, P.J., Galloway, J.M., 1998. Water quality of two streams near Yellowstone Park, Wyoming, following the 1988 Clover-Mist wildfire. Environ. Geol. 36, 127–136. https://doi.org/10.1007/s002540050328 Gibson, J.J., Yi, Y., Birks, S.J., 2016. Isotope-based partitioning of streamflow in the oil sands region, northern Alberta: Towards a monitoring strategy for assessing flow sources and water quality controls. J. Hydrol. Reg. Stud. 5, 131–148. https://doi.org/10.1016/j.ejrh.2015.12.062 Glozier, N.E., Pippy, K., Levesque, L., Ritcey, A., Armstrong, B., Tobin, O., Cooke, C.A., Conly, M., Dirk, L., Epp, C., Gue, A., Hazewinkel, R., Keet, E., Lindeman, D., Maines, J., Syrgiannis, J., Su, M., Tumber, V., 2018. Surface Water Quality of the Athabasca, Peace and Slave Rivers and Riverine Waterbodies within the Peace-Athabasca Delta. Oil Sands Monit. Progr. Tech. Rep. Ser. No. 1.4. 64pp. Government of Canada, 2019. Canadian Environmental Quality Guidelines Summary Table 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 [WWW Document]. URL http://st-ts.ccme.ca/ (accessed 9.20.18). Gresswell, R.E., 1999. Fire and Aquatic Ecosystems in Forested Biomes of North America. Trans. Am. Fish. Soc. 128, 193–221. https://doi.org/10.1577/15488659(1999)128<0193:FAAEIF>2.0.CO;2 Hanes, C.C., Wang, X., Jain, P., Parisien, M.-A., Little, J.M., Flannigan, M.D., 2018. Fireregime changes in Canada over the last half century. Can. J. For. Res. 49, 256–269. https://doi.org/10.1139/cjfr-2018-0293 Herrero, H.S., Díaz Lozada, J.M., García, C.M., Szupiany, R.N., Best, J., Pagot, M., 2018. The influence of tributary flow density differences on the hydrodynamic behavior of a confluent meander bend and implications for flow mixing. Geomorphology 304, 99–112. https://doi.org/10.1016/j.geomorph.2017.12.025 Hohner, A.K., Rhoades, C.C., Wilkerson, P., Rosario-Ortiz, F.L., 2019. Wildfires alter forest watersheds and threaten drinking water quality. Acc. Chem. Res. 52, 1234–1244. https://doi.org/10.1021/acs.accounts.8b00670 Inbar, M., Tamir, M., Wittenberg, L., 1998. Runoff and erosion processes after a forest fire in Mount Carmel, a Mediterranean area. Geomorphology 24, 17–33. https://doi.org/10.1016/S0169-555X(97)00098-6 Ireson, A.M., Barr, A.G., Johnstone, J.F., Mamet, S.D., van der Kamp, G., Whitfield, C.J., Michel, N.L., North, R.L., Westbrook, C.J., DeBeer, C., Chun, K.P., Nazemi, A., Sagin, J., 2015. The changing water cycle: the Boreal Plains ecozone of Western Canada. Wiley Interdiscip. Rev. Water 2, 505–521. https://doi.org/10.1002/wat2.1098 Jolly, W.M., Cochrane, M. a., Freeborn, P.H., Holden, Z. a., Brown, T.J., Williamson, G.J., Bowman, D.M.J.S., 2015. Climate-induced variations in global wildfire danger from 1979 to 2013. Nat. Commun. 6, 1–11. https://doi.org/10.1038/ncomms8537 Kelly, E.N., Schindler, D.W., St. Louis, V.L., Donald, D.B., Vladicka, K.E., 2006. Forest fire increases mercury accumulation by fishes via food web restructuring and increased mercury inputs. Proc. Natl. Acad. Sci. 103, 19380–19385. https://doi.org/10.1073/pnas.0609798104 Lane, P.N.J., Sheridan, G.J., Noske, P.J., Sherwin, C.B., 2008. Phosphorus and nitrogen exports from SE Australian forests following wildfire. J. Hydrol. 361, 186–198. https://doi.org/10.1016/j.jhydrol.2008.07.041 Langhans, C., Smith, H.G., Chong, D.M.O., Nyman, P., Lane, P.N.J., Sheridan, G.J., 2016. A model for assessing water quality risk in catchments prone to wildfire. J. Hydrol. 534, 407– 426. https://doi.org/10.1016/j.jhydrol.2015.12.048 Larsen, I.J., MacDonald, L.H., Brown, E., Rough, D., Welsh, M.J., Pietraszek, J.H., Libohova, Z., de Dios Benavides-Solorio, J., Schaffrath, K., 2009. Causes of Post-Fire Runoff and Erosion: Water Repellency, Cover, or Soil Sealing? Soil Sci. Soc. Am. J. 73, 1393. https://doi.org/10.2136/sssaj2007.0432 Mast, M.A., Murphy, S.F., Clow, D.W., Penn, C.A., Sexstone, G.A., 2016. Water-quality response to a high-elevation wildfire in the Colorado Front Range. Hydrol. Process. 30, 1811–1823. https://doi.org/10.1002/hyp.10755 Mell, W.E., Manzello, S.L., Maranghides, A., Butry, D., Rehm, R.G., 2010. The wildland urban interface fire problem - current approaches and research needs. Int. J. Wildl. Fire 19, 238. https://doi.org/10.1071/wf07131 Minshall, G.W., Brock, J.T., Andrews, D.A., Robinson, C.T., 2001. Water quality, substratum and biotic responses of five central Idaho (USA) streams during the first year following the Mortar Creek fire. Int. J. Wildl. Fire 10, 185–199. https://doi.org/10.1071/WF01017 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 Moody, J.A., Martin, D.A., 2001. Initial hydrologic and geomorphic response following a wildfire in the Colorado front range. Earth Surf. Process. Landforms 26, 1049–1070. https://doi.org/10.1002/esp.253 Neary, D.G., Ryan, K.C., Debano, L.F., 2005. Wildland Fire in Ecosystems Effects of Fire on Soil and Water, United States Department of Agriculture Gen. Tech. Rep. Ogden, Utah. https://doi.org/10.1021/ac102195x Olefeldt, D., Turetsky, M.R., Blodau, C., 2013. Altered Composition and Microbial versus UVMediated Degradation of Dissolved Organic Matter in Boreal Soils Following Wildfire. Ecosystems 16, 1396–1412. https://doi.org/10.1007/s10021-013-9691-y Oliver, A.A., Reuter, J.E., Heyvaert, A.C., Dahlgren, R.A., 2012. Water quality response to the Angora Fire, Lake Tahoe, California. Biogeochemistry. https://doi.org/10.1007/s10533-0119657-0 Owens, P.N., Giles, T.R., Petticrew, E.L., Leggat, M.S., Moore, R.D., Eaton, B.C., 2013. Muted responses of streamflow and suspended sediment flux in a wildfire-affected watershed. Geomorphology 202, 128–139. https://doi.org/10.1016/j.geomorph.2013.01.001 Prepas, E.E., Burke, J.M., Chanasyk, D.S., Smith, D.W., Putz, G., Gabos, S., Chen, W., Millions, D., Serediak, M., 2003. Impact of wildfire on discharge and phosphorus export from the Sakwatamau watershed in the Swan Hills, Alberta, during the first two years. J. Environ. Eng. Sci. https://doi.org/10.1139/s03-036 Radeloff, V.C., Helmers, D.P., Kramer, H.A., Mockrin, M.H., Alexandre, P.M., Bar-Massada, A., Butsic, V., Hawbaker, T.J., Martinuzzi, S., Syphard, A.D., Stewart, S.I., 2018. Rapid growth of the US wildland-urban interface raises wildfire risk. Proc. Natl. Acad. Sci. 115, 3314–3319. https://doi.org/10.1073/pnas.1718850115 Reale, J.K., Van Horn, D.J., Condon, K.E., Dahm, C.N., 2015. The effects of catastrophic wildfire on water quality along a river continuum. Freshw. Sci. 34, 1426–1442. https://doi.org/10.1086/684001 Rhoades, C.C., Entwistle, D., Butler, D., 2011. The influence of wildfire extent and severity on streamwater chemistry, sediment and temperature following the Hayman Fire, Colorado. Int. J. Wildl. Fire 20, 430. https://doi.org/10.1071/wf09086 Rust, A.J., Hogue, T.S., Saxe, S., McCray, J., 2018. Post-fire water-quality response in the western United States. Int. J. Wildl. Fire 27, 203–216. https://doi.org/10.1071/wf17115 Santín, C., Doerr, S.H., Otero, X.L., Chafer, C.J., 2015. Quantity, composition and water contamination potential of ash produced under different wildfire severities. Environ. Res. 142, 297–308. https://doi.org/10.1016/j.envres.2015.06.041 Schlesinger, W., 2007. Biogeochemistry: an analysis of global change, 2nd Ed. ed. Academic Press, London. Schoennagel, T., Balch, J.K., Brenkert-Smith, H., Dennison, P.E., Harvey, B.J., Krawchuk, M. a., Mietkiewicz, N., Morgan, P., Moritz, M. a., Rasker, R., Turner, M.G., Whitlock, C., 2017. Adapt to more wildfire in western North American forests as climate changes. Proc. Natl. Acad. Sci. 114, 4582–4590. https://doi.org/10.1073/pnas.1617464114 Scott, D., Versfeld, D.B., Lesch, W., 1998. Erosion and sediment yield in relation to afforestation and fire in the mountains of the western cape province, south africa. South African Geogr. J. 80, 52–59. https://doi.org/10.1080/03736245.1998.9713644 Shakesby, R.A., Doerr, S.H., 2006. Wildfire as a hydrological and geomorphological agent. Earth-Science Rev. 74, 269–307. https://doi.org/10.1016/j.earscirev.2005.10.006 Silins, U., Bladon, K.D., Kelly, E.N., Esch, E., Spence, J.R., Stone, M., Emelko, M.B., Boon, S., 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 Wagner, M.J., Williams, C.H.S., Tichkowsky, I., 2014. Five-year legacy of wildfire and salvage logging impacts on nutrient runoff and aquatic plant, invertebrate, and fish productivity. Ecohydrology 7, 1508–1523. https://doi.org/10.1002/eco.1474 Smith, H.G., Sheridan, G.J., Lane, P.N.J., Nyman, P., Haydon, S., 2011. Wildfire effects on water quality in forest catchments: A review with implications for water supply. J. Hydrol. 396, 170–192. https://doi.org/10.1016/j.jhydrol.2010.10.043 Stone, M., Emelko, M.B., Droppo, I.G., Silins, U., 2011. Biostabilization and erodibility of cohesive sediment deposits in wildfire-affected streams. Water Res. 45, 521–534. https://doi.org/10.1016/j.watres.2010.09.016 Temnerud, J., Bishop, K., 2005. Spatial variation of streamwater chemistry in two Swedish boreal catchments: Implications for environmental assessment. Environ. Sci. Technol. 39, 1463–1469. https://doi.org/10.1021/es040045q Thurton, D., 2017. Fort McMurray seeing big spike in water treatment costs. Can. Broadcast. Corp. https//www.cbc.ca/news/canada/edmonton/fort-mcmurray-wildfire-water-treatmentcosts-contaminants-1.3973249. Townsend, S.A., Douglas, M.M., 2000. The effect of three fire regimes on stream water quality, water yield and export coefficients in a tropical savanna (northern Australia). J. Hydrol. https://doi.org/10.1016/S0022-1694(00)00165-7 Westerling, A.L., Hidalgo, H.G., Cayan, D.R., Swetnam, T.W., 2006. Warming and earlier spring increase Western U.S. forest wildfire activity. Science (80-. ). https://doi.org/10.1126/science.1128834 Wilkinson, S.L., Moore, P.A., Flannigan, M.D., Wotton, B.M., Waddington, J.M., 2018. Did enhanced afforestation cause high severity peat burn in the Fort McMurray Horse River wildfire? Environ. Res. Lett. 13, 014018. Writer, J.H., Hohner, A., Oropeza, J., Schmidt, A., Cawley, K., Rosario-Ortiz, F.L., 2014. Water treatment implications after the High Park Wildfire in Colorado. J. Am. Water Works Assoc. 106, 85–86. https://doi.org/10.5942/jawwa.2014.106.0055. 663 664 665 8. Tables Table 1 Drainage area and burn metrics at water quality stations operating on rivers draining burned (B) and unburned (U) areas following the 2016 Fort McMurray wildfire. 666 9. Figures 667 668 669 670 Figure 1 Map of the lower Athabasca River in northeastern Alberta showing the extent and severity of the May 2016 Fort McMurray wildfire (see supplementary material for methods) and water monitoring stations. Inset (a.) shows monitoring station clustering around Fort McMurray. 671 672 673 674 Figure 2 Log-scaled historic flow quantiles and 2016 mean daily hydrographs and hyetographs of monitored hydrometric stations (Q) and their closest meteorological stations (P) of unburned (top panel) and burned watersheds in the lower Athabasca River of northeastern Alberta. Daily precipitation events ≥10 mm and five subsequent days are highlighted in yellow. 675 676 677 678 Figure 3 Hourly time series of sonde turbidity and specific conductivity, and discharge in unburned (top panel) and burned river stations in the lower Athabasca River region following the 2016 Fort McMurray wildfire. Precipitation events (daily ≥10 mm + 5 days) are highlighted in yellow. 679 680 681 682 683 Figure 4 Schematic of sonde turbidity and specific conductivity covariance, mean daily flow and precipitation events (≥10 mm and five subsequent days) measured at an unburned control station and five downstream stations within burned landscapes. All graphs share the same axis metrics and time periods as the control station graph. LB and RB denote left bank and right bank of the river. *denote four, 15-minute measurements beyond the scale on the covariance axes. 684 685 686 687 Figure 5 Box and whisker plots (median-mean [dashed]-25/75-10/90 percentiles-outliers) of Athabasca River watershed yields of selected chemicals from an upstream unburned station and a downstream burned station from daily sampling (May–August) following the 2016 Fort McMurray wildfire. 688 689 690 691 Figure 6 Time series of mean daily concentrations of selected chemicals collected by automated and grab samples for unburned (top panel) and burned river stations in the lower Athabasca River region following the 2016 Fort McMurray wildfire. Precipitation events (daily ≥ 10 mm + 5 days) are highlighted in yellow. 692 693 694 695 Figure 7 Log-scaled concentration-discharge relationships of suspended sediment and total calcium (Ca) of historic and post-fire data from three rivers draining burned watersheds in the lower Athabasca River region following the 2016 Fort McMurray wildfire. 95% prediction intervals and linear fit lines on historic data are also shown. 696 697 698 Figure 8 Photos of high-order (left panel; Horse River) and low-order (right panel; Little Fisheries Creek) rivers draining recently burned watersheds and discharging into the Athabasca River just upstream of Fort McMurray after a June 2016 precipitation event. (Credit: S. Hustins). Station Athabasca R. at Grand Rapids (U) Athabasca R. u/s Ft. McMurray (B) Clearwater R. near the mouth (B) Hangingstone R. at Ft. McMurray (B) Watershed area (km2) Mean watershed slope (%) Burned area: watershed area (%) High severity burned area: burned area (%) 94,464 98,013 31,936 903 3.3 3.3 3.4 3.5 0 0.4 10.2 21.5 45.6 52.1 57.1 Highlights • • • • Large rivers show wildfire water quality signatures after precipitation events Continuous monitoring captures short-term wildfire impacts on river water quality Suspended sediment concentrations after fire were higher than historic records Incomplete mixing within large rivers impacts detection of altered water quality Declaration of interests ☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. ☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: