Effects of projected climate on the hydrodynamic and sediment transport regime of the lower Athabasca River in Alberta , Canada

Watershed Hydrology and Ecology Research Division, Water and Climate Impacts Research Centre (W‐CIRC), Environment and Climate Change Canada, Victoria, BC, Canada Watershed Hydrology and Ecology Research Division, Canada Centre for Inland Waters, Environment and Climate Change Canada, Burlington, ON, Canada Correspondence Yonas B. Dibike, Environment and Climate Change Canada, Watershed Hydrology and Ecology Research Division, Water and Climate Impacts Research Centre (W‐CIRC), Victoria, BC, Canada. Email: yonas.dibike@canada.ca

flow as well as sediment and constituent chemical transport in river systems will most likely be affected by a changing climate. Some of the effects could be through changes in the magnitude and timing of seasonal mean as well as extreme river discharges, sediment inflow, flow velocity, and depth, which would in turn affect the available environmental flow, shear stress, erodibility, and transport capacity of rivers (Thodsen, Hasholt, & Kjaersgaard, 2008).
There is a growing body of evidence that climate change is having a significant impact on the sediment loads and transport in rivers. Comparing data collected in the 1970s with those from the 1990s, Amsler and Drago (2009) showed that recent increases in precipitation and run-off across parts of the Parana-Paraguay system in South America have caused increased erosion and sediment mobilization and indicated that climate change has been strongly affecting the hydro-sedimentological regime of the river network.
Using a GIS-based model under a climate-change scenario, Asselman, Middelkoop, and Van Dijk (2003) investigated the potential effects of changes in climate and land use on the mobilization of fine sediment and the net transport of wash load from the upstream basin to the lower Rhine delta. Their research indicated that erosion rates will increase in the Alps and decrease in the German part of the basin as a result of the changing climate and land use. Modelling climate induced changes in suspended sediment transport for two Danish river catchments, Thodsen et al. (2008) also found that suspended sediment transport increases during winter months as a result of the increase in river discharge caused by enhanced precipitation, and decreases during summer and early autumn months when precipitation also decreases. Similarly, Praskievicz (2016) investigated the potential impacts of climate change on streamflow and suspended-sediment transport for snowmelt-dominated rivers in the interior Pacific Northwest and indicate that climate change is likely to amplify the annual cycle of river discharge and simulated changes in suspended-sediment transport that generally follow the changes in streamflow.
The current and projected future states of flow in the Athabasca watershed has been actively investigated by a number of recent studies. Some of the studies that were based on analysis of observed streamflow data in the region have shown statistically significant decreasing trends in streamflow, particularly in recent decades (Bawden, Linton, Burn, & Prowse, 2014;Sauchyn, Luckman, & St-Jacques, 2015). However, using a correlation model between river flow and climate variables to reconstruct long-term (>100 years) natural modes of river discharge, Chen and Grasby (2014) did not find true long-term declines of the annual flow in the Athabasca River basin (ARB). The findings of Rood, Stupple, and Gill (2015) from centurylong records also contrast with interpretations from the above shortterm studies and emphasize the need for sufficiently long time series for hydrologic trend analysis. Peters, Atkinson, Monk, Tenenbaum, and Baird (2013)  With respect to sediment transport, Conly, Crosley, and Headley (2002) determined the contribution of the upstream boundary and tributaries in the annual load of sediments in the lower Athabasca River (LAR) and found that suspended sediment derived from main stem and tributary sources between Fort McMurray and Embarras account for 18% of the mean annual load of the Athabasca River with the remaining originating upstream of Fort McMurray. A recent study by Shrestha and Wang (2018) used the Soil and Water Assessment Tool with future climate projections over the ARB and show a potential increase in soil erosion rate due to climate change is greater than reported soil formation rates in the region. Studies that have attempted numerical modelling of flow and sediment transport in the LAR found it to be challenging due to the complex morphology and seasonality of the flow regime. Andrishak, Abarca, Wojtowicz, and Hicks (2008) and Pietroniro et al. (2011) made early attempts to numerically model the flow in LAR using one-dimensional (1D) models that incorporated simplified rectangular sections to represent channel geometry. More recently, Shakibaeinia et al. (2016; and Kashyap, Dibike, Shakibaeinia, Prowse, and Droppo (2016) developed an integrated numerical modelling framework (1D and two-dimensional) for simulation of flow and sediment transport covering larger portions of the LAR using detailed surveyed bathymetric data.
Although there are a number of studies that have investigated the potential impacts of climate change on the hydrologic (discharge) regimes of the LAR, none have examined the implications of the altered flow regimes on the hydrodynamic and sediment transport characteristics of the river. Therefore, this study investigates the potential impacts of future climate on hydrodynamic and sediment transport regime of the LAR by employing the MIKE-11 1D flow and sediment transport model. While the hydrodynamic and sediment transport model used for this study was calibrated/validated using historical observed discharge and sediment inflow data (Shakibaeinia, Dibike, Kashyap, Prowse, & Droppo, 2017), the corresponding future scenario data are derived from a recent study by Eum et al. (2017). Eum et al. (2017) and Dibike, Eum, and Prowse (2018) investigated the potential hydrologic response of the ARB to projected changes in future climate using the Variable Infiltration Capacity (VIC) process-based and distributed hydrologic model (Liang, Lettenmaier, Wood, & Burges, 1994). The climatic forcings for the VIC hydrologic model were derived from a selected set of GCMs from the latest Coupled Model Intercomparison Project (CMIP5) and statistically downscaled to a higher (10 km) spatial resolution. A subset of the VIC simulated river discharge scenario data corresponding to the baseline period of 1970-1999 (1980s), and the two future periods of 2040-2069 (2050s) and 2070-2099 (2080s)

| Site description
The Athabasca River, with a 156,000 km 2 drainage area, originates from the Columbia glacier in Jasper National Park and flows approximately 1,500 km north-eastward to Peace Athabasca Delta (PAD) and Lake Athabasca. The hydrodynamic and sediment transport scenario simulation is conducted over the~200 km reach of the LAR starting from below the city of Fort McMurray and extending to Old Fort which is located few kilometres upstream of the river discharging into the PAD (Figure 1). This river reach is characterized as meandering and braided with vegetated islands and alternating sand bars as the river and many of its tributaries cuts through the McMurray formations where bitumen can be found close to the earth surface. Major tributaries within the LAR reach include the Steepbank, Ells, MacKay, Muskeg, and Firebag Rivers. Mean daily temperatures in the LAR range between approximately −20°C in January and around 15°C in July while the mean annual precipitation in the region is <500 mm with over 60% occurring as rainfall and the remainder as snowfall (Conly et al., 2002). The mean annual streamflow at the

| River bathymetry data
The river bathymetry data for the LAR are obtained by combining different legacy data sets with a high-resolution (0.5 m) bed elevation data between Fort McMurray and Old Fort surveyed by Environment Canada using a Geoswath sonar sensor Shakibaeinia et al., 2017). The topography of floodplains and islands are reproduce using high resolution (5 m) light detection and ranging (LiDAR) data along the LAR banks from Alberta Environment and Parks that was further processed into Digital Elevation Model (DEM) by Environment and Climate Change Canada, and Digital Elevation Model data of the region from Geobase (2012). The data from all these sources were combined to construct a continuous bathymetry for the LAR main channel and adjacent flood plains with a resolution ranging from 10 to 25 m. The data were then interpolated on the 1D cross-sections along the LAR (200 cross-sections with~1 km intervals) to construct the required model geometry.

| Historical hydrometric and sediment data
The historical hydrometric (flow rates and water levels) and sediment data used as boundary conditions as well as for the purpose of model   selected based on the ranking of the CMIP5 models, which differs by region, that is carried out by the Pacific Climate Impact Consortium to provide the widest spread (range) in projected future climate for smaller subsets of the full ensemble (Cannon, 2015). CNRM represents the closest scenario to CMIP5 multimodel ensemble mean whereas CanESM is the farthest from the first selected GCM (i.e., CNRM) corresponding to higher projected increases in precipitation and temperature. Moreover, as the GCMs data are at coarser resolution (200-300 km) and as they are also known to have seasonal biases compared with the observed climate for the historical period, a widely  (Wood, Leung, Sridhar, & Lettenmaier, 2004), was applied to spatially downscale the GCMs' outputs based on the 10-km resolution Australian National University thin plate spline (ANUSPLIN) algorithm based gridded observed data (Hopkinson et al., 2011). Also, only two of the emission scenarios, namely, the RCP4.5, which is a stabilization scenario that achieves the goal of limiting emission and radiative forcings, and the RCP8.5, which is an emission scenario that greenhouse gas increases as usual until 2100, were considered for the hydrologic simulation (Eum et al., 2017). Therefore, a total of four sets of the ARB VIC hydrologic model projections corresponding to two GCMs (CNRM and CanESM), and two emission scenarios (RCP4.5 and RCP8.5) are employed in this study.
where t is the time, x is the streamwise distance, h(x, t) is the water height, is the friction slope, here given by Manning equation as: where n is Manning's coefficient and R is the hydraulic radius calculated

| Sediment transport
To model the transport of fine sediments, the Advection-Dispersion (AD) and Cohesive Sediment Transport modules of MIKE-11 are used.
The AD module is based on the 1D conservation of mass (of dissolved or suspended materials). The Cohesive Sediment Transport module is coupled with AD module and is used to describe transport of suspended fine sediments. The erosion/deposition is considered as a sink/source term of the AD equations. The areal averaged 1D AD equation used in MIKE-11 is given by: in which C is the concentration, D is dispersion coefficient, K is linear decay coefficient, S is source/sink concentration, and q is lateral inflow. The two primary source/sink terms are sediment deposition (S d ) and erosion (S e ). When the bed shear stress, τ b, is less than the critical shear stress for deposition, τ cd, the particles and flocs in suspension begin to deposit onto the bed. By contrast, the river bed begins to erode when the bed shear stress, τ b, exceeds the critical shear stress for erosion, τ ce . The deposition and erosion rates S d , S e are described by the Van Rijn equations (1984): where W s is the sediments settling velocity, and E 0 and n are the Jan-01 Jan-02 Jan-03 Jan-04 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Jan-10 Jan-11 Jan-00 Jan-01 Jan-02 Jan-03 Jan-04 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Jan-10 Jan    future periods (2050s and 2080s) are presented in Figure 6 for a loca-

| Projected changes in exceedance probability
The scenario simulation results are also presented in terms of probability of exceedance describing the likelihood of the daily mean value of a specified flow variable being exceeded in a given time period. rating curves, and model parameters. However, the hydrodynamic and sediment transport modelling approach applied in this study using hydrologic projection corresponding to the two climate models (CNRM and CanEMS) and the two emissions scenarios (RCP4.5 and RCP8.5) depict the general direction of potential changes in the flow and sediment transport regime of the LAR.
In general, climate change is projected to cause increasing precipitation and temperature in the Athabasca watershed that will, in turn, alter the hydrologic regime in the LAR system. Through hydrodynamic and sediment transport simulation in the LAR, this study found that, by the end of this century, there will be a corresponding potential increase in flow velocity and water level leading to an overall increase in sediment load and transport in the LAR and to the PAD compared with the contemporary baseline period. Implications of such potential changes in the transport characteristics of the river system to the mobilization and transport of various chemical constituents and their effects on the region's aquatic ecosystems are subjects of other ongoing investigations.
Gupta of Alberta Environment and Parks (AEP) for their involvement at the different stage of this research project. The authors acknowledge Dr. Fay Hicks for providing some cross section data, Tom Carter and Jennifer Pesklevits for collecting and processing the GeoSwath data for the lower Athabasca River, and Boyang Jiao in data preparation and model simulation. The authors also acknowledge Dr. Roderick Hazewinkel from AEP for facilitating access to the LiDAR data used for preparing the bathymetry for the Athabasca River and its flood plains.