Mine reclamation in the Athabasca oil sands region Canada, is required by law where companies must reconstruct disturbed landscapes into functioning ecosystems such as forests, wetlands and lakes that existed in the Boreal landscape prior to mining. Winter is a major hydrological factor in this region as snow covers the landscape for 5 to 6 months and is ~25% of the annual precipitation, yet few studies have explored the influence of winter processes on the hydrology of constructed watersheds. One year (2017-2018) of intensive snow hydrology measurements are supplemented with six years (2013-2018) of meteorological measurements from the constructed Sandhill Fen Watershed to: 1) understand snow accumulation and redistribution, snowmelt timing, rate and partitioning, 2) apply a physically-based model for simulating winter processes on hillslopes and 3) evaluate the impact of soil prescriptions and climate change projections on winter processes in reclaimed systems. The 2017-2018 snow season was between November and April and SWE ranged between 40-140 mm. Snow distribution was primarily influenced by topography with little influence of snow trapping from developing vegetation. Snow accumulation was most variable on hillslopes and redistribution was driven by slope position, with SWE greatest at the base of slopes and decreased towards crests. Snowmelt on hillslopes was controlled by slope aspect, as snow declined rapidly on west and south-facing slopes, compared to east and north-facing slopes. Unlike results previously reported on constructed uplands, snowmelt runoff from uplands was much less (~30%), highlighting the influence of different construction materials. Model simulations indicate that antecedent soil moisture and soil temperature have a large influence on partitioning snowmelt over a range of observed conditions. Under a warmer and wetter climate, average annual peak SWE and snow season duration could decline up to 52 % and up to 61 days, respectively while snowmelt runoff ceases completely under the warmest scenarios. Results suggest considerable future variability in snowmelt runoff from hillslopes, yet soil properties can be used to enhance vertical or lateral flows.

Abstract Melting seasonal ground ice (SGI) in western Boreal Plains (WBP) peatlands can reduce the available energy at the surface by reducing potential evapotranspiration (PET). PET often exceeds annual precipitation in the WBP. Including this effect in hydrological models may be important in assessing water deficits. However, SGI melt and the timing of ice‐free conditions vary spatially, which suggests PET spatial variability could be influenced by SGI. Understanding this potential linkage can help improve site scale PET in peatland hydrological models. The objectives of this paper were (a) to quantify the effect of ice thickness and melt rate on peatland PET; (b) quantify the spatial variability of SGI thickness and melt rate across spatial scales; and (c) assess how/if spatial variability in SGI thickness/melt rate affects site scale PET. Results from the sensitivity analysis indicated that SGI thickness had a bigger impact on reducing PET compared with the melt rate. Two SGI thickness values were used that were observed on site: 0.32 m, which was measured in a more treed area, and 0.18 m, which was in a more open area. The 0.32 m had an average PET reduction of 14 mm (±0.7), over the month of May, compared with 9 mm (±1 mm) when there was 0.18 m of SGI, which are 13.7 and 8.8% reductions, respectively. SGI thickness and melt rate, both exhibited large‐ and small‐scale spatial variability. At the large scale, spatial patterns in SGI thickness appeared to be influenced by extensive shading from the adjacent hillslopes. Small scale, SGI thickness may be a function of tree proximity and the snowpack. Finally, net radiation, rather than SGI, appeared to be the main driver behind PET spatial variability. This work enhances our conceptual understanding of the role of SGI in WBP peatlands. Future work can use the findings to better inform peatland hydrological models, allowing for better representation of peatlands in regional‐scale models.