Summary and synthesis of Changing Cold Regions Network (CCRN) 1 research in the interior of western Canada – Part 2: Future change in 2 cryosphere, vegetation, and hydrology 3

The interior of western Canada, like many similar cold mid- to high-latitude regions worldwide, is 38 undergoing extensive and rapid climate and environmental change, which may accelerate in the coming 39 decades. Understanding and predicting changes in coupled climate–land–hydrological systems are crucial 40 to society, yet limited by lack of understanding of changes in cold region process responses and 41 interactions, along with their representation in most current generation land surface and hydrological 42 models. It is essential to consider the underlying processes and base predictive models on the proper 43 physics, especially under conditions of non-stationarity where the past is no longer a reliable guide to the 44 future and system trajectories can be unexpected. These challenges were forefront in the recently 45 completed Changing Cold Regions Network (CCRN), which assembled and focused a wide range of multi- 46 disciplinary expertise to improve the understanding, diagnosis, and prediction of change over the cold 47 interior of western Canada. CCRN advanced knowledge of fundamental cold region ecological and 48 hydrological processes through observation and experimentation across a network of highly instrumented research basins and other sites. Significant efforts were made to improve the functionality and process 1 representation, based on this improved understanding, within the fine-scale Cold Regions Hydrological 2 Modelling (CRHM) platform and the large-scale Modélisation Environmentale Communautaire (MEC) – 3 Surface and Hydrology (MESH) model. These models were, and continue to be, applied under past and 4 projected future climates, and under current and expected future land and vegetation cover 5 configurations to diagnose historical change and predict possible future hydrological responses. This 6 second of two articles synthesizes the nature and understanding of cold region processes and Earth 7 system responses to future climate, as advanced by CCRN. These include changing precipitation and 8 moisture feedbacks to the atmosphere; altered snow regimes, changing balance of snowfall and rainfall, 9 and glacier loss; vegetation responses to climate and the loss of ecosystem resilience to wildfire and 10 disturbance; thawing permafrost and its influence on landscapes and hydrology; groundwater storage and 11 cycling, and its connections to surface water; and stream and river discharge as influenced by the various 12 drivers of hydrological change. Collective insights, expert elicitation, and model application are used to 13 provide a synthesis of this change over the CCRN region for the late-21 st century. and may lead to more basal ice formation, producing complex runoff responses in spring. Follow-on hillslope-scale analysis by Coles and 47

research basins and other sites. Significant efforts were made to improve the functionality and process 1 representation, based on this improved understanding, within the fine-scale Cold Regions Hydrological 2 Modelling (CRHM) platform and the large-scale Modélisation Environmentale Communautaire (MEC) -3 Surface and Hydrology (MESH) model. These models were, and continue to be, applied under past and 4 projected future climates, and under current and expected future land and vegetation cover 5 configurations to diagnose historical change and predict possible future hydrological responses. This 6 second of two articles synthesizes the nature and understanding of cold region processes and Earth 7 system responses to future climate, as advanced by CCRN. These include changing precipitation and 8 moisture feedbacks to the atmosphere; altered snow regimes, changing balance of snowfall and rainfall, 9 and glacier loss; vegetation responses to climate and the loss of ecosystem resilience to wildfire and 10 disturbance; thawing permafrost and its influence on landscapes and hydrology; groundwater storage and 11 cycling, and its connections to surface water; and stream and river discharge as influenced by the various 12 drivers of hydrological change. Collective insights, expert elicitation, and model application are used to 13 provide a synthesis of this change over the CCRN region for the late-21 st century. 14 1. Introduction and objective 15 16 The interior of western Canada is a region undergoing rapid, widespread, and severe hydro-climatic and 17 environmental change. This region is emblematic of the scientific and societal challenges in cold regions 18 around the world where snow, ice, and frozen soils dominate water cycling processes. Parts of western 19 and northern Canada  precipitation (P) toward more rain and less snow, earlier snowmelt and decreasing extent, duration, and 23 maximum depth of seasonal snow cover, retreating glaciers, warming and thawing permafrost, declining 24 freshwater ice cover period, and an earlier spring freshet. Against this backdrop of change, western 25 Canada has been subjected to a series of recent, and in some instances record-breaking, extreme events 26 such as floods, droughts, and wildfires. Human interventions and land and water management have also 27 affected the environment and river systems, with infrastructure developments such as dams, diversions, 28 and irrigation networks, along with industrialization, agricultural development, and urbanization, thereby 29 altering natural ecosystems and water cycling. Future projections of warmer climate, altered P phase and 30 patterns, and more extreme events (Bush and Lemmen, 2019; Stewart et al., 2019), together with 31 increasing human pressures, indicate that the region will continue to undergo rapid change to conditions 32 never before experienced, posing difficult management and decision-making challenges (e.g., Razavi et 33 al., 2020). 34 35 Improved understanding and prediction of the changes in coupled climate-land-hydrological systems are 36 crucial for managing land and water systems, and informing governance and policy direction here and in 37 other similar regions globally. The processes of change in cold regions are manifold and complex, and 38 there is significant uncertainty with the prediction of future change. Often, modelling and projections of 39 hydrological change are based on over-simplistic or empirical approaches and models that fail to 40 adequately capture the interconnected process drivers and responses. It is unclear to what extent the 41 model structures and parameterizations are valid under highly non-stationary conditions, and hence 42 whether the results are meaningful under future climates and land and vegetation cover states. There 43 has been much speculation about how cold regions will change, but, in many cases, this has not been 44 based on appropriate process understanding, which is itself limited. observational tools to understand, diagnose and predict interactions amongst the cryospheric, ecologic, 4 hydrologic, and climatic components of the changing Earth system at multiple scales. Its specific 5 geographic focus has been on the cold interior of western Canada, and in particular, the two major river 6 systems of the region -the Saskatchewan and Mackenzie River Basins (Fig. 1). The overall science 7 objectives of CCRN were to: 8 1. Document and evaluate observed Earth system change, including hydrological, ecological, 9 cryospheric and atmospheric components over a range of scales from local observatories to 10 biome and regional scales; 11 2. Improve understanding and diagnosis of local-scale change by developing new and integrative 12 knowledge of Earth system processes, incorporating these processes into a suite of process-based 13 integrative models, and using the models to better understand Earth system change; 14 3. Improve large-scale atmospheric and hydrological models for river basin-scale modelling and 15 prediction to better account for the changing Earth system and its atmospheric feedbacks; and 16 4. Analyze and predict regional and large-scale variability and change, focusing on the governing 17 factors for the observed trends and variability in large-scale aspects of the Earth system and their 18 representation in current models, and the projections of regional scale effects of Earth system 19 change on climate, land and water resources. 20 21 Key to the success of the network was the ability to observe and diagnose change across the region, and 22 hence provide a platform of data (e.g., see https://essd.copernicus.org/articles/special_issue901.html) 23 and scientific insights to inform model development and application for the analysis and prediction of 24 change. A multiscale observatory was developed, based where possible on existing experimental sites 25 with historical data records (Fig. 1), and this formed the heart of the program, enabling process responses 26 and interactions to be monitored across the different ecological regions, and at the scales of small river 27 basins and major river systems. In conjunction with the experimental and observational program, 28 modelling research aimed at improving the capability of fine and large-scale models to represent key cold 29 region processes, and to diagnose the complex and interacting factors underlying the observed changes 30 over the CCRN region. Finally, these models have begun to be used, in conjunction with expert elicitation, 31 to examine likely future system trajectories for the purposes of informing management and policy and 32 addressing other stakeholder concerns. In doing so, CCRN assembled and focused a wide range and depth 33 of multi-disciplinary expertise to address the network's aims and to develop insights into the process 34 controls across the CCRN domain. 35 36 This article draws together the expert understanding and process insights from CCRN, together with 37 modelling results at different scales, to examine the key drivers of change and to highlight the most likely 38 anticipated future system trajectories across the interior of western Canada. This follows Part 1 (Stewart 39 et al., 2019), which synthesized CCRN's collective assessments of future climate conditions and the 40 associated seasonal patterns, along with particular P-and temperature-related phenomena. The specific 41 objective of this second article is to illustrate how these changes in the climate system will manifest as 42 changes in land and vegetation cover, cryospheric states, and hydrological cycling. 43 44 The article is organized as follows: Section 2 provides a brief overview of CCRN's geographic domain and 45 the two major river basins. Section 3 examines a number of different cold region processes, their 46 interactions and responses to climate, and their influence on water cycling. This highlights complexities 47 that most Earth system models fail to capture. Section 4 briefly describes the advancements in fine-scale 48 https://doi.org/10.5194/hess-2020-491 Preprint. Discussion started: 17 October 2020 c Author(s) 2020. CC BY 4.0 License. and large-scale process-based hydrological models during CCRN, along with their application for the 1 diagnosis and prediction of change, while Section 5 provides a synthesis of this change over the CCRN 2 region for the 21 st century. Section 6 provides concluding remarks and guidance for further research. 3 2. Ecological regions and river systems of the interior of western Canada 4 5 The interior of western Canada spans a wide range of climatic, ecological, and physiographic regions ( Fig.  6 1), and has many of the physical attributes common to cold regions worldwide (Woo et al., 2008). This 7 includes extensive areas of permafrost and seasonally frozen ground, snow and ice cover through a large 8 part of the year, and water cycling that is driven largely by seasonal patterns of energy availability. The 9 principal river systems include the Saskatchewan and Mackenzie Rivers and their respective 406,000 km 2 10 and 1.8 million km 2 drainage basins (Fig. 1). These encompass Prairie, Boreal (including Taiga), Tundra, 11 and Cordillera landscapes (CEC, 1997). 12 13 The Saskatchewan River originates in the Rocky Mountains of Alberta and Montana, and flows through 14 the province of Saskatchewan and into Manitoba, discharging into Lake Winnipeg. Most of the flow 15 originates in the mountains, which provide roughly 80% of total discharge (Pomeroy et al., 2005). The 16 basin is mostly situated within the Prairies, a key agricultural region, and Boreal Plain; the transition 17 between these ecological regions is dynamic and largely coincides with an annual water balance threshold 18 where P equals potential evapotranspiration (PET), with a moisture surplus to the north and deficit to the 19 south (Ireson et al., 2015). In the southern and central portions of the basin, part of the Palliser Triangle, 20 the climate is among the most arid in Canada (Szeto, 2007 2014). In the plains region, the basin includes several very large lakes, and a large portion of the area is 1 covered by smaller lakes and wetlands (Woo et al., 2008). Climate conditions are cool, with considerable 2 intra-and inter-annual variability in air temperature, and the region is a source area for cold, continental 3 air masses (Szeto et al., 2008). The basin is a globally important resource that affects the welfare of people 4 throughout the western hemisphere and globally, yet the ecological, hydrological, and climatological 5 regimes are changing rapidly and are threatened by global warming and human impacts (RIFWP, 2013). 6 While the majority of the river basin is largely undisturbed, local impacts on river flows and ecosystems 7 arise in the headwaters, due to operation of the Bennett Dam on the Peace River, and in downstream 8 areas, for instance, due to operations of the Athabasca oil sands. 9 10 Over this region, past changes in stream and river discharge have exhibited a trend towards earlier spring 11 freshet and river ice breakup and an increase in winter discharge in many northern basins (DeBeer et al. snowmelt to more mixed and rainfall-driven regimes (Burn and Whitfield, 2018), and in spite of warming 15 spring air temperatures, delayed spring streamflow in some areas of the southern Arctic . 16 Naturalized flows (after accounting for the changes due to reservoir operations and water withdrawals) 17 of the South Saskatchewan River have exhibited a steady decline since the early-20 th century, with late 18 summer volumes declining at a greater rate than the annual discharge (Pomeroy et al., 2009). Flows in 19 the Mackenzie River since the early 1970s have shown a shift in timing of peak flows of several days, an 20 increase in maximum discharge of about 3,000 m 3 /s, and a rise in winter base flows . 21 3. Process interactions, changes, and their influence on water cycling 22 23 Field-based observations and experimentation across the network of WECC observatories ( Fig. 1) and at 24 other sites has provided key insights on process interactions and responses. Here we summarize these 25 insights for several important hydrological and ecological processes. 26 27 3.1 Precipitation recycling and evapotranspiration 28 29 P and evapotranspiration (ET) are important terms in the water cycle and even minor shifts in their relative 30 magnitudes can have critical impacts on surface water availability, streamflow, and groundwater storage. 31 Recent changes in P over western Canada have shown regional and seasonal variations, with annual and 32 winter increases in volume in the north, and more significant winter decreases in the southern interior 33 (Vincent et  where an increase in P is likely to increase ET, but the increase in P itself could also be a result of increasing 6 land ET and stronger moisture recycling (Trenberth, 1999;Dirmeyer et al., 2009). In future, under a 7 warming climate, earlier disappearance of the seasonal snow cover will act to increase regional ET in 8 spring as a result of the reduction in surface albedo, increase in net radiation to the ground surface, 9 increase in overall surface temperature, thaw of frozen ground, and increase in exposure of wet soils. 10 Shorter ice cover duration, especially in more northern lakes, will lead to increased lake evaporation and 11 will therefore also play an important role in providing local moisture sources to downwind regions. These continues into summer, and by July and August, simulated future P decreases in these parts of the region 27 (most of the Saskatchewan River Basin), due in part to the decrease in soil moisture and surface water 28 availability in the antecedent spring months. Although there is a simulated increase in moisture recycling 29 in the warm season, the excess of ET over P is associated with an increase in atmospheric moisture 30 divergence (i.e., transport out of the region). 31 32 Changes in ET also occur as a consequence of land cover and vegetation changes. Vegetation cover in and Ryan (1997) showed that, due to physiological limitations to transpiration in Boreal needleleaf trees, 38 they have much lower ET rates than deciduous species, even when soil water is abundant. This is 39 consistent with observations at the Boreal Ecosystem Research and Monitoring Sites (BERMS) flux towers 40 ( Fig. 1, site 7), showing a mature aspen stand with higher ET than a mature black spruce, which had higher 41 ET than a jack pine stand (Fig. 3). Kljun et al (2006) attributed these differences to a combination of type 42 of tree species, topography and soil type. Very young forest stands have also been shown to have much 43 lower rates of ET than older stands (Granger and Pomeroy, 1997). Thus, shifts in Boreal Forest 44 composition and structure, from coniferous to deciduous or mixed-wood, or from black spruce to jack 45 pine (discussed in Sect. 3.4 below), will have potentially large, but species-specific effects on regional ET. balance regionally. The expansion of shrubs in northern tree line and Tundra environments will likely 10 increase regional ET in the snow-free period.  20 21 Over western Canada during the past several decades there has been a widespread reduction in snow 22 depth, snow cover extent, and seasonal duration, with a shorter snow cover period of between one to 23 two months, mostly due to earlier melt in spring (Brown et  in snow regime, including i) a greater fraction of P in the form of rain as opposed to snowfall, especially 26 during shoulder seasons, at lower elevations, and in more southerly locations, ii) more frequent rain-on-27 snow events, iii) warmer and wetter snowfall, iv) more mid-winter melt events as air temperature crosses 28 the freezing point more frequently, and v) earlier spring melt and snow cover depletion (Fig 4). This will 29 also cause distinct changes in runoff, with further transition from snowmelt to rainfall-dominated regimes. 30 The transitions from snowfall to rain and from snow-dominated to rain-dominated hydrological systems 31 are particularly sensitive where and when conditions are relatively warm and large amounts of P occur 32 near 0°C ( formation, producing complex runoff responses in spring. Follow-on hillslope-scale analysis by Coles and 47 McDonnell (2018) found evidence for filling of micro-and meso-depressions on the slope, followed by 1 macro-scale, whole-slope spilling. While surface topography is relatively unimportant under unfrozen 2 conditions on low relief and high infiltrability Prairie sites, surface topography was of critical importance 3 for connectivity and runoff generation when the ground was frozen during the brief, annual snowmelt 4

Snow regime change and snow-vegetation interactions
pulse. Under climate warming, losing this brief period of surface topographic control on runoff generation 5 could have large implications for hillslope runoff, depending on basal ice formation, among other factors. 6 7 Warming can also lead to other important, and sometimes unanticipated, responses in snow 8 accumulation, redistribution, and ablation processes (Fig. 4). Earlier onset of spring melt of the seasonal 9 snow cover shifts snowmelt timing to conditions of lower incoming solar radiation (Pavlovskii et al., 2019). 10 Paradoxically, this can lead to a reduction in daily and seasonal average ablation rates and a longer overall Marmot Creek indicate the reduction of blowing snow transport and sublimation with warming of up to 18 5°C reduces the redistribution by transport by up to 50% and losses from sublimation by up to about 30%. 19 This would also have important, but at present, poorly understood consequences on the redistribution of 20 snow, the variability and patterns of SWE over the landscape, and the timing and rate of snow cover 21 depletion (e.g., DeBeer and Pomeroy, 2017). Suppression of blowing snow would lead to a more uniform 22 spatial distribution and thus more rapid decline of snow-covered area that could not be compensated for 23 by the variability in melt energy (Schirmer and Pomeroy, 2020). 24 25 Snow-vegetation interactions further affect hydrological responses, and the impacts of vegetation change 26 can equal or exceed those due to climate alone (Rasouli et al., 2019). A conceptual summary is shown in 27 Fig. 4. With rising temperatures, warmer and wetter intercepted snow is more likely to fall to the ground 28 instead of remaining in the forest canopy, where it would otherwise mostly sublimate. Snowfall 29 interception efficiency is relatively insensitive to air temperature (Hedstrom and Pomeroy, 1998) and thus 30 warming is unlikely to lead to large changes in initial interception amounts. But retention of the 31 intercepted snow load is highly temperature dependent (Ellis et al., 2010) and so warming promotes faster 32 unloading and a lower sublimation loss. This acts in combination with reduced wind transport of snow on 33 the ground to offset reductions in SWE due to direct warming effects . Forest 34 canopy structure, density, and species composition also significantly influence interception loss. Thinning 35 of existing forest cover, reduction in leaf area index (LAI), and transition from coniferous to deciduous 36 species, which are expected as a result of increasing human and natural disturbance and wildfire (Sect. 37 3.4), will lead to greater surface snow accumulation due to the reduction in canopy interception and 38 sublimation, but at the same time will expose more of the snow surface to increasing net radiation and 39 an accompanying increase in ablation rates. 40 41 In open, windswept environments dominated by short vegetation such as grasses, crops, and shrubs, climate change is expected to further exacerbate the current imbalance and lead to additional retreat 13 (Clarke et al., 2015). 14 15 Mass balance (the net gain or loss of snow and ice averaged over the glacier surface) responds directly to 16 climate perturbations, whereas glacier extent, form, and flow patterns exhibit delayed and modified 17 responses to mass balance changes (e.g., Clarke et al., 2015). Glacier responses are also influenced by 18 secondary factors such as temperature effects on ice flow and meltwater availability at the glacier bed, 19 which affects glacier sliding. In general, warmer air temperatures lead to greater specific ablation rates 20 and a longer melt season, and may reduce accumulation depending on the area-elevation distribution of 21 individual glaciers and the nature of P changes. Many glaciers and icefields in the CCRN region receive 22 snowfall year round at high elevations and some rainfall in the summer. With climate warming, the 23 proportion of rainfall events increases and the late summer snowline moves to higher reaches of the 24 glaciers, exposing firn and bare ice, which melt faster than snow due to their lower albedo. Dust, 25 impurities, and algae in the snow and ice become more concentrated on glacier surfaces as a consequence reductions in accumulation zone extent can lead to rapid glacier disintegration, and even complete 32 disappearance. Glacier fragmentation and detachment of tributary ice streams leads to loss of ice supply 33 to lower reaches, which can then become stagnant and melt out. 34 35 There are other important glacier-climate feedbacks. Energy balance conditions shift in response to 36 glacier retreat; for example, ice-free marginal areas and valley walls contribute turbulent energy supply 37 and longwave radiation fluxes to the glacier, and these fluxes can be enhanced as glaciers thin and retreat, 38 increasing ablation rates. The presence of glacial ice helps to regulate local climates and preserve cold 39 conditions. As reduced snow accumulation leads to a reduction in glacier mass balance, so a reduction in 40 glacier extent leads to a reduction in snow accumulation, given that the glacier surface, which is ≤ 0°C, 41 helps retain snow cover (Marshall et al., 2011). 42 43 Projections of future glacier change indicate that glaciers in the Rocky Mountains will lose roughly half 44 their total area and volume by mid-century, and as much as 90% or more by the end of the 21 st century 45 under a 'business as usual' (RCP8.5) climate scenario (Clarke et al., 2015). By mid-century, many valley 46 glaciers will have retreated substantially up-valley, and by late in the century even high elevation glaciers 47 and icefield plateaus will be greatly reduced or will have disappeared entirely (Fig. 5). Even the Columbia 1 Icefield, the largest and among the highest elevation ice masses in the Rocky Mountains, is projected to 2 disintegrate into several small vestigial patches of ice near the tops of the highest peaks by the late-21 st 3 century. There are not comparable studies for the glaciated regions of the Mackenzie Mountains, 4 Northwest Territories, but the observed patterns of recent change are similar to glaciers in the Rockies 5 and the future change is expected to be similar. 6 7 From a hydrological perspective, as glacier loss progresses, glacier wastage contributions and enhanced 8 ablation will increase glacial contributions to discharge towards "peak water" (Huss and Hock, 2018), 9 followed by a decline in glacier runoff due to loss of ice-covered area, even with further warming and 10 increasing specific ablation rates. It remains uncertain where, when, and over what scales this will occur, 11 although some studies have indicated that peak water has already passed in parts of southwestern Canada 12 ( the buffering effect that glacial storage can provide for discharge variations (e.g., during drought years) in 17 the mountain headwaters will become increasingly diminished. 18 19 3.4 Northern vegetation, wildfire, and loss of ecological resilience 20 21 Ecosystem change can have profound effects on hydrological response and land-atmosphere feedbacks, 22 yet the complexity of expected change and the associated uncertainty are often overlooked in 23 hydrological projections. Across the CCRN region, contemporary climate change is already having direct 24 impacts on northern ecosystems, defined here as including the southern Boreal Forest and its transition 25 with the Prairies, and the Cordillera. The interior of western Canada has been identified as a region of 26 maximum ecological sensitivity (Bergengren et al., 2011). Forests in the southern Boreal region of western 27 Canada have shown signs of declining productivity and increasing mortality associated with drought stress increases in tree density, with changes in growth form from krummholz to erect tree form. 40 41 Climate change alters terrestrial ecosystems broadly through changes to: 1) composition (vegetation, 42 soils, and wildlife), 2) configuration and disturbance patterns, and 3) function. This includes structural 43 changes to the current vegetation (above-and below-ground biomass, plant density, canopy height, LAI, 44 and rooting depth); changes to land cover distribution patterns (resulting from changes in the disturbance 45 regime and changes in competition, colonization, ecosystem resilience and vegetation succession 46 following disturbance); and functional changes (surface albedo, snow accumulation and melt, soil freeze 47 and thaw, ET, ecosystem productivity, decomposition, biogeochemical cycling, and wildlife habitat 2004). While fire has been a foundational process in the functioning and ecology of the Boreal Forest for 18 more than 5,000 years, an increase in the frequency of high-intensity fires, coupled with a warming 19 climate, may weaken ecosystem resilience and disrupt the historically stable cycles of forest succession. 20 The result may be a regime shift from one plant community to another and from one stability domain to 21 another ( . There is consensus that in northern forests, fire frequency and 31 severity will continue to increase (Rogers et al., 2020). 32 33 Projections of future wildfire-induced ecosystem change in the Boreal Forest are challenging and highly 34 uncertain. Increasing fire will result in a younger forest, widespread replacement of black spruce stands, 35 and higher proportions of deciduous broadleaf species or jack pine (e.g., Johnstone et al., 2010a), with 36 greater change in the south than the north. CCRN developed a plausible scenario of post-fire replacement 37 of evergreen needleleaf forest (ENF) with deciduous broadleaf forest (DBF) across the Boreal Forest , as 38 described in the Appendix, for the purpose of use in hydrological model future projections (Fig. 6). 39 Although this is simply a scenario, and not a projection with an associated confidence level, the resulting 40 forest change due to increasing wildfire is potentially great. For both the mid and late-century periods, 41 there is a considerable reduction in DBF across the southern parts of the Boreal Plain, as a result of 42 increasing fire and the conversion of forest to grassland. Farther north and west, in the Taiga Plain, the  43 Shield, and the Western Cordillera, there is extensive and progressive replacement of ENF with DBF as a 44 result of both climate and fire-driven changes in forest succession. In reality, DBF and jack pine stands 45 tend to be more resilient to depth, and thermokarst and subsidence, altering supra-permafrost layer storage, flow paths, and lake 33 development (Sect. 3.5). 34 35 In addition to the forest cover change scenario, CCRN developed a plausible scenario of 21 st century shrub 36 expansion into tundra, grassland, and barren areas, described in the Appendix and shown in Fig. 7. While 37 there is uncertainty and this does not represent a confident projection, prolific shrub growth over the 38 Boreal and Taiga Cordillera, the Southern Arctic, and the Taiga Shield ecological regions is expected. The 39 gradual expansion northward is evident through the increase in shrub cover along the northern part of 40 the Mackenzie River basin and the movement of this growth zone to higher latitudes later in the century. 41 42 3.5 Permafrost thaw as a driver of landscape change and hydrologic rerouting 43 44 Climate warming has led to warming and increased thaw depth of permafrost across northern Canada 45 (Smith, 2011), with associated changes in characteristics of seasonally-frozen soils (e.g., timing of freezing 46 and thawing, frequency of freeze-thaw cycles, depth of frost, etc.). In southerly locations where 47 permafrost is discontinuous, shallow, and relatively warm (i.e., at or near the freezing point depression), 1 there has been widespread thawing and degradation of permafrost, with increasing supra-permafrost 2 layer thickness-including both the active layer (seasonally frozen) and the talik (perennially thawed) 3 . As a result of warming and shallower re-freeze depths during winter, active layer 4 thickness has been decreasing. Where ice-rich soils occur, there has been active thermokarst canopy thinning due to fire, disease, or other disturbance allows for an increase in local solar energy input 31 and leads to preferential ground thaw (Fig. 8). A local depression forms in the relatively impermeable 32 frost table and underlying permafrost table. Such thaw depressions introduce a hydraulic gradient that 33 directs subsurface flow towards them so that thaw depressions soon become local areas of elevated soil 34 moisture content. Since the thermal conductivity of wet soil is far more than that of dry soil, the vertical 35 conduction of energy to the thaw depressions increases due to the increased moisture content, a nd as a 36 result, a positive feedback is initiated which accelerates the thaw of the disturbed areas. Wet conditions 37 prevent trees from re-establishing and a new, isolated flat bog is formed. Many areas within the Taiga 38 Plain are highly susceptible to thaw through this process (e.g., Gibson et al., 2020) and widespread 39 replacement of forest-covered peat plateaus by wetlands is expected over the coming decades. A caveat 40 is that these ecosystems represent some of the strongest ecosystem-protected permafrost, so 41 undoubtedly a portion of permafrost peatland will linger, but this will depend on the degree of warming 42 and also fire (Stralberg et al., 2020). 43 44 The loss of permafrost is impacting water cycling across the northern parts of the CCRN region. Land 45 surface subsidence and the collapse of peat plateaus to wetlands in the Taiga Plain alters drainage  46 networks, surface and groundwater storage distribution, and the transit of water across the landscape 47 (  19 20 Over much of the Prairies and the Boreal Plain, groundwater discharge from shallow sand and gravel 21

Groundwater interactions and Prairie wetland processes
aquifers sustains year-round base flow in some small streams and can be an important component of the 22 water balance of wetlands and of some lakes. Groundwater is thus important with respect to local water 23 resources and in maintaining surface hydrological connectivity and ecosystem function. Groundwater 24 provides rural water supplies and in some cases municipal supplies (Peach and Wheater, 2014), and whilst 25 it is not used as a major source for irrigation water outside of the south-central parts of Manitoba, an 26 issue facing some parts of the Prairies is the increasing reliance on groundwater as water demand rises 27 and surface water becomes over-allocated (Council of Canadian Academies, 2009). Regional-scale 28 groundwater depletion is not common in Canada, unlike other parts of North America (Rodell et al., 2018), 29 but there have been numerous examples of isolated, human induced local-scale depletion in Alberta (e.g., 30 Munroe, 2015). The water- Groundwater processes are closely linked to the water regime (i.e., hydroperiod) of wetlands. Prairie 13 wetlands occur in the form of shallow marshes ("sloughs" or "potholes") with little accumulation of 14 organic matter, whereas Boreal wetlands primarily occur as peatlands. The spatial transition from Prairie 15 marshes to Boreal peatlands is coincident with the transitional ecotone between the Prairie and Boreal 16 Plain  table in surrounding uplands during dry  27 periods. In contrast, the fen sheds water quickly to streams during wet periods when the water table rises  28 above the peat surface. Long-term studies in northern Alberta have shown that the type of glacial 29 sediments has a large influence on the groundwater exchange and runoff generation from the peatlands 30 (e.g., Devito et al., 2017). 31 32 Groundwater replenishment to deeper aquifers is restricted by the low permeability of overlying layers of 33 clay, clay-rich glacial till, and shale, and by the position of the aquifers within larger regional groundwater 34 flow systems (Cummings et al., 2012). In the Prairies, replenishment rates to confined aquifers generally 35 range from a few mm to a few tens of mm per year (van der Kamp and Hayashi, 1998). Recharge to the 36 water table represents a residual in the water balance and is highly sensitive to changes in the balance 37 between P and ET; however, replenishment to deep aquifers is not sensitive to variations of the water 38 table and therefore responds slowly to climate change. 39 40 In the Western Cordillera the interaction of groundwater with surface waters is in many ways different after snowmelt or rainfall events. However, this is generally followed by a slower recession and the 1 remaining storage allows these aquifers to sustain stable base flow during the rest of the year when there 2 is little recharge (Hayashi, 2020). The high topographic relief, together with significant heterogeneity in 3 bedrock and surficial deposits, influences patterns of vertical and lateral groundwater flow and recharge 4 and discharge processes. At lower elevations, aquifers include glacial and alluvial deposits of highly 5 permeable sands and gravels that drape mountainsides and underlie valley bottoms, usually 10s to 100 m 6 thick, but in some instances up to several hundred meters in thickness (Toop and de la Cruz, 2002). These 7 store larger quantities of water and provide a reliable supply for municipal and industrial uses. In 8 floodplain areas, the water table is usually near the ground surface and fluctuates with river levels. 9 Although mountain aquifers are able to buffer base flow against climate warming and associated changes 10 in surface water availability (e.g., Paznekas and Hayashi, 2016), anecdotal evidence has indicated that they 11 cannot sustain high flows in drought years, such as in 2015 when the spring-summer discharge of the Bow 12 River fell to about half its median rate at Banff, and to less than 10% at its mouth. 13 14 15 Due to the complexity in process responses to climate and anthropogenic change in the CCRN domain and 16 other cold regions, there is significant uncertainty associated with model projections of future 17 hydrological change. While all models have limitations, detailed process-based models can yield 18 important insights into interactions and feedbacks, and large-scale models can be used with careful 19 selection of possible scenarios to quantify likely effects of future change. Here we describe CCRN's efforts 20

Process-based modelling of change in CCRN
to improve model process representation, diagnose past change, and predict future change. 21 22 4.1 Fine-scale diagnostic and predictive modelling 23 24 Based on field studies and understanding from the WECC observatories, efforts were directed primarily 25 at improving functionality and expanding the capability of handling complex cold region processes within 26 CRHM ( using results from the North American Regional Climate Change Assessment Program (NARCCAP) 4 consisting of 11 regional climate models driven by outputs from multiple global climate models (GCMs) 5 for the SRES A2 emission scenario (see Rasouli et al., 2019). Hydrological responses to changing 6 vegetation, soils, and land cover were examined using current and expected future states of the basins. framework that facilitates inter-comparison of alternative algorithms and models (e.g., land surface 18 schemes and routing schemes), and can be applied over large river basins. 19 20 Over the course of CCRN, major advancements in the MESH system were made in terms of basic 21 operability, scalability, and parallelization, as well as in its ability to handle sloping and complex terrain, climates at a 10 km resolution, incorporating these advancements in process and water management 30 representation, to examine changes in regional hydrology and river flows. Forcing data included WATCH 31 and ERA-Interim products with bias correction using regional datasets such as the combined Global 32 Environmental to run the models for full future assessment. Scenario results are currently pending, but some preliminary 40 insights are discussed below. 41 42 43 New understanding and insight into process sensitivity, interactions, and responses (Sect. 3), together 44 with expert elicitation and process-based modelling (Sect. 4), have allowed more scientifically-informed 45 projections of future ecological, cryospheric, and hydrological change than have hitherto been available. Here, these are brought together, informed by the new research results from CCRN, to develop a summary 1 picture largely applicable to the late-21 st century (Fig. 9). 2 3

Synthesis of future change and hydrological responses
Future climate is expected to lead to profound changes in land cover and vegetation. In the mountain 4 regions, one of the most striking changes will be the loss of glaciers. The lower parts of many glaciers will 5 have disappeared within decades or less, while upland icefields may persist, but in a much diminished 6 state. By the late-21 st century only vestigial remnants of the former ice cover and small glaciers in 7 favorable locations for ice preservation will likely remain. Over a much larger part of the CCRN domain, 8 and of greater magnitude of change, will be the response of vegetation and forest ecosystems to climate 9 change and climate-induced disturbances. At northern and alpine tundra and tree line ecotones, shrub 10 growth and expansion in tundra will continue and is expected to accelerate over the latter half of the 21 st 11 century. A northern and upward shift in tree line is likely but will occur more slowly and be far less 12 pronounced than for shrub expansion. Across the contiguous Boreal Forest, the major transition will be 13 the loss of ENF and major expansion of DBF and jack pine forest stands, wetlands (in the north), and to a 14 lesser extent, grasslands (e.g., in valley bottom areas of the Cordillera). Permafrost thaw and collapse of 15 permafrost-underlain spruce forest and peat plateaus will accelerate over vast parts of the Taiga Plain. At 16 the southern Boreal-Prairie ecotone and over the Boreal Plain, northward expansion of deciduous shrubs 17 and concomitant loss of deciduous and mixed-wood forest will continue, leading to the expansion of 18 grassland in these areas into the late-21 st century. 19 20 In addition, human activities, land-water management practices, and changes in agricultural cropping 21 patterns will further alter landscapes. These are likely to be most pronounced in the Prairie and southern 22 Boreal parts of the CCRN region. Climate warming will further drive changes in crop mix and spatial 23 patterns, with new crops such as corn becoming more widespread, and northward expansion of other 24 crops such as canola, wheat, and soy (Hannah et al., 2020). Climatic and land suitability limitations will 25 restrict how, where, and the timescales over which this occurs. For example, parts of southern Alberta 26 will experience more extreme heat and heat stress days above 30°C, resulting in declining crop production 27 even with sufficient moisture. In Saskatchewan, work by Coles et al. (2017) has suggested for planted 28 hillslopes, measured decreased snowfall, snowmelt runoff, and spring soil water content is affecting 29 agricultural productivity through increased dependence on growing season precipitation, likely 30 accentuating the future impact of droughts. Areas vulnerable to drought, such as the Palliser Triangle of 31 southern Alberta and Saskatchewan, and where soils have low moisture storage capacity, will most likely 32 undergo conversion to pasture and grassland as arable agriculture becomes non-viable. Other areas may 33 require irrigation to remain viable, and with agricultural expansion and more water-intensive forms of 34 crop production, there will be increased irrigation demand (Council of Canadian Academies, 2013) and 35 possibly a need for more reservoirs. The northward expansion of agriculture will occur in nodes as 36 infrastructure and roads develop, and be limited by the suitability of soils. Another major change in parts 37 of the agricultural zone is the artificial drainage of wetlands, which has various impacts on runoff, erosion, 38 sediment transport, groundwater recharge, and water quality Shook et al., 2015). 39 While recent polices have been implemented to limit drainage (or minimize the impacts), the trend will 40 likely continue, especially in wetter regions to the east and in the face of hydro-climatic change resulting 41 in more spring and summer flooding (Stewart et al., 2019), although the potential exists for wetland 42 restoration to mitigate these effects. 43 44 The combined changes in climate, vegetation, soils, and land cover will have major effects on hydrology. 45 CRHM outputs show that the loss of cold in the CCRN region is expected to cause dramatic shifts in the 46 timing, variability, and volume of streamflow, and even more profoundly, on the processes generating 47 streamflow. There is sometimes compensation by changing vegetation, but also instances where 48 https://doi.org/10.5194/hess-2020-491 Preprint. Discussion started: 17 October 2020 c Author(s) 2020. CC BY 4.0 License. vegetation and soil change enhance the magnitude of climate change impacts on hydrology. Summary 1 results from the CRHM applications at several observatory basins in different ecological regions are 2 provided in Table 1. Results for a number of other basins are pending. These studies show a tendency 3 for increasing total discharge and earlier spring freshet in these headwater basins, as a result of warmer 4 and wetter late-21 st century conditions, but mixed trends in SWE and peak discharge rates. Within 5 Marmot Creek, anticipated warming will cause basin-wide peak SWE to decline by about 30 to 40%, but 6 by as much 90% in some parts of the basin, with valley bottoms becoming almost entirely snow-free, and 7 an accompanying shift in snow cover depletion of up to six weeks. Yet the increase in P leads to a roughly 8 20% increase in total discharge. Farther north at Wolf Creek, where conditions are colder, climate change 9 impacts on snow regime are projected to be less severe and vegetation change (expansion of forest and 10 shrub tundra) is projected to have a compensatory influence. Here, a statistically insignificant increase in 11 SWE due to vegetation increase in the alpine zone was found to offset the statistically significant decrease 12 in SWE due to climate change. At high elevations in Wolf and Marmot Creeks, CHRM results indicate that 13 vegetation/soil changes moderate the impact of climate change on peak SWE, the timing of peak SWE, 14 evapotranspiration, and annual runoff volume. However, at medium elevations, these changes intensify 15 the impact of climate change, further decreasing peak SWE and sublimation. At Havikpak Creek near the 16 Taiga-Tundra transition, where significant expansion of shrubs is expected, maximum SWE will increase 17 as a result of increasing P and reduced blowing snow redistribution and sublimation. This is expected to 18 double the volume of discharge, and significantly increase spring freshet volume, snowmelt rates and 19 peak discharge rates. 20 21 CRHM was also applied to the Bow (~7824 km 2 ) and Elbow (~1192 km 2 ) River Basins above the city of 22 Calgary, AB, and run to diagnose the hydrological effects of forest disturbance in these basins in the 23 context For the larger Saskatchewan and Mackenzie River systems, the results of MESH simulations over the 35 Saskatchewan and Mackenzie River Basins indicate that future climate conditions will lead to considera ble 36 shifts in discharge timing, magnitude, and variability. The results are provisional and do not yet fully 37 account for changing landscapes and vegetation, but initial MESH climate production runs indicate there 38 is likely to be a shift in timing of spring hydrograph rise and peak flows of nearly two weeks earlier by mid-39 century, and as much as one month by late-century. Fine-scale MESH runs on the mountain-sourced Bow 40 and Elbow River Basins, driven by WRF, and with adjustments for slope, aspect and elevation, were able 41 to capture the main river hydrographs well and demonstrate how this forward shift in freshet is a result 42 of a transition to much more rainfall-runoff generation as rainfall increases and snowpacks decline in the 43 late-21 st century (Tessema et al., 2020). The MESH models of the Saskatchewan and Mackenzie River 44 Basins further show that increasing P across the CCRN region of interest is not offset by increasing ET, and 45 overall flow volume increases by as much as 40% by the end of the century. Low flows in winter become 46 slightly higher in magnitude but with more inter-annual variability, and there is a likely considerable 47 increase in spring freshet volume and peak flows. By late-century these spring flows, on average, will 48 increase by a factor of 1.5 to 2; the greater variability and higher peak flows at most locations along the 1 river network will greatly increase the risk of spring flooding. This is likely to stress human water 2 management systems and reservoir operations, as river discharge regimes may be altered far beyond the 3 historical flow ranges, seasonality, and variability under which these systems were designed and 4 operated. 5 6. Concluding remarks 6 7 This article reports results of the multi-disciplinary CCRN, which has examined recent and future 8 ecological, cryospheric, and hydrological change in relation to projected 21 st century climatic change over 9 the interior of western and northern Canada. Key insights into the mechanisms and interactions of Earth 10 surface process responses are presented, gained from a network of highly instrumented and intensively 11 studied experimental observatories. This provided the ability to observe and diagnose change across the 12 region, while the sites acted as a testbed for developing and improving predictive models. CCRN activities 13 also involved improving cold region process representation within the CRHM fine-scale and MESH large-14 scale modelling systems. Application of the fine-scale modelling system has been used to diagnose recent 15 change in selected basins, and the nature of future change. Broader application of the fine-scale and 16 large-scale models under future climate and land cover scenarios, representing mid-and late-21 st century 17 conditions, is currently underway with support of the Global Water Futures program. 18 19 In general, insights from expert elicitation and preliminary modelling indicate that the region will continue 20 to undergo widespread environmental change as a result of warmer temperatures and changing P 21 regimes. This will predominantly involve continued loss of snow and ice, thawing of permafrost, major 22 ecosystem change and an increase in the occurrence and magnitude of wildfire, and a shift from nival and 23 glacial to more rainfall-driven pluvial runoff regimes. However, some of the process responses are non-24 trivial and highly complex. To understand the trajectories of different northern ecological, cryospheric, 25 and hydrological systems under climate change, the details of these processes and their interactions are 26 very important. This can have unanticipated and sometime surprising outcomes that simple models or 27 extrapolations will fail to capture. Many current generation land surface schemes and hydrological 28 models do not handle a dynamic landscape where vegetation, glaciers, permafrost distribution, etc. are 29 transient, and there is large uncertainty in their application under a non-stationary hydro-climatic regime. 30 Human interventions also have a large influence through activities such as forest disturbance, agricultural 31 and forest land management, water abstractions for consumptive use, diversions, and reservoir 32 operations, which further alter ecological and hydrological systems. 33 34 Another critical issue relates, in part, to long-term data acquisition and organization. Climate monitoring 35 and observation are key to understanding its variability and trends, and for providing input to land surface 36 and hydrological models, yet this is a major challenge in cold regions. Forcing data remains the largest 37 source of uncertainty for historical simulations. In Canada, and especially in its alpine and northern 38 regions, there is a sparse observational network, with problems related to station automation and major 39 challenges associated with the measurement of solid P (Rasmussen et al., 2012), thus requiring high 40 priority to expanding the network and to better measuring snowfall (Bush and Lemmen, 2019). 41 42 Finally, we note that modelling at multiple scales is advantageous for more fully examining Earth system 43 behaviour and responses. While all models have limitations, detailed process-based models can yield 44 important insights into interactions and feedbacks, and large-scale models can be used with careful 45 selection of possible scenarios to quantify likely effects of future change. The CRHM and MESH modelling 46 platforms provide a unique capability to represent the complex, energy-dominated processes that control 1 cold regions hydrology. However, while further work is underway on scenario analysis, there are also 2 continuing needs for the development of flexible and robust models with the capability to capture cold 3 region processes and bridge scales from local to regional to large basin-scale. 4 Appendix: Developing Future Land-Cover Maps for Hydrologic Modelling 5 6 This Appendix describes our approach to generate future land-cover scenarios for hydrologic modeling, 7 based on observational and modelling studies, and expert elicitation. The scenarios were developed for 8 use in the MESH hydrologic model, to address the question: What is the potential for vegetation changes 9 to affect 21 st century streamflow in the Saskatchewan and Mackenzie River basins? The approach 10 generated future scenarios by applying a realistic change signal to the current MESH land-cover map. 11 12 The The RFCT analysis did not include four of the MESH PFTs (Wetlands, Water, Ice, or Urban). Consequently, 23 it was necessary to limit the changes in fractional coverage to seven CLASS PFTs (Deciduous Broadleaf  24 Forest (DBF), Evergreen Needleleaf Forest (ENF), Mixedwood Forest (MWF, SK Basin only), Cropland, 25 Grassland, Shrubland, Tundra, and Barren). The Shrubland and Tundra PFTs were identical to Grassland 26 except for height and leaf area index. In addition, the RFCT represented prairie Grassland and Cropland 27 as one vegetation class, so that it was not possible to represent changes due to competition between the 28 two. 29 30 The resulting unmodified RFCT change signals for 2005 to 2040 and 2005 to 2085 represent the land-cover 31 changes that would be expected if climate was the only factor limiting vegetation migration. In reality, 32 vegetation migration is also limited by the rates of colonization, and in some cases, by additional 33 constraints such as the need for wildfire as a trigger. We used expert knowledge to eliminate unrealistic 34 changes from the RFCT change signal, retaining only changes that were deemed to be plausible over the 35 21 st century. The plausible changes are listed in Table A1, with associated conditions and constraints. For 36 land-cover changes that normally occur only after wildfire (ENF to Grassland and ENF to DBF, Table A1),  37 the analysis added two further constraints. The area burned was estimated assuming a prescribed fire-38 return interval which varied with latitude (  above the northern and alpine tree lines, Shrubland expansion into Tundra. 7 8 The strategy of applying a RFCT change signal to the current land-cover map, with modifications based on 9 constraints from expert knowledge, has several advantages over using the RFCT projections directly. It 10 anchors the projections to the current land-cover map, potentially increasing their realism. It eliminates 11 changes that are implausible over the modelling time frame (21 st century). It integrates wildfire as a 12 trigger for changes that most often occur after fire. And it preserves the characteristic patchiness of the 13 boreal forest mosaic. Note that the resulting land-cover projections are intended for use in hydrologic 14 modelling only; at best, they represent an informed guess of the likely changes. Caution is advised against 15 using them in other applications. 16 Data availability 17 18 Data are available through the cited sources throughout the text. 19 Author contributions 20 21 Chris DeBeer led the organization and writing of the article with significant input from all co-authors on 22 aspects of modelling, analysis, review, figures, interpretation and writing. 23 Competing interests 24 25 The authors declare that they have no conflict of interest. 26 Acknowledgements 27 28 We gratefully acknowledge financial support from the Natural Sciences and Engineering Research   Gordon, R. R., and Eckert, C.