4.2.2 Snow and hydrology variables

Lead Authors: Lawrence Mudryk (CRD/ECCC) and Rajesh Shrestha (WHERD/ECCC)

4.2.2.1 Simulated snow data

Snow is a complex climate variable and can be simulated in a variety of model types (e.g., fully coupled AO/ESM models, offline land surface models, and reanalyses) as a component of the surface which can interact with incoming radiation, and which couples the lowest layer of the atmosphere to the soil both thermodynamically and hydrologically. Snow may exist on land surfaces, ice surfaces or both depending on the model type.

There are three key physical effects related to snow typically simulated within a climate model:

1. Snow cover modifies the absorption of radiation by the surface, since its high albedo reflects more incoming shortwave radiation than most other types of surface cover. This effect depends on location and season since the amount of incoming radiation varies according to these, as well as on the condition of the snow: melt and snow aging processes can reduce snow albedo somewhat, and deposition of black carbon (“soot”) or other dark contaminating materials (e.g., organic debris) on the snowpack can alter the albedo more substantially.
2. Snow mediates the exchange of heat between the atmosphere and the soil layer underneath the snow by slowing the transfer of heat between the two. In typical winter conditions the air above the snow pack is colder than the soil layer, such that the snow acts to keep the soil warmer. This effect will depend on the thickness of the snow pack and on internal characteristics that affect its thermal diffusivity such as density, liquid water content, stratigraphy.
3. The snowpack acts as a water reservoir, accumulating frozen precipitation under winter conditions and releasing it during melt. In snow-dominated regions, this stored water typically melts over a reasonably short period and is responsible for springtime surges in runoff and river discharge, as well as seasonal contributions to the soil water levels.

On the land surface, snow is typically modelled as several separate layers (e.g., 1-5, although complex offline snow models may use 10-100) using the same geometry as the atmospheric grid. Sometimes it is considered to be a topmost “soil layer”. On sea ice, snow has not typically been modelled explicitly since sea ice is mobile and it has historically been too computationally expensive to model individual ice floes (nor are heuristic parametrizations, for example, a snow depth related to sea ice surface roughness usually employed). More recently snow on sea ice has been modelled in a multi-step way, using ice motion derived from model output in a second implementation that then deposits snow on the moving sea ice.

Snow is treated as a component of the global hydrological cycle within climate models and in general, all snow-related output stems from calculating how much snow mass is contained within a grid cell. This mass of snow is usually expressed as its equivalent height in water spread uniformly across either the entire grid cell or across the land-covered portion of the grid cell (“snow water equivalent”). The mass is calculated by tracking the balance between accumulation and ablative losses (melt, sublimation), and depending on the complexity of the model, various processes related to vertical mass transfer between layers, and horizontal transfer of mass between adjacent cells (very infrequently). From a practical perspective, all other snow output is derived from this key quantity. Estimates of snow depth are computed by also modelling the density of snow layers combined with the information on mass. The proportion of the surface covered by snow is also typically parametrized as a function of the mass of snow within the cell (e.g., as fully covered above some mass threshold and as a decreasing fraction of all land cover below, although a range of formulations exist and in mountainous regions they may include additional dependencies).

Considerations and limitations:

Typical considerations required for model ensemble analysis of other physical climate variables also apply to snow since uncertainty in snow projections stem from a combination of different model biases and internal climate variability. Therefore, the current practice is to use the CMIP or CORDEX ensembles for several emission scenarios to obtain large-scale or regional climate information.

The output frequency of specific variables will differ by model and choice of experiments (e.g., historical vs future scenarios) for both CMIP6 and CMIP5. For CMIP6 snow related variables, the typically provided variables are snow cover fraction (snc) and surface snow amount (snw), followed by snow depth (snd) and snowfall (prsn). For some models, some of the variables will be available at daily frequency.

Information on snow cover from CMIP5 over Canada is summarized in Canada’s Changing Climate Report (Bush and Lemmen, 2019) for RCP2.6, RCP4.5, and RCP8.5. Presently, statistically downscaled or bias corrected simulations for snow for the Canadian North are not available. The following is summarizing the current knowledge on models’ skill.

Successive generations of climate model ensembles have improved in their ability to represent historical snow extent on average (i.e. the ensemble mean) between CMIP3 through CMIP6, however, there remains a persistent amount of spread in skill among the models (Mudryk et al., 2020 and references therein). This spread in historical snow extent continues to cause spread in assessments of simulated snow-albedo feedbacks (Thackeray et al., 2021). The ability of climate models to reproduce the historical snow mass climatology is less constrained because estimates of historical snow mass from gridded products have relatively high uncertainty compared to snow extent, especially in mountain regions (Mudryk et al. 2015). This stems from the fact that snow observations are sampled at limited spatial and temporal frequency compared to their variability, although recent work has both narrowed the spread (Pulliainen et al. 2020) and better assessed the accuracy of gridded snow mass products (Mortimer et al. 2020) across Northern Hemisphere nonalpine regions. The uncertainty in the overall mass balance means that climatological rates of melt and sublimation are also uncertain (and it is difficult to isolate biases in one versus the other). The overall uncertainty in the primary snowpack balance related to accumulation and ablation also means that biases stemming from higher-order effects (vegetation-snow interactions, wind redistribution) can be difficult to isolate. Given the above, seasonal and spatial biases in snow water equivalent, snow depth, and snow cover fraction can be expected to reflect not only model-specific biases in temperature and precipitation as they vary by region and season, but also the combination of parametrization uncertainties specific to a particular model.

The ability of climate models to simulate historical trends in snow extent has also improved between CMIP3 and CMIP6, in part due a better understand of trends in historical datasets (Brown and Derksen 2013, Hori et al., 2017, Mudryk et al. 2020). For snow mass, despite poor constraints on the seasonal balance of accumulation and ablation, monthly trends in this quantity are better estimated. This is because the trends are controlled by a combination of more slowly evolving boundary forcings (e.g., anthropogenic aerosols, GHGs, volcanic forcing and solar variability), and synoptic scale variability in temperature and precipitation, which models simulate more or less reasonably. Similarly, in offline models or reanalysis, historically observed temperature, precipitation, and other observational data are used to simulate the snowpack, and these variables already contain the time-varying signals related to boundary forcing and synoptic-scale variability, producing snow mass estimates with similar temporal components.

Sensitivity experiments indicate that hemispheric snow extent, like sea ice extent, responds as a fast component of the cryosphere, and therefore, doesn’t depend on the rate of warming (Mudryk et al., 2020), or prior snow extent conditions (hysteresis).This fast response suggests that accurate projections of hemispheric snow extent should result from accurate projections of global temperatures. The fidelity of such projections on regional or local scales is less well established since historical evaluation on these scales is limited due to the presences of increased natural variability (Mudryk et al., 2017). Sensitivity of snow mass to temperature is also less established.

References - snow:

Brown, R.D. and C. Derksen, 2013: Is Eurasian October snow cover extent increasing? Environmental Research Letters, 8(2), 024006.

Bush, E. and D.S. Lemmen (editors), 2019: Canada’s Changing Climate Report. Government of Canada. Ottawa, ON. 444 pp.

Hori, M., K. Sugiura, K. Kobayashi, T. Aoki, T. Tanikawa, K. Kuchiki, M. Niwano, and H. Enomoto, 2017: A 38-year (1978–2015) Northern Hemisphere daily snow cover extent product derived using consistent objective criteria from satellite-borne optical sensors. Remote Sensing of Environment, 191, 402-418.

Mortimer, C., L. Mudryk, C. Derksen, K. Luojus, R. Brown, R. Kelly, and M. Tedesco, 2020: Evaluation of long-term Northern Hemisphere snow water equivalent products. The Cryosphere, 14(5), 1579–1594, doi:10.5194/tc-14-1579-2020.

Mudryk, L.R., C. Derksen, P.J. Kushner, and R. Brown, 2015: Characterization of Northern Hemisphere Snow Water Equivalent Datasets, 1981–2010. Journal of Climate, 28(20), 8037-8051.

Mudryk, L.R., P.J. Kushner, C. Derksen, and C. Thackeray, 2017: Snow cover response to temperature in observational and climate model ensembles. Geophysical Research Letters, 44(2), 919-926, doi:10.1002/2016GL071789.

Mudryk, L., M. Santolaria-Otín, G. Krinner, M. Ménégoz, C. Derksen, C. Brutel-Vuilmet, M. Brady, and R. Essery, 2020: Historical Northern Hemisphere snow cover trends and projected changes in the CMIP6 multi-model ensemble. The Cryosphere, 14(7), 2495-2514, doi:10.5194/tc-14-2495-2020.

Pulliainen, J., K. Luojus, C. Derksen, L. Mudryk, J. Lemmetyinen, M. Salminen, J. Ikonen, M. Takala, J. Cohen, T. Smolander, and J. Norberg, 2020: Patterns and trends of Northern Hemisphere snow mass from 1980 to 2018. Nature, 581(7808), 294-298, doi:10.1038/s41586-020-2258-0.

Thackeray, C.W., A. Hall, M.D. Zelinka, and C.G. Fletcher, 2021: Assessing Prior Emergent Constraints on Surface Albedo Feedback in CMIP6. Journal of Climate, 34(10), 3889-3905.

4.2.2.2 Simulated streamflow data

Simulated streamflow data are usually obtained from watershed to river basin scale hydrologic models, which are considered suitable for detailed analyses (e.g., seasonal water availability, extreme events). Watershed processes in cold regions are mainly dominated by spring snowmelt, which often lead to the largest annual discharge events. The presence of permafrost, and frozen soils that limit infiltration rates lead to unique challenges in modelling cold regions hydrologic regime. Several models, such as Variable Infiltration Capacity (VIC) (Liang et al., 1994; Hamman et al., 2018), Modelisation Environmentale Communautaire – Surface and Hydrology (MESH) (Pietroniro et al., 2007), the Canadian Hydrological Model (CHM) (Marsh et al., 2020), and Soil and Water Assessment Tool (SWAT) (Arnold et al., 1998, 2012) among others, were developed to take into account those specific requirements of cold climates. Wang et al. (2021) offers a recent review of the work done for modelling watershed and river basin processes in cold climate regions. Along with streamflow, the hydrologic models provide simulations of a range of hydrologic variables, such as evapotranspiration, snow accumulation, snowmelt, infiltration, soil moisture, and surface and subsurface runoff. The hydrologic models are set up at various spatial resolutions (typically ranging from 100 m to 10 km) and calibrated to reproduce watershed/sub-watershed scale streamflow, which may be supplemented by calibration of SWE, evaporation, etc.

Required meteorological forcings for hydrologic models include a range of variables (e.g., minimum and maximum temperature, precipitation, relative humidity, surface pressure, wind speed, incoming shortwave radiation and incoming longwave radiation, etc.), with historical simulations obtained by driving the model with observations or gridded observation products, reanalyses, etc. For GCM driven simulations, it is customary to employ a downscaling technique for a finer spatial-scale representation of climatic variables, which may include statistical and/or dynamical downscaling technique in combination with a bias correction method. This procedure is necessary because the GCMs do not provide sufficient resolution for hydrologic model simulation, as well as representing local-level extremes (e.g., wet and dry conditions), which are the important issues for water resources management.

A number of studies have used downscaled CMIP5 GCMs for simulating historical and future streamflow in northern Canada. These include, climate change impacts on inland waterway transport in the Mackenzie River (Scheepers et al., 2018), impacts of 1.5 and 2.0° C warming on river discharge into the Hudson Bay (MacDonald et al., 2018), uncertainty of climate change impacts due to hydrologic model components over two Nordic Quebec catchments (Troin et al., 2018), climatic controls on future hydrologic changes in the Liard river basin (Shrestha et al., 2019), and simulation of 90-year (1981–2070) spatially distributed freshwater fluxes from the Arctic basins (Stadnyk et al., 2021). However, modelled streamflow from these studies are not readily accessible. Historical and projected future streamflow and runoff based on CMIP5 GCMs are available for downloading from the Pacific Climate Impacts Consortium data portal (https://www.pacificclimate.org/data/daily-gridded-meteorological-datasets). The dataset is only available for the Peace, Fraser and Columbia basins in western/southern Canada (Schoeneberg and Schnorbus 2021) and further details are given in Tables 3.2 and 3.3.

At Ouranos, PAVICS offers a suite of tools constructed in Python, named Raven, to streamline the analysis of climate change's impacts on hydrology (https://pavics.ouranos.ca/hydrology.html#a). The tool permits to run hydrological model simulations on remote servers or with remote input data as well as to calibrate hydrological models on a remote server. The website presents an example on how to run the GR4J-CemaNeige hydrological model. Other three models can be run in the same way: HBV-E, HMETS, MOHYSE.

Considerations and limitations:

Similar to other hydro-climatic variables, future hydrologic projections are subject to different sources of uncertainties including anthropogenic forcings (emission scenarios), natural variability of the climate system, GCM structure and downscaling methods (e.g., Giuntoli et al., 2018; Hattermann et al., 2018). Additionally, hydrologic model structure and parameterization, such as the representation of permafrost and glacier affect the streamflow projections (Bring et al., 2016). Generally, uncertainties due GCM structure and greenhouse gas concentration and anthropogenic forcings are considered to be the most important sources of uncertainties, and it is a common practice to incorporate multiple RCP or emissions scenarios, and an ensemble of GCMs in projecting hydrologic impacts of climate change (e.g., Shrestha et al., 2014; 2019; MacDonald et al., 2018).

Besides basin-scale hydrologic models, global/regional hydrology models or land surface models (e.g., Schewe et al., 2014; Bring et al., 2017; Gädeke et al., 2020) provide CMIP5 GCM driven streamflow simulations over historical and future periods in northern Canada. These models are typically of coarser resolution (~0.5°), and unlike basin-scale hydrologic models they are generally not calibrated to reproduce the basin specific streamflow, which can lead to considerable uncertainty. Likewise, for the northern Canada domain, RCM land surface simulations also provide runoff variables, for example, RCMs that participated in the Coordinated Regional Climate Downscaling Experiment (CORDEX) (Giorgi et al., 2009) over North America (NA-CORDEX: na-cordex.org) and Arctic (Arctic-CORDEX: https://climate-cryosphere.org/activities/polar-cordex/arctic). CORDEX has both hindcast (ERA-Interim and GCM-driven historical simulations) and CMIP5 scenarios (RCP4.5, RCP8.5 simulations), with a spatial resolution of 0.22°(~25 km) or 0.44°(~50 km). Some GCMs/RCMs have been coupled with routing schemes to simulated streamflow, for example, projections of 2 °C vs. high warming flood-generating mechanisms across Canada (Teufel and Sushama, 2021), simulation of future changes in flood hazards across Canada (Gaur et al., 2018). However, as in the case of other variables, bias in the RCMs also affects the streamflow simulations. Nevertheless, the simulations from the global/regional hydrology models and RCMs offer useful information regarding the direction of climate change driven streamflow changes over large areas.

References - Streamflow data:

Arnold, J.G., R. Srinivasan, R.S. Muttiah, and J.R. Williams, 1998: Large Area Hydrologic Modeling and Assessment Part I: Model Development1. JAWRA Journal of the American Water Resources Association, 34(1), 73-89, https://doi.org/10.1111/j.1752-1688.1998.tb05961.x.

Arnold, J.G., D.N. Moriasi, P.W. Gassman, K.C. Abbaspour, M.J. White, R. Srinivasan, C. Santhi, R.D. Harmel, A. van Griensven, M.W. Van Liew, N. Kannan and M.K. Jha, 2012: SWAT: Model use, calibration, and validation. Transactions of the ASABE, 55(4), 1491-1508., https://doi.org/10.13031/2013.42256.

Bring, A., A. Shiklomanov, and R.B. Lammers, 2017: Pan-Arctic river discharge: Prioritizing monitoring of future climate change hot spots. Earth's Future, 5(1), 72-92, https://doi.org/10.1002/2016EF000434.

Cannon, A.J., S.R. Sobie, and T.Q. Murdock, 2015: Bias Correction of GCM Precipitation by Quantile Mapping: How Well Do Methods Preserve Changes in Quantiles and Extremes? Journal of Climate, 28(17), 6938-6959, https://doi.org/10.1175/JCLI-D-14-00754.1.

Gädeke, A., V. Krysanova, A. Aryal, J. Chang, M. Grillakis, N. Hanasaki, A. Koutroulis, Y. Pokhrel, Y. Satoh, S. Schaphoff, H.M. Schmied, T. Stacke, Q. Tang, Y. Wada, and K. Thonicke, 2020: Performance evaluation of global hydrological models in six large Pan-Arctic watersheds. Climatic Change, 163(3), 1329-1351, https://doi.org/10.1007/s10584-020-02892-2.

Gaur, A., A. Gaur, and S.P. Simonovic, 2018: Future Changes in Flood Hazards across Canada under a Changing Climate. Water, 10(10), 1441., https://doi.org/10.3390/w10101441.

Giorgi, F., C. Jones, and G.R. Asrar, 2009: Addressing climate information needs at the regional level: the CORDEX framework. World Meteorological Organization (WMO) Bulletin, 58(3), 175.

Hamman, J.J., B. Nijssen, T.J. Bohn, D.R. Gergel, and Y. Mao, 2018: The Variable Infiltration Capacity model version 5 (VIC-5): infrastructure improvements for new applications and reproducibility. Geoscientific Model Development, 11(8), 3481-3496, https://doi.org/10.5194/gmd-11-3481-2018.

Liang, X., D.P. Lettenmaier, E.F. Wood, and S.J. Burges, 1994: A simple hydrologically based model of land surface water and energy fluxes for general circulation models. Journal of Geophysical Research: Atmospheres, 99(D7), 14415-14428, https://doi.org/10.1029/94JD00483.

MacDonald, M.K., T.A. Stadnyk, S.J. Déry, M. Braun, D. Gustafsson, K. Isberg, and B. Arheimer, 2018: Impacts of 1.5 and 2.0 °C Warming on Pan-Arctic River Discharge Into the Hudson Bay Complex Through 2070. Geophysical Research Letters, 45(15), 7561-7570, https://doi.org/10.1029/2018GL079147.

Marsh, C. B., J.W. Pomeroy, and H.S. Wheater, 2020: The Canadian Hydrological Model (CHM) v1.0: a multi-scale, multi-extent, variable-complexity hydrological model – design and overview, Geoscientific Model Development, 13(1), 225-247, https://doi.org/10.5194/gmd-13-225-2020.

Pietroniro, A.,V. Fortin, N. Kouwen, C. Neal, R. Turcotte, B. Davison, D. Verseghy, E.D. Soulis,R. Caldwell, N. Evora, and P. Pellerin, 2007: Development of the MESH modelling system for hydrological ensemble forecasting of the Laurentian Great Lakes at the regional scale. Hydrology and Earth System Sciences, 11(4), 1279-1294, https://doi.org/10.5194/hess-11-1279-2007.

Scheepers, H., J. Wang, T.Y. Gan, and C.C. Kuo, 2018: The impact of climate change on inland waterway transport: Effects of low water levels on the Mackenzie River. Journal of Hydrology, 566, 285–298, https://doi.org/10.1016/j.jhydrol.2018.08.059.

Schewe, J., J. Heinke, D. Gerten, I. Haddeland, N.W. Arnell, D.B. Clark, R. Dankers, S. Eisner, B.M. Fekete, F.J. Colón-González, S.N. Gosling, H. Kim, X. Liu, Y. Masaki, F.T. Portmann, Y. Satoh, T. Stacke, Q. Tang, Y. Wada, D. Wisser, T. Albrecht, K. Frieler, F. Piontek, L. Warszawski, and P. Kabat, 2014: Multimodel assessment of water scarcity under climate change. Proceedings of the National Academy of Sciences, 111(9), 3245-3250, https://doi.org/10.1073/pnas.1222460110.

Schoeneberg, A.T., and M.A. Schnorbus, 2021: Exploring the Strength and Limitations of PCIC’s CMIP5 Hydrologic Scenarios. Pacific Climate Impacts Consortium, University of Victoria, https://www.pacificclimate.org/sites/default/files/publications/Revised_Hydro_Scenarios_ENV_Water_Use_Allocation_Report_21Jun2021.pdf.

Shrestha, R.R., A.J. Cannon, M.A. Schnorbus, and H. Alford, 2019: Climatic Controls on Future Hydrologic Changes in a Subarctic River Basin in Canada. Journal of Hydrometeorology, 20(9), 1757-1778., https://doi.org/10.1175/JHM-D-18-0262.1.

Shrestha, R.R., M.A. Schnorbus, A.T. Werner, F.W. Zwiers, 2014: Evaluating hydroclimatic change signals from statistically and dynamically downscaled GCMs and hydrologic models. Journal of Hydrometeorology, 15(2), 844-860. doi:10.1175/JHM-D-13-030.1.

Stadnyk, T. A., A. Tefs, M. Broesky, S.J. Déry, P.G. Myers, N.A. Ridenour, K. Koenig, L. Vonderbank, and D. Gustafssonand, 2021: Changing freshwater contributions to the Arctic: A 90-year trend analysis (1981–2070). Elementa: Science of the Anthropocene, 9(1), 00098, https://doi.org/10.1525/elementa.2020.00098.

Teufel, B., and L. Sushama, 2021: 2 °C vs. High Warming: Transitions to Flood-Generating Mechanisms across Canada. Water, 13(11), 1494, https://doi.org/10.3390/w13111494.

Troin, M., R. Arsenault, J.-L. Martel, and F. Brissette, 2018: Uncertainty of Hydrological Model Components in Climate Change Studies over Two Nordic Quebec Catchments. Journal of Hydrometeorology, 19(1), 27-46., https://doi.org/10.1175/JHM-D-17-0002.1.

Wang, J., N. K. Shrestha, M.A. Delavar, T.W. Meshesha, and S.N. Bhanja, 2021: Modelling Watershed and River Basin Processes in Cold Climate Regions: A Review. Water, 13(4), 518. https://doi.org/10.3390/w13040518.

Werner, A.T., M.A. Schnorbus, R.R. Shrestha, A.J. Cannon, F.W. Zwiers, G. Dayon, and F. Anslow, 2019: A long-term, temporally consistent, gridded daily meteorological dataset for northwestern North America. Scientific Data, 6(1), 1-16., https://doi.org/10.1038/sdata.2018.299.