Abstract

A numerical atmospheric boundary layer model, based on higher-order turbulence closure assumptions, is developed and used to simulate the local advection of momentum, heat, and moisture during the melt of patchy snow covers over a 10-km horizontal domain. The coupled model includes solution of the mass continuity equation, the horizontal and vertical momentum equations, an E−ε turbulence model, an energy equation, and a water vapor conservation equation. Atmospheric buoyancy is accounted for, and a land surface energy balance model is implemented at the lower boundary.

Model integrations indicate that advective processes occurring at local scales produce nonlinear horizontal variations in surface fluxes. Under conditions of the numerical experiments, the energy available to melt snow-covered regions has been found to increase by as much as 30% as the area of exposed vegetation increases upwind of the snow cover. The melt increase is found to vary in a largely linear fashion with decreasing snow-covered area for snow-covered areas greater than 25% and in a strongly nonlinear fashion below that value. Decreasing the ratio of patch size to total area, or increasing the patchiness, of the snow cover also leads to nonlinear increases in the energy available to melt the snow. In the limit of a snow cover composed of small patches, melt energy is found to increase linearly as the fractional snow-covered area decreases. In addition, for the purpose of computing grid-average surface fluxes during snowmelt in regional atmospheric models, the results of this study indicate that separate energy balance computations can be performed over the snow-covered and vegetation-covered regions, and the resulting fluxes can be weighted in proportion to the fractional snow cover to allocate the total energy flux partitioning within each surface grid cell.

Abstract

Local, regional, and global atmospheric, hydrologic, and ecologic models used to simulate weather, climate, land surface moisture, and vegetation processes all commonly represent their computational domains by a collection of finite areas or grid cells. Within each of these cells three fundamental features are required to describe the evolution of seasonal snow cover from the end of winter through spring melt. These three features are 1) the within-grid snow water equivalent (SWE) distribution, 2) the gridcell melt rate, and 3) the within-grid depletion of snow-covered area. This paper defines the exact mathematical interrelationships among these three features and demonstrates how knowledge of any two of them allows generation of the third. During snowmelt, the spatially variable subgrid SWE depth distribution is largely responsible for the patchy mosaic of snow and vegetation that develops as the snow melts. Applying the melt rate to the within-grid snow distribution leads to the exposure of vegetation, and the subgrid-scale vegetation exposure influences the snowmelt rate and the grid-averaged surface fluxes. By using the developed interrelationships, the fundamental subgrid-scale features of the seasonal snow cover evolution and the associated energy and moisture fluxes can be simulated using a combination of remote sensing products that define the snow-covered area evolution and a submodel that appropriately handles the snowmelt computation. Alternatively, knowledge of the subgrid SWE distribution can be used as a substitute for the snow-covered area information.

Abstract

To improve the depiction of autumn through spring land–atmosphere interactions and feedbacks within regional and global weather, climate, and hydrologic models, a Subgrid SNOW Distribution (SSNOWD) submodel that explicitly includes subgrid snow-depth and snow-cover variability has been developed. From both atmospheric and hydrologic perspectives, the subgrid snow-depth distribution is an important quantity to account for within large-scale models. In the natural system, these subgrid snow-depth distributions are largely responsible for the mosaic of snow-covered and snow-free areas that develop as the snow melts, and the impacts of these fractional areas must be quantified in order to realistically simulate grid-averaged surface fluxes. SSNOWD's formulation incorporates observational studies showing that snow distributions can be described by a lognormal distribution and the snow-depth coefficient of variation. Using an understanding of the physical processes that lead to these observed snow-depth variations, a global distribution of nine subgrid snow-depth-variability categories was developed, and coefficient-of-variation values were assigned to each category based on published measurements. In addition, SSNOWD adopts the physically realistic approach of performing separate surface-energy-balance calculations over the snow-covered and snow-free portions of each model grid cell and weighing the resulting fluxes according to these fractional areas. Using a climate version of the Regional Atmospheric Modeling System (ClimRAMS) over a North American domain, SSNOWD was compared with a snow-cover formulation similar to those currently used in most general circulation models. The simulations indicated that accounting for snow-distribution variability has a significant impact on snow-cover evolution and associated energy and moisture fluxes.

Abstract

A methodology for assimilating ground-based and remotely sensed snow data within a snow-evolution modeling system (SnowModel) is presented. The data assimilation scheme (SnowAssim) is consistent with optimal interpolation approaches in which the differences between the observed and modeled snow values are used to constrain modeled outputs. The calculated corrections are applied retroactively to create improved fields prior to the assimilated observations. Thus, one of the values of this scheme is the improved simulation of snow-related distributions throughout the entire snow season, even when observations are only available late in the accumulation and/or ablation periods. Because of this, the technique is particularly applicable to reanalysis applications. The methodology includes the ability to stratify the assimilation into regions where either the observations and/or model has unique error properties, such as the differences between forested and nonforested snow environments. The methodologies are introduced using synthetic data and a simple simulation domain. In addition, the model is applied over NASA’s Cold Land Processes Experiment (CLPX), Rabbit Ears Pass, Colorado, observation domain. Simulations using the data assimilation scheme were found to improve the modeled snow water equivalent (SWE) distributions, and simulated SWE displayed considerably more realistic spatial heterogeneity than that provided by the observations alone.

Abstract

SnowModel is a spatially distributed snow-evolution modeling system designed for application in landscapes, climates, and conditions where snow occurs. It is an aggregation of four submodels: MicroMet defines meteorological forcing conditions, EnBal calculates surface energy exchanges, SnowPack simulates snow depth and water-equivalent evolution, and SnowTran-3D accounts for snow redistribution by wind. Since each of these submodels was originally developed and tested for nonforested conditions, details describing modifications made to the submodels for forested areas are provided. SnowModel was created to run on grid increments of 1 to 200 m and temporal increments of 10 min to 1 day. It can also be applied using much larger grid increments, if the inherent loss in high-resolution (subgrid) information is acceptable. Simulated processes include snow accumulation; blowing-snow redistribution and sublimation; forest canopy interception, unloading, and sublimation; snow-density evolution; and snowpack melt. Conceptually, SnowModel includes the first-order physics required to simulate snow evolution within each of the global snow classes (i.e., ice, tundra, taiga, alpine/mountain, prairie, maritime, and ephemeral). The required model inputs are 1) temporally varying fields of precipitation, wind speed and direction, air temperature, and relative humidity obtained from meteorological stations and/or an atmospheric model located within or near the simulation domain; and 2) spatially distributed fields of topography and vegetation type. SnowModel’s ability to simulate seasonal snow evolution was compared against observations in both forested and nonforested landscapes. The model closely reproduced observed snow-water-equivalent distribution, time evolution, and interannual variability patterns.

Abstract

In many applications, a realistic description of air temperature inversions is essential for accurate snow and glacier ice melt, and glacier mass-balance simulations. A physically based snow evolution modeling system (SnowModel) was used to simulate 8 yr (1998/99–2005/06) of snow accumulation and snow and glacier ice ablation from numerous small coastal marginal glaciers on the SW part of Ammassalik Island in SE Greenland. These glaciers are regularly influenced by inversions and sea breezes associated with the adjacent relatively low temperature and frequently ice-choked fjords and ocean. To account for the influence of these inversions on the spatiotemporal variation of air temperature and snow and glacier melt rates, temperature inversion routines were added to MircoMet, the meteorological distribution submodel used in SnowModel. The inversions were observed and modeled to occur during 84% of the simulation period. Modeled inversions were defined not to occur during days with strong winds and high precipitation rates because of the potential of inversion breakup. Field observations showed inversions to extend from sea level to approximately 300 m MSL, and this inversion level was prescribed in the model simulations. Simulations with and without the inversion routines were compared. The inversion model produced air temperature distributions with warmer lower-elevation areas and cooler higher-elevation areas than without inversion routines because of the use of cold sea-breeze-based temperature data from underneath the inversion. This yielded an up to 2 weeks earlier snowmelt in the lower areas and up to 1–3 weeks later snowmelt in the higher-elevation areas of the simulation domain. Averaged mean annual modeled surface mass balance for all glaciers (mainly located above the inversion layer) was −720 ± 620 mm w.eq. yr⁻¹ (w.eq. is water equivalent) for inversion simulations, and −880 ± 620 mm w.eq. yr⁻¹ without the inversion routines, a difference of 160 mm w.eq. yr⁻¹. The annual glacier loss for the two simulations was 50.7 × 10⁶ and 64.4 × 10⁶ m³ yr⁻¹ for all glaciers—a difference of ∼21%. The average equilibrium line altitude (ELA) for all glaciers in the simulation domain was located at 875 and 900 m MSL for simulations with or without inversion routines, respectively.

Abstract

Arctic snow presence, absence, properties, and water amount are key components of Earth’s changing climate system that incur far-reaching physical and biological ramifications. Recent dataset and modeling developments permit relatively high-resolution (10-km horizontal grid; 3-h time step) pan-Arctic snow estimates for 1979–2009. Using MicroMet and SnowModel in conjunction with land cover, topography, and 30 years of the NASA Modern-Era Retrospective Analysis for Research and Applications (MERRA) atmospheric reanalysis data, a distributed snow-related dataset was created including air temperature, snow precipitation, snow-season timing and length, maximum snow water equivalent (SWE) depth, average snow density, snow sublimation, and rain-on-snow events. Regional variability is a dominant feature of the modeled snow-property trends. Both positive and negative regional trends are distributed throughout the pan-Arctic domain, featuring, for example, spatially distinct areas of increasing and decreasing SWE or snow season length. In spite of strong regional variability, the data clearly show a general snow decrease throughout the Arctic: maximum winter SWE has decreased, snow-cover onset is later, the snow-free date in spring is earlier, and snow-cover duration has decreased. The domain-averaged air temperature trend when snow was on the ground was 0.17°C decade⁻¹ with minimum and maximum regional trends of −0.55° and 0.78°C decade⁻¹, respectively. The trends for total number of snow days in a year averaged −2.49 days decade⁻¹ with minimum and maximum regional trends of −17.21 and 7.19 days decade⁻¹, respectively. The average trend for peak SWE in a snow season was −0.17 cm decade⁻¹ with minimum and maximum regional trends of −2.50 and 5.70 cm decade⁻¹, respectively.

Abstract

A gridded linear-reservoir runoff routing model (HydroFlow) was developed to simulate the linkages between runoff production from land-based snowmelt and icemelt processes and the associated freshwater fluxes to downstream areas and surrounding oceans. HydroFlow was specifically designed to account for glacier, ice sheet, and snow-free and snow-covered land applications. Its performance was verified for a test area in southeast Greenland that contains the Mittivakkat Glacier, the local glacier in Greenland with the longest observed time series of mass-balance and ice-front fluctuations. The time evolution of spatially distributed gridcell runoffs required by HydroFlow were provided by the SnowModel snow-evolution modeling system, driven with observed atmospheric data, for the years 2003 through 2010. The spatial and seasonal variations in HydroFlow hydrographs show substantial correlations when compared with observed discharge coming from the Mittivakkat Glacier area and draining into the adjacent ocean. As part of its discharge simulations, HydroFlow creates a flow network that links the individual grid cells that make up the simulation domain. The collection of networks that drain to the ocean produced a range of runoff values that varied most strongly according to catchment size and percentage and elevational distribution of glacier cover within each individual catchment. For 2003–10, the average annual Mittivakkat Glacier region runoff period was 200 ± 20 days, with a significant increase in annual runoff over the 8-yr study period, both in terms of the number of days (30 days) and in volume (54.9 × 10⁶ m³).

Abstract

Runoff magnitudes, the spatial patterns from individual Greenland catchments, and their changes through time (1960–2010) were simulated in an effort to understand runoff variations to adjacent seas and to illustrate the capability of SnowModel (a snow and ice evolution model) and HydroFlow (a runoff routing model) to link variations in terrestrial runoff with ocean processes and other components of Earth’s climate system. Significant increases in air temperature, net precipitation, and local surface runoff lead to enhanced and statistically significant Greenland ice sheet (GrIS) surface mass balance (SMB) loss. Total Greenland runoff to the surrounding oceans increased 30%, averaging 481 ± 85 km³ yr⁻¹. Averaged over the period, 69% of the runoff to the surrounding seas originated from the GrIS and 31% came from outside the GrIS from rain and melting glaciers and ice caps. The runoff increase from the GrIS was due to an 87% increase in melt extent, 18% from increases in melt duration, and a 5% decrease in melt rates (87% + 18% − 5% = 100%). In contrast, the runoff increase from the land area surrounding the GrIS was due to a 0% change in melt extent, a 108% increase in melt duration, and an 8% decrease in melt rate. In general, years with positive Atlantic multidecadal oscillation (AMO) index equaled years with relatively high Greenland runoff volume and vice versa. Regionally, runoff was greater from western than eastern Greenland. Since 1960, the data showed pronounced runoff increases in west Greenland, with the greatest increase occurring in the southwest and the lowest increase in the northwest.

Abstract

A vegetation-protruding-above-snow parameterization for earth system models was developed to improve energy budget calculations of interactions among vegetation, snow, and the atmosphere in nonforested areas. These areas include shrublands, grasslands, and croplands, which represent 68% of the seasonally snow-covered Northern Hemisphere land surface (excluding Greenland). Snow depth observations throughout nonforested areas suggest that mid- to late-winter snowpack depths are often comparable or lower than the vegetation heights. As a consequence, vegetation protruding above the snow cover has an important impact on snow-season surface energy budgets. The protruding vegetation parameterization uses disparate energy balances for snow-covered and protruding vegetation fractions of each model grid cell, and fractionally weights these fluxes to define grid-average quantities. SnowModel, a spatially distributed snow-evolution modeling system, was used to test and assess the parameterization. Simulations were conducted during the winters of 2005/06 and 2006/07 for conditions of 1) no protruding vegetation (the control) and 2) with protruding vegetation. The spatial domain covered Colorado, Wyoming, and portions of the surrounding states; 81% of this area is nonforested. The surface net radiation, energy, and moisture fluxes displayed considerable differences when protruding vegetation was included. For shrubs, the net radiation, sensible, and latent fluxes changed by an average of 12.7, 6.9, and −22.7 W m⁻², respectively. For grass and crops, these fluxes changed by an average of 6.9, −0.8, and −7.9 W m⁻², respectively. Daily averaged flux changes were as much as 5 times these seasonal averages. As such, the new parameterization represents a major change in surface flux calculations over more simplistic and less physically realistic approaches.