• Baker, D. G., , R. H. Skaggs, , and D. L. Ruschy, 1991: Snow depth required to mask the underlying surface. J. Appl. Meteor., 30 , 387392.

    • Search Google Scholar
    • Export Citation
  • Baker, D. G., , D. L. Ruschy, , R. H. Skaggs, , and D. B. Wall, 1992: Air temperature and radiation depressions associated with a snow cover. J. Appl. Meteor., 31 , 247254.

    • Search Google Scholar
    • Export Citation
  • Barnes, S. L., 1973: Mesoscale objective map analysis using weighted time series observations. NOAA Tech. Memo. ERL NSSL-62, National Severe Storms Laboratory, Norman, OK, 60 pp.

  • Bromwich, D. H., , R. L. Fogt, , K. I. Hodges, , and J. E. Walsh, 2007: A tropospheric assessment of the ERA-40, NCEP, and JRA-25 global reanalyses in the polar regions. J. Geophys. Res., 112 .D10111, doi:10.1029/2006JD007859.

    • Search Google Scholar
    • Export Citation
  • Brown, R. D., 2000: Northern Hemisphere snow cover variability and change, 1915–97. J. Climate, 13 , 23392355.

  • Brown, R. D., , and R. O. Braaten, 1998: Spatial and temporal variability of Canadian monthly snow depths, 1946–1995. Atmos.–Ocean, 36 , 3745.

    • Search Google Scholar
    • Export Citation
  • Dewey, K. F., 1977: Daily maximum and minimum temperature forecasts and the influence of snow cover. Mon. Wea. Rev., 105 , 15941597.

  • Dirnhirm, I., , and F. D. Eaton, 1975: Some characteristics of the albedo of snow. J. Appl. Meteor., 14 , 375379.

  • Durre, I., , and J. M. Wallace, 2001: The warm season dip in diurnal temperature range over the eastern United States. J. Climate, 14 , 354360.

    • Search Google Scholar
    • Export Citation
  • Dyer, J., , and T. L. Mote, 2006: Spatial variability and trends in observed snow depth over North America. Geophys. Res. Lett., 33 .L16503, doi:10.1029/2006GL027258.

    • Search Google Scholar
    • Export Citation
  • Ellis, A. W., , and D. J. Leathers, 1998: A quantitative approach to evaluating the effects of snow cover on cold airmass temperatures across the U.S. Great Plains. Wea. Forecasting, 13 , 688701.

    • Search Google Scholar
    • Export Citation
  • Ellis, A. W., , and D. J. Leathers, 1999: Analysis of cold airmass temperature modification across the U.S. Great Plains as a consequence of snow depth and albedo. J. Appl. Meteor., 38 , 696711.

    • Search Google Scholar
    • Export Citation
  • Frei, A., , and D. A. Robinson, 1999: Northern Hemisphere snow extent: Regional variability 1972–1994. Int. J. Climatol., 19 , 15351560.

    • Search Google Scholar
    • Export Citation
  • Groisman, P. Ya, , T. R. Karl, , and R. W. Knight, 1994: Observed impact of snow cover on the heat balance and the rise of continental spring temperatures. Science, 263 , 198200.

    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77 , 437471.

  • Kanamitsu, M., , W. Ebisuzaki, , J. Woollen, , S-K. Yang, , J. J. Hnilo, , M. Fiorino, , and G. L. Potter, 2002: NCEP–DOE AMIP-II Reanalysis (R-2). Bull. Amer. Meteor. Soc., 83 , 16311643.

    • Search Google Scholar
    • Export Citation
  • Kistler, R., and Coauthors, 2001: The NCEP–NCAR 50-Year Reanalysis: Monthly means CD-ROM and documentation. Bull. Amer. Meteor. Soc., 82 , 247267.

    • Search Google Scholar
    • Export Citation
  • Leathers, D. J., , A. W. Ellis, , and D. A. Robinson, 1995: Characteristics of temperature depressions associated with snow cover across the northeast United States. J. Appl. Meteor., 34 , 381390.

    • Search Google Scholar
    • Export Citation
  • Namias, J., 1963: Surface–atmosphere interactions as fundamental causes of drought and other climatic fluctuations. Proc. Rome Symp., Changes of Climate, Arid Zone Research, Rome, Italy, UNESCO and WMO, 345–359.

  • Namias, J., 1985: Some empirical evidence for the influence of snow cover on temperature and precipitation. Mon. Wea. Rev., 113 , 15421553.

    • Search Google Scholar
    • Export Citation
  • Robinson, D. A., 1989: Construction of a United States historical snow data base. Proc. 45th Eastern Snow Conf., Lake Placid, NY, Eastern Snow Conference, 50–59.

  • Robinson, D. A., , and G. Kukla, 1985: Maximum surface albedo of seasonally snow-covered lands in the Northern Hemisphere. J. Climate Appl. Meteor., 24 , 402411.

    • Search Google Scholar
    • Export Citation
  • Robinson, D. A., , and A. Frei, 1997: A Northern Hemisphere snow cover climatology using satellite information. Proc. 10th Conf. on Applied Climatology, Reno, NV, Amer. Meteor. Soc., 24–28.

  • Robinson, D. A., , K. F. Dewey, , and R. R. Heim, 1993: Global snow cover monitoring: An update. Bull. Amer. Meteor. Soc., 74 , 16891696.

  • Serreze, M. C., , J. R. Key, , J. E. Box, , J. A. Maslanik, , and K. Steffen, 1998: A new monthly climatology of global radiation for the Arctic and comparisons with NCEP–NCAR reanalysis and ISCCP-C2 fields. J. Climate, 11 , 121136.

    • Search Google Scholar
    • Export Citation
  • Shepard, D., 1968: A two-dimensional interpolation function for irregularly-spaced data. Proc. 23rd ACM National Conf., Association of Computing Machinery, 517–524.

  • Strack, J. E., , R. A. Pielke Sr., , and J. Adegoke, 2003: Sensitivity of model-generated daytime surface heat fluxes over snow to land-cover changes. J. Hydrometeor., 4 , 2442.

    • Search Google Scholar
    • Export Citation
  • U.S. Department of Commerce, cited. 2003: TD-3200 Cooperative Summary of the Day. National Climatic Data Center, NOAA. [Available online at http://www.ncdc.noaa.gov.].

  • U.S. Department of Commerce, cited. 2007: DS-3505 Integrated Surface Hourly. Global Hourly Surface Observations National Climatic Data Center, NOAA. [Available online at http://www.ncdc.noaa.gov.].

  • Walsh, J. E., , and B. Ross, 1988: Sensitivity of 30-day dynamical forecasts to continental snow cover. J. Climate, 1 , 739754.

  • Walsh, J. E., , D. R. Tucek, , and M. R. Peterson, 1982: Seasonal snow cover and short-term climatic fluctuations over the United States. Mon. Wea. Rev., 110 , 14741486.

    • Search Google Scholar
    • Export Citation
  • Walsh, J. E., , W. H. Jasperson, , and B. Ross, 1985: Influences of snow cover and soil moisture on monthly air temperature. Mon. Wea. Rev., 113 , 756768.

    • Search Google Scholar
    • Export Citation
  • Willmott, C. J., , C. M. Rowe, , and W. D. Philpot, 1985: Spheremap. University of Delaware. [Available online at http://climate.geog.udel.edu/~climate/html_pages/software2.html.].

  • Wojcik, G. S., , and D. S. Wilks, 1992: Temperature forecast biases associated with snow cover in the Northeast. Wea. Forecasting, 7 , 501506.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    MODIS visible satellite image from 1710 UTC 3 Dec 2006 with that day’s min temperatures (Tmin, °C) and snow depth (cm, contours) from cooperative observing stations across the Midwest. A lower albedo is evident in the forested areas from southeast to central Missouri, including some areas with up to 30 cm of snow.

  • View in gallery

    Relative max air temperature (Tmax, °C) vs 2-cm snow depth increments for days with noon solar elevation angle of 15° to 30° and cloud cover less than 25%. Only locations in the top quartile of max snow-covered albedo are included. Mean values are labeled. Bars show ±1 std error.

  • View in gallery

    As in Fig. 2 but for Tmin.

  • View in gallery

    As in Fig. 2 but for obs Tmax.

  • View in gallery

    Depression of Tmax (Tmax-dep) vs 2-cm snow depth increments for days and locations by quartile of noon solar elevation angle. Only days with cloud cover less than 25% and locations in the top quartile of max snow-covered albedo are included. Mean values are labeled. Bars show ±1 std error.

  • View in gallery

    The Tmax-dep vs 2-cm snow depth increments for cloud cover from 1200 UTC (a) NCAR–NCEP reanalysis and (b) an objective analysis of station cloud fractional coverage. Lines represent cloud cover categories of 0%–25% (solid black), 25%–50% (broken black), 50%–75% (solid gray), and 75%–100% (broken gray). Only days and locations with noon solar elevation angles of 15° to 30° and in the top quartile of max snow-covered albedo are included. Mean values are labeled. Bars show ±1 std error.

  • View in gallery

    The Tmax-dep vs 2-cm snow depth increments for days with Tmax > 0°C (black) or ≤0°C (gray). Only days with noon solar elevation angle of 15° to 30° and locations in the top quartile of maximum snow-covered albedo are included. Bars show ±1 std error.

  • View in gallery

    The Tmax-dep vs 2-cm snow depth increments by quartile of max snow-covered albedo. Only days with noon solar elevation angle of 15° to 30° and cloud cover less than 25% are included. Mean values are labeled. Bars show ±1 std error.

  • View in gallery

    Daily Tmax-dep for snow depth ≥10 cm (numbers) and max snow-covered albedo (shading) on a 1° lat × 1° lon grid. Only days with cloud cover less than 25% and no precipitation are included.

  • View in gallery

    As in Fig. 9 but for daily Tmin depression (Tmin-dep).

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 94 94 13
PDF Downloads 82 82 9

On the Role of Snow Cover in Depressing Air Temperature

View More View Less
  • 1 Climate Research Laboratory, Department of Geography, University of Georgia, Athens, Georgia
© Get Permissions
Full access

Abstract

This study empirically examines the role of snow depth on the depression of air temperature after controlling for effect of temperature changes above the boundary layer. In addition, this study examines the role of cloud cover, solar elevation angle, and maximum snow-covered albedo on the temperature depression due to snow cover. The work uses a new dataset of daily, gridded snow depth, snowfall, and maximum and minimum temperatures for North America from 1960 to 2000 in conjunction with 850-hPa temperature data for the same period from the NCEP–NCAR reanalysis, version 1. The 850-hPa temperatures are used as a control to remove the effect of temperature changes above the boundary layer on surface air temperatures. Findings from an analysis of variance demonstrate that snow cover can result in daily maximum (minimum) temperature depressions on average of 4.5°C (2.6°C) for snow depths greater than 10 cm over the grasslands of central North America, but temperature depressions average only 1.2°C (1.1°C) overall. The temperature depression of snow cover is shown to be reduced by increased cloud cover and decreased maximum albedo, which is indicative of denser forest cover. The role of snow melting on temperature depression is further explored by comparing days with maximum temperatures above or below freezing.

Corresponding author address: Thomas L. Mote, Climate Research Laboratory, Department of Geography, University of Georgia, Athens, GA 30602-2502. Email: tmote@uga.edu

Abstract

This study empirically examines the role of snow depth on the depression of air temperature after controlling for effect of temperature changes above the boundary layer. In addition, this study examines the role of cloud cover, solar elevation angle, and maximum snow-covered albedo on the temperature depression due to snow cover. The work uses a new dataset of daily, gridded snow depth, snowfall, and maximum and minimum temperatures for North America from 1960 to 2000 in conjunction with 850-hPa temperature data for the same period from the NCEP–NCAR reanalysis, version 1. The 850-hPa temperatures are used as a control to remove the effect of temperature changes above the boundary layer on surface air temperatures. Findings from an analysis of variance demonstrate that snow cover can result in daily maximum (minimum) temperature depressions on average of 4.5°C (2.6°C) for snow depths greater than 10 cm over the grasslands of central North America, but temperature depressions average only 1.2°C (1.1°C) overall. The temperature depression of snow cover is shown to be reduced by increased cloud cover and decreased maximum albedo, which is indicative of denser forest cover. The role of snow melting on temperature depression is further explored by comparing days with maximum temperatures above or below freezing.

Corresponding author address: Thomas L. Mote, Climate Research Laboratory, Department of Geography, University of Georgia, Athens, GA 30602-2502. Email: tmote@uga.edu

Keywords: Temperature; Snow; Albedo

1. Introduction

The role of snow cover on lower tropospheric temperatures has taken on increased importance as regions of the interior of North America have witnessed decreased snow cover extent concurrent with increased atmospheric temperatures beginning in the late 1980s (e.g., Robinson et al. 1993; Groisman et al. 1994; Robinson and Frei 1997; Frei and Robinson 1999; Brown 2000; Dyer and Mote 2006). To understand to what extent increased atmospheric temperatures are a result of thinning snow cover, it is critical to understand the role of changing snow depth on lower atmospheric temperatures.

Namias (1963) was among the first to examine the influence of snow cover on boundary layer temperatures. He calculated that a change in surface albedo (from 0.2 to 0.8) with the addition of snow cover would warm a boundary layer 1 km thick by approximately 7.5°C. Walsh et al. (1982) examined temperature departures because of snow cover as a deviation from the expected temperatures for a given 700-hPa height pattern. They found that snow cover accounts for approximately 10%–20% of the variance in monthly temperature over the United States. Large-scale snow extent, measured by the number of degrees of latitude of snow cover, accounted for 1°–2°C in mean monthly temperatures for locations across the interior of the conterminous United States, according to Walsh et al. (1985). Dewey (1977) and Wojcik and Wilks (1992) examined the temperature depression because of snow cover as a forecast bias problem. Dewey (1977) examined the forecast error of model output statistics (MOS) for one event in 1977 over South Dakota and found temperature depressions of 6°–10°C in maximum temperatures and 4°–5°C in minimum temperatures. While examining the effect of snow cover on forecasts using the National Center for Atmospheric Research (NCAR) Community Forecast Model, Walsh and Ross (1988) found that surface air temperatures depressed 5°–10°C in regions with greater snow cover.

Baker et al. (1992) examined surface temperatures and radiation components during 11 winters at a University of Minnesota station in St. Paul. Baker et al. (1992) noted that the temperature depression was driven primarily by the increased outgoing shortwave radiation because of the higher albedo of the snow-covered surface. They concluded that snow cover of less than 10 cm depressed 2-m air temperatures by 6.4°C on average while a snow depth greater than 10 cm depressed temperatures by 8.4°C. Leathers et al. (1995) conducted one of the few studies outside the plains, identifying a 6°C depression in maximum temperatures and 5°C depression in minimum temperatures on days with at least 2.5 cm of snow cover in the northeast United States. The primary limitation of both studies was that they did not isolate whether the temperature differences were solely because of snow cover or partially as a result of the tendency for snow cover to coincide with days that would be cold otherwise. Ellis and Leathers (1998, 1999) examined how snow cover modifies airmass temperatures through the use of a snowpack energy and mass balance model. By examining temperatures in simulations with snow cover removed, they found daily maximum temperatures 10°–15°C higher and minimum temperatures 1°–2°C higher. Moreover, Ellis and Leathers (1999) found little effect on air temperature comparing a snowpack of 2.5-cm depth versus much greater snow depths. Strack et al. (2003) found differences in modeled sensible heat flux for different land cover types, all with deep snow, as large as 80 W m−2 using the Regional Atmospheric Modeling System (RAMS). The modeled differences in sensible heat flux could lead to differences in daytime temperatures of as much as 6°C (Strack et al. 2003).

In addition to the literature cited, numerous anecdotal examples can be given to illustrate the role of snow cover in depressing air temperatures. One clear example from the recent past (3 December 2006) is shown in Fig. 1. The Moderate Resolution Imaging Spectroradiometer (MODIS) satellite image shown in the figure was taken after the 30 November–1 December 2006 storm that left more than 20 cm of snow over a large swath of Illinois and Missouri. Maximum temperatures on 3 December ranged from −3° to +7°C immediately west of the snow cover and 0° to +9°C in the cloud-free areas east of the snow (Fig. 1). The maximum (minimum) temperature was an average 5.0°C (2.8°C) lower at stations with 20 cm or more of snow cover than at stations without snow cover (Table 1). (Some of the 637 cooperative stations included in Table 1 means were not plotted on Fig. 1 to improve legibility.) While independent sample t tests show significant differences among the means in Table 1 (α = 0.01), the large standard deviations demonstrate that it is difficult to quantify the affect of snow depth on air temperature using a case study approach. Clearly, a more systematic, climatological approach is warranted to address this issue.

This observational study examines the depression of daily mean maximum and minimum surface air temperatures across North America using a unique dataset of gridded snow depths and temperatures. In a departure from much of the previous work, this research incorporates upper-air data in an attempt to isolate the effect of snow cover on maximum and minimum air temperature from any synoptic-scale forcing. This research attempts to address the following questions: What is the effect of snow cover on surface air temperatures when isolated from synoptic-scale influences? How does this effect vary across from the plains to the more densely vegetated eastern North America?

2. Data

Daily surface observations of maximum and minimum temperature, snow depth, snowfall, and precipitation from the United States (U.S. Department of Commerce 2003) and Meteorological Service of Canada (Brown and Braaten 1998; R. Brown 2006, personal communication) are used in this research. All observations were subjected to a quality control (QC) methodology proposed by Robinson (1989). The Robinson QC routine involves comparing a time series of daily snow depth measurements with associated daily snowfall, maximum and minimum temperature, and precipitation data from the same station. The quality control also checks the snow depth data before and after the day tested. Additionally, if the temperature, precipitation, or snowfall data did not meet their own quality control criteria, they were not used to check the consistency of the snow depth observations. Any data that did not meet the quality control criteria were not included in the analysis. Approximately 4.5% of the data was marked as missing in the original datasets or removed by the quality control routine.

All available maximum and minimum temperatures, snow depth, snowfall, and precipitation observations were interpolated to daily 0.25° latitude × 0.25° longitude grids using an inverse-distance spherical interpolation called Spheremap (Willmott et al. 1985), adopted from Shepard’s (1968) Cartesian-based algorithm. The spatial extent of the grids was chosen to include the United States and continental Canada north of 20°N, although only regions north of 35°N are used in this study. Daily 1° latitude × 1° longitude grids were resampled from the quarter-degree grids. The intermediate quarter-degree grids were created to minimize the effect of missing snow depth observations on interpolated grid values. A more complete description of the gridded dataset is available in Dyer and Mote (2006).

National Centers for Environmental Prediction–NCAR (NCEP–NCAR) reanalysis, version 1, data were acquired from the National Oceanic and Atmospheric Administration–Cooperative Institute for Research in Environmental Sciences (NOAA–CIRES) Climate Diagnostics Center (Kalnay et al. 1996; http://www.cdc.noaa.gov). The 850-hPa air temperature and surface pressure data are available at 6-h intervals, but only the 1200 UTC data were extracted. The 850-hPa temperature and surface pressure data are available on a 2.5° latitude × 2.5° longitude grid and were interpolated to the same 1° latitude × 1° longitude grid used for the cooperative station data. Total cloud cover data, expressed as a percent cover, are available on the T62 Gaussian grid (approximately 2° resolution) and were also resampled to a 1° latitude × 1° longitude grid. All resampling of the reanalysis data was done using NOAA’s Ferret software (http://ferret.pmel.noaa.gov).

This NCEP–NCAR reanalysis product used a climatological snow cover prior to 1967 (Kistler et al. 2001). Furthermore, the reanalysis product had a problem where the 1973 snow cover was used for 1974–94. This problem primarily affected “near surface land temperatures during transition seasons” (Kanamitsu et al. 2002). Over the midlatitudes in September–November, the effect of the snow cover problem is approximately 1.6°C (Kanamitsu et al. 2002). The work presented here did not use surface temperature fields from the reanalysis. Because the effect is primarily on near-surface temperatures, rather than 850-hPa temperatures, this should not constitute a problem in the present study.

Another documented problem in the NCEP–NCAR reanalysis product is with the cloud cover fraction in polar environments (Serreze et al. 1998; Bromwich et al. 2007). Because of concerns regarding the cloud cover fraction in the NCEP–NCAR reanalysis product, a second cloud cover dataset was included in this study. The National Climatic Data Center DS3505 aviation routine weather report (METAR) data were used, and the 1200 UTC total cloud fractions from approximately 650 stations (actual number varied over time) were extracted for North America for 1977–2000 (U.S. Department of Commerce 2007). The data were gridded to a 1° latitude × 1° longitude grid using a Barnes objective analysis (Barnes 1973).

3. Methods

Ample anecdotal evidence and numerous studies show that daily maximum and minimum surface temperatures are depressed by the presence of snow cover. However, other factors must be considered when examining the influence of snow on air temperature on a climatological basis. First and foremost, snow cover is more likely to be present on days that would be cold even without snow cover. This study attempts to remove the effect of temperature changes above the boundary layer on the surface air temperature depression because of snow cover. Furthermore, this study addresses the question of how much of the air temperature depression shown in Baker et al. (1992) and Leathers et al. (1995) is because of lower air temperatures above the boundary layer and not the presence of snow cover.

This study attempts to isolate variations in 2-m air temperature that are caused by changes in snow depth rather than changes above the boundary layer. Relative temperatures were calculated from the daily maximum and minimum surface air temperatures at each grid cell using the 1200 UTC 850-hPa temperatures and geopotential heights resampled to the 1° latitude × 1° longitude grid. The dry adiabatic lapse rate was multiplied by the 850-hPa geopotential height above the height of the surface. That product was added to the 850-hPa temperature. The resulting reference temperature was then subtracted from the corresponding maximum (minimum) surface temperature, resulting in the relative maximum (minimum) temperature. This is identical to the approach used by Durre and Wallace (2001). An increase (decrease) in relative temperature, therefore, represents a warming (cooling) of the surface relative to the atmospheric boundary layer (Durre and Wallace 2001). As a simple forecast tool, this relative temperature may be used to approximate the daily maximum surface temperature. Here the concern is not to forecast the temperature, but simply to control for variations in air temperature above the boundary layer.

Results are shown as the relative maximum (minimum) temperatures versus snow depth. The plots incorporate data for all available 1° latitude × 1° longitude grid cells and all days, subject to restrictions regarding cloud cover fraction, solar elevation angle, and maximum snow-covered albedo. Snow depths are plotted to a maximum 30 cm, in most instances. Data were limited for snow depths greater than 30 cm, and little temperature response was observed at greater depths.

For example, in Fig. 2 the relative maximum temperature is −8.9°C for snow depths between 14 and 20 cm and −4.4°C for no snow cover (Fig. 2). The difference between these two values (4.5°C) is the temperature depression resulting from 14–20 cm of snow cover. Some of the results in the following section are shown as the relative temperature versus snow depth; most are shown as the temperature depression versus snow depth.

This approach does assume that changes in snow cover do not affect air temperatures at 850 hPa, which is generally assumed to be above the boundary layer. Walsh and Ross (1988) found that Chemistry–Climate Models (CCM) simulations with a heavy snow cover over eastern North America depressed 850-hPa air temperatures by 1°–2°C, which agrees with Namias’s (1985) empirical study. While the clear implication is that one cannot assume that the 850-hPa temperature is independent of snow cover, the effect on 850-hPa temperatures shown by Namias (1985) and by Walsh and Ross (1988) is substantially smaller than changes in surface air temperature documented in the literature. It is also important to note that Namias (1985) and Walsh and Ross (1988) used extremes in snow cover extent when documenting the effect on 850-hPa temperatures and were not looking at small changes in snow depth. Therefore, this study treats the 850-hPa air temperatures as independent of surface snow cover with the implicit understanding that the resulting surface air temperature depressions may be, at most, 1°–2°C too conservative.

Large areas of northern Canada did not have any consistent observations within a given 1° latitude × 1° longitude cell. Cells without at least one consistent observing station throughout 1960–2000 were removed from the analysis. The snow depth grids did not include any topography in the interpolation and, therefore, are likely biased to the preponderance of stations at low elevation. Additionally, the complex topography in mountainous regions introduces slope and aspect effects that are not accounted for in this study. Finally, the use of the 850-hPa air temperatures assumes that this level lies above the boundary layer. Therefore, any stations with mean surface pressure less than 925 hPa were removed from the analysis. Areas south of 35°N were excluded because of lack of snow cover. Only the months of December–March were examined as these months account for the most extensive snow cover across North America. The result is 266 1° latitude × 1° longitude grid cells included in the analysis that cover North America east of the Rocky Mountains and south of the boreal forest, as well as small portions of the interior west of the United States and Canada. The included regions of Canada are largely confined to the region near the U.S. border, with the exception of the prairie region of Alberta, Saskatchewan, and Manitoba.

The effect of snow cover is largely assumed to be one of the increased albedo of snow-covered surfaces (e.g., Dirnhirn and Eaton 1975). Clearly, even on a small scale the height of vegetation is important in determining what depth of snow cover is necessary to effectively mask the surface (Baker et al. 1991). The role of vegetation is even more critical at the scale of North America; this study includes biomes ranging from short grass prairies to dense coniferous forests. Measurements of maximum snow-covered albedo from Robinson and Kukla (1985) and provided by D. Robinson (2005, personal communication) are included in this study as a control on the effect of vegetation. The snow-covered albedo values are areally weighted, clear-sky surface albedo measured from Defense Meteorological Satellite Program satellite imagery in 1° latitude × 1° longitude cells. Scene brightnesses were converted to surface albedos by linear interpolation between bright and dark snow-covered surfaces with known albedo. The highest values over North America approach 0.82, while much lower albedos in the northeastern United States and eastern Canada, in many cases lower than 0.50, are likely because of the masking of snow-covered ground by the canopy of coniferous forests. Quartiles of maximum snow-covered albedo across North America were determined (0.26–0.41, 0.42–0.57, 0.58–0.70, and 0.71–0.82).

Because the effect of snow cover is largely due to albedo effects, the amount of available solar radiation is also important in understanding the effect of snow depth on air temperature. This study attempts to control for the role of incoming solar radiation by including cloud cover fraction and solar elevation angle data. Solar elevation angle is important not only because of total incoming solar radiation but also because of the anisotropy of snow cover. Snow cover has an increasingly important specular component to reflection at lower solar elevation angles, particularly with elevation angles of 18° or less, and after melt–freeze metamorphism (Dirnhirn and Eaton 1975). Maximum daily solar elevation angles were calculated each day for each grid cell. Quartiles of solar elevation angle were also determined for all available grid cells, but a range of 15° to 30° was used to represent typical midwinter solar elevation angles over grid cells with the most consistent snow cover. In mid-January, nearly all cells north of 40°N latitude are included.

Total cloud cover, expressed as a percentage, was taken from both the NCEP–NCAR reanalysis 1200 UTC grids and the gridded METAR reports of fractional cloud coverage. While these do not account for the type of cloud cover or variations in cloud cover during the day, they do provide a basic means for assessing the role of cloud cover and are appropriate for this study.

4. Discussion

Figures 2 and 3 show the relative maximum (minimum) temperatures versus snow depth for solar elevation angles of 15°–30° and ≤25% cloud cover. Temperature depressions can be determined by comparing the relative temperature at a given depth to the relative temperature at a depth of <0.1 cm. Maximum (minimum) temperature depressions of 5.3°–5.5°C (3.7°–4.1°C) are evident for snow depths of 24–28 cm. The effect of cloud cover and solar angle on the temperature depression is explored more fully later in this discussion. Figures 2 and 3 also illustrate that more than half of the temperature depression with the >20-cm snowpacks is evident with only 4–6 cm of snow cover. The effect of snow cover on temperature depression quickly plateaus at approximately 12 cm. However, the depth where the temperature depression plateaus is highly dependent on the height of the predominant vegetation. The role of vegetation is also explored more fully later in the discussion.

a. Relative versus observed temperatures

Figure 4 shows the observed maximum temperatures. These observed temperature depressions are substantially larger and approximately 15.7°C at snow depths of 22–24 cm (Fig. 4). This is largely because more northerly locations in the study area consistently have both deeper snow and lower temperatures.

To make a direct comparison between relative and observed temperature depression, a restricted geographic domain was selected. The grid cell with the lower left corner of 44°N, 94°W (the grid cell is immediately southwest of Minneapolis–St. Paul, Minnesota) was selected to match the location used in Baker et al. (1992). Events were stratified by no snow cover, snow cover <10 cm (shallow), and snow cover ≥10 cm (deep), similar to Baker et al. (1992). The maximum temperature (observed) was depressed 5.7°C (±0.5°C at one standard error) for shallow snow and 9.5°C (±0.5°C) for deep snow. These values are similar to the maximum temperature depression of 6.5°C found for shallow snow and 8.4°C for deep snow by Baker et al. (1992). The minimum temperature (observed) of the 44°N, 94°W grid cell was depressed 4.6°C (±0.5°C) for shallow snow and 8.9°C (±0.5°C) for deep snow, less than the 6.3°C Baker et al. (1992) found for shallow snow, but near the 8.4°C minimum temperature depression they found for deep snow.

The same grid cell in southern Minnesota has a relative maximum temperature depression of 3.2°C (±0.4°C) for shallow snow and 4.6°C (±0.4°C) for deep snow. The relative minimum temperature depression was 2.4°C (±0.4°C) for shallow snow and 4.0°C (±0.4°C) for deep snow. The use of the relative temperature depression as a control on temperature variability above the boundary layer removes approximately half of the temperature depression compared to using surface air temperature alone. From this example, it appears that roughly half of the depression in temperatures found in Baker et al. (1992) was because snow cover tends to persist on days that would be cold even in the absence of snow cover. However, one must exercise caution in making such a comparison between a 1° latitude × 1° longitude grid cell that includes many stations, both urban and rural, to Baker et al.’s single station located on a farm field inside a major metropolitan area (Minneapolis–St. Paul). It could be instructive to take the original data from Baker et al. (1992) and calculate relative temperatures using the corresponding 850-hPa temperatures from the nearest radiosonde station.

b. Solar elevation, cloud cover, and melt

The unique nature of the gridded dataset allows one to examine other possible influences on the role of snow cover in depressing surface air temperatures. It is possible to examine the data by month and latitude, but both are clearly related to the air temperature depression through changes in solar angle. Therefore, the noon solar elevation angle for each grid cell was calculated each day. The data were then stratified by quartiles of solar elevation angle (0°–24.1°, 24.2°–30.3°, 30.4°–38.7°, and 38.8°–60.6°). Quartiles were used to assure that there would be a sufficient number of observations in each category. The temperature depressions for each quartile of solar elevation angle were calculated by differencing the relative temperatures at a given snow depth from the relative temperature for snow-free conditions (snow depth <0.1 cm). (Figures 5 –9 show temperature depression versus snow depth.) The temperature depression for high solar elevation angles (38.8°–60.6°) is 4.2°C for a 10.1–12.0-cm snow depth, while the depression is 3.4°C for a solar elevation angle of 24.2°–30.3° (Fig. 5). There is some tendency for a greater air temperature depression with high solar elevation angles and, therefore, greater insolation. This should be expected as the shortwave radiation budget accounts for a greater portion of the surface energy budget. However, the lowest quartile of sun angles (0°–24.1°) have larger temperature depressions than the next lowest quartile (24.2°–30.3°) for deep snowpacks (Fig. 5). Perhaps this is due to an increasing specular component of reflection at low solar elevation angles, increasing the effective albedo (Dirnhirn and Eaton 1975). Furthermore, the greater shadowing from even small amounts of protruding vegetation may mean that deeper snowpacks may be necessary to effectively mask the vegetation at low solar elevation angles. This may explain the continued increase in the magnitude of the temperature depression at snow depths >20 cm, as shown for the lowest quartile of solar elevation angles (0°–24.1°) in Fig. 5.

The effect of cloud cover was also examined. To isolate the effect of cloud cover, the dataset was stratified to only examine those days with a noon solar angle of 15°–30° and with relatively open vegetation (i.e., maximum snow-covered albedo in the top quartile of all locations included). For relatively clear-sky conditions (cloud cover ≤25%), there is a clear relationship between snow depth and daily maximum air temperature, even up to snow depths of 20 cm (Figs. 6a,b). However, the relationship is weak for the nearly overcast cases with snow depths >10 cm (Figs. 6a,b). Figure 6 shows that the effect of a deep snowpack on air temperature is significantly greater with ≤25% cloud cover as it is with >75% cloud cover with snow depths of >10 cm in both the reanalysis and gridded station cloud cover datasets (Figs. 6a,b). The extremes in cloud cover do result in significantly different temperature depression responses in both the reanalysis and gridded station cloud cover datasets. The magnitude of the temperature depression is comparable in Figs. 6a,b, with the exception of the deepest snow and the most extensive cloud cover, where a larger temperature depression (approximately 1°C) is evident when using the station cloud cover dataset.

On days with little cloud cover, the shortwave components of the radiation budget should constitute a larger potion of the overall energy budget, and the relatively high albedo of snow should play a larger role in the overall surface energy budget. Conversely, on days with extensive cloud cover, the longwave radiation components should play a larger role, reducing the impact of the high albedo of snow. Intuitively, one could reasonably expect snow to depress temperature more on relatively clear days, as was found here.

Another possible effect considered was the role of snow age and melt. As snow ages, metamorphism increases snow grain sizes and reduces the snow albedo. Furthermore, there is an increased likelihood of soot or other contaminants decreasing the snow albedo. It may be useful to incorporate some measure of snow age as a variable. When examining only those days immediately after a 2-cm snowfall versus all other days, there was a significant increase in temperature depression (α = 0.10). However, the difference was small and more dependent on other factors, such as cloud cover. Perhaps this approach would be more useful if one were to examine individual stations rather than data aggregated to 1° latitude × 1° longitude.

To examine the role of melt on temperature depression, days with maximum temperatures greater than 0°C were assumed to have melting snow. Days with maximum temperatures of less than 0°C were assumed to have no melt, and the few days with a mean temperature in the grid cell of exactly 0°C were omitted. (For days with no snow, clearly one cannot draw a distinction between melt and no melt. However, the days without snow were still stratified to calculate the relative maximum temperatures.) As discussed above, only days with a noon solar elevation angle of 15°–30° and locations with high maximum snow-covered albedos were considered to isolate the effect of melting snow. Figure 7 shows that melting snow actually has a larger effect on air temperature depression, likely because of the latent heat used in melting. The latent heat released by the melting snow is available for warming the overlying air. This effect is evident to snow depths <18 cm. Greater snow depths are difficult to assess in the case of melting snow because of the large range in relative maximum temperatures (Fig. 7).

c. Maximum snow-covered albedo

Throughout this study, the Robinson and Kukla (1985) maximum snow-covered albedo dataset has been used as a proxy for the effect of forest cover. It is a particularly useful proxy in this study, as it directly measures the effect of forest cover and other land surface features, such as lakes and other vegetation, on the surface albedo during periods of deep snow cover. To more fully examine the effects of forest cover, the grid cells were stratified by quartiles of maximum snow-covered albedo (0.26–0.41, 0.42–0.57, 0.58, 0.70, and 0.71–0.82). Figure 8 shows the relative maximum temperature as a function of snow depth for each quartile of maximum snow-covered albedo. The two “brightest,” or least forested, quartiles show a much stronger response than the two most forested quartiles. The most densely forested quartile (maximum snow-covered albedo of 0.26–0.41) shows only about a 1°C maximum air temperature depression for snowpacks of >10 cm compared to snow-free conditions. Alternatively, the least forested quartile (maximum snow-covered albedo of 0.71–0.82) shows a maximum temperature depression as great as 5.1°C for snow depths >24 cm.

The effect of forest cover was also examined spatially for the contiguous 239 grid cells across the eastern United States and Canada. (Some noncontiguous grid cells in western North America were omitted.) The depression of maximum and minimum temperatures for snow depths >10 cm is shown in Figs. 9 and 10, respectively. The numbers in each grid cell indicate how much lower the temperature is for a snowpack >10 cm compared to days with no snow cover, after removing the effect of temperature changes above the boundary layer. The shading in each figure is the Robinson and Kukla (1985) maximum snow-covered albedo.

Maximum temperature depressions of 5° to 7°C are common in the lightly forested Midwest and the plains, while values of 0° to 3°C are more common in the eastern United States and southeast Canada (Fig. 9). Daily minimum air temperatures with snowpacks of >10 cm are depressed by 3° to 6°C in the Midwest and the plains, but only 0° to 4°C in the East. The effect of forest cover is more pronounced on the maximum temperatures than on the minimum temperatures. Again, this is likely because of the greater influence of snow cover on the shortwave radiative budget during the day, particularly when the skyview is relatively free of tree canopy. The average maximum temperature depression for ≤25% cloud cover for the entire study area is approximately 4.5°C compared to snow-free conditions, after removing the effect of temperature changes above the boundary layer. The average minimum temperature depression for the same conditions is approximately 3.8°C.

d. Analysis of variance

To assess the role of snow depth and land cover (i.e., maximum snow-covered albedo) while controlling for cloud cover, solar elevation angle, and melt, a factorial analysis of variance was used. The dependent variable was the relative temperature for each day and grid cell. The independent variable was the snow depth category (2-cm interval). The solar elevation angle and cloud cover were included as covariates. A binary variable (referred to as “melt”) was created from the maximum temperature, where T < 0°C is 0 and T ≥ 0°C is 1. This variable was also included as a covariate to represent the occurrence of snowmelt. Four analysis of variance models were created: 1) maximum relative temperature and all quartiles of maximum snow-covered albedo, 2) maximum relative temperature and the highest quartile of maximum snow-covered albedo, 3) minimum relative temperature and all quartiles of maximum snow-covered albedo, and 4) minimum relative temperature and the highest quartile of maximum snow-covered albedo. The top quartile of maximum snow-covered albedo was considered separately based on its sharply different temperature depression (Fig. 8).

Estimated marginal means of relative temperature were calculated with covariates evaluated at their mean values: solar elevation = 34.1°, cloud fraction = 29.8%, and melt = 0.48. These marginal means are predicted values based on each analysis of variance model. Pairwise differences were computed between the relative temperature at each snow depth category and the relative temperature for no snow. The differences represent the temperature depression for each snow depth category with the covariates held at their mean values.

Table 2 shows the depression of maximum temperature for model 1, which is approximately 1°C for depths of 6 cm and greater. However, when considering only those locations with the highest quartile of maximum snow-covered albedo (model 2), the temperature depression exceeds 4°C for snow depths greater than 20 cm (Table 2). When all depths >10 cm are treated as one category (not shown in table), the depression of maximum temperature from the analysis of variance model is 4.5°C.

Table 3 shows the depression of minimum temperatures for model 3, which are approximately 1°C for depths of 20 cm and greater. During days with snow depth of ≤6 cm, the modeled minimum temperature was actually higher (i.e., a negative temperature depression) than days without snow. This is likely because of latent heat release of melting snow. When the same analysis of variance was conducted using only days with temperatures below 0°C (not shown), the analysis of variance model shows temperatures depressed even with snow depths ≤6 cm. When considering only those locations with the highest quartile of maximum snow-covered albedo, the depression of minimum temperature (model 4; Table 3) in each category is 1.5°–2.0°C less than the depression of maximum temperature (model 2; Table 2). When all depths >10 cm are treated as one category (not shown in table), the depression of minimum temperature is 2.6°C. The findings from the analysis of variance models demonstrate the important role of the land cover on the temperature depression due to snow cover.

5. Conclusions

This study demonstrates that the maximum (minimum) temperature depression due to snow cover >10 cm, once the effects of temperature changes above the boundary layer are removed, averages approximately 4.5°C (2.6°C) for areas in the highest quartile of maximum snow-covered albedo, primarily in the grasslands of central North America. The maximum (minimum) temperature depression is much smaller, 1.2°C (1.1°C), when using an analysis of variance model that estimates the temperature depression for average cloud cover and incorporates the entire study region. However, as stated in the introduction, the temperature depressions found here may be, at most, 1°–2°C too conservative because of the implicit assumption that the snow cover has no influence on temperatures at 850 hPa.

The temperature depression due to snow varies substantially by time, location, and snow depth. The temperature depression was greater in areas with less vegetation, during spring versus midwinter because of the difference in solar angle and total solar insolation. The estimates are generally lower than previously published work, but this is likely because of the tendency to have lower temperatures above the boundary layer on days when snow cover was present. While Walsh et al. (1985) did account for this bias, they only considered snow cover extent, not depth.

The maximum temperature depression under “ideal” conditions (low vegetation, high solar insolation) of 5°–7°C (see Fig. 8, 0.71–0.82 maximum snow-covered albedo, and Fig. 9, in the northern plains) is lower than modeling studies by Ellis and Leathers (1998, 1999; 10°–15°C) for deep snowpacks. Furthermore, this work suggests that Ellis and Leathers (1999) overstate the impact of shallow snowpacks on air temperatures, as they found little difference in temperature depression between shallow (2.5 cm) and deeper snowpacks.

The findings presented here compare more favorably with Strack et al. (2003). The difference between temperature depression for the highest and lowest quartile of maximum snow-covered albedo for deep snow is 4.8°C (Fig. 8). That value is near the maximum difference of 6°C that Strack et al. (2003) found when modeling the temperature difference over deep snow for different land cover types. These findings underline a need for further comparisons of empirical and modeling studies on the role of snow in depressing air temperature.

Acknowledgments

This work was partially supported by NOAA Grant NA04OAR4310169. The author thanks David Robinson of Rutgers University for supplying the maximum snow-covered albedo dataset. The author also thanks Dan Leathers, John Walsh, and two anonymous reviewers for their comments on this manuscript.

REFERENCES

  • Baker, D. G., , R. H. Skaggs, , and D. L. Ruschy, 1991: Snow depth required to mask the underlying surface. J. Appl. Meteor., 30 , 387392.

    • Search Google Scholar
    • Export Citation
  • Baker, D. G., , D. L. Ruschy, , R. H. Skaggs, , and D. B. Wall, 1992: Air temperature and radiation depressions associated with a snow cover. J. Appl. Meteor., 31 , 247254.

    • Search Google Scholar
    • Export Citation
  • Barnes, S. L., 1973: Mesoscale objective map analysis using weighted time series observations. NOAA Tech. Memo. ERL NSSL-62, National Severe Storms Laboratory, Norman, OK, 60 pp.

  • Bromwich, D. H., , R. L. Fogt, , K. I. Hodges, , and J. E. Walsh, 2007: A tropospheric assessment of the ERA-40, NCEP, and JRA-25 global reanalyses in the polar regions. J. Geophys. Res., 112 .D10111, doi:10.1029/2006JD007859.

    • Search Google Scholar
    • Export Citation
  • Brown, R. D., 2000: Northern Hemisphere snow cover variability and change, 1915–97. J. Climate, 13 , 23392355.

  • Brown, R. D., , and R. O. Braaten, 1998: Spatial and temporal variability of Canadian monthly snow depths, 1946–1995. Atmos.–Ocean, 36 , 3745.

    • Search Google Scholar
    • Export Citation
  • Dewey, K. F., 1977: Daily maximum and minimum temperature forecasts and the influence of snow cover. Mon. Wea. Rev., 105 , 15941597.

  • Dirnhirm, I., , and F. D. Eaton, 1975: Some characteristics of the albedo of snow. J. Appl. Meteor., 14 , 375379.

  • Durre, I., , and J. M. Wallace, 2001: The warm season dip in diurnal temperature range over the eastern United States. J. Climate, 14 , 354360.

    • Search Google Scholar
    • Export Citation
  • Dyer, J., , and T. L. Mote, 2006: Spatial variability and trends in observed snow depth over North America. Geophys. Res. Lett., 33 .L16503, doi:10.1029/2006GL027258.

    • Search Google Scholar
    • Export Citation
  • Ellis, A. W., , and D. J. Leathers, 1998: A quantitative approach to evaluating the effects of snow cover on cold airmass temperatures across the U.S. Great Plains. Wea. Forecasting, 13 , 688701.

    • Search Google Scholar
    • Export Citation
  • Ellis, A. W., , and D. J. Leathers, 1999: Analysis of cold airmass temperature modification across the U.S. Great Plains as a consequence of snow depth and albedo. J. Appl. Meteor., 38 , 696711.

    • Search Google Scholar
    • Export Citation
  • Frei, A., , and D. A. Robinson, 1999: Northern Hemisphere snow extent: Regional variability 1972–1994. Int. J. Climatol., 19 , 15351560.

    • Search Google Scholar
    • Export Citation
  • Groisman, P. Ya, , T. R. Karl, , and R. W. Knight, 1994: Observed impact of snow cover on the heat balance and the rise of continental spring temperatures. Science, 263 , 198200.

    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77 , 437471.

  • Kanamitsu, M., , W. Ebisuzaki, , J. Woollen, , S-K. Yang, , J. J. Hnilo, , M. Fiorino, , and G. L. Potter, 2002: NCEP–DOE AMIP-II Reanalysis (R-2). Bull. Amer. Meteor. Soc., 83 , 16311643.

    • Search Google Scholar
    • Export Citation
  • Kistler, R., and Coauthors, 2001: The NCEP–NCAR 50-Year Reanalysis: Monthly means CD-ROM and documentation. Bull. Amer. Meteor. Soc., 82 , 247267.

    • Search Google Scholar
    • Export Citation
  • Leathers, D. J., , A. W. Ellis, , and D. A. Robinson, 1995: Characteristics of temperature depressions associated with snow cover across the northeast United States. J. Appl. Meteor., 34 , 381390.

    • Search Google Scholar
    • Export Citation
  • Namias, J., 1963: Surface–atmosphere interactions as fundamental causes of drought and other climatic fluctuations. Proc. Rome Symp., Changes of Climate, Arid Zone Research, Rome, Italy, UNESCO and WMO, 345–359.

  • Namias, J., 1985: Some empirical evidence for the influence of snow cover on temperature and precipitation. Mon. Wea. Rev., 113 , 15421553.

    • Search Google Scholar
    • Export Citation
  • Robinson, D. A., 1989: Construction of a United States historical snow data base. Proc. 45th Eastern Snow Conf., Lake Placid, NY, Eastern Snow Conference, 50–59.

  • Robinson, D. A., , and G. Kukla, 1985: Maximum surface albedo of seasonally snow-covered lands in the Northern Hemisphere. J. Climate Appl. Meteor., 24 , 402411.

    • Search Google Scholar
    • Export Citation
  • Robinson, D. A., , and A. Frei, 1997: A Northern Hemisphere snow cover climatology using satellite information. Proc. 10th Conf. on Applied Climatology, Reno, NV, Amer. Meteor. Soc., 24–28.

  • Robinson, D. A., , K. F. Dewey, , and R. R. Heim, 1993: Global snow cover monitoring: An update. Bull. Amer. Meteor. Soc., 74 , 16891696.

  • Serreze, M. C., , J. R. Key, , J. E. Box, , J. A. Maslanik, , and K. Steffen, 1998: A new monthly climatology of global radiation for the Arctic and comparisons with NCEP–NCAR reanalysis and ISCCP-C2 fields. J. Climate, 11 , 121136.

    • Search Google Scholar
    • Export Citation
  • Shepard, D., 1968: A two-dimensional interpolation function for irregularly-spaced data. Proc. 23rd ACM National Conf., Association of Computing Machinery, 517–524.

  • Strack, J. E., , R. A. Pielke Sr., , and J. Adegoke, 2003: Sensitivity of model-generated daytime surface heat fluxes over snow to land-cover changes. J. Hydrometeor., 4 , 2442.

    • Search Google Scholar
    • Export Citation
  • U.S. Department of Commerce, cited. 2003: TD-3200 Cooperative Summary of the Day. National Climatic Data Center, NOAA. [Available online at http://www.ncdc.noaa.gov.].

  • U.S. Department of Commerce, cited. 2007: DS-3505 Integrated Surface Hourly. Global Hourly Surface Observations National Climatic Data Center, NOAA. [Available online at http://www.ncdc.noaa.gov.].

  • Walsh, J. E., , and B. Ross, 1988: Sensitivity of 30-day dynamical forecasts to continental snow cover. J. Climate, 1 , 739754.

  • Walsh, J. E., , D. R. Tucek, , and M. R. Peterson, 1982: Seasonal snow cover and short-term climatic fluctuations over the United States. Mon. Wea. Rev., 110 , 14741486.

    • Search Google Scholar
    • Export Citation
  • Walsh, J. E., , W. H. Jasperson, , and B. Ross, 1985: Influences of snow cover and soil moisture on monthly air temperature. Mon. Wea. Rev., 113 , 756768.

    • Search Google Scholar
    • Export Citation
  • Willmott, C. J., , C. M. Rowe, , and W. D. Philpot, 1985: Spheremap. University of Delaware. [Available online at http://climate.geog.udel.edu/~climate/html_pages/software2.html.].

  • Wojcik, G. S., , and D. S. Wilks, 1992: Temperature forecast biases associated with snow cover in the Northeast. Wea. Forecasting, 7 , 501506.

    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

MODIS visible satellite image from 1710 UTC 3 Dec 2006 with that day’s min temperatures (Tmin, °C) and snow depth (cm, contours) from cooperative observing stations across the Midwest. A lower albedo is evident in the forested areas from southeast to central Missouri, including some areas with up to 30 cm of snow.

Citation: Journal of Applied Meteorology and Climatology 47, 7; 10.1175/2007JAMC1823.1

Fig. 2.
Fig. 2.

Relative max air temperature (Tmax, °C) vs 2-cm snow depth increments for days with noon solar elevation angle of 15° to 30° and cloud cover less than 25%. Only locations in the top quartile of max snow-covered albedo are included. Mean values are labeled. Bars show ±1 std error.

Citation: Journal of Applied Meteorology and Climatology 47, 7; 10.1175/2007JAMC1823.1

Fig. 3.
Fig. 3.

As in Fig. 2 but for Tmin.

Citation: Journal of Applied Meteorology and Climatology 47, 7; 10.1175/2007JAMC1823.1

Fig. 4.
Fig. 4.

As in Fig. 2 but for obs Tmax.

Citation: Journal of Applied Meteorology and Climatology 47, 7; 10.1175/2007JAMC1823.1

Fig. 5.
Fig. 5.

Depression of Tmax (Tmax-dep) vs 2-cm snow depth increments for days and locations by quartile of noon solar elevation angle. Only days with cloud cover less than 25% and locations in the top quartile of max snow-covered albedo are included. Mean values are labeled. Bars show ±1 std error.

Citation: Journal of Applied Meteorology and Climatology 47, 7; 10.1175/2007JAMC1823.1

Fig. 6.
Fig. 6.

The Tmax-dep vs 2-cm snow depth increments for cloud cover from 1200 UTC (a) NCAR–NCEP reanalysis and (b) an objective analysis of station cloud fractional coverage. Lines represent cloud cover categories of 0%–25% (solid black), 25%–50% (broken black), 50%–75% (solid gray), and 75%–100% (broken gray). Only days and locations with noon solar elevation angles of 15° to 30° and in the top quartile of max snow-covered albedo are included. Mean values are labeled. Bars show ±1 std error.

Citation: Journal of Applied Meteorology and Climatology 47, 7; 10.1175/2007JAMC1823.1

Fig. 7.
Fig. 7.

The Tmax-dep vs 2-cm snow depth increments for days with Tmax > 0°C (black) or ≤0°C (gray). Only days with noon solar elevation angle of 15° to 30° and locations in the top quartile of maximum snow-covered albedo are included. Bars show ±1 std error.

Citation: Journal of Applied Meteorology and Climatology 47, 7; 10.1175/2007JAMC1823.1

Fig. 8.
Fig. 8.

The Tmax-dep vs 2-cm snow depth increments by quartile of max snow-covered albedo. Only days with noon solar elevation angle of 15° to 30° and cloud cover less than 25% are included. Mean values are labeled. Bars show ±1 std error.

Citation: Journal of Applied Meteorology and Climatology 47, 7; 10.1175/2007JAMC1823.1

Fig. 9.
Fig. 9.

Daily Tmax-dep for snow depth ≥10 cm (numbers) and max snow-covered albedo (shading) on a 1° lat × 1° lon grid. Only days with cloud cover less than 25% and no precipitation are included.

Citation: Journal of Applied Meteorology and Climatology 47, 7; 10.1175/2007JAMC1823.1

Fig. 10.
Fig. 10.

As in Fig. 9 but for daily Tmin depression (Tmin-dep).

Citation: Journal of Applied Meteorology and Climatology 47, 7; 10.1175/2007JAMC1823.1

Table 1.

Means (std dev) of Tmax and Tmin as a function of snow depth on 3 Dec 2006. Means are for 637 cooperative stations in and adjacent to the area affected by the 30 Nov–1 Dec 2006 Midwest snowstorm.

Table 1.
Table 2.

Estimated marginal means from an analysis of variance. Values represent Tmax-dep and the 95% confidence interval (CI). The columns include data for all quartiles of max snow-covered albedo (model 1) and for the highest quartile of snow-covered albedo (model 2). All values are significant at p = 0.05.

Table 2.
Table 3.

As in Table 2 but for Tmin-dep and models 3 and 4.

Table 3.
Save