• Beaudry, P. G., , and Golding D. L. , 1983: Snowmelt during rain-on-snow in coastal British Columbia. Proc. 51st Annual Western Snow Conf., Vancouver, WA, Western Snow Conference, 55–66. [Available online at www.westernsnowconference.org/sites/westernsnowconference.org/PDFs/1983Beaudry.pdf.]

  • Berris, S. N., , and Harr R. D. , 1987: Comparative snow accumulation and melt during rainfall in forested and clear-cut plots in the western Cascades of Oregon. Water Resour. Res., 23, 135142, doi:10.1029/WR023i001p00135.

    • Search Google Scholar
    • Export Citation
  • Birsan, M. V., , Molnar P. , , Burlando P. , , and Pfaundler M. , 2005: Streamflow trends in Switzerland. J. Hydrol., 314, 312329, doi:10.1016/j.jhydrol.2005.06.008.

    • Search Google Scholar
    • Export Citation
  • Blöschl, G., , Kirnbauer R. , , and Gutknecht D. , 1990: Modelling snowmelt in a mountainous river basin on an event basis. J. Hydrol., 113, 207229, doi:10.1016/0022-1694(90)90176-X.

    • Search Google Scholar
    • Export Citation
  • Blöschl, G., , Kirnbauer R. , , and Gutknecht D. , 1991: Distributed snowmelt simulations in an alpine catchment. 1. Model evaluation on the basis of snow cover patterns. Water Resour. Res., 27, 31713179, doi:10.1029/91WR02250.

    • Search Google Scholar
    • Export Citation
  • Christner, J., , and Harr R. D. , 1982: Peak streamflows from the transient snow zone, western Cascades, Oregon. Proc. 50th Annual Western Snow Conf., Reno, NV, Western Snow Conference, 27–38. [Available online at www.westernsnowconference.org/sites/westernsnowconference.org/PDFs/1982Christner.pdf.]

  • Dozier, J., , and Warren S. G. , 1982: Effect on viewing angle on the infrared brightness temperature. Water Resour. Res., 18, 14241434, doi:10.1029/WR018i005p01424.

    • Search Google Scholar
    • Export Citation
  • Ellis, C. R., , Pomeroy J. W. , , Brown T. , , and MacDonald J. , 2010: Simulation of snow accumulation and melt in needleleaf forest environments. Hydrol. Earth Syst. Sci., 14, 925940, doi:10.5194/hess-14-925-2010.

    • Search Google Scholar
    • Export Citation
  • Ellis, C. R., , Pomeroy J. W. , , Essery R. L. H. , , and Link T. E. , 2011: Effects of needleleaf forest cover on radiation and snowmelt dynamics in the Canadian Rocky Mountains. Can. J. For. Res., 41, 608620, doi:10.1139/X10-227.

    • Search Google Scholar
    • Export Citation
  • Essery, R. L., , Pomeroy J. , , Ellis C. , , and Link T. , 2008: Modelling longwave radiation to snow beneath forest canopies using hemispherical photography or linear regression. Hydrol. Processes, 22, 27882800, doi:10.1002/hyp.6930.

    • Search Google Scholar
    • Export Citation
  • Floyd, W., , and Weiler M. , 2008: Measuring snow accumulation and ablation dynamics during rain-on-snow events: Innovative measurement techniques. Hydrol. Processes, 22, 48054812, doi:10.1002/hyp.7142.

    • Search Google Scholar
    • Export Citation
  • Frazer, G. W., , Canham C. D. , , and Lertzman K. P. , 1999: Gap Light Analyzer (GLA): Imaging software to extract canopy structure and gap light transmission indices from true-colour fisheye photographs. Users Manual and Program Documentation, version 2.0, Simon Fraser University and Institute of Ecosystem Studies, 36 pp. [Available online at http://rem-main.rem.sfu.ca/forestry/downloads/Files/GLAV2UsersManual.pdf.]

  • Freudiger, D., , Kohn I. , , Stahl K. , , and Weiler M. , 2013: Large scale analysis of changing frequencies of rain-on-snow events and their impact on floods. Hydrol. Earth Syst. Sci. Discuss., 10, 13 23113 263, doi:10.5194/hessd-10-13231-2013.

    • Search Google Scholar
    • Export Citation
  • Garvelmann, J., , Pohl S. , , and Weiler M. , 2013: From observation to the quantification of snow processes with a time-lapse camera network. Hydrol. Earth Syst. Sci., 17, 14151429, doi:10.5194/hess-17-1415-2013.

    • Search Google Scholar
    • Export Citation
  • Gelfan, A. N., , Pomeroy J. W. , , and Kuchment L. S. , 2004: Modeling forest cover influences on snow accumulation, sublimation and melt. J. Hydrometeor., 5, 785803, doi:10.1175/1525-7541(2004)005<0785:MFCIOS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Grömping, U., 2007: Estimators of relative importance in linear regression based on variance decomposition. Amer. Stat., 61, 139147, doi:10.1198/000313007X188252.

    • Search Google Scholar
    • Export Citation
  • Hamlet, A. F., , Mote P. W. , , Clark M. P. , , and Lettenmeier D. P. , 2005: Effects of temperature and precipitation variability on snowpack trends in the western United States. J. Climate, 18, 45454561, doi:10.1175/JCLI3538.1.

    • Search Google Scholar
    • Export Citation
  • Hardy, J. P., , Melloh R. A. , , Koenig G. , , Marks D. , , Winstral A. , , and Pomeroy J. W. , , and Link T. , 2004: Solar radiation transmission through conifer canopies. Agric. For. Meteor., 126, 257270, doi:10.1016/j.agrformet.2004.06.012.

    • Search Google Scholar
    • Export Citation
  • Harr, R. D., 1981: Some characteristics and consequences of snowmelt during rainfall in western Oregon. J. Hydrol., 53, 277304, doi:10.1016/0022-1694(81)90006-8.

    • Search Google Scholar
    • Export Citation
  • Harr, R. D., 1986: Effects of clearcutting on rain-on-snow runoff in western Oregon: A new look on old studies. Water Resour. Res., 22, 10951100, doi:10.1029/WR022i007p01095.

    • Search Google Scholar
    • Export Citation
  • Heron, R., , and Woo M. K. , 1978: Snowmelt computations for a high Arctic site. Proc. 35th Eastern Snow Conf., Hanover, NH, Eastern Snow Conference, 162172.

  • Jones, J. A., 2000: Hydrologic processes and peak discharge response to forest removal, regrowth and roads in 10 small experimental basins, western Cascades, Oregon. Water Resour. Res., 36, 26212642, doi:10.1029/2000WR900105.

    • Search Google Scholar
    • Export Citation
  • Jost, G., , Weiler M. , , Gluns D. R. , , and Alila Y. , 2007: The influence of forest and topography on snow accumulation and melt at the watershed-scale. J. Hydrol., 347, 101115, doi:10.1016/j.jhydrol.2007.09.006.

    • Search Google Scholar
    • Export Citation
  • Kattelmann, R. C., 1997: Flooding from rain-on-snow events in the Sierra Nevada. IAHS Publ.,239, 59–65.

  • Koivusalo, H., , and Kokkonen T. , 2002: Snow processes in a forest clearing and in a coniferous forest. J. Hydrol., 262, 145165, doi:10.1016/S0022-1694(02)00031-8.

    • Search Google Scholar
    • Export Citation
  • Köplin, N., , Schäder B. , , Vivirolli D. , , and Weingartner R. , 2013: Seasonality and magnitude of floods in Switzerland under future climate change. Hydrol. Processes, 28, 2567–2578, doi:10.1002/hyp.9757.

    • Search Google Scholar
    • Export Citation
  • Link, T. E., , and Marks D. , 1999a: Point simulation of seasonal snow cover dynamics beneath boreal forest canopies. J. Geophys. Res., 104, 27 84127 857, doi:10.1029/1998JD200121.

    • Search Google Scholar
    • Export Citation
  • Link, T. E., , and Marks D. , 1999b: Distributed simulation of snowcover mass- and energy-balance in the boreal forest. Hydrol. Processes, 13, 24392452, doi:10.1002/(SICI)1099-1085(199910)13:14/15<2439::AID-HYP866>3.0.CO;2-1.

    • Search Google Scholar
    • Export Citation
  • Link, T. E., , Marks D. , , and Hardy J. , 2004: A deterministic method to characterize canopy radiative transfer properties. Hydrol. Processes, 18, 35833594, doi:10.1002/hyp.5793.

    • Search Google Scholar
    • Export Citation
  • Liston, G. E., , and Elder K. , 2006: A distributed snow-evolution modeling system (SnowModel). J. Hydrometeor., 7, 217234, doi:10.1175/JHM486.1.

    • Search Google Scholar
    • Export Citation
  • Male, D. H., , and Granger R. J. , 1981: Snow surface energy exchange. Water Resour. Res., 17, 609627, doi:10.1029/WR017i003p00609.

  • Marks, D., , Dozier J. , , and Davis R. E. , 1992: Climate and energy exchanges at the snow surface in the alpine region of the Sierra Nevada: 1. Meteorological measurements and monitoring. Water Resour. Res., 28, 30293042, doi:10.1029/92WR01482.

    • Search Google Scholar
    • Export Citation
  • Marks, D., , Kimball J. , , Tingey D. , , and Link T. , 1998: The sensitivity of snowmelt processes to climate conditions and forest cover during rain-on-snow: A case study of the 1996 Pacific Northwest flood. Hydrol. Processes, 12, 15691587, doi:10.1002/(SICI)1099-1085(199808/09)12:10/11<1569::AID-HYP682>3.0.CO;2-L.

    • Search Google Scholar
    • Export Citation
  • Marks, D., , Domingo J. , , Susong D. , , Link T. , , and Garen D. , 1999: A spatially distributed energy balance snowmelt model for application in mountain basins. Hydrol. Processes, 13, 19351959, doi:10.1002/(SICI)1099-1085(199909)13:12/13<1935::AID-HYP868>3.0.CO;2-C.

    • Search Google Scholar
    • Export Citation
  • Marks, D., , Link T. , , Winstral A. , , and Garen D. , 2001: Simulating snowmelt processes during rain-on-snow over a semi-arid mountain basin. Ann. Glaciol., 32, 195202, doi:10.3189/172756401781819751.

    • Search Google Scholar
    • Export Citation
  • Marsh, P., , Onclin C. , , Russel M. , , and Pohl S. , 2003: Effects of shrubs on snow processes in the vicinity of the Arctic tree line in NW Canada. Proc. 14th Int. Northern Research Basins Symp. and Workshop, Kangerlussuaq, Greenland, University of Copenhagen, 113117.

  • Martin, E., , and Lejeune Y. , 1998: Turbulent fluxes above the snow surface. Ann. Glaciol., 26, 179183.

  • Mazurkiewicz, A. B., , Callery D. G. , , and McDonnell J. J. , 2008: Assessing the controls of the snow energy balance and water available for runoff in a rain-on-snow environment. J. Hydrol., 354, 1–14, doi:10.1016/j.jhydrol.2007.12.027.

    • Search Google Scholar
    • Export Citation
  • McCabe, G. J., , Hay L. E. , , and Clark M. P. , 2007: Rain-on-snow events in the western United States. Bull. Amer. Meteor. Soc.,88, 319–328, doi:10.1175/BAMS-88-3-319.

  • Moore, R. D., 1983: On the use of bulk aerodynamic formulae over melting snow. Nord. Hydrol., 14, 193206.

  • Morris, E. M., 1989: Turbulent fluxes over snow and ice. J. Hydrol., 105, 205233, doi:10.1016/0022-1694(89)90105-4.

  • Pohl, S., , Marsh P. , , and Liston G. E. , 2006a: Spatial–temporal variability in turbulent fluxes during spring snowmelt. Arct. Antarct. Alp. Res., 38, 136146, doi:10.1657/1523-0430(2006)038[0136:SVITFD]2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Pohl, S., , Marsh P. , , and Pietrino A. , 2006b: Spatial–temporal variability in solar radiation during spring snowmelt. Nord. Hydrol., 37, 119.

    • Search Google Scholar
    • Export Citation
  • Pohl, S., , Garvelmann J. , , Wawerla J. , , and Weiler M. , 2014: Potential of an innovative low cost sensor network to understand the spatial and temporal dynamics of a mountain snow cover. Water Resour. Res., doi:10.1002/2013WR014594, in press.

    • Search Google Scholar
    • Export Citation
  • Pomeroy, J. W., , Gray D. M. , , Hedstrom M. , , and Janowicz J. R. , 2002: Prediction of seasonal snow accumulation in cold climate forests. Hydrol. Processes, 16, 35433558, doi:10.1002/hyp.1228.

    • Search Google Scholar
    • Export Citation
  • Pomeroy, J. W., , Toth B. , , Granger R. M. , , Hedstrom N. R. , , and Essery R. , 2003: Variation in surface energetics during snowmelt in a sub-arctic mountain catchment. J. Hydrometeor., 4, 702719, doi:10.1175/1525-7541(2003)004<0702:VISEDS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Pomeroy, J. W., , Marks D. , , Link T. , , Ellis C. , , Hardy J. , , Rowlands A. , , and Granger R. , 2009: The impact of coniferous forest temperature on incoming longwave radiation to melting snow. Hydrol. Processes, 23, 25132525, doi:10.1002/hyp.7325.

    • Search Google Scholar
    • Export Citation
  • Ranzi, R., , and Rosso R. , 1991: A physically based approach to modeling distributed snowmelt in a small alpine catchment. IAHS Publ.,205, 141–150.

  • Rössler, O., , Froidevaux P. , , Börst U. , , Rickli R. , , Martius O. , , and Weingartner R. , 2013: Retrospective analysis of a non-forecasted rain-on-snow flood in the Alps—A matter of model-limitations or unpredictable nature? Hydrol. Earth Syst. Sci. Discuss., 10, 12 86112 904, doi:10.5194/hessd-10-12861-2013.

    • Search Google Scholar
    • Export Citation
  • Satterlund, D., 1979: An improved equation for estimating long-wave radiation from the atmosphere. Water Resour. Res., 15, 16491650, doi:10.1029/WR015i006p01649.

    • Search Google Scholar
    • Export Citation
  • Sicart, E., , Pomery J. W. , , Essery R. L. , , Hardy J. E. , , Link T. E. , , and Marks D. , 2004: A sensitivity study of daytime net radiation during snowmelt to forest canopy and atmospheric conditions. J. Hydrometeor., 5, 744784.

    • Search Google Scholar
    • Export Citation
  • Singh, P., , Spitzbart G. , , Hübl H. , , and Weinmeister H. W. , 1997: Hydrological response of snowpack under rain-on-snow events: A field study. J. Hydrol., 202, 120, doi:10.1016/S0022-1694(97)00004-8.

    • Search Google Scholar
    • Export Citation
  • Stenberg, P., , Linder S. , , Smolander H. , , and Flower-Ellis J. , 1994: Performance of the LAI-2000 plant canopy analyzer in estimating leaf area index of some Scots pine stands. Tree Physiol., 14, 981995, doi:10.1093/treephys/14.7-8-9.981.

    • Search Google Scholar
    • Export Citation
  • Storck, P., 2000: Trees snow and flooding: An investigation of forest canopy effects on snow accumulation and melt at the plot and watershed scales in the Pacific Northwest. Water Resources Series Tech. Rep. 161, Dept. of Civil and Environmental Engineering, University of Washington, Seattle, WA, 176 pp.

  • Strasser, U., , and Etchevers P. , 2005: Simulation of daily discharges for the upper Durance catchment (French Alps) using subgrid parametrization for topography and a forest canopy climate model. Hydrol. Processes, 19, 23612373, doi:10.1002/hyp.5889.

    • Search Google Scholar
    • Export Citation
  • Strasser, U., , and Marke T. , 2010: ESCIMO.spread —A spreadsheet-based point snow surface energy balance model to calculate hourly snow water equivalent and melt rates for historical and changing climate conditions. Geosci. Model Dev., 3, 643652, doi:10.5194/gmd-3-643-2010.

    • Search Google Scholar
    • Export Citation
  • Strasser, U., , Warscher M. , , and Liston G. E. , 2011: Modeling snow-canopy processes on an idealized mountain. J. Hydrometeor., 12, 663677, doi:10.1175/2011JHM1344.1.

    • Search Google Scholar
    • Export Citation
  • Sui, J., , and Koehler G. , 2001: Rain-on-snow induced flood events in southern Germany. J. Hydrol., 252, 205220, doi:10.1016/S0022-1694(01)00460-7.

    • Search Google Scholar
    • Export Citation
  • Susong, D., , Marks D. , , and Garen D. ,1999: Methods for developing time-series climate surface to drive topographically distributed energy- and water-balance models. Hydrol. Processes,13, 2003–2021, doi:10.1002/(SICI)1099-1085(199909)13:12/13<2003::AID-HYP884>3.0.CO;2-K.

  • Tetens, V. O., 1930: Ueber einige meteorologische Begriffe (in German). Z. Geophys., 6, 297309.

  • Tribbeck, M. J., , Gurney R. , , Morris E. M. , , and Pearson D. W. C. , 2004: A new Snow-SVAT to simulate the accumulation and ablation of seasonal snow beneath a forest canopy. J. Glaciol., 50, 171182, doi:10.3189/172756504781830187.

    • Search Google Scholar
    • Export Citation
  • U.S. Army Corps of Engineers, 1956: Snow hydrology: Summary report of the snow investigations. North Pacific Division Army Corps of Engineers, Portland, OR, 437 pp. [Available online at ftp://ftp.wcc.nrcs.usda.gov/wntsc/H&H/snow/SnowHydrologyCOE1956.pdf.]

  • Welles, J. M., , and Norman J. M. , 1991: Instrument for indirect measurement of canopy architecture. Agron. J., 83, 818825, doi:10.2134/agronj1991.00021962008300050009x.

    • Search Google Scholar
    • Export Citation
  • Whitaker, A. C., , and Sugiyama H. , 2005: Seasonal snowpack dynamics and runoff in a cool temperate forest: Lysimeter experiment in Niigata, Japan. Hydrol. Processes, 19, 41794200, doi:10.1002/hyp.6059.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    The three study basins in the Black Forest, southwestern Germany. Snow monitoring stations (SnoMoS) located at open sites are indicated by circles, and SnoMoS located in the forest are indicated by triangles. The polar plot shows the spatial distribution (elevation and aspect) of the individual SnoMoS sites used for the study.

  • View in gallery

    Discharge measured at the outlet of the study catchments, air temperature from a mean altitude SnoMoS within each catchment ,and precipitation measured at representative weather stations for the basins.

  • View in gallery

    Hourly values of observed surface EB components and the total EB during ROS 1. The individual locations are shown in gray lines, the average in black.

  • View in gallery

    Hourly values of observed surface EB components and the total EB during ROS 2. The individual locations are shown in gray lines, the average in black. Since the event occurred during nighttime only, there was no shortwave energy input.

  • View in gallery

    Hourly values of observed surface EB components and the total EB during the clear sky period. The individual locations are shown in gray lines, the average in black.

  • View in gallery

    Proportion of the individual EB terms on the total EB observed. Forest data are shown in gray boxes.

  • View in gallery

    Observed accumulated positive energy amounts and associated potential snowmelt at the open and forested study sites during the two ROS events and the clear sky period. Gray (black) lines show the individual sites (the average).

  • View in gallery

    Variability of average melt rates calculated from the available energy for the two ROS events and the clear sky period.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 37 37 16
PDF Downloads 17 17 8

Variability of Observed Energy Fluxes during Rain-on-Snow and Clear Sky Snowmelt in a Midlatitude Mountain Environment

View More View Less
  • 1 Chair of Hydrology, University of Freiburg, Freiburg, Germany
© Get Permissions
Full access

Abstract

Hourly observations of 65 snow monitoring stations were used to investigate the spatiotemporal variability of the surface energy balance during snowmelt in the Black Forest region of southwestern Germany. The study focuses on two rain-on-snow (ROS) events in December 2012 and a clear sky period at the beginning of March 2013 using the same study locations. ROS and clear sky were chosen since they are completely different snowmelt conditions in terms of energy exchanges and dynamics. The results show that snowmelt was dominated by turbulent exchanges at the open field sites and by both turbulent exchanges and net longwave radiation in the forest during ROS. The energy available for snowmelt can be almost identical at open and forest locations during ROS, and a constant energy flux even during night was directed toward the snowpack. During the clear sky conditions, net shortwave radiation was the dominating term in the open, whereas net shortwave and net longwave radiation were most important in the forest. A diurnal signal with positive energy balance during daylight and negative energy balance in the night was observed, with considerably reduced energy available for snowmelt in the forest. Furthermore, the stratified sampling design revealed the strong influence of the canopy and the topography at the locations on the observed energy fluxes. Elevation, aspect, and leaf area index (LAI) were the most important predictor variables during ROS, whereas aspect and LAI were most influential during the clear sky period. The study highlights the distinct spatial variability of the individual energy balance terms over a relatively small area during the differing snowmelt conditions.

Corresponding author address: Jakob Garvelmann, Albert-Ludwigs-Universität Freiburg, Fahnenbergplatz, 79098 Freiburg, Germany. E-mail: jakob.garvelmann@hydrology.uni-freiburg.de

Abstract

Hourly observations of 65 snow monitoring stations were used to investigate the spatiotemporal variability of the surface energy balance during snowmelt in the Black Forest region of southwestern Germany. The study focuses on two rain-on-snow (ROS) events in December 2012 and a clear sky period at the beginning of March 2013 using the same study locations. ROS and clear sky were chosen since they are completely different snowmelt conditions in terms of energy exchanges and dynamics. The results show that snowmelt was dominated by turbulent exchanges at the open field sites and by both turbulent exchanges and net longwave radiation in the forest during ROS. The energy available for snowmelt can be almost identical at open and forest locations during ROS, and a constant energy flux even during night was directed toward the snowpack. During the clear sky conditions, net shortwave radiation was the dominating term in the open, whereas net shortwave and net longwave radiation were most important in the forest. A diurnal signal with positive energy balance during daylight and negative energy balance in the night was observed, with considerably reduced energy available for snowmelt in the forest. Furthermore, the stratified sampling design revealed the strong influence of the canopy and the topography at the locations on the observed energy fluxes. Elevation, aspect, and leaf area index (LAI) were the most important predictor variables during ROS, whereas aspect and LAI were most influential during the clear sky period. The study highlights the distinct spatial variability of the individual energy balance terms over a relatively small area during the differing snowmelt conditions.

Corresponding author address: Jakob Garvelmann, Albert-Ludwigs-Universität Freiburg, Fahnenbergplatz, 79098 Freiburg, Germany. E-mail: jakob.garvelmann@hydrology.uni-freiburg.de

1. Introduction

Rapid snowmelt during rain-on-snow (ROS) conditions has the potential to produce fairly large flood events (Beaudry and Golding 1983; Marks et al. 1998; McCabe et al. 2007) on different scales (Sui and Koehler 2001; Freudiger et al. 2013). Studies have reported an increase in winter air temperatures and liquid precipitation in the Northern Hemisphere (e.g., Birsan et al. 2005; Hamlet et al. 2005). Therefore, midwinter ROS will likely also become an important subject in cold climates and high alpine regions in a changing climate. Simulations by Köplin et al. (2013) have suggested an increase in ROS events for Switzerland in the future. ROS typically occurs after a cold period, allowing the accumulation of a snow cover, when warm, moist air masses with liquid precipitation and positive air temperatures cause rapid snowmelt. Usually high antecedent soil moisture conditions allow a very efficient runoff generation (Jones 2000) and rain and snowmelt together can produce greater floods than any runoff, resulting from rainfall or snowmelt alone (Harr 1981; Kattelmann 1997; Singh et al. 1997; Marks et al. 1998).

Most studies have focused on modeling experiments using few point measurements to study the snowmelt energetics during ROS and other snowmelt conditions since it is very challenging to measure the spatial distribution of the snowmelt energy balance (EB) components in complex terrain (Marks et al. 1992). Understanding the spatial variability of snowmelt, however, is crucially important in order to distribute point measurements to broader scales, to validate distributed snowmelt models, and to accurately predict melt rates—especially in flood forecasting applications. Furthermore, relative differences in the snowmelt EB are dependent on the spatial distribution of the snow cover and the micrometeorological conditions during the melt event (Storck 2000). The exchange of energy at the snow–atmosphere interface is most important for snowmelt. Net shortwave and longwave radiation and the turbulent fluxes of sensible and latent heat are generally the most important components of the snowmelt EB, and these terms can be highly spatially variable owing to topography and vegetation (Martin and Lejeune 1998; Marks et al. 1999; Pohl et al. 2006a,b). Generally, shortwave radiation is the most important term during clear sky snowmelt (U.S. Army Corps of Engineers 1956). Owing to overcast conditions and often considerable wind speeds during ROS, net shortwave radiation becomes less important, and turbulent exchanges are the dominating terms of the snowmelt EB (Harr 1981; Marks et al. 1998; Storck 2000). Furthermore, the influence of net longwave radiation, especially under a vegetation canopy, becomes an important term due to usually positive air temperatures (Berris and Harr 1987). Energy advected from rainfall is another energy source. However, despite considerable amounts of energy that can be released by the refreezing of rain in cold snow packs, Koivusalo and Kokkonen (2002) report a small influence of this energy source on snowmelt. Numerous studies have shown the influence of a vegetation cover on the snowmelt EB (e.g., Pomeroy et al. 2003; Hardy et al. 2004; Liston and Elder 2006; Ellis et al. 2010). There are melt accelerating and delaying factors in the forest (Link and Marks 1999a,b; Tribbeck et al. 2004). Shading of the forest canopy and reduction of wind speed are the most important melt delaying factors, whereas increased thermal emission of the vegetation accelerates snowmelt and becomes a crucial factor for the EB within forests (Link et al. 2004; Sicart et al. 2004; Essery et al. 2008; Pomeroy et al. 2009). Whitaker and Sugiyama (2005) observed considerable micrometeorological differences in air temperature, relative humidity (RH), wind speed, and shortwave radiation between open and forested areas. The importance and spatial variability of those processes strongly depends on the structure of the vegetation cover, topographic positions, and meteorological conditions (Pomeroy et al. 2002; Jost et al. 2007). Furthermore, studies have reported reduced melt rates under the forest canopy compared to clear-cut areas and therefore higher peak flows from areas with a high proportion of open areas (Christner and Harr 1982; Harr 1986). For the spring snowmelt, modeling results of Link and Marks (1999a,b) indicate that a dense forest canopy can delay the snowmelt up to three weeks compared to open areas because of reduced incoming solar radiation. Studies have focused on basin scale ROS runoff simulations (Blöschl et al. 1990) and the investigation of the physical processes within the snowpack (Singh et al. 1997), as well as the quantification of the snow cover processes during ROS with observation techniques (Floyd and Weiler 2008). Nevertheless, there is a lack of knowledge about the spatiotemporal variability of the snowmelt EB, in particular for ROS conditions. Furthermore, the influence of a vegetation cover on the snowmelt energetics under ROS conditions is poorly understood. Consequently, the prediction of melt rates and runoff generation during ROS is still very challenging since extensive observations are necessary for a correct simulation of the relevant processes (Rössler et al. 2013). A common approach is to use available climatic information from a few point measurements together with interpolation or model simulation to get spatial distributed information. It is well known that using direct measurements of the meteorological conditions in the forest can considerably improve the simulation of melt rates as well as the amount and timing of meltwater releases in hydrological modeling (Marsh et al. 2003; Strasser and Etchevers 2005).

This study presents an event-based analysis of the spatiotemporal variability and the relative importance of the individual snowmelt EB components during ROS and clear sky snowmelt conditions in a midelevation mountain environment where ROS is a commonly occurring hydrological event. The available continuous measurements at fixed open and forested locations allow for an exact comparison of the energetic conditions during different melt events at a multitude of points. Furthermore, the spatially distributed sampling design allowed quantifying the influence of the topography and the forest on the surface EB during the differing snowmelt conditions without spatial interpolation of the individual point measurements.

2. Study area and experimental setup

The study was carried out in three research basins in the Black Forest, a midlatitude medium elevation mountain range with a seasonal snow cover in the southwest of Germany. Elevations in the area range from 400 up to 1500 m MSL. Average winter air temperatures range from 4.1°C in the lower parts to −2.1°C in the highest elevations. Mean annual precipitation ranges from approximately 900 mm in the lower parts to about 1950 mm in the higher regions. Prevailing main wind direction in the area is westerly. On average, 27% are open areas used for grazing and haymaking. Only 3% of the catchment areas are covered by human settlements. The remaining 70% of the areas are covered by forest. The forests in the region are about 80% coniferous (spruce, fir, pine) and 20% deciduous (beech, birch, oak).

A network consisting of a total of 65 snow monitoring stations was established for the whole 2012/13 winter season. The snow monitoring station (SnoMoS) is a standalone measurement system able to measure snow depth, air temperature, relative humidity, incoming shortwave radiation, surface temperature, barometric pressure, and wind speed/precipitation—basically, all major climatic variables in hourly intervals needed to calculate a complete surface energy balance. A comprehensive description of the SnoMoS can be found in Pohl et al. (2014). A stratified sampling design was used to cover a wide range of elevations and exposures (Fig. 1). To specifically investigate the influence of the vegetation cover on the snowmelt beneath, pairs of SnoMoS were generally installed in close proximity to each other with one being located underneath the canopy while the other was situated on an adjacent open field site. Additional information about weather and snow conditions could be retrieved from a time-lapse camera network established in the study basins as presented by Garvelmann et al. (2013). Manual snow surveys were conducted frequently, at least every two weeks, at different elevations and exposures at forested and nonforested locations in order to estimate the snow densities, and therefore the snow water equivalents (SWE), from the continuous snow depth measurements in the study area.

Fig. 1.
Fig. 1.

The three study basins in the Black Forest, southwestern Germany. Snow monitoring stations (SnoMoS) located at open sites are indicated by circles, and SnoMoS located in the forest are indicated by triangles. The polar plot shows the spatial distribution (elevation and aspect) of the individual SnoMoS sites used for the study.

Citation: Journal of Hydrometeorology 15, 3; 10.1175/JHM-D-13-0187.1

Shown in Fig. 2 are the average air temperatures, precipitation, and measured streamflows in the three study basins during December 2012. Streamflow was measured hourly at official gauging stations at the catchment outlets operated by the governmental flood forecast agency. There was snow accumulation at the beginning of December, with the last snowfall on 13 December 2012 resulting in SWE of up to 300 mm. During the night of 15 December 2012, masses of warm air with precipitation hit the region, resulting in areawide snowmelt and a distinct rise of streamflow in the study basins. There was intermittent precipitation with moderate intensity averaging 0.6 mm h−1 and a total precipitation amount averaging 58.9 mm during the next 4.5 days. The study period of ROS 1 was chosen between 0000 central European time (CET) 15 December and 2300 CET 16 December 2012 (48 h), the main phase of the event. During this event snow melted almost completely below 700 m MSL, but substantial amounts of SWE persisted in the higher elevations. Therefore, the measurements from the SnoMoS in the Kinzig basin, the catchment with the lowest elevation, were not used for the analysis of ROS 2, starting 5 days after the end of the previous ROS event with lower air temperatures and a few centimeters of new snow, when air temperatures did rise again and another rain event with high precipitation intensity averaging 1.6 mm h−1 and average total precipitation of 31.3 mm during one night caused substantial snowmelt at all elevations, producing fairly high streamflows with return periods of up to 20 years in the region. The study period for ROS 2 was chosen between 1800 CET 22 December and 0800 CET 23 December 2012 (15 h).

Fig. 2.
Fig. 2.

Discharge measured at the outlet of the study catchments, air temperature from a mean altitude SnoMoS within each catchment ,and precipitation measured at representative weather stations for the basins.

Citation: Journal of Hydrometeorology 15, 3; 10.1175/JHM-D-13-0187.1

The snow accumulation in the second half of February 2013 resulted in season peak snow depths with up to 750 mm SWE at the end of the month. On 2 March 2013, the atmospheric conditions changed to high pressure, and a couple of days with positive daytime temperatures and clear skies resulted in a significant depletion of the snow cover. The considered study period was from 0000 CET 4 March to 2300 CET 5 March 2013 (48 h).

3. Methods

a. Energy balance calculations

The components of the EB at the snow surface were either directly measured by the SnoMoS or calculated from the measured climatic variables at each location using physically based empirical equations. A comprehensive review of energy exchange at the snow surface can be found in Male and Granger (1981). Hourly measurements of incoming shortwave radiation, air temperature, RH, surface temperature, wind speed, precipitation, and barometric pressure have been used to calculate the EB at the snow surface:
e1
where is the measured incoming shortwave radiation, the albedo of the snow, and are the incoming and outgoing longwave radiation, is the sensible heat flux, the latent heat flux, and the advective energy input by liquid precipitation. The individual terms of the EB are expressed in watts per square meter. Energy fluxes directed to the surface are designated as positive while energy fluxes directed away from the surface are considered to be negative in the computations. Table 1 provides an overview of the applied EB estimations, whereas the following section describes the EB calculations in detail. The term EB is used in the following for the energy balance at the snow surface.
Table 1.

Overview of directly measured factors and variables needed to calculate an energy balance at the snow surface and additional needed information and physical-based empirical equations.

Table 1.

Incoming shortwave radiation was measured horizontally directly by the SnoMoS. Therefore, a geometric inclination correction was carried out for the direct component of measured incoming shortwave radiation for locations on tilted terrain to account for local slope and aspect. The measured total incoming shortwave radiation was divided in its direct and diffuse component using a cloudiness factor similar to Ranzi and Rosso (1991). Cloudiness was derived manually from time-lapse images taken in each catchment using the Octa scheme for fractional cloud cover. Because of the short time scale maximum of 48 h, a constant albedo value of 0.7 was used for the snow cover at the open and forested sites in the calculations for the two melt conditions (ROS and clear sky) in order to obtain the net shortwave energy component. This value was based on the average of numerous albedo measurements carried out at open field sites around the respective study periods using a handheld albedometer. Using the same value for open and forested sites is a simplification since snow albedo is usually lower in the forest owing to needles and other debris on the surface, but we could not rely on the measurements taken in the forest.

Incoming longwave radiation for open locations was calculated following an empirical approach presented by Satterlund (1979) using air temperature, atmospheric vapor pressure, and cloudiness:
e2
with the cloudiness factor and an emissivity function ϵ accounting for atmospheric vapor pressure in millibars derived from RH and the saturated air vapor pressure, air temperature in kelvin, and an empirical constant b with a value of 2016:
e3
The thermal energy emission from the adjacent terrain has not been considered in the presented longwave calculations. Incoming longwave radiation for the locations under the forest canopy was calculated following the approach from Essery et al. (2008):
e4
where is the sky-view factor estimated from hemispherical images, the Stefan–Boltzmann constant, and the canopy temperature. In the approach presented here, the canopy temperature is assumed to be equal to the local air temperature. Longwave radiation emission of the snow surface was calculated from the measured surface temperatures:
e5
with the surface temperature in kelvin, the Stefan–Boltzmann constant, and the emissivity of snow with a value of 0.985 (Dozier and Warren 1982).
The sensible heat and latent heat fluxes were calculated with the bulk transfer approach presented by Heron and Woo (1978) using measurements of wind speed, air temperature, snow surface temperature, and RH:
e6
e7
where is the measured wind speed, is the air density, the heat capacity of the air, the latent heat of vaporization, the barometric pressure, the atmospheric vapor pressure derived from RH and the saturated air vapor pressure, and the water vapor saturation pressure at the snow surface both calculated with the Tetens formula (Tetens 1930) from air temperature and surface temperature, respectively. The eddy heat diffusivity was calculated dependent on the atmospheric stability conditions. The bulk Richardson number () was calculated with measured wind speed , the gravitational constant , the air temperature , the snow surface temperature , and the aerodynamic roughness height of 2.0 mm for the open and forested sites, a value used in other snow studies for open terrain (Moore 1983; Morris 1989; Pohl et al. 2006a):
e8
The bulk Richardson number was used to distinguish among stable, unstable, or neutral atmospheric conditions. For stable conditions the following calculation was used:
e9
whereas for unstable conditions the equation is as shown below:
e10
For neutral atmospheric conditions equals , where is the aerodynamic eddy diffusivity calculated as
e11
with being the von Kármán constant, the instrument height, and the aerodynamic roughness height of 2.0 mm.
The advective heat input from liquid precipitation refreezing within the snow cover was calculated for surface temperatures below the freezing point. It is assumed that the temperature of the rainwater is equal to the air temperature:
e12
where is the density of water, is the specific heat capacity of water, P the precipitation (m s−1), and the latent heat of fusion. For surface temperatures at the freezing point, the sensible heat energy released by the cooling of the rainwater is calculated as follows:
e13
Precipitation data from established weather stations were used and evenly distributed precipitation input over the study basins was used in the calculations.Owing to the highly variable forest canopy conditions and high pre-event moisture levels in the canopy, interception of rainwater in the canopy was not considered.

b. Data analysis

The energy fluxes of the EB terms as well as the sum of the fluxes during the two ROS events and the clear sky period were calculated for every SnoMoS location with an hourly resolution. The periods with positive energy ( > 0) were used to calculate the accumulated energy amounts potentially available for snowmelt at the observation sites. These values and the potential melt rates calculated from the available positive energy were used to investigate the snowmelt variability among the locations. The influence of the vegetation cover was investigated by separately looking at open and forest locations. Effective leaf area index (LAI) was used for characterization of the forest canopy at the study locations. Conifer and deciduous canopy were not further distinguished between in the calculations. The influence of topography and vegetation was further investigated by comparing the average values of the EB terms and observed during the three study periods against the predictor’s elevation, aspect, slope, and LAI. The terrain parameters of the SnoMoS locations were derived from a 10-m digital elevation model. The aspect was considered as northing with 0 indicating south, 1 equal to north, and 0.5 applying similarly to east and west. Hemispherical images were used to determine the effective LAI and the sky-view factor at the forest locations using the average effective LAI value calculated with the approaches presented by Welles and Norman (1991) and Stenberg et al. (1994) implemented into the Gap Light Analyzer software (Frazer et al. 1999). Effective LAI at the SnoMoS locations ranges from 1.4 to 2.6 for conifer and from 0.9 to 1.7 for deciduous stands. Linear regression analysis was applied in order to identify the influence of the predictors on the average energy values of the individual EB terms and their sum. The linear regressions were calculated separately using the data from the open and forested sites for the two ROS events and the clear sky period. Multiple linear regressions (MLRs) were used to further investigate the influence of topography and vegetation on the measured energy values by using all predictors. To analyze the gradients of each predictor, we also calculated the relative importance of each predictor in the MLRs using the approach of Grömping (2007), which is based on a proportional marginal variance decomposition method that is more stable and robust than calculating the partial correlation coefficient. An F test was performed in order to determine the significance of the individual correlations. A significance level of 5% (p value < 0.05) was applied.

For the ROS conditions the average energy flux values of 29 open field sites could be used for the comparison with elevation. Since two locations were established on flat terrain, only 27 open field sites could be analyzed for aspect. For the investigation of below-canopy conditions 31 locations could be used. During the clear sky conditions 26 open sites could be used for elevation analysis (24 for aspect), while 30 forest sites were available.

4. Results

a. Spatiotemporal variability of the surface energy balance components

1) ROS

The energy fluxes during the ROS 1 event are shown in Fig. 3. Despite cloudy weather, the diurnal signal of net shortwave radiation was evident at the open locations and with reduced energy amounts under the canopy as well. At the open sites net longwave radiation was positive at most stations at the beginning, followed by a decreasing trend. The available net longwave energy was constantly positive in the forest. The development of individual measurements was similar with an apparent variability among the locations. The observed turbulent fluxes were both mostly positive at virtually all individual stations throughout the whole event. Reduced energy exchanges were observed at the stations in the forest. The variability among the different sites for open and forested locations was rather high throughout the whole event. Additional positive energy amounts were added from precipitation during the event. These energy inputs were very similar at the open and forest locations. The input was especially high at the beginning of the event when rain started. At the forest sites was constantly positive. This behavior was also observed for the bulk of the open sites at the beginning and middle of the event. Toward the end, the energy inputs were getting negative at the open sites. The highest positive energy inputs were observed around noon. Variability among the sites was more pronounced for the open locations.

Fig. 3.
Fig. 3.

Hourly values of observed surface EB components and the total EB during ROS 1. The individual locations are shown in gray lines, the average in black.

Citation: Journal of Hydrometeorology 15, 3; 10.1175/JHM-D-13-0187.1

The picture was quite similar for ROS 2 (Fig. 4): was constantly positive at both open and forest locations. The variability among the stations was more pronounced at the open sites. Since the event occurred during nighttime only, there was no shortwave energy input, and therefore no diurnal signal was evident. Net longwave radiation was very similar at the open and forest locations with the difference that most of the open locations had negative longwave values at the very beginning before they became constantly positive. The variability among individual open and forest locations was apparent, with a very equal range of values. The turbulent fluxes were again both providing positive energy input to the snowpack, with reduced energy inputs from turbulent exchanges being observed under the forest canopy. The variability among the stations was again fairly high for both sensible and latent heat flux, especially at the open sites. However, this time the variability in the forest was quite high, especially due to a limited number of stations that showed markedly increased energy inputs compared to the bulk of the forest stations. The energy inputs from rainfall were very similar at the open and forest locations. The clusters evident in the data were representing the stations from the two study basins having different precipitation inputs during the event.

Fig. 4.
Fig. 4.

Hourly values of observed surface EB components and the total EB during ROS 2. The individual locations are shown in gray lines, the average in black. Since the event occurred during nighttime only, there was no shortwave energy input.

Citation: Journal of Hydrometeorology 15, 3; 10.1175/JHM-D-13-0187.1

2) Clear sky

For the clear sky conditions, the observations of the EB terms and are presented in Fig. 5. A clear diurnal signal was evident for net shortwave radiation at the open sites and with reduced energy under the forest canopy. The variability among the stations was rather high. There was also a diurnal trend in the net longwave radiation data with, on average, more negative values in the night and less negative or even positive values during daylight at open sites. In the forest the net longwave values were, on average, more positive and the diurnal signal is not as pronounced. On average, the behavior of the turbulent fluxes was quite different compared to the observed ROS conditions. The average sensible heat flux was positive, whereas the average latent heat flux was negative. The same was observed in the forest with overall reduced energy amounts. The high variability among the locations has to be pointed out here too. Furthermore, while the sensible heat flux was constantly positive at the open and forest sites, the observed latent heat fluxes follow no clear trend. Especially during daylight hours at the open sites, variability was extremely high with some stations that showed positive while others showed negative energy values. For a diurnal signal was evident for the energy values observed at both open and forested sites, and was positive during daylight for both. In the forest a dampened signal with more positive energy was evident, while the difference between daylight and night hours was more distinct at the open sites.

Fig. 5.
Fig. 5.

Hourly values of observed surface EB components and the total EB during the clear sky period. The individual locations are shown in gray lines, the average in black.

Citation: Journal of Hydrometeorology 15, 3; 10.1175/JHM-D-13-0187.1

b. Percentage of the surface energy balance terms during ROS and clear sky

The average of the energy flux values for each location during the two ROS events and the clear sky period were used to calculate the percentages of the individual components of at the open and forest locations and are shown as boxplots to highlight the variability among the sites (Fig. 6). The variability among the observations was again clearly evident. In particular, the spatial variability of net shortwave radiation and net longwave radiation was rather high. Unfortunately, ROS 2 was only an overnight event, and therefore the variability of net shortwave radiation could only be studied during ROS 1.

Fig. 6.
Fig. 6.

Proportion of the individual EB terms on the total EB observed. Forest data are shown in gray boxes.

Citation: Journal of Hydrometeorology 15, 3; 10.1175/JHM-D-13-0187.1

The distinct variability was observed for clear sky in particular. Furthermore, the increased spatial variability of the turbulent exchanges, especially the sensible heat flux, during ROS conditions can be clearly seen. Finally, Fig. 6 indicates that the amount of energy advected from rainfall to the snow were spatially fairly homogenous and similar for both events.

c. Accumulated positive energy amounts and potential snowmelt

The potential snowmelt can be calculated from the available positive energy. To do this, the (latent) heat of fusion for snow is used to determine the amount of SWE melt with respect to the available positive energy, an approach commonly used in energy balance models for snowmelt (e.g., Strasser and Marke 2010).

1) ROS

The accumulated positive energy amounts and respective accumulated potential snowmelt are presented in Fig. 7 for the two ROS events and the clear sky period for the open and forest locations. For ROS 1, the average available energy and, thus, average potential melt rates were very similar for open and forest locations with a slightly higher average energy input (16.7%) at the open sites. It is interesting to note that the accumulation of energy available for melt was generally fairly constant over time at both open and forested sites. This behavior was also observed for ROS 2. However, this time the average available energy was 31.1% higher for the open location than for the forest. It has to be mentioned that during the 15-h period of ROS 2, higher (open) or very similar (forest) amounts of energy were available compared to the much longer-lasting (2 days) ROS 1 event. The variability among the observations was again clearly pronounced. The average accumulated energy amounts for ROS 1 were 4.0 MJ m−2 (maximum 9.6 MJ m−2) at the open sites and 3.5 MJ m−2 (maximum 8.4 MJ m−2) in the forest. These values correspond to average potential snowmelt amounts of 12.0 and 11.0 mm of SWE at the open and forest locations, respectively. The maximum accumulated potential snowmelt was 29.0 mm for open and 25.0 mm for the forest. For ROS 2 average accumulated energy amounts were 5.0 MJ m−2 (maximum 13.8 MJ m−2) in the open and 4.4 MJ m−2 (maximum 11.0 MJ m−2) in the forest corresponding to 15.0 mm (maximum 41.4 mm) and 13.3 mm (maximum 33.0 mm) of potential snowmelt, respectively. The potential snowmelt values calculated from the available positive energy at the snow surface were compared to observed SWE losses at the SnoMoS stations. It became evident that, on average, the calculated snowmelt amounts were fairly close to the observed average SWE melt at open and forested locations. Observed melt of the snowpack during ROS 1 was, on average, 13.9 mm at the open stations and 11.5 mm in the forest and for ROS 2 13.2 mm at the open station and 6.7 mm in the forest.

Fig. 7.
Fig. 7.

Observed accumulated positive energy amounts and associated potential snowmelt at the open and forested study sites during the two ROS events and the clear sky period. Gray (black) lines show the individual sites (the average).

Citation: Journal of Hydrometeorology 15, 3; 10.1175/JHM-D-13-0187.1

2) Clear sky

The estimated accumulated positive energy amounts for the clear sky period are shown in the bottom panel of Fig. 7. On average, 74.5% more energy was available for snowmelt at the open compared to the forested sites. Furthermore, the higher maximum available energy at the open sites can be seen clearly. A distinct diurnal signal was clearly evident, especially for the open locations indicated by the steep increase in the accumulated energy amounts during the daylight hours and stagnation during night (negative energy and possible refreezing). The diurnal variation was not as distinct in the forest, and the overall melt energy amounts were considerably lower. The average accumulated positive energy amounts reached values of 7.4 MJ m−2 (22.0 mm potential melt) for the open locations and 2.1 MJ m−2 (6.0 mm potential melt) for the forested sites. Maximum accumulated energy amounts were 15.4 MJ m−2 (46.0 mm potential melt) and 11.1 MJ m−2 (33.0 mm potential melt) for the open and forested sites, respectively. Average observed melt of the snowpack during clear sky was 9.4 mm at the open stations and 3.7 mm in the forest.

3) Potential melt rates during ROS and the clear sky conditions

Figure 8 shows the variability of potential melt rates at the SnoMoS locations, calculated from the energy balance observed at the open and forested locations. On average, potential melt rates were always higher at the open locations. During ROS 1 potential melt rates were very similar at open and forest locations. It is very interesting to note that the highest potential melt rates were observed during ROS 2, not during the clear sky melt period. The lowest potential melt rates were observed in the forest under clear sky conditions. The absolute average difference between the potential melt rates at the open and forested sites was largest during the clear sky period and smallest for ROS 1.

Fig. 8.
Fig. 8.

Variability of average melt rates calculated from the available energy for the two ROS events and the clear sky period.

Citation: Journal of Hydrometeorology 15, 3; 10.1175/JHM-D-13-0187.1

d. Influence of topography and vegetation on the surface energy balance terms

Table 2 shows the coefficients of correlation for the linear regressions between the average values of the EB terms and the predictor elevation, aspect, and LAI. Table 3 presents the coefficients of determination for the MLR model using elevation, aspect, slope, and LAI as predictor variables for all the available SnoMoS locations together. Additionally, Table 3 shows the proportions of variance explained by the individual predictors.

Table 2.

Coefficients of correlation for the three study periods from the linear regression analysis. Significant correlations (p value < 0.05) are shown in bold: represents net shortwave radiation; the net longwave radiation, () the turbulent fluxes of sensible (latent) heat, Qm the energy advected by liquid precipitation, and is the sum of the EB terms.

Table 2.
Table 3.

Total (bold) and relative explained variances of the MLR analysis using elevation, aspect, slope, and effective LAI as predictors for , , , , , and .

Table 3.

1) ROS

Aspect of the open locations had no influence on the measured net shortwave radiation during the ROS conditions. Surprisingly, decreasing net shortwave radiation with increasing elevation was observed during ROS 1. LAI was significantly correlated with net shortwave radiation observed at the forest locations. Net longwave radiation was significantly decreasing with increasing elevation at the open locations during ROS 1 and ROS 2. During ROS 1 elevation was a good predictor for the net longwave radiation in the forest. For ROS 2 no correlation of net longwave radiation with elevation, aspect, or LAI was found for the forest locations. For both ROS events, significant correlations were found between the turbulent fluxes and the aspect of the open and forest locations. Furthermore, LAI was significantly correlated with sensible and latent heat in the forest. Elevation had no strong influence on the observed turbulent fluxes at the open sites. The best predictor for the rainfall energy was elevation at both open and forest locations during ROS 1; for ROS 2 no significant correlations were found. Slope was not an important predictor for the EB terms.

The variance explained by the MLR model was 29% and 36% for the average observed during ROS 1 and ROS 2. Elevation and aspect together (27%) explained most of the variance for ROS 1, whereas aspect and effective LAI (34%) explained most of the observed variance during ROS 2. Significant correlations were also found for the MLR model for the EB terms. For both ROS events, the same predictor variables explained most of the total variance. Elevation and LAI were most important for net longwave radiation. Aspect and LAI explained most of the variance for both turbulent fluxes. Elevation was most influential on the observed energy input by precipitation. Net shortwave radiation was only observed during ROS 1 and the analysis shows that this term is markedly influenced by the land cover (LAI).

2) Clear sky

During clear sky a significant correlation was found for aspect and net shortwave radiation at the open sites. South-facing terrain received increased shortwave energy input compared to north-facing terrain. This correlation was not found for the forest locations. Under the forest canopy net shortwave radiation was significantly correlated with the LAI. Net longwave radiation increased with increasing elevation at the open as well as the forest locations. The best predictor for the observed average values of both turbulent fluxes was elevation in the forest, whereas only the sensible heat flux was significantly correlated with elevation at the open sites. No significant correlation with aspect was found for either sensible heat flux or the turbulent heat flux during the clear sky conditions.

For the clear sky period, the MLR model explained 43% of the observed variance, with aspect (24%) and LAI (13%) being the most important variables. Aspect and LAI explained most of the variance of net shortwave radiation, whereas LAI alone was most important for net longwave radiation. Elevation and LAI explained most of the sensible heat variance, whereas elevation and slope were the most important predictor variables for latent heat.

5. Discussion

During both melt conditions, the observed differences among the individual SnoMoS locations for the individual EB terms and , and thus potential snowmelt, are fairly high at both open and forested SnoMoS locations, indicating a well-pronounced spatial variability of the energy balance at the snow surface due to vegetation and topographic influences in the study area.

a. ROS

During the observed ROS conditions, the energy was almost continuously directed to the snow surface, as indicated by the constantly positive total energy (Figs. 3 and 4). Studies have shown the dominance of the turbulent fluxes during ROS (Berris and Harr 1987; Marks et al. 1998), but net radiation has also been reported as being the largest contributor (Mazurkiewicz et al. 2008). In our study, net shortwave radiation played only a minor role and accounted on average for 25.7% of the EB at the open field sites during ROS 1. It has to be mentioned at this point again that the results of ROS 1 and ROS 2 are somewhat biased by the fact that ROS 2 was an overnight event only. Therefore, no shortwave input was recorded during ROS 2, a fact that influences the overall energy balance and the emphasis of the individual terms. The negative correlations found for the average net shortwave radiation measured at the open sites and elevation can be explained with a denser cloud cover in the higher elevations during the ROS event. During ROS the diffuse component of shortwave radiation is dominating owing to overcast sky, and therefore no significant correlation was found for net shortwave radiation with aspect for both open and forested locations; the MLR model also showed no influence of aspect. The density of the forest canopy had, even under overcast conditions, an influence on the shortwave radiation reaching the forest floor. Therefore, a significantly negative correlation for net shortwave radiation with LAI was found. However, net shortwave radiation accounted, on average, for only 3.2% of during ROS 1 in the forest. The presented data show that, on average, both turbulent energy fluxes became positive at the onset of ROS and were dominating throughout the event. In the open, the turbulent fluxes accounted for, on average, 67.0% of during ROS 1 and 71.6% during ROS 2. Marks et al. (1998, 2001) reported that 60%–90% of the energy available for snowmelt came from turbulent exchanges. Net longwave radiation accounted only for a very small proportion (−2.0%) of the total energy balance.

In the forest, turbulent exchanges (ROS 1 = 31.8%, ROS 2 = 47.0%) and the net longwave component (ROS 1 = 55.1%, ROS 2 = 38.8%) were the dominating energy sources during ROS. The importance of longwave radiation in the forest can be explained by the additional emission of longwave energy from the forest vegetation itself. However, no significant correlation for net longwave with LAI at the forest locations was found. This could be explained by the fact that denser vegetation covers generally block more atmospheric longwave radiation while emitting more longwave radiation themselves (Essery et al. 2008). The negative correlations for net longwave with elevation observed in the forest and in the open can be explained by the decline of air temperature used in the computations of incoming longwave radiation with elevation during the ROS conditions. Aspect had no influence on net longwave during ROS at the open and forest locations. The influence of elevation and forest cover is also clearly evident from the MLR model for both ROS events. Decreased turbulent exchanges under the canopy compared to open sites were observed throughout and can be explained by the reduction of wind speed in the forest since the turbulent fluxes are strongly correlated with wind speed (Marks et al. 1998; Susong et al. 1999). In fact, data from the SnoMoS were analyzed and revealed average wind speed reductions of about 80% for coniferous and about 44% for deciduous forest locations. Consequently, decreasing sensible and latent energy values with increasing LAI were observed in the forest. During ROS this negative correlation was particularly strong for the turbulent fluxes and, since they are important terms of the EB during those conditions, showed a significant negative correlation with LAI. Furthermore, a significant negative correlation with aspect was found for the turbulent fluxes at the open and forest sites, and the MLR model also showed that aspect explained substantially the variance of the turbulent exchanges, meaning that energy amounts were lower at north-exposed slopes compared to south-facing locations. This particular behavior can be explained by the dominant southwest wind direction observed during the two ROS events, causing higher wind speeds in more wind exposed locations with southern orientation. The influence of wind direction on the turbulent fluxes was also shown by Pohl et al. (2006a). Especially during ROS 2, the influence of wind speed was more pronounced. Average wind speed was 1.4 m s−1 at the open sites and 0.8 m s−1 at the forested sites during ROS 1, while average wind speeds of 2.5 and 1.3 m s−1 were measured at the open and forested sites during ROS 2, respectively. The contribution of the energy advected by rainfall was, on average, 9.3% of at the open sites during ROS 1 and ROS 2, while the contribution was 9.9% and 14.2% in the forest during the two ROS events. The influence of this term strongly depends on the initial temperature of the snow cover and the air temperature. Because of refreezing of rainwater in a cold snowpack, high amounts of latent energy are released, contributing to the warming of the snow cover. Once the snowpack is isothermal at 0°C, additional energy is only gained owing to the temperature difference between precipitation and snow temperature. This influence can be seen at the beginning of ROS 1, especially at the open sites, when surface temperatures were clearly below 0 and high amounts of advective energy were observed. Contrary to other studies (Beaudry and Golding 1983; Berris and Harr 1987; Marks et al. 1998) that have reported significantly less snowmelt from forested areas compared to open areas during ROS, the results of this study show, on average, only slightly decreased potential melt rates at forested compared to open areas. It is also important to note that was almost permanently positive at the open as well as at the forest locations during ROS. Owing to constant positive energy flux to the snowpack, a continuous snowmelt, even during nighttime hours, is possible. This is an important issue for the generation of stormflow runoff since large areas with differing land uses can potentially contribute meltwater continuously to streamflow during ROS.

To better understand the spatial variability, we looked at the characteristics of stations with the highest and lowest observed overall melt energy values. During both ROS events, open field sites at midelevations showed the most positive accumulated energy values. This can be explained by a combination of climatic factors, as these locations show considerable wind speeds along with higher air temperatures. The lowest energy amounts available for snowmelt at open sites were observed at very wind-sheltered high-elevation sites. In the forest the highest accumulated positive energy was observed at forest locations with a relatively low tree and canopy density, where very high wind speeds have frequently been observed. These locations were mostly deciduous tree stands or located in higher elevations. The lowest energy amounts have been observed at very wind-sheltered conifer forest locations with a high canopy density.

b. Clear sky

A diurnal signal is, as expected, clearly apparent during the clear sky conditions at the open sites, with positive energy values during daylight and negative energy values in the night. Net shortwave radiation was the absolutely dominating term accounting for, on average, 199.7% of . The impact of shortwave radiation is constantly increasing in the course of the winter season since sun angles are getting constantly higher. Therefore, a direct comparison of net shortwave radiation during the two study periods (ROS in December 2012 versus clear sky in March 2013) is hardly possible. In the forest, the diurnal signal is still evident but considerably dampened compared to the open location. Nevertheless, net shortwave radiation accounted for 60.9% of in the forest. During the clear sky conditions, net shortwave radiation was very significantly correlated with aspect at the open locations, with the highest energy amounts at south-facing locations. Thus, the incident angle of shortwave radiation has the most important influence on the EB during these conditions at open field sites, as was found in other studies (e.g., Blöschl et al. 1991; Marks et al. 1999; Pohl et al. 2006b; Ellis et al. 2011). Net longwave radiation accounted for, on average, 25.5% of under the canopy. At the open sites this value was considerably lower (−144.9%). Net longwave radiation was mainly negative, with some positive hours during daylight, at the open sites and less negative or even positive at the forested sites. During clear sky the incoming component of longwave radiation is fairly low at the open sites, whereas the emission of longwave energy from the snowpack to the atmosphere is fairly high owing to cold nighttime air temperatures. In the forest the longwave emission from the trees toward the snow cover counteracts and often outweighs the longwave energy losses from the snow surface. The warming up of tree trunks due to solar heating is not captured by the used approach, and therefore, longwave emission by the trees might have been underestimated during the clear sky conditions. A significant positive correlation with elevation was found for net longwave radiation at the open and forest locations. This resulted from an inversion of air temperatures during this period in the region. Aspect had no influence on net longwave during the clear sky period at the open and forest locations. The turbulent exchanges accounted for on average, 45.2% and 14.6% of at the open and forested sites, respectively. The differences between open and forest are not as pronounced compared to the ROS conditions because of overall lower wind speeds in the considered clear sky period. The sensible heat flux was mostly positive, whereas negative values were observed for the latent heat flux. During the clear sky period no influence of aspect was found, but elevation had a significant influence with higher turbulent exchanges at higher elevations and lower values in the valley (also evident in the MLR model results). This can be explained with higher wind speeds observed at the higher elevations and the inversion conditions affecting air temperatures at that time, causing higher temperature gradients between the snow and the overlying air masses and therefore increased turbulent exchanges for higher elevations. For the MLR model showed that aspect and effective LAI were the most important predictor variables for the clear sky period.

Distinct differences in the snowmelt energetics have been observed, on average, at open and forest locations with increased positive energy amounts and therefore higher potential melt observed at the open locations. Reduced melt rates under the forest canopy have also been shown by Link and Marks (1999b), Gelfan et al. (2004), and Strasser et al. (2011). During the clear sky period, a clear diurnal signal was observed for at open areas providing positive energy during daylight for snowmelt, whereas during nighttime no energy was available for snowmelt, while becomes negative. The same can be observed in the forest along with overall lower potential snowmelt rates. The potential snowmelt calculated from the available positive energy was considerably higher than the actual observed snowmelt. The lower observed snowmelt values can be explained by the fact that the snow cover was relatively cold (average snow surface temperatures of −5.7° and −3.5°C at the open and forest sites) at the beginning of the investigated clear sky period, and much of the initial excess surface energy was used to erase the internal snowpack energy deficit.

During the clear sky period, maximum accumulated energy amounts were observed at south-facing open hillslopes, whereas minimum energy was available on steep, north-facing hillslopes located at high elevations. In the forest, the location with the absolute maximum accumulated energy is a south-facing, deciduous forest. The other locations with high accumulated energy amounts are relatively low density forests or deciduous forests also on south-facing or level terrain. The minimum accumulated energy amounts were observed at north-facing terrain or at conifer forest locations with a high tree density.

c. Validation of the observed energy fluxes available for snowmelt

A direct comparison of the observed SWE melt and the potential SWE melt calculated from the available positive energy at the individual SnoMoS sites was conducted as a way of validation of the results. The analysis showed average standard errors of 1.9 mm at the open and 1.1 mm at the forested sites for ROS 1. During ROS 2 the average standard error was 9.4 mm at the open and 7.8 mm at the forested sites. For the clear sky period the average standard error was 9.9 and 1.0 mm for the open and forested sites, respectively. This difference can be explained with some uncertainties in the observed SWE values since we had few snow density measurements available, especially around the second ROS event. As mentioned, the positive energy during the clear sky period was mainly used to warm up the snow cover, especially in the open areas, thus explaining the difference in observed versus simulated snowmelt.

6. Conclusions

A high-density snow observation network has been used to determine and analyze the energy balance (EB) at the snow surface during midwinter rain-on-snow (ROS) conditions and, for comparison, a radiation-driven clear sky period in early spring at a high spatial and temporal resolution in a midlatitude, forested, medium-elevation mountain range environment with a seasonal snow cover. The importance of the individual EB terms and the dynamics of have been investigated for numerous open and forested locations with an event-based analysis for the ROS and the clear sky conditions, respectively. The turbulent fluxes of sensible and latent heat dominated the surface EB at the open locations, whereas both turbulent exchanges and net longwave radiation were the dominating factors in the forest during the two ROS events in December 2012. For comparison, the EB was, as expected, heavily dominated by net shortwave radiation at the open locations and net shortwave and net longwave radiation together in the forest during a clear sky period at the beginning of March 2013. Surprisingly, accumulated energy available for snowmelt was, on average, very similar at open and forested locations during ROS conditions in the study catchments. The study clearly showed that the spatial variability among the individual locations is very high, especially during ROS 2. The study results further revealed that a constant energy flux was directed toward the snowpack during ROS, whereas during clear sky, a clear diurnal signal with positive energy values during daylight and negative energy values in the night was observed at the open and, to a lesser extent, at the forested sites. The average energy available for snowmelt was considerably reduced under the forest canopy for clear skies, and the variability among the locations very well pronounced. These observations are highly relevant for the generation of runoff during differing snowmelt conditions. The constant positive energy fluxes available for snowmelt during ROS may cause, together with the liquid precipitation, a very rapid and efficient stormflow runoff generation. During ROS, elevation and aspect had the strongest influence on the EB at the open locations, whereas in the forest, the structure of the vegetation (LAI) and aspect are the most influential factors for the energy fluxes. For the clear sky conditions, the aspect of a location had the strongest influence on the EB at the open sites. In the forest, the structure of the vegetation rather than the topography (elevation and aspect) of the location seems to have the strongest influence on the energy fluxes. Combining all the observations together in a MLR model elevation, aspect, and effective leaf area index (LAI) were the most important predictor variables during ROS, and aspect and LAI were most important during the clear sky period. Aspect was important because of the main wind direction during ROS and incoming solar radiation during the clear sky period.

The study clearly shows the importance of the turbulent fluxes of sensible and latent heat for snowmelt during ROS as well as the radiation-dominated snowmelt during clear sky periods in early spring and highlight the considerable spatial variability of the individual EB terms over the 274 km2 area studied. Topography (elevation, aspect) and vegetation (LAI) are important factors strongly influencing the spatial distribution of snowmelt during ROS and clear sky conditions and should be integrated in any hydrological model. Therefore, the results are an important contribution toward an improved understanding of the spatiotemporal variability of snowmelt. The findings of the presented study have the potential to considerably improve the simulation of snowmelt and the generation of runoff, especially during ROS conditions in forested mountain environments.

Acknowledgments

We thank the German Research Foundation (DFG) for the funding of the project “Field Observations and Modelling of Spatial and Temporal Variability of Processes Controlling Basin Runoff during Rain on Snow Events.” Furthermore, we would also like to thank our field assistants Franziska Zieger, Daniel Günther, and Denis Blümel for their valuable help in the field and the technicians Emil Blattmann and Lukas Neuhaus. Finally, we thank Tim Reid and an anonymous reviewer for their valuable comments and suggestions.

REFERENCES

  • Beaudry, P. G., , and Golding D. L. , 1983: Snowmelt during rain-on-snow in coastal British Columbia. Proc. 51st Annual Western Snow Conf., Vancouver, WA, Western Snow Conference, 55–66. [Available online at www.westernsnowconference.org/sites/westernsnowconference.org/PDFs/1983Beaudry.pdf.]

  • Berris, S. N., , and Harr R. D. , 1987: Comparative snow accumulation and melt during rainfall in forested and clear-cut plots in the western Cascades of Oregon. Water Resour. Res., 23, 135142, doi:10.1029/WR023i001p00135.

    • Search Google Scholar
    • Export Citation
  • Birsan, M. V., , Molnar P. , , Burlando P. , , and Pfaundler M. , 2005: Streamflow trends in Switzerland. J. Hydrol., 314, 312329, doi:10.1016/j.jhydrol.2005.06.008.

    • Search Google Scholar
    • Export Citation
  • Blöschl, G., , Kirnbauer R. , , and Gutknecht D. , 1990: Modelling snowmelt in a mountainous river basin on an event basis. J. Hydrol., 113, 207229, doi:10.1016/0022-1694(90)90176-X.

    • Search Google Scholar
    • Export Citation
  • Blöschl, G., , Kirnbauer R. , , and Gutknecht D. , 1991: Distributed snowmelt simulations in an alpine catchment. 1. Model evaluation on the basis of snow cover patterns. Water Resour. Res., 27, 31713179, doi:10.1029/91WR02250.

    • Search Google Scholar
    • Export Citation
  • Christner, J., , and Harr R. D. , 1982: Peak streamflows from the transient snow zone, western Cascades, Oregon. Proc. 50th Annual Western Snow Conf., Reno, NV, Western Snow Conference, 27–38. [Available online at www.westernsnowconference.org/sites/westernsnowconference.org/PDFs/1982Christner.pdf.]

  • Dozier, J., , and Warren S. G. , 1982: Effect on viewing angle on the infrared brightness temperature. Water Resour. Res., 18, 14241434, doi:10.1029/WR018i005p01424.

    • Search Google Scholar
    • Export Citation
  • Ellis, C. R., , Pomeroy J. W. , , Brown T. , , and MacDonald J. , 2010: Simulation of snow accumulation and melt in needleleaf forest environments. Hydrol. Earth Syst. Sci., 14, 925940, doi:10.5194/hess-14-925-2010.

    • Search Google Scholar
    • Export Citation
  • Ellis, C. R., , Pomeroy J. W. , , Essery R. L. H. , , and Link T. E. , 2011: Effects of needleleaf forest cover on radiation and snowmelt dynamics in the Canadian Rocky Mountains. Can. J. For. Res., 41, 608620, doi:10.1139/X10-227.

    • Search Google Scholar
    • Export Citation
  • Essery, R. L., , Pomeroy J. , , Ellis C. , , and Link T. , 2008: Modelling longwave radiation to snow beneath forest canopies using hemispherical photography or linear regression. Hydrol. Processes, 22, 27882800, doi:10.1002/hyp.6930.

    • Search Google Scholar
    • Export Citation
  • Floyd, W., , and Weiler M. , 2008: Measuring snow accumulation and ablation dynamics during rain-on-snow events: Innovative measurement techniques. Hydrol. Processes, 22, 48054812, doi:10.1002/hyp.7142.

    • Search Google Scholar
    • Export Citation
  • Frazer, G. W., , Canham C. D. , , and Lertzman K. P. , 1999: Gap Light Analyzer (GLA): Imaging software to extract canopy structure and gap light transmission indices from true-colour fisheye photographs. Users Manual and Program Documentation, version 2.0, Simon Fraser University and Institute of Ecosystem Studies, 36 pp. [Available online at http://rem-main.rem.sfu.ca/forestry/downloads/Files/GLAV2UsersManual.pdf.]

  • Freudiger, D., , Kohn I. , , Stahl K. , , and Weiler M. , 2013: Large scale analysis of changing frequencies of rain-on-snow events and their impact on floods. Hydrol. Earth Syst. Sci. Discuss., 10, 13 23113 263, doi:10.5194/hessd-10-13231-2013.

    • Search Google Scholar
    • Export Citation
  • Garvelmann, J., , Pohl S. , , and Weiler M. , 2013: From observation to the quantification of snow processes with a time-lapse camera network. Hydrol. Earth Syst. Sci., 17, 14151429, doi:10.5194/hess-17-1415-2013.

    • Search Google Scholar
    • Export Citation
  • Gelfan, A. N., , Pomeroy J. W. , , and Kuchment L. S. , 2004: Modeling forest cover influences on snow accumulation, sublimation and melt. J. Hydrometeor., 5, 785803, doi:10.1175/1525-7541(2004)005<0785:MFCIOS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Grömping, U., 2007: Estimators of relative importance in linear regression based on variance decomposition. Amer. Stat., 61, 139147, doi:10.1198/000313007X188252.

    • Search Google Scholar
    • Export Citation
  • Hamlet, A. F., , Mote P. W. , , Clark M. P. , , and Lettenmeier D. P. , 2005: Effects of temperature and precipitation variability on snowpack trends in the western United States. J. Climate, 18, 45454561, doi:10.1175/JCLI3538.1.

    • Search Google Scholar
    • Export Citation
  • Hardy, J. P., , Melloh R. A. , , Koenig G. , , Marks D. , , Winstral A. , , and Pomeroy J. W. , , and Link T. , 2004: Solar radiation transmission through conifer canopies. Agric. For. Meteor., 126, 257270, doi:10.1016/j.agrformet.2004.06.012.

    • Search Google Scholar
    • Export Citation
  • Harr, R. D., 1981: Some characteristics and consequences of snowmelt during rainfall in western Oregon. J. Hydrol., 53, 277304, doi:10.1016/0022-1694(81)90006-8.

    • Search Google Scholar
    • Export Citation
  • Harr, R. D., 1986: Effects of clearcutting on rain-on-snow runoff in western Oregon: A new look on old studies. Water Resour. Res., 22, 10951100, doi:10.1029/WR022i007p01095.

    • Search Google Scholar
    • Export Citation
  • Heron, R., , and Woo M. K. , 1978: Snowmelt computations for a high Arctic site. Proc. 35th Eastern Snow Conf., Hanover, NH, Eastern Snow Conference, 162172.

  • Jones, J. A., 2000: Hydrologic processes and peak discharge response to forest removal, regrowth and roads in 10 small experimental basins, western Cascades, Oregon. Water Resour. Res., 36, 26212642, doi:10.1029/2000WR900105.

    • Search Google Scholar
    • Export Citation
  • Jost, G., , Weiler M. , , Gluns D. R. , , and Alila Y. , 2007: The influence of forest and topography on snow accumulation and melt at the watershed-scale. J. Hydrol., 347, 101115, doi:10.1016/j.jhydrol.2007.09.006.

    • Search Google Scholar
    • Export Citation
  • Kattelmann, R. C., 1997: Flooding from rain-on-snow events in the Sierra Nevada. IAHS Publ.,239, 59–65.

  • Koivusalo, H., , and Kokkonen T. , 2002: Snow processes in a forest clearing and in a coniferous forest. J. Hydrol., 262, 145165, doi:10.1016/S0022-1694(02)00031-8.

    • Search Google Scholar
    • Export Citation
  • Köplin, N., , Schäder B. , , Vivirolli D. , , and Weingartner R. , 2013: Seasonality and magnitude of floods in Switzerland under future climate change. Hydrol. Processes, 28, 2567–2578, doi:10.1002/hyp.9757.

    • Search Google Scholar
    • Export Citation
  • Link, T. E., , and Marks D. , 1999a: Point simulation of seasonal snow cover dynamics beneath boreal forest canopies. J. Geophys. Res., 104, 27 84127 857, doi:10.1029/1998JD200121.

    • Search Google Scholar
    • Export Citation
  • Link, T. E., , and Marks D. , 1999b: Distributed simulation of snowcover mass- and energy-balance in the boreal forest. Hydrol. Processes, 13, 24392452, doi:10.1002/(SICI)1099-1085(199910)13:14/15<2439::AID-HYP866>3.0.CO;2-1.

    • Search Google Scholar
    • Export Citation
  • Link, T. E., , Marks D. , , and Hardy J. , 2004: A deterministic method to characterize canopy radiative transfer properties. Hydrol. Processes, 18, 35833594, doi:10.1002/hyp.5793.

    • Search Google Scholar
    • Export Citation
  • Liston, G. E., , and Elder K. , 2006: A distributed snow-evolution modeling system (SnowModel). J. Hydrometeor., 7, 217234, doi:10.1175/JHM486.1.

    • Search Google Scholar
    • Export Citation
  • Male, D. H., , and Granger R. J. , 1981: Snow surface energy exchange. Water Resour. Res., 17, 609627, doi:10.1029/WR017i003p00609.

  • Marks, D., , Dozier J. , , and Davis R. E. , 1992: Climate and energy exchanges at the snow surface in the alpine region of the Sierra Nevada: 1. Meteorological measurements and monitoring. Water Resour. Res., 28, 30293042, doi:10.1029/92WR01482.

    • Search Google Scholar
    • Export Citation
  • Marks, D., , Kimball J. , , Tingey D. , , and Link T. , 1998: The sensitivity of snowmelt processes to climate conditions and forest cover during rain-on-snow: A case study of the 1996 Pacific Northwest flood. Hydrol. Processes, 12, 15691587, doi:10.1002/(SICI)1099-1085(199808/09)12:10/11<1569::AID-HYP682>3.0.CO;2-L.

    • Search Google Scholar
    • Export Citation
  • Marks, D., , Domingo J. , , Susong D. , , Link T. , , and Garen D. , 1999: A spatially distributed energy balance snowmelt model for application in mountain basins. Hydrol. Processes, 13, 19351959, doi:10.1002/(SICI)1099-1085(199909)13:12/13<1935::AID-HYP868>3.0.CO;2-C.

    • Search Google Scholar
    • Export Citation
  • Marks, D., , Link T. , , Winstral A. , , and Garen D. , 2001: Simulating snowmelt processes during rain-on-snow over a semi-arid mountain basin. Ann. Glaciol., 32, 195202, doi:10.3189/172756401781819751.

    • Search Google Scholar
    • Export Citation
  • Marsh, P., , Onclin C. , , Russel M. , , and Pohl S. , 2003: Effects of shrubs on snow processes in the vicinity of the Arctic tree line in NW Canada. Proc. 14th Int. Northern Research Basins Symp. and Workshop, Kangerlussuaq, Greenland, University of Copenhagen, 113117.

  • Martin, E., , and Lejeune Y. , 1998: Turbulent fluxes above the snow surface. Ann. Glaciol., 26, 179183.

  • Mazurkiewicz, A. B., , Callery D. G. , , and McDonnell J. J. , 2008: Assessing the controls of the snow energy balance and water available for runoff in a rain-on-snow environment. J. Hydrol., 354, 1–14, doi:10.1016/j.jhydrol.2007.12.027.

    • Search Google Scholar
    • Export Citation
  • McCabe, G. J., , Hay L. E. , , and Clark M. P. , 2007: Rain-on-snow events in the western United States. Bull. Amer. Meteor. Soc.,88, 319–328, doi:10.1175/BAMS-88-3-319.

  • Moore, R. D., 1983: On the use of bulk aerodynamic formulae over melting snow. Nord. Hydrol., 14, 193206.

  • Morris, E. M., 1989: Turbulent fluxes over snow and ice. J. Hydrol., 105, 205233, doi:10.1016/0022-1694(89)90105-4.

  • Pohl, S., , Marsh P. , , and Liston G. E. , 2006a: Spatial–temporal variability in turbulent fluxes during spring snowmelt. Arct. Antarct. Alp. Res., 38, 136146, doi:10.1657/1523-0430(2006)038[0136:SVITFD]2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Pohl, S., , Marsh P. , , and Pietrino A. , 2006b: Spatial–temporal variability in solar radiation during spring snowmelt. Nord. Hydrol., 37, 119.

    • Search Google Scholar
    • Export Citation
  • Pohl, S., , Garvelmann J. , , Wawerla J. , , and Weiler M. , 2014: Potential of an innovative low cost sensor network to understand the spatial and temporal dynamics of a mountain snow cover. Water Resour. Res., doi:10.1002/2013WR014594, in press.

    • Search Google Scholar
    • Export Citation
  • Pomeroy, J. W., , Gray D. M. , , Hedstrom M. , , and Janowicz J. R. , 2002: Prediction of seasonal snow accumulation in cold climate forests. Hydrol. Processes, 16, 35433558, doi:10.1002/hyp.1228.

    • Search Google Scholar
    • Export Citation
  • Pomeroy, J. W., , Toth B. , , Granger R. M. , , Hedstrom N. R. , , and Essery R. , 2003: Variation in surface energetics during snowmelt in a sub-arctic mountain catchment. J. Hydrometeor., 4, 702719, doi:10.1175/1525-7541(2003)004<0702:VISEDS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Pomeroy, J. W., , Marks D. , , Link T. , , Ellis C. , , Hardy J. , , Rowlands A. , , and Granger R. , 2009: The impact of coniferous forest temperature on incoming longwave radiation to melting snow. Hydrol. Processes, 23, 25132525, doi:10.1002/hyp.7325.

    • Search Google Scholar
    • Export Citation
  • Ranzi, R., , and Rosso R. , 1991: A physically based approach to modeling distributed snowmelt in a small alpine catchment. IAHS Publ.,205, 141–150.

  • Rössler, O., , Froidevaux P. , , Börst U. , , Rickli R. , , Martius O. , , and Weingartner R. , 2013: Retrospective analysis of a non-forecasted rain-on-snow flood in the Alps—A matter of model-limitations or unpredictable nature? Hydrol. Earth Syst. Sci. Discuss., 10, 12 86112 904, doi:10.5194/hessd-10-12861-2013.

    • Search Google Scholar
    • Export Citation
  • Satterlund, D., 1979: An improved equation for estimating long-wave radiation from the atmosphere. Water Resour. Res., 15, 16491650, doi:10.1029/WR015i006p01649.

    • Search Google Scholar
    • Export Citation
  • Sicart, E., , Pomery J. W. , , Essery R. L. , , Hardy J. E. , , Link T. E. , , and Marks D. , 2004: A sensitivity study of daytime net radiation during snowmelt to forest canopy and atmospheric conditions. J. Hydrometeor., 5, 744784.

    • Search Google Scholar
    • Export Citation
  • Singh, P., , Spitzbart G. , , Hübl H. , , and Weinmeister H. W. , 1997: Hydrological response of snowpack under rain-on-snow events: A field study. J. Hydrol., 202, 120, doi:10.1016/S0022-1694(97)00004-8.

    • Search Google Scholar
    • Export Citation
  • Stenberg, P., , Linder S. , , Smolander H. , , and Flower-Ellis J. , 1994: Performance of the LAI-2000 plant canopy analyzer in estimating leaf area index of some Scots pine stands. Tree Physiol., 14, 981995, doi:10.1093/treephys/14.7-8-9.981.

    • Search Google Scholar
    • Export Citation
  • Storck, P., 2000: Trees snow and flooding: An investigation of forest canopy effects on snow accumulation and melt at the plot and watershed scales in the Pacific Northwest. Water Resources Series Tech. Rep. 161, Dept. of Civil and Environmental Engineering, University of Washington, Seattle, WA, 176 pp.

  • Strasser, U., , and Etchevers P. , 2005: Simulation of daily discharges for the upper Durance catchment (French Alps) using subgrid parametrization for topography and a forest canopy climate model. Hydrol. Processes, 19, 23612373, doi:10.1002/hyp.5889.

    • Search Google Scholar
    • Export Citation
  • Strasser, U., , and Marke T. , 2010: ESCIMO.spread —A spreadsheet-based point snow surface energy balance model to calculate hourly snow water equivalent and melt rates for historical and changing climate conditions. Geosci. Model Dev., 3, 643652, doi:10.5194/gmd-3-643-2010.

    • Search Google Scholar
    • Export Citation
  • Strasser, U., , Warscher M. , , and Liston G. E. , 2011: Modeling snow-canopy processes on an idealized mountain. J. Hydrometeor., 12, 663677, doi:10.1175/2011JHM1344.1.

    • Search Google Scholar
    • Export Citation
  • Sui, J., , and Koehler G. , 2001: Rain-on-snow induced flood events in southern Germany. J. Hydrol., 252, 205220, doi:10.1016/S0022-1694(01)00460-7.

    • Search Google Scholar
    • Export Citation
  • Susong, D., , Marks D. , , and Garen D. ,1999: Methods for developing time-series climate surface to drive topographically distributed energy- and water-balance models. Hydrol. Processes,13, 2003–2021, doi:10.1002/(SICI)1099-1085(199909)13:12/13<2003::AID-HYP884>3.0.CO;2-K.

  • Tetens, V. O., 1930: Ueber einige meteorologische Begriffe (in German). Z. Geophys., 6, 297309.

  • Tribbeck, M. J., , Gurney R. , , Morris E. M. , , and Pearson D. W. C. , 2004: A new Snow-SVAT to simulate the accumulation and ablation of seasonal snow beneath a forest canopy. J. Glaciol., 50, 171182, doi:10.3189/172756504781830187.

    • Search Google Scholar
    • Export Citation
  • U.S. Army Corps of Engineers, 1956: Snow hydrology: Summary report of the snow investigations. North Pacific Division Army Corps of Engineers, Portland, OR, 437 pp. [Available online at ftp://ftp.wcc.nrcs.usda.gov/wntsc/H&H/snow/SnowHydrologyCOE1956.pdf.]

  • Welles, J. M., , and Norman J. M. , 1991: Instrument for indirect measurement of canopy architecture. Agron. J., 83, 818825, doi:10.2134/agronj1991.00021962008300050009x.

    • Search Google Scholar
    • Export Citation
  • Whitaker, A. C., , and Sugiyama H. , 2005: Seasonal snowpack dynamics and runoff in a cool temperate forest: Lysimeter experiment in Niigata, Japan. Hydrol. Processes, 19, 41794200, doi:10.1002/hyp.6059.

    • Search Google Scholar
    • Export Citation
Save