• Alcott, T. I., , and W. J. Steenburgh, 2013: Orographic influences on a Great Salt Lake–effect snowstorm. Mon. Wea. Rev., 141, 24322450, doi:10.1175/MWR-D-12-00328.1.

    • Search Google Scholar
    • Export Citation
  • Andersson, T., , and S. Nilsson, 1990: Topographically induced convective snowbands over the Baltic Sea and their precipitation distribution. Wea. Forecasting, 5, 299312, doi:10.1175/1520-0434(1990)005<0299:TICSOT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Bluestein, H. B., 1993: Synoptic-Dynamic Meteorology in Midlatitudes. Vol. II, Observations and Theory of Weather Systems, Oxford University Press, 594 pp.

  • Bohren, C. F., , and B. A. Albrecht, 2010: Atmospheric Thermodynamics. Oxford University Press, 402 pp.

  • Buisan, S. T., , M. E. Earle, , J. L. Collado, , J. Kochendorfer, , J. Alastrué, , M. Wolff, , C. D. Smith, , and J. I. López-Moreno, 2016: Assessment of the underestimation of snowfall accumulation by tipping bucket gauges used operationally by the Spanish national weather service. Atmos. Meas. Tech. Discuss., doi:10.5194/amt-2016-197, in press.

    • Search Google Scholar
    • Export Citation
  • Cairns, M. M., , R. Collin, , T. Cylke, , M. Deutschendort, , and D. Mercer, 2001: A lake-effect snowfall in western Nevada—Part I: Synoptic setting and observations. 18th Conf. on Weather Analysis and Forecasting/14th Conf. on Numerical Weather Prediction, Fort Lauderdale, FL, Amer. Meteor. Soc., 5.5. [Available online at https://ams.confex.com/ams/WAF-NWP-MESO/techprogram/paper_22957.htm.]

  • Carpenter, D. M., 1993: The lake effect of the Great Salt Lake: Overview and forecast problems. Wea. Forecasting, 8, 181193, doi:10.1175/1520-0434(1993)008<0181:TLEOTG>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Conrick, R., , H. D. Reeves, , and S. Zhong, 2015: The dependence of QPF on the choice of boundary- and surface-layer parameterization for a lake-effect snowstorm. J. Appl. Meteor. Climatol., 54, 11771190, doi:10.1175/JAMC-D-14-0291.1.

    • Search Google Scholar
    • Export Citation
  • German Meteorological Society (DGM), 2014: Europäischer Meteorologischer Kalender 2015 (European Meteorological Calendar 2015). Gebr. Borntraeger Verlagsbuchhandlung. [Available online at http://www.borntraeger-cramer.de/9783443010799.]

  • Ek, M. B., , K. E. Mitchell, , Y. Lin, , E. Rogers, , P. Grunmann, , V. Koren, , G. Gayno, , and J. D. Tarpley, 2003: Implementation of Noah land surface model advances in the National Centers for Environmental Prediction operational mesoscale Eta model. J. Geophys. Res., 108, 8851, doi:10.1029/2002JD003296.

    • Search Google Scholar
    • Export Citation
  • Goodison, B. E., , P. Y. T. Louie, , and D. Yang, 1998: WMO solid precipitation measurement intercomparison. WMO Instruments and Observing Methods Rep. 67, WMO/TD-872, WMO, 212 pp.

  • Hjelmfelt, M. R., 1992: Orographic effects in simulated lake-effect snowstorms over Lake Michigan. Mon. Wea. Rev., 120, 373377, doi:10.1175/1520-0493(1992)120<0373:OEISLE>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Holroyd, E. W., 1971: Lake-effect cloud bands as seen from weather satellites. J. Atmos. Sci., 28, 11651170, doi:10.1175/1520-0469(1971)028<1165:LECBAS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Huggins, A. W., , D. E. Kingsmill, , and M. M. Cairns, 2001: A lake effect snowfall in western Nevada—Part II: Radar characteristics and quantitative precipitation estimates. 18th Conf. on Weather Analysis and Forecasting/14th Conf. on Numerical Weather Prediction, Fort Lauderdale, FL, Amer. Meteor. Soc., 5.6. [Available online at https://ams.confex.com/ams/WAF-NWP-MESO/techprogram/paper_22944.htm.]

  • Jiang, Q., , J. D. Doyle, , and R. B. Smith, 2005: Blocking, descent and gravity waves: Observations and modelling of a MAP northerly föhn event. Quart. J. Roy. Meteor. Soc., 131, 675701, doi:10.1256/qj.03.176.

    • Search Google Scholar
    • Export Citation
  • Kain, J. S., 2004: The Kain–Fritsch convective parameterization: An update. J. Appl. Meteor., 43, 170181, doi:10.1175/1520-0450(2004)043<0170:TKCPAU>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Kindap, T., 2010: A severe sea-effect snow episode over the city of Istanbul. Nat. Hazards, 54, 707723, doi:10.1007/s11069-009-9496-7.

    • Search Google Scholar
    • Export Citation
  • Laird, N. F., , J. E. Walsh, , and D. A. R. Kristovich, 2003: Model simulations examining the relationship of lake-effect morphology to lake shape, wind direction, and wind speed. Mon. Wea. Rev., 131, 21022111, doi:10.1175/1520-0493(2003)131<2102:MSETRO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Laird, N. F., , J. Desrochers, , and M. Payer, 2009a: Climatology of lake-effect precipitation events over Lake Champlain. J. Appl. Meteor. Climatol., 48, 232250, doi:10.1175/2008JAMC1923.1.

    • Search Google Scholar
    • Export Citation
  • Laird, N. F., , R. Sobash, , and N. Hodas, 2009b: The frequency and characteristics of lake-effect precipitation events associated with the New York State Finger Lakes. J. Appl. Meteor. Climatol., 48, 873886, doi:10.1175/2008JAMC2054.1.

    • Search Google Scholar
    • Export Citation
  • Laird, N. F., , R. Sobash, , and N. Hodas, 2010: Climatological conditions of lake-effect precipitation events associated with the New York State Finger Lakes. J. Appl. Meteor. Climatol., 49, 10521062, doi:10.1175/2010JAMC2312.1.

    • Search Google Scholar
    • Export Citation
  • Laird, N. F., , N. D. Metz, , L. Gaudet, , C. Grasmick, , L. Higgins, , C. Loeser, , and D. A. Zelinsky, 2016: Climatology of cold season lake-effect cloud bands for the North American Great Lakes. Int. J. Climatol., doi:10.1002/joc.4838, in press.

    • Search Google Scholar
    • Export Citation
  • Lang, U., , R. Schick, , and G. Schroder, 2010: The Decision Support System BodenseeOnline for hydrodynamics and water quality in Lake Constance. Advances in Decision Support Systems, G. Devlin, Ed., InTech, doi:10.5772/39391.

  • Lavoie, R. L., 1972: A mesoscale numerical model of lake-effect storms. J. Atmos. Sci., 29, 10251040, doi:10.1175/1520-0469(1972)029<1025:AMNMOL>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Lin, Y.-L., , R. D. Farley, , and H. D. Orville, 1983: Bulk parameterization of the snow field in a cloud model. J. Climate Appl. Meteor., 22, 10651092, doi:10.1175/1520-0450(1983)022<1065:BPOTSF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Lofgren, B. M., , and Y. Zhu, 2000: Surface energy fluxes on the Great Lakes based on satellite-observed surface temperatures 1992 to 1995. J. Great Lakes Res., 26, 305314, doi:10.1016/S0380-1330(00)70694-0.

    • Search Google Scholar
    • Export Citation
  • Markowski, P., , and Y. Richardson, 2010: Mesoscale Meteorology in Midlatitudes. Wiley-Blackwell, 430 pp.

  • Minder, J. R., , T. W. Letcher, , L. S. Campbell, , P. G. Veals, , and W. J. Steenburgh, 2015: The evolution of lake-effect convection during landfall and orographic uplift as observed by profiling radars. Mon. Wea. Rev., 143, 44224442, doi:10.1175/MWR-D-15-0117.1.

    • Search Google Scholar
    • Export Citation
  • Mlawer, E. J., , S. J. Taubman, , P. D. Brown, , M. J. Iacono, , and S. A. Clough, 1997: Radiative transfer for inhomogeneous atmospheres: RRTM, a validated correlated-k model for the longwave. J. Geophys. Res., 102, 16 66316 682, doi:10.1029/97JD00237.

    • Search Google Scholar
    • Export Citation
  • Muller, R. A., 1966: Snowbelts of the Great Lakes. Weatherwise, 19, 248255, doi:10.1080/00431672.1966.10544204.

  • Nagata, W. S., 1987: On the structure of a convergent cloud band over the Japan Sea in winter: A prediction experiment. J. Meteor. Soc. Japan, 65, 871883.

    • Search Google Scholar
    • Export Citation
  • Nakanishi, M., , and H. Niino, 2004: An improved Mellor–Yamada level-3 model with condensation physics: Its design and verification. Bound.-Layer Meteor., 112, 131, doi:10.1023/B:BOUN.0000020164.04146.98.

    • Search Google Scholar
    • Export Citation
  • Nicosia, D. J., and et al. , 1999: A flash flood from a lake-enhanced rainband. Wea. Forecasting, 14, 271288, doi:10.1175/1520-0434(1999)014<0271:AFFFAL>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Niziol, T. A., 1987: Operational forecasting of lake effect snowfall in western and central New York. Wea. Forecasting, 2, 310321, doi:10.1175/1520-0434(1987)002<0310:OFOLES>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Niziol, T. A., , W. R. Snyder, , and J. S. Waldstreicher, 1995: Winter weather forecasting throughout the eastern United States. Part IV: Lake effect snow. Wea. Forecasting, 10, 6177, doi:10.1175/1520-0434(1995)010<0061:WWFTTE>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Norris, J. G., , G. Vaughan, , and D. M. Schultz, 2013: Snowbands over the English Channel and Irish Sea during cold-air outbreaks. Quart. J. Roy. Meteor. Soc., 139, 17471761, doi:10.1002/qj.2079.

    • Search Google Scholar
    • Export Citation
  • Onton, D. J., , and W. J. Steenburgh, 2001: Diagnostic and sensitivity studies of the 7 December 1998 Great Salt Lake–effect snowstorm. Mon. Wea. Rev., 129, 13181338, doi:10.1175/1520-0493(2001)129<1318:DASSOT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Peace, R. L., , and R. B. Sykes, 1966: Mesoscale study of a lake effect snow storm. Mon. Wea. Rev., 94, 495507, doi:10.1175/1520-0493(1966)094<0495:MSOALE>2.3.CO;2.

    • Search Google Scholar
    • Export Citation
  • Rasmussen, R., and et al. , 2012: How well are we measuring snow: The NOAA/FAA/NCAR Winter Precipitation Test Bed. Bull. Amer. Meteor. Soc., 93, 811829, doi:10.1175/BAMS-D-11-00052.1.

    • Search Google Scholar
    • Export Citation
  • Reiner, G., 2015: Der Bodensee und seine Mitschuld am Schneechaos (The Lake Constance and its complicity in the snow chaos). Accessed 19 May 2016, Vorarlberger Nachrichten. [Available online at http://www.vorarlbergernachrichten.at/sonderthemen/2013/02/08/.]

  • Riffler, M., , G. Lieberherr, , and S. Wunderle, 2015: Lake surface water temperatures of European Alpine lakes (1989–2013) based on the Advanced Very High Resolution Radiometer (AVHRR) 1 km data set. Earth Syst. Sci. Data, 7, 117, doi:10.5194/essd-7-1-2015.

    • Search Google Scholar
    • Export Citation
  • Saito, K., , M. Murakami, , T. Matsuo, , and H. Mizuno, 1996: Sensitivity experiments on the orographic snowfall over the mountainous region of northern Japan. J. Meteor. Soc. Japan, 74, 797813.

    • Search Google Scholar
    • Export Citation
  • Schultz, D. M., , D. S. Arndt, , D. J. Stensrud, , and J. W. Hanna, 2004: Snowbands during the cold-air outbreak of 23 January 2003. Mon. Wea. Rev., 132, 827842, doi:10.1175/1520-0493(2004)132<0827:SDTCOO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Schumacher, R. S., , D. M. Schultz, , and J. A. Knox, 2015: Influence of terrain resolution on banded convection in the lee of the Rocky Mountains. Mon. Wea. Rev., 143, 13991416, doi:10.1175/MWR-D-14-00255.1.

    • Search Google Scholar
    • Export Citation
  • Skamarock, W. C., and et al. , 2008: A description of the Advanced Research WRF version 3. NCAR Tech. Note NCAR/TN-475+STR, 113 pp., doi:10.5065/D68S4MVH.

  • Steenburgh, W. J., 2003: One hundred inches in one hundred hours: Evolution of a Wasatch Mountain winter storm cycle. Wea. Forecasting, 18, 10181036, doi:10.1175/1520-0434(2003)018<1018:OHIIOH>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Steenburgh, W. J., , and D. J. Onton, 2001: Multiscale analysis of the 7 December 1998 Great Salt Lake–effect snowstorm. Mon. Wea. Rev., 129, 12961317, doi:10.1175/1520-0493(2001)129<1296:MAOTDG>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Steenburgh, W. J., , S. F. Halvorson, , and D. J. Onton, 2000: Climatology of lake-effect snowstorms of the Great Salt Lake. Mon. Wea. Rev., 128, 709727, doi:10.1175/1520-0493(2000)128<0709:COLESO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Steiger, S. M., and et al. , 2013: Circulations, bounded weak echo regions, and horizontal vortices observed within long-lake-axis-parallel–lake-effect storms by the Doppler on Wheels. Mon. Wea. Rev., 141, 28212840, doi:10.1175/MWR-D-12-00226.1.

    • Search Google Scholar
    • Export Citation
  • Umek, L., 2016: Lake and orographic effects on precipitation at Lake Constance: Case study and numerical sensitivity experiments. M.S. thesis, Institute of Atmospheric and Cryospheric Sciences, University of Innsbruck, 69 pp. [Available online at http://resolver.obvsg.at/urn:nbn:at:at-ubi:1-3615.]

  • Veals, P. G., , and W. J. Steenburgh, 2015: Climatological characteristics and orographic enhancement of lake-effect precipitation east of Lake Ontario and over the Tug Hill Plateau. Mon. Wea. Rev., 143, 35913609, doi:10.1175/MWR-D-15-0009.1.

    • Search Google Scholar
    • Export Citation
  • Wilken, G. R., 1997: A lake-effect snow in Arkansas. Tech. Attachment SR/SSD 97-21, NSFO Little Rock, AR. [Available online at http://www.srh.noaa.gov/topics/attach/html/ssd97-21.htm.]

  • Wilson, J. W., 1977: Effect of Lake Ontario on precipitation. Mon. Wea. Rev., 105, 207214, doi:10.1175/1520-0493(1977)105<0207:EOLOOP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Wright, D. M., , D. J. Posselt, , and A. L. Steiner, 2013: Sensitivity of lake-effect snowfall to lake ice cover and temperature in the Great Lakes region. Mon. Wea. Rev., 141, 670689, doi:10.1175/MWR-D-12-00038.1.

    • Search Google Scholar
    • Export Citation
  • Yeager, K. N., , W. J. Steenburgh, , and T. I. Alcott, 2013: Contributions of lake-effect periods to the cool-season hydroclimate of the Great Salt Lake basin. J. Appl. Meteor. Climatol., 52, 341362, doi:10.1175/JAMC-D-12-077.1.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Lake Constance region with locations of automatic weather stations used in this study marked by black diamonds. Abbreviations and height above MSL for the stations as follows: Alberschwende (ALB, 715 m), Altenrhein (ALT, 399 m), Bregenz (BRE, 424 m), Dornbirn (DOR, 407 m), Friedrichshafen (FRI, 417 m), Kressbronn (KRE, 450 m), Lindau (LIN, 400 m), Meersburg (MEE, 399 m), and Rohrspitz (ROH, 395 m). The blue and the black crisscrosses mark the locations of vertical profiles shown in Fig. 10. Topography as color contours with an increment of 100 m based on ETOPO1 data with 1 arc-min resolution. Thin black lines denote political boundaries.

  • View in gallery

    (a) Precipitation accumulated over 36 h from 0000 to 1200 UTC of the following day measured by AWS at Meersburg (MEE), Lindau (LIN), Bregenz (BRE), and Alberschwende (ALB) for several events with stationary banded precipitation (see Fig. 1 for the location of the stations). The rather long accumulation period accounts for the fact that some events lasted until the next day. (b) Difference between the daily mean LST measured in the harbor of Bregenz at a depth of 0.5 m and the air temperature at 2 m AGL in Bregenz averaged over the same 36-h period as in (a).

  • View in gallery

    (a) Domains of WRF simulations. Horizontal grid spacing annotated in bottom-left corner, except for innermost domain (x = 0.9 km). WRF terrain elevation of innermost domain as gray shading with 200-m contour interval (scale at right) for (b) CTL, (c) NLs and NLr, (d) FLAT, (e) NoBF, and (f) NoDT simulations. See Table 1 for a short description of the simulations. Thick black lines denote shorelines. Thin black lines mark political boundaries. A white triangle in (b) represents the location of the Valluga radar site.

  • View in gallery

    (a) Air temperature at 2 m AGL (°C, solid line) and relative humidity (%, dashed line) and (b) accumulated precipitation (mm, solid line) and pressure reduced to mean sea level (hPa, dashed line) measured by the AWS in Bregenz (see location in Fig. 1) between 0600 UTC 6 Feb and 0000 UTC 9 Feb 2013. Passage of a cold front occurred at about 1500 UTC 6 Feb. The period with a stationary snowband in extension of Lake Constance visible in the radar imagery is indicated by the gray shading.

  • View in gallery

    Synoptic situation at 0600 UTC 8 Feb 2013 based on ECMWF analysis data. Geopotential height (m) at 700 hPa as purple contour lines with 40-m interval (every second line labeled), mean relative humidity (%) between 850 and 700 hPa as color contours (scale at right), and wind at 700 hPa as wind barbs. Half barbs, full barbs, and triangles represent 2.5, 5, and 25 m s−1, respectively. A red star indicates the location of Lake Constance.

  • View in gallery

    Maximum projection of radar reflectivity derived from a series of PPI scans at multiple elevation angles ranging from −2° to 60° as color contours (dBZ, scale at right) at (a) 0645, (b) 0800, (c) 0900, and (d) 1015 UTC 8 Feb 2013. Data are taken from the Valluga radar (see location in Fig. 3b). Terrain elevation shown as gray contour lines with an increment of 200 m, lake shoreline as thick black line, and political boundaries as thin black lines. Black triangle (diamond) identifies AWS Lindau (Bregenz).

  • View in gallery

    (a) Radar-estimated accumulated precipitation as color contours and accumulated precipitation measured with rain gauges as circles with color shading (mm, scale at right) from 0640 to 1200 UTC (0700–1200 UTC for stations with hourly temporal resolution) 8 Feb 2013. The radar-estimated precipitation is based on radar data with a 5-min temporal resolution. (b) Accumulated precipitation as in (a) valid for CTL from 0700 to 1200 UTC 8 Feb 2013. Terrain elevation shown as brown contour lines with an increment of 200 m. Thick black lines denote lake shorelines. Thin black lines identify political boundaries. A dashed rectangle in (b) identifies lake-effect subdomain for calculation of spatial average and maximum value of accumulated precipitation (cf. Table 1).

  • View in gallery

    Observations from several AWS in the Lake Constance region (see locations and abbreviations in Fig. 1) from 0300 to 1500 UTC 8 Feb 2013: (a) accumulated precipitation in mm and (b) wind direction in degrees. Red (blue) dots and triangles in (b) denote stations on the southern (northern) shore. The wind direction in (b) is only plotted for wind speeds exceeding 2 m s−1. Gray shading identifies the period with a snowband visible in radar imagery. Elevation of each AWS in m MSL is given in the legend.

  • View in gallery

    Precipitation of the control simulation accumulated over the previous hour in mm as color contours (scale at right), horizontal wind vectors at 10 m AGL plotted at every sixth grid point [reference vector shown in (a)], and model terrain elevation as brown contour lines with increment of 200 m at (a) 0700, (b) 0800, (c) 0900, (d) 1000, (e) 1100, and (f) 1200 UTC 8 Feb 2013.

  • View in gallery

    Vertical profiles of θ, , and taken from the control simulation (a) over Lake Constance and (b) over land west of the lake at 0600 (solid) and 1100 UTC (dashed) 8 Feb 2013. Location of vertical profiles is shown in Fig. 1. Red shading in (a) indicates the layer with absolute unstable stratification at 0600 UTC. The LST in the numerical model is set to 276.5 K.

  • View in gallery

    Horizontal wind at 700 m MSL as vectors plotted at every sixth grid point [reference vector shown in (a)], and horizontal divergence of wind at the same height as color contours (in 10−4 s−1, scale at right) at 0900 UTC 8 Feb 2013 for the simulations (a) CTL, (b) FLAT, (c) FLAT-NLs, (d) NLs, (e) LST-3K, (f) LST+3K, (g) FLAT-LST+3K, (h) NLr, (i) NoBF, (j) NoBF-NLs, (k) NoDT, and (l) NoDT-NLs. See Table 1 for a short description of the simulations. Model topography shown as brown contours with an interval of 200 m. Gray shading indicates areas below the model terrain.

  • View in gallery

    LCL, LFC, and LNB as a function of height, calculated for the same grid point over Lake Constance as the profile shown in Fig. 10a prior to the snowband formation at 0600 UTC 8 Feb 2013. Location of grid point is shown in Fig. 1. The abscissa represents heights relative to the starting level z.

  • View in gallery

    Vertical cross section A–B along the northern (German) shore of Lake Constance at 0900 UTC 8 Feb 2015 for (a) CTL and (b) NLs simulation: conditional instability () as blue shading; absolute instability () as red shading; potential temperature θ as thin black contour lines with 0.5-K increments; total hydrometeor mixing ratio as thick green contour lines for 0.1, 0.4, and 0.6 g kg−1; and wind vectors for the wind component parallel to the transect [reference vectors for horizontal and vertical wind speed in bottom-right corner of (a)]. Location of cross section is shown in the inset of (a). Gray hatched rectangle below the model topography in (a) denotes intersection of cross section with the water body.

  • View in gallery

    As in Fig. 13, but valid for CTL and cross sections perpendicular to long axis of the lake: (a) cross section C1–D1 and (b) cross section C2–D2. Locations of cross sections are shown in the inset in (b).

  • View in gallery

    As in Fig. 7b, but for the simulations (a) CTL, (b) FLAT, (c) FLAT-NLs, (d) NLs, (e) LST-3K, (f) LST+3K, (g) FLAT-LST+3K, (h) NLr, (i) NoBF, (j) NoBF-NLs, (k) NoDT, and (l) NoDT-NLs. A dashed rectangle in (a) denotes the lake-effect subdomain for the calculation of spatially averaged and maximum value of accumulated precipitation. See Table 1 for a short description of the simulations and statistics for the lake-effect subdomain.

  • View in gallery

    As in Fig. 13, but for the NoDT simulation at 0800 UTC.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 83 83 7
PDF Downloads 65 65 6

Lake and Orographic Effects on a Snowstorm at Lake Constance

View More View Less
  • 1 Institute of Atmospheric and Cryospheric Sciences, University of Innsbruck, Innsbruck, Austria
© Get Permissions
Full access

Abstract

This is one of the first case studies of a snowstorm at Lake Constance, located between Austria, Germany, and Switzerland, which assesses the influence of the lake and the orography on the generation of heavy precipitation. The analysis is based on surface and radar observations and numerical simulations with the Weather Research and Forecasting (WRF) Model. On 8 February 2013, a rather stationary and banded radar reflectivity pattern was observed during postfrontal conditions with northwesterly flow. The associated snowband affected the downstream shore and the adjacent mountainous region with 36 mm of precipitation within 5 h at the shore. Surface observations show a convergence in the wind field over the lake during the period of banded precipitation. The control simulation captures the formation of a convergence line and a snowband near the shoreline and over the downstream orography. A lake-induced, low-level conditionally unstable layer is essential for the snowband formation. Orographically and thermally induced convergence provides the lifting to release conditional instability and to trigger convection. Orographic enhancement of precipitation occurs downstream of the lake. Sensitivity experiments with modified orography, land use, and lake surface temperature show that the lake is a crucial factor controlling the amount and distribution of snowfall. However, neither the lake nor the orography alone would have been able to form a snowband. This study highlights the complex interaction between lake and orographic effects and shows that Lake Constance is large enough to impact the formation of precipitation.

Denotes Open Access content.

Corresponding author address: Lukas Umek, Institute of Atmospheric and Cryospheric Sciences, University of Innsbruck, Innrain 52f, Innsbruck, Austria. E-mail: lukas.umek@gmail.com

Abstract

This is one of the first case studies of a snowstorm at Lake Constance, located between Austria, Germany, and Switzerland, which assesses the influence of the lake and the orography on the generation of heavy precipitation. The analysis is based on surface and radar observations and numerical simulations with the Weather Research and Forecasting (WRF) Model. On 8 February 2013, a rather stationary and banded radar reflectivity pattern was observed during postfrontal conditions with northwesterly flow. The associated snowband affected the downstream shore and the adjacent mountainous region with 36 mm of precipitation within 5 h at the shore. Surface observations show a convergence in the wind field over the lake during the period of banded precipitation. The control simulation captures the formation of a convergence line and a snowband near the shoreline and over the downstream orography. A lake-induced, low-level conditionally unstable layer is essential for the snowband formation. Orographically and thermally induced convergence provides the lifting to release conditional instability and to trigger convection. Orographic enhancement of precipitation occurs downstream of the lake. Sensitivity experiments with modified orography, land use, and lake surface temperature show that the lake is a crucial factor controlling the amount and distribution of snowfall. However, neither the lake nor the orography alone would have been able to form a snowband. This study highlights the complex interaction between lake and orographic effects and shows that Lake Constance is large enough to impact the formation of precipitation.

Denotes Open Access content.

Corresponding author address: Lukas Umek, Institute of Atmospheric and Cryospheric Sciences, University of Innsbruck, Innrain 52f, Innsbruck, Austria. E-mail: lukas.umek@gmail.com

1. Introduction

In recent years, several heavy snowfall events affected the southeastern shore of Lake Constance, located between Austria, Germany, and Switzerland (Fig. 1). Radar imagery revealed rather stationary and banded precipitation downstream of the lake. The structures resembled wind-parallel precipitation bands orientated roughly parallel to the major axis of the lake [i.e., type-I snowbands according to the classification of Niziol et al. (1995)]. Some of these events occurred rather surprisingly, as neither high amounts of snowfall nor banded precipitating structures were forecasted by the Austrian national weather service. It is suspected that the lake and the surrounding terrain play a major role in generating precipitation. However, a detailed study that supports this hypothesis and identifies the relevant precipitation mechanisms is lacking. This paper tries to fill this gap.

Fig. 1.
Fig. 1.

Lake Constance region with locations of automatic weather stations used in this study marked by black diamonds. Abbreviations and height above MSL for the stations as follows: Alberschwende (ALB, 715 m), Altenrhein (ALT, 399 m), Bregenz (BRE, 424 m), Dornbirn (DOR, 407 m), Friedrichshafen (FRI, 417 m), Kressbronn (KRE, 450 m), Lindau (LIN, 400 m), Meersburg (MEE, 399 m), and Rohrspitz (ROH, 395 m). The blue and the black crisscrosses mark the locations of vertical profiles shown in Fig. 10. Topography as color contours with an increment of 100 m based on ETOPO1 data with 1 arc-min resolution. Thin black lines denote political boundaries.

Citation: Monthly Weather Review 144, 12; 10.1175/MWR-D-16-0032.1

In a preliminary unpublished analysis, the authors have identified several potential lake-effect events between 2008 and 2014 (Umek 2016). Figure 2 shows six events based on observed precipitation accumulated over 36 h at four automatic weather stations (AWS), as well as the difference between the daily mean lake surface temperature (LST) measured in the harbor of Bregenz at a depth of 0.5 m and the mean 2-m air temperature in Bregenz. The town of Bregenz (capital of Vorarlberg, which is the westernmost federal state of Austria) and the mountainous region to the southeast (Bregenzerwald, Fig. 1) experienced high amounts of precipitation, while other stations around the lake received far less precipitation in the same period of time. Some events were rather short or the rain gauges were not located at the precipitation maximum (i.e., at the center of the band). Both effects allowed the incident to appear less intense (e.g., 22–23 November 2008). This paper presents a detailed case study of the 8 February 2013 based on observations from AWS and radar as well as numerical simulations. The event is characterized by over 35 mm of accumulated precipitation in 6 h and up to 62 mm in 36 h. This heavy precipitation caused an extraordinary traffic chaos in Bregenz (Reiner 2015). An impressive photograph of the convective clouds and the associated snowfall, which occurred over the lake also on the next day of the event, is shown in German Meteorological Society (DGM) (2014).

Fig. 2.
Fig. 2.

(a) Precipitation accumulated over 36 h from 0000 to 1200 UTC of the following day measured by AWS at Meersburg (MEE), Lindau (LIN), Bregenz (BRE), and Alberschwende (ALB) for several events with stationary banded precipitation (see Fig. 1 for the location of the stations). The rather long accumulation period accounts for the fact that some events lasted until the next day. (b) Difference between the daily mean LST measured in the harbor of Bregenz at a depth of 0.5 m and the air temperature at 2 m AGL in Bregenz averaged over the same 36-h period as in (a).

Citation: Monthly Weather Review 144, 12; 10.1175/MWR-D-16-0032.1

The preliminary analysis showed that these events typically occur after the passage of a cold front, resulting in a lower air temperature than the LST (Fig. 2b). Such a near-surface temperature gradient implies a positive surface sensible heat flux. Therefore, it is suspected that the lake has a significant impact on the formation of precipitation bands by destabilizing and moistening the planetary boundary layer (e.g., Markowski and Richardson 2010). Hence, our hypothesis is that Lake Constance is able to produce lake-effect precipitation, a well-known phenomenon usually caused by significantly larger lakes, for instance the Great Lakes (e.g., Muller 1966; Peace and Sykes 1966; Niziol 1987; Niziol et al. 1995; Laird et al. 2016) or the Great Salt Lake (e.g., Carpenter 1993; Steenburgh et al. 2000; Onton and Steenburgh 2001; Steenburgh 2003; Alcott and Steenburgh 2013). Lake-effect precipitation is also documented over other water bodies worldwide, like the Sea of Japan (e.g., Nagata 1987), the Baltic Sea (e.g., Andersson and Nilsson 1990), the Black Sea (e.g., Kindap 2010), and the British Isles (e.g., Norris et al. 2013).

Lake Constance is surrounded by complex terrain with the Swabian Jura to the north, the Black Forest to the northwest, and the first foothills of the Alps south and east of the lake (Fig. 3b). Especially the latter may contribute to the formation of precipitation. For example, the topography rises from about 400 m above mean sea level (MSL) at the shore to about 1000 m MSL at the peak of the Pfänder, a mountain immediately east of Bregenz (Fig. 1). In this study we denote the shoreline between Bregenz and Meersburg, Germany, as the northern shore (cf. Fig. 1).

Fig. 3.
Fig. 3.

(a) Domains of WRF simulations. Horizontal grid spacing annotated in bottom-left corner, except for innermost domain (x = 0.9 km). WRF terrain elevation of innermost domain as gray shading with 200-m contour interval (scale at right) for (b) CTL, (c) NLs and NLr, (d) FLAT, (e) NoBF, and (f) NoDT simulations. See Table 1 for a short description of the simulations. Thick black lines denote shorelines. Thin black lines mark political boundaries. A white triangle in (b) represents the location of the Valluga radar site.

Citation: Monthly Weather Review 144, 12; 10.1175/MWR-D-16-0032.1

The combined effects of the lake and its surrounding orography in forming and distributing precipitation have been studied only by a few authors (e.g., Hjelmfelt 1992; Saito et al. 1996; Onton and Steenburgh 2001; Alcott and Steenburgh 2013). Enhancement of lake-effect precipitation by gently sloping terrain or rather low mountains downstream of water bodies is mentioned in several studies (e.g., Lavoie 1972; Wilson 1977; Niziol et al. 1995; Yeager et al. 2013; Minder et al. 2015; Veals and Steenburgh 2015). It is conceivable that the banded precipitation at Lake Constance is generated by a combination of lake effects (e.g., latent and sensible heat input to the planetary boundary layer) and orographic effects, such as flow deflection and forced lifting by the surrounding terrain.

For the formation of lake-effect snowstorms in the Great Lakes area, the difference between the LST and the temperature at 850 hPa has to be greater than the dry adiabatic temperature decrease (Holroyd 1971). This threshold often appears in literature and is also used for other lakes (e.g., the Great Salt Lake), but with the 700-hPa level instead of the 850-hPa level because of the higher elevation of the Great Salt Lake (Steenburgh et al. 2000).

Steenburgh et al. (2000) illustrated that lake-effect events at the Great Salt Lake and the Great Lakes typically occur for postfrontal westerly to northerly flow at 700 hPa following the passage of an upper-level trough and the associated low-level cold front. Further, a few important conditions need to be met for the formation of lake-effect snowbands: (i) a conditionally unstable low-level temperature lapse rate within 150 hPa of the surface, (ii) the absence of capping inversions or stable layers within 150 hPa of the surface, (iii) a directional wind shear smaller than 60° in the layer forming the snowband, and (iv) the presence of a low-level convergence over the lake (Steenburgh et al. 2000).

Lake Constance is located at an elevation of 394 m MSL and covers an area of about 540 km2 with a major axis of about 60 km and a mean depth of 90 m. The longest possible fetch over the water body is approximately 50–60 km for northwesterly winds. For the Great Lakes, Bluestein (1993) mentions a minimal fetch of 80 km needed for lake-effect precipitation. In contrast, research on lake-effect precipitation at smaller lakes, comparable to the size of Lake Constance, has been conducted by Cairns et al. (2001) and Huggins et al. (2001) for Lake Tahoe (490 km2, 35-km fetch) and Pyramid Lake (487 km2, 40-km fetch), by Wilken (1997) for Bull Shoals Lake (183 km2, 78-km fetch), by Schultz et al. (2004) for Lake Kentucky (650 km2, 40–60-km fetch), by Laird et al. (2009a) for Lake Champlain (1127 km2, 190-km length), and by Laird et al. (2009b, 2010) for the New York State Finger Lakes (up to 175 km2 and 64-km length).

The difference in surface roughness over water and land can influence the formation of lake-effect precipitation due to mesoscale convergence on the downwind right (left) shore in the Northern (Southern) Hemisphere (Markowski and Richardson 2010). The frictional contrast along the shoreline can, therefore, enhance precipitation downstream of the lake (Lavoie 1972; Nicosia et al. 1999). However, other studies found that frictional-induced convergence does not play a dominant role for the formation of lake-effect precipitation (Onton and Steenburgh 2001).

So far, the potential for lake-effect precipitation at Lake Constance has not been shown in peer-reviewed literature. Therefore, this study aims to clarify the impact of the lake and the surrounding terrain on the formation of precipitation downstream of Lake Constance for one specific event.

Data used in this study and the setup of the numerical model are described in section 2. In section 3, the event of 8 February 2013 is analyzed based on observations. Section 4 presents the results of the control run. The numerical sensitivity experiments are shown in section 5. Finally, results are discussed in section 6 and conclusions are drawn in section 7.

2. Data and methods

a. Observations

Observations of temperature, humidity, precipitation, wind speed, as well as wind direction gathered by AWS around Lake Constance are used for the case study and for the verification of the numerical simulations. Observations from AWS located in Austria are provided by the national weather service [Zentralanstalt für Meteorologie und Geodynamik (ZAMG)] with a temporal resolution of 10 min. Precipitation data from rain gauges are provided by the hydrographic service of Vorarlberg with hourly resolution. Additional observations are contributed by MeteoGroup for AWS in Germany and Switzerland with 10-min and hourly resolution. Precipitation is measured with heated tipping-bucket rain gauges without windshields by ZAMG and MeteoGroup while the hydrographic service of Vorarlberg mainly uses heated weighing rain gauges without windshields. Measurements of solid precipitation are affected by several errors, including wetting loss, evaporational loss, and undercatch due to wind, which is considered to have the highest impact (Goodison et al. 1998). The error caused by snowfall undercatch depends on the wind speed and is considered to be on the order of 20%–50% (Rasmussen et al. 2012). With a wind speed of about 6 m s−1 the snowfall undercatch error can be as large as 70% (Buisan et al 2016). However, we did not apply a systematic correction to the precipitation data since we do not know the exact wind speed at the gauge orifices. Hence, it is conceivable that the observed precipitation presented in the following sections may be affected by an undercatch error larger than 20% at some of the stations.

The temperature of Lake Constance is measured in the harbor of Bregenz at a depth of 0.5 m by the hydrographic service of Vorarlberg. Their website1 allows for public access to a more than 15-yr-long time series of daily mean values, which we used to compile Fig. 2. On 7, 8, and 9 February 2013, the corresponding temperatures were 4.0°, 3.4°, and 2.6°C, respectively. To check the representativeness of these point measurements for a larger area of the lake we compared them to an independent satellite-based dataset of daily mean LST for a location near the center of the lake at 47.6097°N, 9.4272°E (Riffler et al. 2015). The corresponding values are 4.3° and 4.1°C on 31 January and 10 February 2013, respectively. Unfortunately, no data are available for the period between these two days. Furthermore, data of LST predicted by a hydrodynamic lake model (Lang et al. 2010) have been retrieved from the decision support system BodenseeOnline.2 The corresponding values vary spatially between 2.5° and 4.5°C at 1200 UTC 8 February 2013. All the above-mentioned values are within about 1°C of the mean daily water temperature measured in the harbor of Bregenz on 8 February 2013. Hence, for the sake of simplicity, we use a constant value of 3.4°C for the LST in the remainder of this study.

Radar data are provided by the Austrian aviation weather service (Austro Control) as a maximum projection of reflectivity from a series of PPI scans at multiple elevation angles ranging from −2° to 60° at a 5-min resolution. The pixel size of the data is approximately 850 m 250 m over the Lake Constance area. The radar site is located on top of a mountain (Valluga, 2809 m MSL), approximately 50 km to the southeast of Bregenz (white triangle in Fig. 3b). As the lowest radar beam is located about 1600 m above the lake surface (at approximately 2000 m MSL), shallow precipitating clouds may not be captured by the radar.

b. Setup of the numerical model

Numerical simulations are performed with the Weather Research and Forecasting (WRF) Model, version 3.6.1. All simulations are based on the Advanced Research WRF core (Skamarock et al. 2008) and four two-way-nested domains (Fig. 3a) with 79 vertical levels. The lowest half-level is located approximately 26 m above ground level (AGL) and the vertical resolution increases from 60 m near ground to 270 m at the model top. The horizontal grid spacing of the four domains is 24.3, 8.1, 2.7, and 0.9 km, respectively.

The model uses the Lin et al. (1983) microphysics parameterization, the Mellor–Yamada–Nakanishi–Niino (MYNN) level-2.5 boundary layer scheme (Nakanishi and Niino 2004), the Noah land surface model (Ek et al. 2003), and the Rapid Radiative Transfer Model for long- and shortwave radiation (Mlawer et al. 1997). The MYNN scheme is chosen as it produced more accurate results compared to other boundary layer parameterizations for a lake-effect snowstorm at Lake Erie (Conrick et al. 2015). Conrick et al. (2015) showed that the choice of the boundary layer parameterization may have a great impact on the surface heat flux and, hence, also on lake-effect precipitation. However, this sensitivity to the boundary layer scheme is not assessed in the present study. The Lin et al. (1983) single-moment microphysics parameterization is chosen because preliminary investigations showed that the best matching spatial distribution of precipitation is produced by this parameterization (Umek 2016). The two outer domains use the Kain–Fritsch convective parameterization (Kain 2004). Operational analyses from the European Centre for Medium-Range Weather Forecasts (ECMWF) are used as initial and boundary conditions. Lateral boundary conditions are available at a 6-h interval. The model is initialized at 0000 UTC 8 February 2013 and run for 24 h.

Several sensitivity simulations are conducted and compared to the control simulation (CTL), similar to the approach of Alcott and Steenburgh (2013). The sensitivity simulations involve modifications to the terrain surrounding Lake Constance, the land-use characteristics, and the LST in the two innermost model domains.

The LST for Lake Constance and other lakes in the domain assigned by the WRF preprocessing system (WPS) is unrealistic. Hence, for grid points representing inland lakes the LST is manually set to 276.5 K, in agreement with the value observed in the harbor of Bregenz on 8 February 2013 (see section 2a). For two sensitivity experiments the LST is either increased (LST+3K) or decreased (LST-3K) by three kelvins in order to gain insight into the influence of the LST on the formation of precipitation and to address the uncertainty inherent to the measurement of the LST in the harbor of Bregenz.

To study the impact of Lake Constance as a source of latent and sensible heat, the lake is removed in several numerical simulations. It is replaced by land points representing either barren soil (NLs) or pasture (NLr). These different categories are used in order to assess the impact of sudden changes in surface roughness at the shoreline of the lake. The roughness length of the barren soil land-use category was manually set to a value appropriate for water surfaces (z0 = 10−4 m). Hence, NLs represents a no-lake simulation with a rather smooth surface. In contrast, NLr is characterized by a rather rough surface. Here, the lake is replaced by pasture ( between 0.05 and 0.15 m), which is the most frequent land-use category of the surrounding coastal region. Variables such as soil moisture and temperature, skin temperature, albedo, and others are bilinearly interpolated from adjacent land points to former water grid points to obtain initial values at new land points. The topography is identical to CTL (Fig. 3c).

In some of the sensitivity experiments the terrain is modified in three different ways. First, parts of the orography that rise above the surface height of Lake Constance (394 m MSL) are limited to the elevation of the lake for the flat simulation (FLAT; Fig. 3d) and flat-no-lake-smooth simulation (FLAT-NLs), while lower terrain is preserved. One simulation was conducted with flat terrain, but the LST of Lake Constance increased by 3 K (FLAT-LST+3K) in order to assess the influence of the LST. Second, the upstream mountainous terrain in the area of the Black Forest and the Swabian Jura surmounting Lake Constance is removed in the no-Black-Forest simulation (NoBF; Fig. 3e) and the no-Black-Forest-no-lake-smooth simulation (NoBF-NLs). Third, the downstream mountains eastward of the Rhine Valley are removed in the no-downstream-terrain simulation (NoDT; Fig. 3f) and the no-downstream-terrain-no-lake-smooth simulation (NoDT-NLs). In all three cases, the terrain height in the corresponding region is changed to the surface height of Lake Constance in the two innermost domains. Additionally, a transition zone of 15 grid points blending from modified to unmodified topography is introduced in order to avoid steep slopes.

As in Alcott and Steenburgh (2013) and Schumacher et al. (2015), land-use, soil, and vegetation properties are retained in the simulations with altered topography. However, soil temperature, soil moisture, and skin temperature are changed using implicit relationships of the WPS based on terrain elevation. As the removal of orography introduces an additional volume of air, the properties of this air volume are derived from the ECMWF initial conditions assuming a moist-adiabatic lapse rate for grid points below the ECMWF model surface.

Table 1 describes the main differences as well as the abbreviations of the simulations and summarizes maximum and mean values of predicted and radar-derived accumulated precipitation.

Table 1.

Description of numerical simulations and predicted mean (domain averaged) and maximum accumulated precipitation for the period 0700–1200 UTC 8 Feb 2013. The domain used for the calculations of mean and maximum precipitation is shown in Fig. 7b. Values are given as absolute numbers and in percent relative to the control simulation CTL. For comparison, the radar-estimated precipitation is listed in the last row.

Table 1.

3. Observational case study

A case study of a heavy snowfall event, which occurred on 8 February 2013, with stationary, banded precipitation structures downstream of Lake Constance is presented. The snowband was visible in radar imagery from 0640 to 1200 UTC and was located in the southeastern part of Lake Constance where it affected the city of Bregenz and the Bregenzerwald.

a. Synoptic situation

A cold front passed the Lake Constance area about 40 h before the first banded precipitation was visible in radar imagery. With the arrival of the cold front at about 1500 UTC 6 February 2013, the 2-m air temperature started to decrease and dropped below the freezing point (Fig. 4a). The cold front caused widespread precipitation with about 25 mm at Bregenz between 1800 UTC 6 February and 1100 UTC 7 February 2013 (Fig. 4b).

Fig. 4.
Fig. 4.

(a) Air temperature at 2 m AGL (°C, solid line) and relative humidity (%, dashed line) and (b) accumulated precipitation (mm, solid line) and pressure reduced to mean sea level (hPa, dashed line) measured by the AWS in Bregenz (see location in Fig. 1) between 0600 UTC 6 Feb and 0000 UTC 9 Feb 2013. Passage of a cold front occurred at about 1500 UTC 6 Feb. The period with a stationary snowband in extension of Lake Constance visible in the radar imagery is indicated by the gray shading.

Citation: Monthly Weather Review 144, 12; 10.1175/MWR-D-16-0032.1

At 0600 UTC 8 February 2013, shortly before the snowband became visible in radar imagery, the synoptic situation at 700 hPa was characterized by a cutoff low over Germany and Poland (Fig. 5). The Lake Constance region was located southwest of this low pressure system. The associated northerly flow (8 m s−1) caused low-level cold-air advection in the target area. The air mass over the northern Alpine foreland was moister than the air south of the Alps, partly as a result of foehnlike subsidence in the lee of the Alps (Jiang et al. 2005). Other events with stationary, banded precipitation investigated in a preliminary work (Fig. 2) are characterized by similar synoptic conditions, including a ridge over the North Atlantic and a cutoff low at 700 hPa over northern Europe.

Fig. 5.
Fig. 5.

Synoptic situation at 0600 UTC 8 Feb 2013 based on ECMWF analysis data. Geopotential height (m) at 700 hPa as purple contour lines with 40-m interval (every second line labeled), mean relative humidity (%) between 850 and 700 hPa as color contours (scale at right), and wind at 700 hPa as wind barbs. Half barbs, full barbs, and triangles represent 2.5, 5, and 25 m s−1, respectively. A red star indicates the location of Lake Constance.

Citation: Monthly Weather Review 144, 12; 10.1175/MWR-D-16-0032.1

b. Lake–air temperature difference

On 8 February 2013, the daily mean 2-m air temperature at Bregenz was −1.3°C and the LST was 3.4°C, leading to a mean temperature difference of 4.7°C (see section 2a for representativeness of this LST value). According to the ECMWF analysis, the temperature at 850 hPa (1400 m MSL) over Lake Constance was between −8°and −9°C. Hence, the average temperature decrease in the 1000-m-deep layer between the lake surface and 850 hPa was between 11.5° and 12.5°C, which exceeds the dry adiabatic lapse rate (10°C km−1). Therefore, the event fulfills the lapse rate criterion of Holroyd (1971) for the occurrence of lake-effect snow. Other criteria, following Steenburgh et al. (2000), were not examined due to the lack of upper-air observations in the area, but will be examined in the next chapter based on data from the control simulation.

c. Evolution and distribution of precipitation

During the whole event, precipitating clouds were moving from northwest to southeast over the Lake Constance area. At 0640 UTC a rather stationary precipitation band evolved over the eastern part of the lake with its upstream edge located over the lake close to the shoreline near Lindau (triangle in Fig. 6a). The structure had a length of about 15 km, a width of about 5 km, and was slightly changing its position in time (Figs. 6a–c). It was characterized by radar reflectivities above 30 dBZ, which represent the highest values in the area after 0800 UTC, when the widespread precipitation began to decay. The band reached its maximal reflectivity (39 dBZ) at 0900 UTC (Fig. 6c). After 1000 UTC the band widened to about 10 km (Fig. 6d) and started to weaken at 1130 UTC. After 1200 UTC no stationary precipitation structures could be observed in the radar data, while other minor transient precipitating structures were still present in the region.

Fig. 6.
Fig. 6.

Maximum projection of radar reflectivity derived from a series of PPI scans at multiple elevation angles ranging from −2° to 60° as color contours (dBZ, scale at right) at (a) 0645, (b) 0800, (c) 0900, and (d) 1015 UTC 8 Feb 2013. Data are taken from the Valluga radar (see location in Fig. 3b). Terrain elevation shown as gray contour lines with an increment of 200 m, lake shoreline as thick black line, and political boundaries as thin black lines. Black triangle (diamond) identifies AWS Lindau (Bregenz).

Citation: Monthly Weather Review 144, 12; 10.1175/MWR-D-16-0032.1

A ZS relationship of (Alcott and Steenburgh 2013) is used to calculate radar-derived precipitation estimates, with Z denoting the radar reflectivity factor (in mm6 m−3) and S the liquid equivalent snowfall rate (in mm h−1). The corresponding precipitation map suggests that in the period from 0640 to 1200 UTC, 2.5–10 mm fell over the complex terrain east and about 5 mm and less west of the Rhine Valley (Fig. 7a). These values are in good agreement with available gauge observations. However, potential measurement errors due to snowfall undercatch need to be considered, as mentioned in section 2a. The radar-derived precipitation shows an embedded maximum of 20 mm in the area of Bregenz. This is an underestimation, as the AWS Bregenz measured 35.8 mm for the same period (red circle in Fig. 7a). The underestimation may be either caused by (i) the overshooting of the radar beam above the precipitating clouds, (ii) the width of the radar beam being too wide to capture the small scale embedded maximum in precipitation, or (iii) the inaccuracy inherent to the ZS relationship. The amount and distribution of precipitation measured by rain gauges in the complex terrain southeast of Bregenz is in reasonable agreement with the radar-estimated precipitation. Rain gauges measured up to 23 mm from 0700 to 1200 UTC (green circles in Fig. 7a).

Fig. 7.
Fig. 7.

(a) Radar-estimated accumulated precipitation as color contours and accumulated precipitation measured with rain gauges as circles with color shading (mm, scale at right) from 0640 to 1200 UTC (0700–1200 UTC for stations with hourly temporal resolution) 8 Feb 2013. The radar-estimated precipitation is based on radar data with a 5-min temporal resolution. (b) Accumulated precipitation as in (a) valid for CTL from 0700 to 1200 UTC 8 Feb 2013. Terrain elevation shown as brown contour lines with an increment of 200 m. Thick black lines denote lake shorelines. Thin black lines identify political boundaries. A dashed rectangle in (b) identifies lake-effect subdomain for calculation of spatial average and maximum value of accumulated precipitation (cf. Table 1).

Citation: Monthly Weather Review 144, 12; 10.1175/MWR-D-16-0032.1

The highly variable spatial distribution of precipitation is also depicted in Fig. 8a based on selected stations. The corresponding accumulated precipitation together with the distance of each station from the AWS Bregenz is given in Table 2. Maximum precipitation was observed at Bregenz, whereas the AWS Alberschwende, located in the mountainous terrain of the Bregenzerwald, observed less than half of the maximum precipitation, but was still affected by the snowband (Fig. 8a). Precipitation at Bregenz was characterized by variable intensity, especially before 1000 UTC, since the snowband slightly changed its location (cf. Fig. 6).

Fig. 8.
Fig. 8.

Observations from several AWS in the Lake Constance region (see locations and abbreviations in Fig. 1) from 0300 to 1500 UTC 8 Feb 2013: (a) accumulated precipitation in mm and (b) wind direction in degrees. Red (blue) dots and triangles in (b) denote stations on the southern (northern) shore. The wind direction in (b) is only plotted for wind speeds exceeding 2 m s−1. Gray shading identifies the period with a snowband visible in radar imagery. Elevation of each AWS in m MSL is given in the legend.

Citation: Monthly Weather Review 144, 12; 10.1175/MWR-D-16-0032.1

Table 2.

Precipitation in mm accumulated from 0640 to 1200 UTC 8 Feb 2013 at stations depicted in Fig. 8a, together with direction and distance (km) from Bregenz. The location of the station is shown in Fig. 1.

Table 2.

The AWS Bregenz measured heavy snowfall until about 1230 UTC and a slight decrease in intensity afterward, with precipitation continuing until approximately 1330 UTC. However, the signal in radar reflectivity near Bregenz vanished around 1200 UTC as mentioned previously. Hence, it is conceivable that the snowband became too shallow to be captured by the radar after 1200 UTC due to overshooting or blocking of radar beams.

Stations located on both the southern and northern shore of Lake Constance measured far less precipitation (1 mm, dashed lines in Fig. 8a), except for the AWS in Lindau, which measured 7.5 mm between 0900 and 1030 UTC. During that time the upstream edge of the snowband was located close to this station.

d. Wind field around Lake Constance

The low-level flow around the lake was characterized by westerly winds with 2–5 m s−1 after 0400 UTC 8 February 2013 (Fig. 8b). Rohrspitz was affected by a weak southerly outflow (2.5 m s−1) from the Rhine Valley between 0400 and 0600 UTC.

Shortly before the stationary band became visible in the radar, the wind direction measured by stations on the northern shore (e.g., Kressbronn and Lindau) changed to northwest. At the same time, stations on the southern shore still measured westerly wind directions leading to convergence over the lake (Fig. 8b). Stations in the Rhine Valley observed a weak southerly downvalley flow (1.5 m s−1 in Dornbirn, not shown) resulting in a weak low-level outflow near the southern shore of the lake. This outflow may have contributed to flow convergence in the area of Bregenz.

After 1200 UTC the convergence between the northern and southern shore vanished as the wind direction on the northern shore veered back to west (Fig. 8b). At the same time the snowband disappeared in the radar imagery.

4. Control simulation

a. Simulated precipitation structure

The model is initialized at 0000 UTC 8 February 2013 and run for 24 h. The evolution of the hourly precipitation is depicted in Fig. 9 and described below.

Fig. 9.
Fig. 9.

Precipitation of the control simulation accumulated over the previous hour in mm as color contours (scale at right), horizontal wind vectors at 10 m AGL plotted at every sixth grid point [reference vector shown in (a)], and model terrain elevation as brown contour lines with increment of 200 m at (a) 0700, (b) 0800, (c) 0900, (d) 1000, (e) 1100, and (f) 1200 UTC 8 Feb 2013.

Citation: Monthly Weather Review 144, 12; 10.1175/MWR-D-16-0032.1

At 0700 UTC, a first maximum in precipitation (4–6 mm h−1) can be detected downstream of Lake Constance on the windward side of the Pfänder, as well as a banded extension northwest of the maximum along the shore (2 mm h−1, Fig. 9a). Afterward, the banded pattern along the northern shoreline intensifies (Fig. 9b). From 0800 to 1000 UTC the embedded maximum reaches over 15 mm h−1. This represents the maximum intensity of the snowband (Figs. 9c,d).

After 1000 UTC the intensity and spatial extent of the snowband decreases. The embedded maximum is now located slightly on the leeward side of the Pfänder with a value of 8–12 mm h−1 (Fig. 9e). Afterward the intensity decreases further (Fig. 9f) until the snowband vanishes after 1200 UTC, when only weak precipitation structures with intensities less than 4 mm h−1 approach the lake area from a northwest direction. The overall pattern of the snowband resembles a long-lake-axis-parallel band (e.g., as in Steiger et al. 2013; Veals and Steenburgh 2015), located along the northern shoreline for most of the time.

Evaluation of simulated precipitation by means of observed and radar-estimated accumulated precipitation in the area shows reasonable agreement (Fig. 7). The region north- and westward of Lake Constance receives 0–2.5 mm of precipitation in the period from 0700 to 1200 UTC. The precipitation is more intense over the complex terrain in the southern and eastern part of the investigated area, reaching 20 mm. The maximum accumulated precipitation of 46.7 mm (dark red in Fig. 7b) is located on the windward side of the Pfänder, slightly north of Bregenz. It exceeds the measured value at the AWS Bregenz by 10.9 mm (cf. Table 2). However, measurements of solid precipitation by unshielded rain gauges, as used in the target area, may be affected by an undercatch error (see section 2a). The elongated structure of the precipitation band downstream of the Pfänder is not perfectly reproduced by the model, as most of the Bregenzerwald receives less than 10 mm in CTL (Fig. 7).

b. Thermodynamic structure

Vertical gradients of potential temperature θ, equivalent potential temperature , and saturated equivalent potential temperature are examined in order to assess the occurrence of absolute, potential, or conditional instabilities, respectively (e.g., Bohren and Albrecht 2010, 312–313). Two representative locations are considered: one over the lake and one over land westward of the lake (Fig. 1). This specific location over land is considered since it represents the condition unmodified by the lake. The low-level air is advected from this land point by westerly winds across the lake (Fig. 9) where it becomes affected by the heat fluxes from the lake surface. The second location over the lake is chosen since it is close to the location of the banded convection. The corresponding profiles in Fig. 10 illustrate conditions prior to convection (0600 UTC, solid line) as well as during the presence of the snowband (1100 UTC, dashed line).

Fig. 10.
Fig. 10.

Vertical profiles of θ, , and taken from the control simulation (a) over Lake Constance and (b) over land west of the lake at 0600 (solid) and 1100 UTC (dashed) 8 Feb 2013. Location of vertical profiles is shown in Fig. 1. Red shading in (a) indicates the layer with absolute unstable stratification at 0600 UTC. The LST in the numerical model is set to 276.5 K.

Citation: Monthly Weather Review 144, 12; 10.1175/MWR-D-16-0032.1

The stratification over Lake Constance is characterized by an absolute unstable layer () within the lowest 100–200 m AGL (Fig. 10a). The profile over the water body does also show a potentially unstable () and conditionally unstable stratification () from the surface up to 400 m AGL. The general characteristics of the vertical profile over the lake prior to the initiation of convection and during the presence of the snowband are essentially the same (cf. solid and dashed lines in Fig. 10a). The main difference between the two times shown is the deeper conditionally unstable layer over the lake at 1100 UTC, which reaches up to 700 m AGL. Absolute stable stratification () is present above, but without a capping inversion, as the stratification is nearly moist neutral up to 500 hPa. This is in contrast to other studies of lake-effect precipitation (e.g., Steenburgh and Onton 2001; Minder et al. 2015), which indicate a capping inversion limiting the vertical extent of lake-effect convection. On the other hand, vertical profiles over land prior to the snowband formation (0600 UTC) mainly show an absolutely stable stratification with a shallow layer (100 m AGL) of rather weak conditionally unstable stratification (solid lines in Fig. 10b).

The low-level absolute instability over the lake is generated by the relatively warm lake surface. The heat input from the water body to the atmospheric surface layer through latent and sensible heat fluxes (averaged over the lake surface) is approximately 40 W m−2 each. The heat fluxes stay rather constant throughout the simulated period. These values seem reasonable as Lofgren and Zhu (2000) mention a mean latent and sensible heat flux at the much larger Lake Huron during the winter of approximately 50 and 60 W m−2, respectively. In contrast, sensible and latent heat fluxes are both less than 10 W m−2 over the surrounding land before 0900 UTC.

Until 1000 UTC the target area is mostly overcast. However, cloud cover reduces afterward over the region north and west of the lake (in agreement with observations). This increases incoming shortwave radiation to values exceeding 200 W m−2, which, in turn increases sensible and latent heat fluxes over the adjacent land to about 20–50 W m−2. The inhomogeneity in heat fluxes over land and water disappears, leading to weaker land-breeze circulations and, in turn, to a more homogenous convective boundary layer throughout the Lake Constance region (cf. blue dashed lines in Fig. 10). Low-level conditional instability is now present also over land (Fig. 10b), but is weaker and shallower than over the lake.

Convective available potential energy (CAPE) is calculated for surface-based parcels characterized by the temperature, pressure, and humidity taken from the first model level. CAPE is zero until 0600 UTC and only marginal values (20 J kg−1) are produced along the northern shoreline afterward. After 1000 UTC highest CAPE (50 J kg−1) occurs in the area of the snowband close to the northern shoreline. After 1100 UTC the isolated banded maximum of CAPE along the northern shore vanishes as CAPE is also produced over the area north of the lake with values below 70 J kg−1.

c. Low-level wind field

Low-level winds over and around the lake are from southwesterly to westerly directions until 0500 UTC. A downvalley flow in the Rhine Valley is present. After 0600 UTC the 10-m wind direction over the northern shore changes from west to northwest with wind speeds between 1 and 2 m s−1. South of and over the lake, 10-m winds remain westerly with speeds ranging from 2.5 to 5 m s−1 over the land and up to 10 m s−1 over the lake. Based on this westerly wind direction of the low-level flow, the fetch is approximately 20 km (Fig. 9). This westerly flow encounters northwesterly winds near the northern shore where it forms a low-level convergence zone parallel to the shoreline. The difference in wind direction vanishes and a more homogeneous flow from northwest is present above 1000 m MSL. Therefore, winds over and west of the lake exhibit minor directional wind shear as they veer from westerly directions near the surface to northwesterly directions at 1000 m MSL (600 m AGL).

At 0900 UTC, the upstream edge of the convergence zone is not located directly over the lake, but slightly east of it (Fig. 11a). Although the convergence line in CTL is more prominent at other times, this time has been chosen since it shows the strongest differences between the sensitivity experiments (see section 5). The convergence line vanishes after 1200 UTC.

Fig. 11.
Fig. 11.

Horizontal wind at 700 m MSL as vectors plotted at every sixth grid point [reference vector shown in (a)], and horizontal divergence of wind at the same height as color contours (in 10−4 s−1, scale at right) at 0900 UTC 8 Feb 2013 for the simulations (a) CTL, (b) FLAT, (c) FLAT-NLs, (d) NLs, (e) LST-3K, (f) LST+3K, (g) FLAT-LST+3K, (h) NLr, (i) NoBF, (j) NoBF-NLs, (k) NoDT, and (l) NoDT-NLs. See Table 1 for a short description of the simulations. Model topography shown as brown contours with an interval of 200 m. Gray shading indicates areas below the model terrain.

Citation: Monthly Weather Review 144, 12; 10.1175/MWR-D-16-0032.1

Low-level westerly winds south of the lake are caused by flow deflection of the impinging northwesterly synoptic-scale flow at the foothills of the Alps. This, in turn, causes a more westerly flow in the Swiss Plateau region (Fig. 3b). These first orographic features form a ridge with a height of about 800–1000 m MSL within a few kilometers south of the lake (Fig. 1).

d. Triggering of convection

The presence of CAPE near the northern shore indicates a favorable position to initiate convection, but additional lifting is needed in order to trigger convective motion. We hypothesize that topographically- and thermally-induced convergence in the low-level wind field over the downwind portion of the lake provides a lifting mechanism to trigger conditional instability and, hence, convection. On the other hand, the banded structure of convection may also be responsible for the formation of the convergence line visible in Fig. 11a due to compensating lateral inflow into the area of banded convection. In the presented case it is hypothesized that lifting of low-level air initiated convection and that the convective motion itself further enhanced and amplified the convergence line. The sensitivity experiments in section 5 show that a specific topographic setting (i.e., topography south of the lake) is necessary to cause convergence and initiate strong convection.

The stratification over the lake is conditionally and potentially unstable. However, we hypothesize that the potentially unstable layer is too shallow (only a few hundred meters) to explain the convection reaching up to nearly 3 km MSL (see below). Therefore, we focus on parcel theory and the release of conditional instability in the remainder of this section.

Air parcels in the layer below approximately 800 m MSL do reach saturation when being lifted by about 50–250 m. This is illustrated in Fig. 12, which shows the lifting condensation level (LCL), level of free convection (LFC), and level of neutral bouyancy (LNB) as a function of the starting height z of an air parcel. The value for on the abscissa denotes the vertical distance an air parcel at a specific height needs to be lifted in order to reach its LCL, LFC, or LNB, respectively.

Fig. 12.
Fig. 12.

LCL, LFC, and LNB as a function of height, calculated for the same grid point over Lake Constance as the profile shown in Fig. 10a prior to the snowband formation at 0600 UTC 8 Feb 2013. Location of grid point is shown in Fig. 1. The abscissa represents heights relative to the starting level z.

Citation: Monthly Weather Review 144, 12; 10.1175/MWR-D-16-0032.1

Figure 12 depicts the situation prior to the snowband formation. As soon as the LCL of a low-level air parcel is reached, convection is triggered due to the presence of CAPE and the absence of convective inhibition. In other words, the LCL and LFC are nearly identical up to about 700 m MSL (Fig. 12). The LNB is reached after 2 to 1.1 km vertical distance relative to the starting height of the parcel. Therefore, the release of conditional instability enabled low-level air parcels to rise to a height of up to 2–2.5 km MSL at the time and location shown in Fig. 12.

The cross section in Fig. 13a shows a layer of conditional instability (blue shading) and a layer of absolute instability (red shading) over the water body. Vertical velocities over the lake and the adjacent shoreline reach 4 m s−1 and are highest over the foot of the sloping terrain. Strong precipitation falls over the lake and the windward side of the first hill. Convection reaches 2 (3) km MSL in the upstream (downstream) half of the snowband. As mentioned in section 2a, the lowest radar beam is located at approximately 2 km MSL over Lake Constance. Therefore, the model results suggest that an underestimation of radar-derived precipitation values may be due to the radar beam overshooting the shallow precipitating clouds, especially over the upstream part of the snowband.

Fig. 13.
Fig. 13.

Vertical cross section A–B along the northern (German) shore of Lake Constance at 0900 UTC 8 Feb 2015 for (a) CTL and (b) NLs simulation: conditional instability () as blue shading; absolute instability () as red shading; potential temperature θ as thin black contour lines with 0.5-K increments; total hydrometeor mixing ratio as thick green contour lines for 0.1, 0.4, and 0.6 g kg−1; and wind vectors for the wind component parallel to the transect [reference vectors for horizontal and vertical wind speed in bottom-right corner of (a)]. Location of cross section is shown in the inset of (a). Gray hatched rectangle below the model topography in (a) denotes intersection of cross section with the water body.

Citation: Monthly Weather Review 144, 12; 10.1175/MWR-D-16-0032.1

The cross sections in Fig. 14, located at the eastern end of the lake and oriented perpendicular to its long axis, show a transect through the snowband near its upstream (Fig. 14a) and downstream (Fig. 14b) end. At the upstream end, low-level westerly winds are present over the lake. Farther aloft and north of the lake, northwesterly winds are illustrated by a weak horizontal wind component parallel to this transect. Convection and cloud formation is collocated with the convergence zone (Fig. 14a).

Fig. 14.
Fig. 14.

As in Fig. 13, but valid for CTL and cross sections perpendicular to long axis of the lake: (a) cross section C1–D1 and (b) cross section C2–D2. Locations of cross sections are shown in the inset in (b).

Citation: Monthly Weather Review 144, 12; 10.1175/MWR-D-16-0032.1

At the downstream end, the hydrometeor mixing ratio within the snowband is higher and convection reaches up to 3 km MSL (Fig. 14b). At this location the convergence line is stronger due to higher low-level wind speeds and a southwesterly outflow from the Rhine Valley.

Orographic lifting of the impinging low-level airflow seems to be important to either induce or amplify precipitation over the complex terrain downstream of the lake. Updrafts reaching 2–4 m s−1 over the sloping terrain downwind of the lake indicate that convective orographic precipitation is present during the whole period of banded precipitation and is initiated at rather low levels (Fig. 13a).

In summary, the conditions for lake-effect snow according to Steenburgh et al. (2000) are met in the control run: (i) a low-level conditionally unstable layer is present, but with a vertical extent only up to the lowest 800 m AGL; (ii) a capping inversion is missing; (iii) directional wind shear is 60° over the lake; and (iv) a low-level convergence line is present over the lake.

To assess the influence of the lake and the topography on the formation of precipitation, a series of sensitivity experiments have been conducted. They are presented in the following section.

5. Sensitivity experiments

In this section we try to illustrate various effects on the amount and distribution of precipitation based on sensitivity experiments, as described in section 2b. To quantify the difference in precipitation between the simulations, a spatial averaging of accumulated precipitation from 0700 to 1200 UTC is performed within a subdomain termed lake-effect area. It is indicated by a dashed box in Fig. 15b and covers 50 km × 40 km. Values are given in Table 1 for each model run and Fig. 15 shows the spatial distribution of accumulated precipitation.

Fig. 15.
Fig. 15.

As in Fig. 7b, but for the simulations (a) CTL, (b) FLAT, (c) FLAT-NLs, (d) NLs, (e) LST-3K, (f) LST+3K, (g) FLAT-LST+3K, (h) NLr, (i) NoBF, (j) NoBF-NLs, (k) NoDT, and (l) NoDT-NLs. A dashed rectangle in (a) denotes the lake-effect subdomain for the calculation of spatially averaged and maximum value of accumulated precipitation. See Table 1 for a short description of the simulations and statistics for the lake-effect subdomain.

Citation: Monthly Weather Review 144, 12; 10.1175/MWR-D-16-0032.1

a. Flat terrain

Flattening the terrain (FLAT) results in a negligible amount of spatially averaged precipitation in the lake-effect domain (less than 1 mm). Compared to CTL, the maximum accumulated precipitation reduces by 95% to 2.5 mm (Table 1). Weak precipitation occurs downstream of the southern shore (Fig. 15b) caused by a single cell bound to the lake with a maximum intensity of less than 1.5 mm h−1. The 10-m wind over the lake blows from northwesterly directions, resulting in an overlake fetch of up to 35 km compared to 20 km in the control run. The 10-m wind field indicates a minor convergence line near the southern (i.e., downwind right shore), which is presumably the result of an increase in surface friction at the transition from the lake to the land. Weak precipitation is present south of the lake in the FLAT simulation. Flattening the terrain and removing the lake (FLAT-NLs) reduces the size of this area of weak precipitation. The maximum accumulated precipitation reduces to 1.4 mm (Fig. 15c, Table 1).

Neither of the two simulations shows a pronounced convergence line or an associated snowband. Instead, a rather uniform northwesterly wind field is present (Figs. 11b,c). The FLAT simulation features minor precipitation south of the lake. Hence, the lake alone is not able to produce the observed precipitation structure and intensity with a LST of 276.5 K. Therefore, the LST and the interplay of the lake with the surrounding terrain play important roles in the generation of strong precipitation, as will be shown below.

b. Influence of the lake as a heat and moisture source

In the NLs simulation, the lake as a heat and moisture source is removed and replaced by barren, sparsely vegetated land with a roughness length appropriate for water surfaces. The mean (maximum) value of accumulated precipitation is reduced by 24% (76%) relative to CTL (Table 1). The strongest precipitation occurs on the windward side of the Pfänder and at a small convergence zone at the Rhine Valley exit (5–10 mm), whereas precipitation over the Bregenzerwald is greatly reduced (Fig. 15d). No significant convergence can be found over the area formerly occupied by the lake (Fig. 11d).

The vertical stratification in NLs does not show a low-level absolute unstable stratification, because the lake as a heat source is missing. Regions showing low-level conditional instability are present over areas formerly occupied by the lake, but with a vertical extent only up to 200 m AGL and weaker intensity than in CTL (cf. Figs. 13a,b). Further, conditional instability is not triggered. Instead, rather stratiform orographic clouds form over the mountains downstream of the lake (Fig. 13b). Hence, without the lake, stable instead of convective orographic precipitation occurs.

To address the uncertainty inherent to the point measurement of the LST in the harbor of Bregenz and the influence of the LST on the formation of precipitation, two sensitivity experiments were conducted with increased and decreased LST in the model. When the LST is decreased by 3 K to 273.5 K compared to CTL (LST-3K), no convergence line and banded precipitation forms over and downstream of the lake (Figs. 11e and 15e). Only an isolated maximum in accumulated precipitation of 13.5 mm (Table 1) on the windward side of the Pfänder can be identified. A LST of 273.5 K exceeds the daily mean 2-m temperature over the lake by only 2 K. Mean latent (sensible) heat fluxes over the lake decrease to less than 20 (15) W m−2 compared to CTL (approximately 40 W m−2 each). Conditionally and absolutely unstable stratification is present over the lake, but with a lower vertical extent compared to CTL. Conditional instability is released on the windward slope of the Pfänder, resulting in weak convective orographic precipitation. The LNB of air parcels over the lake is at about 1 to 1.6 km MSL and, hence, significantly lower than in CTL (2–2.5 km MSL).

When the LST is increased by 3 K to 279.5 K (LST+3K), a snowband develops with a similar size as in CTL, but with a higher intensity. The maximum of accumulated precipitation increases by 61% to 75.3 mm (Fig. 15f, Table 1). Latent (sensible) heat fluxes reach a mean value of 80 (70) W m−2 over the lake. The absolute unstable layer over the lake increases in depth to about 300–400 m AGL. The layer showing conditional instability has a comparable depth as in CTL, but shows a stronger negative vertical gradient of . Conditional instability is triggered with convection reaching up to a height of 2.5–3.5 km MSL. The wind at 10 m AGL is from westerly directions over the lake and comparable to the control run, leading to a similar overlake fetch of approximately 20 km. Low-level winds have a similar strength than in CTL, but the northwesterly flow across the northern shore is able to penetrate toward the lake center. Consequently, the location of the snowband is shifted to a more midlake position (i.e., from Bregenz toward the exit of the Rhine Valley). This shift may be the result of a stronger thermally induced circulation and, hence, convergence over the lake. The snowband persists until 1800 UTC, with a short interruption of about 1 h around noon. This is an increase in the snowband lifetime of about 5 h compared to CTL.

One simulation is conducted with flat terrain but the LST increased by 3 K to 279.5 K (FLAT-LST+3). This simulation shows a maximum of accumulated precipitation of 5.7 mm (Fig. 15g, Table 1). Although not much in absolute terms, this is more than a doubling compared to the FLAT simulation presented in section 5a and shows that the lake is capable of generating a weak-to-moderate amount of lake-effect precipitation even with no topography surrounding the lake. However, the lake alone is not able to create a pronounced convergence line without the presence of the surrounding topography.

c. Sensitivity to surface roughness

Differences in hourly and total accumulated precipitation between the simulations with the lake replaced by smooth barren soil (NLs) and by comparatively rough pasture (NLr) are negligible (Figs. 15d,h). The mean (maximum) accumulated precipitation in the lake-effect domain is 2.8 (11.2) mm for NLs and 2.8 (12.6) mm for NLr (Table 1). The near-surface wind speed over the area formerly occupied by the lake is slightly higher (up to 1.5 m s−1) in NLs than in NLr. As expected, this is due to the smaller roughness length in the area formerly occupied by the lake in NLs compared to NLr (see section 2b).

d. Sensitivity to upstream orography

Removing the upstream terrain of the Black Forest and the Swabian Jura (NoBF) causes hardly any change in the simulation until 0600 UTC. However, from 0700 to 1300 UTC a snowband forms with a more midlake position than in CTL (Fig. 15i). The maximum accumulated precipitation downstream of the lake decreases to 26.4 mm (−43%), but the mean accumulated precipitation increases to 5.3 mm (%, Table 1). This increase is caused by a moister impinging airflow. It is conceivable that the reduced moisture in CTL compared to NoBF results from orographic precipitation above the hilly terrain upstream of the lake, which does not occur in NoBF. Further, the impinging northwesterly flow is not disturbed by upstream orography and, hence, is up to 4 m s−1 stronger than in CTL (Fig. 11i). Stronger winds together with the higher low-level moisture content lead to increased orographic precipitation, especially northeast of the lake (cf. Figs. 15a,i). The wind direction at 10 m AGL over the lake and, therefore, the overlake fetch is comparable to CTL.

The intensity and length of the convergence line increases after 0900 UTC, when it covers about two-thirds of the lake in streamwise direction (Fig. 11i). It spans nearly the whole lake after 1000 UTC, implying a greater extent of the snowband than in CTL. Further, the lifetime of the band is 1 h longer than in CTL.

We hypothesize that the decrease in maximum accumulated precipitation is caused by the change of the snowband position. The convergence line intersects with the downstream shoreline at a position slightly more southward than in CTL (cf. Figs. 11a,i). The terrain at this position is less steep and less high and, therefore, orographic enhancement is smaller, leading to lower precipitation rates.

Removing the lake together with the upstream topography (NoBF-NLs) leads to the disappearance of the convergence line and the snowband (Figs. 11j and 15j). The maximum accumulated precipitation decreases by 74%, whereas the mean accumulated precipitation is 16% higher than in CTL (Table 1). This increase in the spatial average is mainly attributed to slightly higher precipitation over the first foothills east of the lake, the southern part of the lake, and at the Rhine Valley exit. It is conceivable that at least part of this increase is caused by enhanced orographic lifting and precipitation due to a stronger low-level flow. The latter results from removing the upstream orography.

e. Sensitivity to downstream orography

After the removal of the downstream mountains east of the Rhine Valley (NoDT), the convergence line and the associated snowband still forms. The convergence line extends much farther downstream since no complex terrain disturbs it (Fig. 11k). The magnitude of the convergence (cf. Figs. 11a,k), the 10-m wind direction over the lake (i.e., overlake fetch), and the vertical stratification over the lake are similar to CTL (cf. Figs. 13a and 16).

Fig. 16.
Fig. 16.

As in Fig. 13, but for the NoDT simulation at 0800 UTC.

Citation: Monthly Weather Review 144, 12; 10.1175/MWR-D-16-0032.1

The convergence line develops around 0600 UTC and vanishes after 1200 UTC. Conditional instability is released around 0600 UTC, which triggers convection even without the presence of downstream mountains (Fig. 16). However, the maximum accumulated precipitation in the lake-effect subdomain reduces to 19 mm (−59%, Table 1, Fig. 15k). The snowband exists from 0700 to 1200 UTC, with a short break between 1000 and 1100 UTC. The peak intensity is less than 8 mm h−1.

In contrast to CTL and NoBF, the highest values of accumulated precipitation do not occur near Bregenz but about 10 km farther downstream (cf. Figs. 15a,k). There, a 200-m-deep cold pool is present, which is illustrated in Fig. 16 by an increased number of isentropes near the ground at a distance 25 km on the abscissa. The cold pool is generated by a combination of nocturnal cooling in the large artificial basin of the modified topography and by the extrapolation of the air temperature below the lowest level of the ECMWF analysis when preparing the initial conditions with the WPS. Therefore, the presence of the cold pool is rather unrealistic. After 1000 UTC the cold pool breaks up due to the increasing incoming shortwave radiation and the simulation shows more reasonable results in the area formerly influenced by the cold pool. The convergence line loses strength and the precipitation intensity of the snowband decreases after 1000 UTC.

When the downstream mountains are removed together with the lake (NoDT-NLs), the convergence line and the snowband both disappear completely (Figs. 11l and 15l). Consequently, the mean (maximum) accumulated precipitation decreases by 75% (82%) compared to CTL (Table 1).

6. Discussion

The sensitivity experiments have supported the hypothesis that the lake is a crucial factor for the formation of snowbands over and downwind of Lake Constance. On 8 February 2013, however, it was not the effect of the lake alone, but a synergistic interaction between lake and orographic effects that caused the high amount of snowfall. This is in agreement with Alcott and Steenburgh (2013), who state that, especially with smaller water bodies where fetch is limited, a superposition of lake- and orographically induced convergence may have a particularly large impact on the formation of lake-effect precipitation. Wright et al. (2013) also note the interaction between lake effects and local orographic enhancement of convergence in regions with complex terrain.

In all sensitivity runs with modified nonflat terrain and the lake present, a snowband forms with its upstream edge located over the lake or slightly inland from the northern shoreline. Therefore, the topography south of the lake seems to be essential for the formation of the snowband. As in Onton and Steenburgh (2001), frictional-induced convergence played a minor role during the investigated event. In the simulation with flat terrain the pure lake effect is too weak to produce significant precipitation, although the lake seems to cause weak precipitation south of the lake which slightly increases precipitation compared to the FLAT-NLs simulation (cf. Figs. 15b and 15c). The precipitation generated by the lake increases if the LST is increased by 3 K in the FLAT-LST+3K simulation (Fig. 15g). The 10-m wind field of the FLAT simulation and even more of FLAT-LST+3K show a roughness-induced convergence zone over the downstream shorelinge and shows qualitative similarities to the idealized simulations presented by Laird et al. (2003).

The presence of a shallow absolutely unstable layer over the lake alone is thought to be not sufficient enough for triggering moist convection. The location of the upstream portion of the snowband changes over time. It is located over the lake, as well as over land (e.g., Fig. 13a) where no absolutely unstable low-level layer is present. Figure 14a shows that despite the presence of a low-level absolutely and conditionally unstable layer no convection is initiated over the lake. Therefore, it is hypothesized that some additional forced uplift of up to 250 m was necessary to release conditional instability. Several processes contributed to this lifting:

  1. Thermally induced convergence, as the relatively warmer lake causes a deflection of the low-level winds.
  2. Orographically induced convergence, as the foothills of the larger scale Alpine topography south of the lake cause a low-level flow deflection from northwesterly to westerly directions, leading to westerly winds over the lake surface.
  3. Forced lifting by the mountains downstream of the lake is an important factor for enhancing precipitation and determining its peak intensity near the shore, but not for triggering the formation of the snowband over the lake.

The relative importance of the thermally and orographically induced convergence could not be quantified due to nonlinear interaction of these effects. However, it became clear that neither lake nor terrain effects alone were strong enough to produce a snowband. Alcott and Steenburgh (2013) note that at the Great Salt Lake some events are dominated by lake effects (Onton and Steenburgh 2001; Steenburgh and Onton 2001), while others are caused by an interaction between the orography and the lake. This is considered to be true for Lake Constance as well and topographic effects are hypothesized to dominate at least for the presented event, since the contrast between the LST and the 2-m air temperature of about 5 K was not sufficiently large to enable strong thermally induced convergence. However, the simulations with different LST showed that a sensitivity on the LST is evident. Therefore, a proper initialization of the LST in the numerical model seems crucial as it has a great impact on the formation of precipitation. An increase in the LST by 3 K causes a stronger land breeze, leading to a more midlake position of the snowband. Additionally, the vertical gradient within the conditionally unstable layer above the lake is stronger than in the control simulation. Beside the sensitivity to the prescribed LST, the microphysical parameterization also introduces some uncertainty in the quantitative precipitation forecast (Umek 2016).

A removal of the lake causes the convergence line and the snowband to vanish completely. Hence, it is unlikely that the convergence is solely induced by the surrounding topography. However, it is hard to disentangle the interaction between convergence and banded convection, as the convergence line may be either the cause or an effect of the banded convection. It is hypothesized that the conditional instability generated by the lake determines a favorable position for initiation of shallow convection. Subsequently, lateral inflow of low-level conditionally unstable air maintains and amplifies the low-level convergence and convective motion.

The model underestimates the snowfall amount over the Bregenzerwald compared with observed precipitation in that area (Fig. 7). High amounts of precipitation in this mountainous region may have been caused by (i) advection of hydrometeors generated by the snowband or (ii) orographic effects due to forced ascent of the impinging flow. Most likely a combination of the two effects was responsible for precipitation amounts exceeding 20 mm within a period of 6 h in the Bregenzerwald on 8 February 2013. However, we did not perform an analysis of hydrometeor trajectories to clarify the role of advection.

Such a trajectory analysis was conducted by Alcott and Steenburgh (2013). They state that at the Great Salt Lake the advection of hydrometeors to complex terrain downstream of the lake may be the strongest contribution to increased precipitation in that area. Their study showed maximum values of precipitation at higher altitudes rather than at the valley bottom. The event considered herein is characterized by high amounts of precipitation from the downstream shore up the windward side of the Pfänder mountain, suggesting that orographic uplift is more important for determining the location and magnitude of the snowfall maximum than hydrometeor advection. Figure 13a shows the strongest vertical velocities over the foot of the sloping terrain. Therefore, orographic lifting may be a possible mechanism to invigorate lake-effect convection as proposed and hypothesized by Minder et al. (2015, their Fig. 2a). Although Minder et al. (2015) found that this process was not a major factor in the events they investigated. Orographic lifting is thought to be the governing factor for generating the localized maximum in Bregenz. This is confirmed by the NoDT simulation, which shows a snowband but also a reduction of the maximum accumulated precipitation within the lake-effect subdomain of about 60% (Table 1). Therefore, orographic lifting over the downstream mountains is considered to be an amplification factor for precipitation, but not a necessary condition for triggering convection. If the downstream mountains are removed together with the lake, no snowband forms, emphasizing the importance of low-level instabilities caused by the lake.

Besides the previously mentioned lake-surface-to-air temperature contrast, the snowband position seems to be governed also by the strength of the low-level wind field. In CTL, winds on the northern shore are rather weak (2 m s−1), whereas westerly flow over the lake and the southern shore is stronger than 5 m s−1 during the snowband period. Hence, the low-level convergence line and the associated snowband is shifted toward the northern shore. Removing the upstream terrain increases the wind speed along the northern shore and moves the snowband to a more midlake position. Another factor that may influence the position of the snowband is the presence of an outflow from the Rhine Valley south of the lake that leads to convergence and lifting of the impinging airflow. The convergence at the valley exit causes an isolated maximum in precipitation in the simulations without the lake (e.g., Figs. 15d,h,j).

7. Conclusions

This study assessed the impact of Lake Constance and the surrounding orography on the formation of a snowstorm. Observations illustrated the formation of a rather stationary, banded precipitation structure between about 0700 and 1200 UTC 8 February 2013, which caused a high spatial variability in accumulated precipitation downstream of the lake. The snowband was characterized by a length of up to 30 km and a width of 5–10 km. During the period of banded precipitation, weather stations recorded a convergence in the wind field over the lake with westerly winds on the southern shore and northwesterly winds on the northern shore. The lake provided a moisture and heat source for the atmospheric boundary layer with the lake surface being about 5 K warmer than the atmosphere at 2 m AGL.

The results of the numerical simulations and the analysis of observational data indicate that Lake Constance is a crucial factor for the formation of heavy wintertime precipitation in the town of Bregenz and the downstream mountainous region. However, in this case study the lake was not the only factor; a complex interaction of several physical processes determined the formation of the snowband and the magnitude of precipitation. These processes include the following:

  • destabilization of low-level air by the lake surface due to latent and sensible heat fluxes;
  • deflection of low-level airflow by the larger scale Alpine orography west of the target area;
  • formation of a thermally and topographically induced low-level convergence line along the lake and downstream, capable of lifting air parcels to their LFC and triggering shallow convection over the lake; and
  • enhancement of precipitation by orographic lifting downstream of the lake.
The sensitivity experiments revealed the following:
  • The generation of heavy precipitation is determined by the interaction of lake and orographic processes. Neither the lake nor the orography alone would have been able to provide strong enough forcing to trigger convection and to form a snowband.
  • The location of the convergence line and, hence, the position of the snowband is determined by the strength of the lake and orographic forcing (i.e., the strength of the heat and moisture fluxes and the strength and deflection of the impinging flow). Increasing the lake surface temperature and removing the upstream terrain both shifted the convergence line closer to a midlake position.

Finally, an obvious question is to which extent our findings can be transferred to other precipitation events at Lake Constance. Especially, as the investigated event is characterized by a rather low lake–air temperature difference compared to other potential lake-effect events (Fig. 2). Therefore, it is hypothesized that during other events the lake may play a more important role.

A preliminary analysis of a second case (17 January 2013; Fig. 2) based on observations and simulations revealed similar results compared to the presented event. Nevertheless, only few events with banded, stationary precipitation downstream of Lake Constance during wintertime have been identified so far and even fewer have been investigated thoroughly up to now. Therefore, a climatological study to determine the synoptic and mesoscale conditions leading to snowbands downstream of the lake is desirable. In this context it is worth mentioning that the events listed in Fig. 2 are characterized by similar synoptic conditions; however, the sample size is rather small. Moreover, an analysis of air parcel trajectories could help to gain further insight into the role of flow deflection by the upstream terrain for the formation of the convergence line as well as to quantify the heat and moisture uptake of the airflow over the lake. Finally, hydrometeor trajectories may be helpful to clarify the role of advection for the distribution of precipitation downstream of the lake.

Acknowledgments

The Austrian national weather service ZAMG and specifically Susanne Drechsel are acknowledged for supporting the work and providing data. We are indebted to Rudolf Kaltenböck (Austro Control) for providing radar data. We thank Ralf Grabher (Hydrographic Service Vorarlberg) for providing precipitation data from rain gauges in Vorarlberg and for measurements of water temperatures from the harbor of Bregenz. Thomas Wolf (Landesanstalt für Umwelt, Messungen und Naturschutz Baden-Württemberg) supported us with data from a hydrodynamic lake model. MeteoGroup is acknowledged for providing data from automatic weather stations in Germany and Switzerland. We are grateful to Lukas Lehner for fruitful discussions and contributions to the preliminary climatology of lake-effect precipitation events. This work was supported by the Austrian Federal Ministry of Science, Research and Economy (BMWF) as part of the Uni-Infrastrukturprogramm of the Research Focal Point Scientific Computing at the University of Innsbruck.

REFERENCES

  • Alcott, T. I., , and W. J. Steenburgh, 2013: Orographic influences on a Great Salt Lake–effect snowstorm. Mon. Wea. Rev., 141, 24322450, doi:10.1175/MWR-D-12-00328.1.

    • Search Google Scholar
    • Export Citation
  • Andersson, T., , and S. Nilsson, 1990: Topographically induced convective snowbands over the Baltic Sea and their precipitation distribution. Wea. Forecasting, 5, 299312, doi:10.1175/1520-0434(1990)005<0299:TICSOT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Bluestein, H. B., 1993: Synoptic-Dynamic Meteorology in Midlatitudes. Vol. II, Observations and Theory of Weather Systems, Oxford University Press, 594 pp.

  • Bohren, C. F., , and B. A. Albrecht, 2010: Atmospheric Thermodynamics. Oxford University Press, 402 pp.

  • Buisan, S. T., , M. E. Earle, , J. L. Collado, , J. Kochendorfer, , J. Alastrué, , M. Wolff, , C. D. Smith, , and J. I. López-Moreno, 2016: Assessment of the underestimation of snowfall accumulation by tipping bucket gauges used operationally by the Spanish national weather service. Atmos. Meas. Tech. Discuss., doi:10.5194/amt-2016-197, in press.

    • Search Google Scholar
    • Export Citation
  • Cairns, M. M., , R. Collin, , T. Cylke, , M. Deutschendort, , and D. Mercer, 2001: A lake-effect snowfall in western Nevada—Part I: Synoptic setting and observations. 18th Conf. on Weather Analysis and Forecasting/14th Conf. on Numerical Weather Prediction, Fort Lauderdale, FL, Amer. Meteor. Soc., 5.5. [Available online at https://ams.confex.com/ams/WAF-NWP-MESO/techprogram/paper_22957.htm.]

  • Carpenter, D. M., 1993: The lake effect of the Great Salt Lake: Overview and forecast problems. Wea. Forecasting, 8, 181193, doi:10.1175/1520-0434(1993)008<0181:TLEOTG>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Conrick, R., , H. D. Reeves, , and S. Zhong, 2015: The dependence of QPF on the choice of boundary- and surface-layer parameterization for a lake-effect snowstorm. J. Appl. Meteor. Climatol., 54, 11771190, doi:10.1175/JAMC-D-14-0291.1.

    • Search Google Scholar
    • Export Citation
  • German Meteorological Society (DGM), 2014: Europäischer Meteorologischer Kalender 2015 (European Meteorological Calendar 2015). Gebr. Borntraeger Verlagsbuchhandlung. [Available online at http://www.borntraeger-cramer.de/9783443010799.]

  • Ek, M. B., , K. E. Mitchell, , Y. Lin, , E. Rogers, , P. Grunmann, , V. Koren, , G. Gayno, , and J. D. Tarpley, 2003: Implementation of Noah land surface model advances in the National Centers for Environmental Prediction operational mesoscale Eta model. J. Geophys. Res., 108, 8851, doi:10.1029/2002JD003296.

    • Search Google Scholar
    • Export Citation
  • Goodison, B. E., , P. Y. T. Louie, , and D. Yang, 1998: WMO solid precipitation measurement intercomparison. WMO Instruments and Observing Methods Rep. 67, WMO/TD-872, WMO, 212 pp.

  • Hjelmfelt, M. R., 1992: Orographic effects in simulated lake-effect snowstorms over Lake Michigan. Mon. Wea. Rev., 120, 373377, doi:10.1175/1520-0493(1992)120<0373:OEISLE>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Holroyd, E. W., 1971: Lake-effect cloud bands as seen from weather satellites. J. Atmos. Sci., 28, 11651170, doi:10.1175/1520-0469(1971)028<1165:LECBAS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Huggins, A. W., , D. E. Kingsmill, , and M. M. Cairns, 2001: A lake effect snowfall in western Nevada—Part II: Radar characteristics and quantitative precipitation estimates. 18th Conf. on Weather Analysis and Forecasting/14th Conf. on Numerical Weather Prediction, Fort Lauderdale, FL, Amer. Meteor. Soc., 5.6. [Available online at https://ams.confex.com/ams/WAF-NWP-MESO/techprogram/paper_22944.htm.]

  • Jiang, Q., , J. D. Doyle, , and R. B. Smith, 2005: Blocking, descent and gravity waves: Observations and modelling of a MAP northerly föhn event. Quart. J. Roy. Meteor. Soc., 131, 675701, doi:10.1256/qj.03.176.

    • Search Google Scholar
    • Export Citation
  • Kain, J. S., 2004: The Kain–Fritsch convective parameterization: An update. J. Appl. Meteor., 43, 170181, doi:10.1175/1520-0450(2004)043<0170:TKCPAU>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Kindap, T., 2010: A severe sea-effect snow episode over the city of Istanbul. Nat. Hazards, 54, 707723, doi:10.1007/s11069-009-9496-7.

    • Search Google Scholar
    • Export Citation
  • Laird, N. F., , J. E. Walsh, , and D. A. R. Kristovich, 2003: Model simulations examining the relationship of lake-effect morphology to lake shape, wind direction, and wind speed. Mon. Wea. Rev., 131, 21022111, doi:10.1175/1520-0493(2003)131<2102:MSETRO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Laird, N. F., , J. Desrochers, , and M. Payer, 2009a: Climatology of lake-effect precipitation events over Lake Champlain. J. Appl. Meteor. Climatol., 48, 232250, doi:10.1175/2008JAMC1923.1.

    • Search Google Scholar
    • Export Citation
  • Laird, N. F., , R. Sobash, , and N. Hodas, 2009b: The frequency and characteristics of lake-effect precipitation events associated with the New York State Finger Lakes. J. Appl. Meteor. Climatol., 48, 873886, doi:10.1175/2008JAMC2054.1.

    • Search Google Scholar
    • Export Citation
  • Laird, N. F., , R. Sobash, , and N. Hodas, 2010: Climatological conditions of lake-effect precipitation events associated with the New York State Finger Lakes. J. Appl. Meteor. Climatol., 49, 10521062, doi:10.1175/2010JAMC2312.1.

    • Search Google Scholar
    • Export Citation
  • Laird, N. F., , N. D. Metz, , L. Gaudet, , C. Grasmick, , L. Higgins, , C. Loeser, , and D. A. Zelinsky, 2016: Climatology of cold season lake-effect cloud bands for the North American Great Lakes. Int. J. Climatol., doi:10.1002/joc.4838, in press.

    • Search Google Scholar
    • Export Citation
  • Lang, U., , R. Schick, , and G. Schroder, 2010: The Decision Support System BodenseeOnline for hydrodynamics and water quality in Lake Constance. Advances in Decision Support Systems, G. Devlin, Ed., InTech, doi:10.5772/39391.

  • Lavoie, R. L., 1972: A mesoscale numerical model of lake-effect storms. J. Atmos. Sci., 29, 10251040, doi:10.1175/1520-0469(1972)029<1025:AMNMOL>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Lin, Y.-L., , R. D. Farley, , and H. D. Orville, 1983: Bulk parameterization of the snow field in a cloud model. J. Climate Appl. Meteor., 22, 10651092, doi:10.1175/1520-0450(1983)022<1065:BPOTSF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Lofgren, B. M., , and Y. Zhu, 2000: Surface energy fluxes on the Great Lakes based on satellite-observed surface temperatures 1992 to 1995. J. Great Lakes Res., 26, 305314, doi:10.1016/S0380-1330(00)70694-0.

    • Search Google Scholar
    • Export Citation
  • Markowski, P., , and Y. Richardson, 2010: Mesoscale Meteorology in Midlatitudes. Wiley-Blackwell, 430 pp.

  • Minder, J. R., , T. W. Letcher, , L. S. Campbell, , P. G. Veals, , and W. J. Steenburgh, 2015: The evolution of lake-effect convection during landfall and orographic uplift as observed by profiling radars. Mon. Wea. Rev., 143, 44224442, doi:10.1175/MWR-D-15-0117.1.

    • Search Google Scholar
    • Export Citation
  • Mlawer, E. J., , S. J. Taubman, , P. D. Brown, , M. J. Iacono, , and S. A. Clough, 1997: Radiative transfer for inhomogeneous atmospheres: RRTM, a validated correlated-k model for the longwave. J. Geophys. Res., 102, 16 66316 682, doi:10.1029/97JD00237.

    • Search Google Scholar
    • Export Citation
  • Muller, R. A., 1966: Snowbelts of the Great Lakes. Weatherwise, 19, 248255, doi:10.1080/00431672.1966.10544204.

  • Nagata, W. S., 1987: On the structure of a convergent cloud band over the Japan Sea in winter: A prediction experiment. J. Meteor. Soc. Japan, 65, 871883.

    • Search Google Scholar
    • Export Citation
  • Nakanishi, M., , and H. Niino, 2004: An improved Mellor–Yamada level-3 model with condensation physics: Its design and verification. Bound.-Layer Meteor., 112, 131, doi:10.1023/B:BOUN.0000020164.04146.98.

    • Search Google Scholar
    • Export Citation
  • Nicosia, D. J., and et al. , 1999: A flash flood from a lake-enhanced rainband. Wea. Forecasting, 14, 271288, doi:10.1175/1520-0434(1999)014<0271:AFFFAL>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Niziol, T. A., 1987: Operational forecasting of lake effect snowfall in western and central New York. Wea. Forecasting, 2, 310321, doi:10.1175/1520-0434(1987)002<0310:OFOLES>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Niziol, T. A., , W. R. Snyder, , and J. S. Waldstreicher, 1995: Winter weather forecasting throughout the eastern United States. Part IV: Lake effect snow. Wea. Forecasting, 10, 6177, doi:10.1175/1520-0434(1995)010<0061:WWFTTE>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Norris, J. G., , G. Vaughan, , and D. M. Schultz, 2013: Snowbands over the English Channel and Irish Sea during cold-air outbreaks. Quart. J. Roy. Meteor. Soc., 139, 17471761, doi:10.1002/qj.2079.

    • Search Google Scholar
    • Export Citation
  • Onton, D. J., , and W. J. Steenburgh, 2001: Diagnostic and sensitivity studies of the 7 December 1998 Great Salt Lake–effect snowstorm. Mon. Wea. Rev., 129, 13181338, doi:10.1175/1520-0493(2001)129<1318:DASSOT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Peace, R. L., , and R. B. Sykes, 1966: Mesoscale study of a lake effect snow storm. Mon. Wea. Rev., 94, 495507, doi:10.1175/1520-0493(1966)094<0495:MSOALE>2.3.CO;2.

    • Search Google Scholar
    • Export Citation
  • Rasmussen, R., and et al. , 2012: How well are we measuring snow: The NOAA/FAA/NCAR Winter Precipitation Test Bed. Bull. Amer. Meteor. Soc., 93, 811829, doi:10.1175/BAMS-D-11-00052.1.

    • Search Google Scholar
    • Export Citation
  • Reiner, G., 2015: Der Bodensee und seine Mitschuld am Schneechaos (The Lake Constance and its complicity in the snow chaos). Accessed 19 May 2016, Vorarlberger Nachrichten. [Available online at http://www.vorarlbergernachrichten.at/sonderthemen/2013/02/08/.]

  • Riffler, M., , G. Lieberherr, , and S. Wunderle, 2015: Lake surface water temperatures of European Alpine lakes (1989–2013) based on the Advanced Very High Resolution Radiometer (AVHRR) 1 km data set. Earth Syst. Sci. Data, 7, 117, doi:10.5194/essd-7-1-2015.

    • Search Google Scholar
    • Export Citation
  • Saito, K., , M. Murakami, , T. Matsuo, , and H. Mizuno, 1996: Sensitivity experiments on the orographic snowfall over the mountainous region of northern Japan. J. Meteor. Soc. Japan, 74, 797813.

    • Search Google Scholar
    • Export Citation
  • Schultz, D. M., , D. S. Arndt, , D. J. Stensrud, , and J. W. Hanna, 2004: Snowbands during the cold-air outbreak of 23 January 2003. Mon. Wea. Rev., 132, 827842, doi:10.1175/1520-0493(2004)132<0827:SDTCOO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Schumacher, R. S., , D. M. Schultz, , and J. A. Knox, 2015: Influence of terrain resolution on banded convection in the lee of the Rocky Mountains. Mon. Wea. Rev., 143, 13991416, doi:10.1175/MWR-D-14-00255.1.

    • Search Google Scholar
    • Export Citation
  • Skamarock, W. C., and et al. , 2008: A description of the Advanced Research WRF version 3. NCAR Tech. Note NCAR/TN-475+STR, 113 pp., doi:10.5065/D68S4MVH.

  • Steenburgh, W. J., 2003: One hundred inches in one hundred hours: Evolution of a Wasatch Mountain winter storm cycle. Wea. Forecasting, 18, 10181036, doi:10.1175/1520-0434(2003)018<1018:OHIIOH>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Steenburgh, W. J., , and D. J. Onton, 2001: Multiscale analysis of the 7 December 1998 Great Salt Lake–effect snowstorm. Mon. Wea. Rev., 129, 12961317, doi:10.1175/1520-0493(2001)129<1296:MAOTDG>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Steenburgh, W. J., , S. F. Halvorson, , and D. J. Onton, 2000: Climatology of lake-effect snowstorms of the Great Salt Lake. Mon. Wea. Rev., 128, 709727, doi:10.1175/1520-0493(2000)128<0709:COLESO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Steiger, S. M., and et al. , 2013: Circulations, bounded weak echo regions, and horizontal vortices observed within long-lake-axis-parallel–lake-effect storms by the Doppler on Wheels. Mon. Wea. Rev., 141, 28212840, doi:10.1175/MWR-D-12-00226.1.

    • Search Google Scholar
    • Export Citation
  • Umek, L., 2016: Lake and orographic effects on precipitation at Lake Constance: Case study and numerical sensitivity experiments. M.S. thesis, Institute of Atmospheric and Cryospheric Sciences, University of Innsbruck, 69 pp. [Available online at http://resolver.obvsg.at/urn:nbn:at:at-ubi:1-3615.]

  • Veals, P. G., , and W. J. Steenburgh, 2015: Climatological characteristics and orographic enhancement of lake-effect precipitation east of Lake Ontario and over the Tug Hill Plateau. Mon. Wea. Rev., 143, 35913609, doi:10.1175/MWR-D-15-0009.1.

    • Search Google Scholar
    • Export Citation
  • Wilken, G. R., 1997: A lake-effect snow in Arkansas. Tech. Attachment SR/SSD 97-21, NSFO Little Rock, AR. [Available online at http://www.srh.noaa.gov/topics/attach/html/ssd97-21.htm.]

  • Wilson, J. W., 1977: Effect of Lake Ontario on precipitation. Mon. Wea. Rev., 105, 207214, doi:10.1175/1520-0493(1977)105<0207:EOLOOP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Wright, D. M., , D. J. Posselt, , and A. L. Steiner, 2013: Sensitivity of lake-effect snowfall to lake ice cover and temperature in the Great Lakes region. Mon. Wea. Rev., 141, 670689, doi:10.1175/MWR-D-12-00038.1.

    • Search Google Scholar
    • Export Citation
  • Yeager, K. N., , W. J. Steenburgh, , and T. I. Alcott, 2013: Contributions of lake-effect periods to the cool-season hydroclimate of the Great Salt Lake basin. J. Appl. Meteor. Climatol., 52, 341362, doi:10.1175/JAMC-D-12-077.1.

    • Search Google Scholar
    • Export Citation
Save