• Bell, G. D., and L. F. Bosart, 1988: Appalachian cold-air damming. Mon. Wea. Rev.,116, 137–161.

  • ——, and ——, 1989: The large-scale atmospheric structure accompanying New England coastal frontogenesis and associated North American east coast cyclogenesis. Quart. J. Roy. Meteor. Soc.,115, 1133–1146.

  • Bosart, L. F., and J. P. Cussen Jr., 1973: Gravity wave phenomena accompanying East Coast cyclogenesis. Mon. Wea. Rev.,101, 5445–5454.

  • ——, and F. Sanders, 1986: Mesoscale structure in the megalopolitan snowstorm of 11–12 February 1983. Part III: A large amplitude gravity wave and coastal frontogenesis. J. Atmos. Sci.,43, 924–939.

  • ——, and A. Seimon, 1988: A case study of an unusually intense atmospheric gravity wave. Mon. Wea. Rev.,116, 1857–1886.

  • ——, and G. M. Lackmann, 1995: Post-landfall tropical cyclone reintensification in a weakly-baroclinic environment: A case study of Hurricane David (September 1979). Mon. Wea. Rev.,123, 3268–3291.

  • ——, M. A. Bedrick, W. E. Bracken, M. J. Dickinson, G. J. Hakim, D. M. Schultz, and K. R. Tyle, 1996: Large-scale antecedent conditions associated with the 12–14 March 1993 cyclone (“Superstorm ’93”) over eastern North America. Mon. Wea. Rev.,124, 1865–1891.

  • Brunk, I., 1949: The pressure pulsation of 11 April 1944. J. Meteor.,6, 181–187.

  • Clark, J. D., 1983: The GOES user’s guide. United States Department of Commerce, 7-38–7-39.

  • Cressman, G. P., 1959: An operational objective analysis system. Mon. Wea. Rev.,87, 367–374.

  • Davis, C. A., E. D. Grell, and M. A. Shapiro, 1996: The balanced dynamical nature of a rapidly intensifying oceanic cyclone. Mon. Wea. Rev.,124, 3–26.

  • Ferguson, H. L., 1967: Mathematical and synoptic aspects of a small-scale wave disturbance over the lower Great Lakes area. J. Appl. Meteor.,6, 523–529.

  • Forbes, P. A., G. S., R. A. Anthes, and D. W. Thomson, 1987: Synoptic and mesoscale aspects of an Appalachian ice storm associated with cold-air damming. Mon. Wea. Rev.,115, 564–591.

  • Hakim, G. J., L. F. Bosart, and D. Keyser, 1995: The Ohio Valley wave-merger cyclogenesis event of 25–26 January 1978. Part I:Multiscale case study. Mon. Wea. Rev.,123, 2663–2692.

  • ——, D. Keyser, and L. F. Bosart, 1996: The Ohio Valley wave-merger cyclogenesis event of 25–26 January 1978. Part II: Diagnosis using quasigeostrophic potential vorticity inversion. Mon. Wea. Rev.,124, 2176–2205.

  • Hoskins, B. J., M. E. McIntyre, and A. W. Robertson, 1985: On the use and significance of isentropic potential vorticity maps. Quart. J. Roy. Meteor. Soc.,111, 877–946.

  • Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-year reanalysis project. Bull. Amer. Meteor. Soc.,77, 437–471.

  • Kaplan, M. L., S. E. Koch, Y.-L. Lin, R. P. Weglarz, and R. A. Rozumalski, 1997: Numerical simulations of a gravity wave event over CCOPE. Part I: The role of geostrophic adjustment in mesoscale jetlet formation. Mon. Wea. Rev.,125, 1185–1211.

  • Keshishian, L. G., and L. F. Bosart 1987: A case study of extended East Coast frontogenesis. Mon. Wea. Rev.,115, 100–117.

  • Koch, S. E., and R. E. Golus, 1988: A mesoscale gravity wave event observed during CCOPE. Part I: Multiscale statistical analysis of wave characteristics. Mon. Wea. Rev.,116, 2527–2544.

  • ——, and C. O’Handley, 1997: Operational forecasting and detection of mesoscale gravity waves. Wea. Forecasting,12, 253–281.

  • ——, M. desJardins, and P. J. Kocin, 1983: An interactive Barnes objective map analysis scheme for use with satellite and conventional data. J. Climate Appl. Meteor.,22, 1487–1503.

  • Krishnamurti, T. N., and W. J. Moxim, 1971: On the parameterization of convective and nonconvective latent heat release. J. Appl. Meteor.,10, 3–13.

  • Lefevre, R. J., and J. W. Nielsen-Gammon, 1995: An objective climatology of mobile troughs in the Northern Hemisphere. Tellus,47A, 638–655.

  • Lin, Y. L., and R. C. Goff, 1988: A study of a mesoscale solitary wave in the atmosphere originating near a region of deep convection. J. Atmos. Sci.,45, 194–205.

  • Lindzen, R. S., and K. K. Tung, 1976: Banded convective activity and ducted gravity waves. Mon. Wea. Rev.,104, 1602–1617.

  • Nielsen-Gammon, J. W., 1995: Dynamical conceptual models of upper-level mobile trough formation: Comparison and application. Tellus,47A, 705–721.

  • ——, and R. J. Lefevre, 1996: Piecewise tendency diagnosis of dynamical processes governing the development of an upper-tropospheric mobile trough. J. Atmos. Sci.,53, 3120–3142.

  • Olson, D. A., N. W. Junker, and B. Korty, 1995: Evaluation of 33 years of quantitative precipitation forecasting at the NMC. Wea. Forecasting,10, 498–511.

  • Pecnick, M. J., and J. A. Young, 1984: Mechanics of a strong sub-synoptic gravity wave deduced from satellite and surface observations. J. Atmos. Sci.,41, 1850–1862.

  • Pokrandt, P. J., G. J. Tripoli, and D. D. Houghton, 1996: Processes leading to the formation of mesoscale waves in the Midwest cyclone of 15 December 1987. Mon. Wea. Rev.,124, 2726–2752.

  • Powers, J. G., 1997: Numerical model simulations of a mesoscale gravity wave event: Sensitivity tests and spectral analyses. Mon. Wea. Rev.,125, 1838–1869.

  • ——, and R. J. Reed, 1993: Numerical simulation of the large-amplitude mesoscale gravity-wave event of 15 December 1987 in the central United States. Mon. Wea. Rev.,121, 2285–2308.

  • Ralph, F. M., M. Crochet, and S. V. Venkateswaran, 1993: Observations of a mesoscale ducted gravity wave. J. Atmos. Sci.,50, 3277–3291.

  • Ramamurthy, M. K., R. M. Rauber, B. P. Collins, and N. K. Malhotra, 1993: A comparative study of large-amplitude gravity-wave events. Mon. Wea. Rev.,121, 2951–2974.

  • Schneider, R. S., 1990: Large-amplitude mesoscale wave disturbances within the intense midwest extratropical cyclone of 15 December 1987. Wea. Forecasting,5, 533–558.

  • Tepper, M., 1954: Pressure jump lines in midwestern United States, January–August 1951. Research Paper No. 37, U.S. Weather Bureau, 70 pp.

  • Tsou, C.-H., P. J. Smith, and P. M. Pauley, 1987: A comparison of adiabatic and diabatic forcing in an intense extratropical cyclone system. Mon. Wea. Rev.,115, 763–786.

  • Uccellini, L. W., and S. E. Koch, 1987: The synoptic setting and possible energy sources for mesoscale wave disturbances. Mon. Wea. Rev.,115, 721–729.

  • ——, and P. J. Kocin, 1987: The interaction of jet streak circulations during heavy snow events along the east coast of the United States. Wea. Forecasting,2, 289–308.

  • Van Tuyl, A. H., and J. A. Young, 1982: Numerical simulation of nonlinear jet streak adjustment. Mon. Wea. Rev.,110, 2038–2054.

  • Vincent, D. G., G. B. Pant, and H. J. Edmon, 1977: Generation of available potential energy of an extratropical cyclone system. Mon. Wea. Rev.,105, 1252–1265.

  • Young, J. A., and C. F. Richards, 1973: A climatological study of surface pressure events at Madison, Wisconsin: Preliminary results relating to possible gravity wave activity. Studies of the Atmosphere using Aerospace Proves, Space Science and Engineering Center 1973 Annual Report, University of Wisconsin—Madison, 172–187. [NTIS COM-74-11626/0.].

  • Zack, J. W., and M. L. Kaplan, 1987: Numerical simulation of the subsynoptic features associated with the AVE-SESAME I case. Part I: The preconvective environment. Mon. Wea. Rev.,115, 2367–2393.

  • View in gallery

    Dominant IGW isochrone analysis for 0700–1900 UTC 4 January 1994. Area affected by “snow bomb” outlined by bold-dashed ellipse. Region of multiple small-amplitude IGW disturbances outlined by bold-dotted ellipse.

  • View in gallery

    Station locator map. Solid dots, triangles, and open triangles denote surface stations, upper-air stations, and WSR-88D sites, respectively. Profiler site at Bloomfield (BMF), Connecticut, is denoted by an open square. Inset shows cross-sectional orientations used in Fig. 29.

  • View in gallery

    National Meteorological Center (NMC, now the National Centers for Environmental Prediction) Eta Model–initialized 1000-hPa heights (solid, every 4 dam) and 1000–500-hPa thickness (dashed, every 6 dam) for (a) 0000, (b) 0600, (c) 1200, and (d) 1800 UTC 4 January 1994. Shading (scaled at left) denotes regions where the Richardson number (Ri) is <1. Richardson number is computed as a layer average over 500–450 (350–300) hPa in (a), (b) [(c), (d)].

  • View in gallery

    As in Fig. 3 except for 500-hPa heights (solid, every 6 dam), 500-hPa absolute vorticity (dashed, every 6 × 10−5 s−1), and 500-hPa vertical motion (ascent only; every 3 × 10−3 hPa s−1 and shaded according to the gray scale at the left of each panel). Representative gridpoint winds plotted with one pennant, full barb, and half-barb denoting 25, 5, and 2.5 m s−1, respectively.

  • View in gallery

    Left column (a, d, g): pressure (solid, every 50 hPa and shaded beginning at 350 hPa according to the gray scale) on the dynamic tropopause (DT) defined by the 1.5 potential vorticity unit (PVU) surface (1.0 PVU = 10−6 K m2 kg−1 s−1) for 1200 UTC 3 January 1994 (top), 0000 UTC 4 January 1994 (middle), and 1200 UTC 4 January 1994 (bottom). Middle column (b, e, h): as in the left column except for potential temperature (solid, every 10 K) on the DT. Right column (c, f, i): as in the left column except for 850-hPa heights (solid, every 3 dam) and equivalent potential temperature (dashed, every 10 K). Winds as in Fig. 4.

  • View in gallery

    Digitized pressure (hPa) versus time (UTC) for Bedford, Massachusetts: (a) 24 h ending 0000 UTC 5 January 1994, and (b) 1-min resolution data for the 2-h period ending 1600 UTC 4 January 1994. Source: A. Jackson of the Air Force Phillips Laboratories, Hanscom Field, Bedford, Massachusetts.

  • View in gallery

    Reconstructed analog wind speed trace for (a) Boston (BOS), Massachusetts, and (b) Providence (PVD), Rhode Island, centered at 1500 (1430) UTC 4 January 1994 for BOS (PVD). Vertical lines denote 5-min intervals. Wind speeds in knots (multiply by 0.515 to convert to m s−1) with horizontal lines every 10 kt. Precipitation type and intensity indicated at BOS along the abscissa.

  • View in gallery

    Reconstructed microbarogram traces for selected stations in the northeastern United States for 4 January 1994. Pressure (hPa) and time (hours) scale as shown with tick marks denoting every 1 hPa and 1 h, respectively. Solid contours denote peak ridge-to-trough pressure decreases (hPa) associated with IGW passage and are shaded accordingly.

  • View in gallery

    National Meteorological Center (NMC, now the National Centers for Environmental Prediction) manually prepared 24-h quantitative precipitation forecast (QPF) for 24-h periods ending (a) 1200 UTC 4 January 1994, (b) 1200 UTC 5 January 1994, and the 24-h update ending (c) 1200 UTC 4 January 1994. Contour intervals are 6.2, 12.5, and 25.0 mm.

  • View in gallery

    Accumulated precipitation (solid, contoured every 1, 5, 10, and 20 mm) for 6-h periods ending (a) 0600, (b) 1200, and (c) 1800 UTC 4 January 1994.

  • View in gallery

    Reconstructed station pressure (solid, hPa; tick marks every 2 hPa on the ordinate) for Elkins (EKN), West Virginia (top), Williamsport (IPT), Pennsylvania (middle), and Syracuse (SYR), New York (bottom), for 4 January 1994. Times (UTC) shown by two-digit numbers along the abscissa; sloping vertical lines every 1 h. Hourly snowfall rates (cm h−1) are shown stippled. Conventional plotting for present weather with temperature and dewpoint temperature in degrees Celsius and winds in m s−1 according to Fig. 4.

  • View in gallery

    (a) Infrared satellite imagery for (a) 0001 UTC, (b) 0301 UTC, (c) 0601 UTC, and (d) 0901 UTC 4 January 1994. Gray scales correspond to operational MB curve (Clark 1983) except that temperatures −31.2° to 1.8°C are linearly stretched. Arrows mark the position of multiple cloud bands discussed in text.

  • View in gallery

    Manually prepared surface analysis for 0600 UTC 4 January 1994. Mean sea level isobars (solid, every 2 hPa). Plotting format of conventional surface observations is as in Figs. 4 and 11.

  • View in gallery

    Sounding in skew T–logp format for Greensboro (GSO), North Carolina, for 0000 (dotted) and 0600 (solid) UTC 4 January 1994 and for Athens (AHN), Georgia, for 0000 (dashed) UTC 4 January 1994. Winds (m s−1, in format of Fig. 4) for GSO at 0000 UTC (left), AHN (center), and GSO at 0600 UTC (right).

  • View in gallery

    National Lightning Detection Network cloud-to-ground lightning flash distribution: (a) 0000–0600 UTC 4 January 1994, (b) 0600–1200 UTC 4 January 1994. Triangle, box, and circle symbols denote first 2 h, middle 2 h, and last 2 h of each 6-h period, respectively.

  • View in gallery

    As in Fig. 13 except for 1200 UTC 4 January 1994.

  • View in gallery

    As in Fig. 14 except for (a) Albany (ALB; solid), New York, and Chatham (CHH; dashed), Massachusetts, at 1200 UTC 4 January 1994, and (b) Stephenville (YJT; solid), Newfoundland, and Sable Island (WSA; dashed), Nova Scotia, at 0000 UTC 5 January 1994. Winds:(a) ALB (left) and CHH (right); (b) YJT (left); and WSA (right).

  • View in gallery

    Base reflectivity from the WSR-88D Doppler radar at Sterling (KWBC), Virginia, for 0908 UTC 4 January 1994. Solid (dashed) contours denote 15 (30) dBZ with reflectivities greater than 30 dBZ shown stippled.

  • View in gallery

    As in Fig. 15 except for (a) 1001, (b) 1201, (c) 1401, and (d) 1601 UTC 4 January 1994. Arrows denote region of high cold cloud tops associated with dominant IGW.

  • View in gallery

    As in Fig. 13 except for 1350 UTC 4 January 1994.

  • View in gallery

    Left column: composite WSR-88D base reflectivities for (a) 1130, (c) 1210, and (e) 1250 UTC 4 January 1994. Contour intervals every 10 dBZ beginning 15 dBZ. Right column: composite 20-min pressure change analysis with solid (dashed) lines denoting pressure falls (rises) contoured every 0.5, 1.0, 2.0, 4.0, and 8.0 hPa with zero contour dashed for (b) 1130, (d) 1210, and (f) 1250 UTC 4 January 1994.

  • View in gallery

    As in Fig. 21 except for 1330, 1410, and 1450 UTC 4 January 1994.

  • View in gallery

    Composite WSR-88D 20-min isochrones of the 15-dBZ base reflectivity contour along the back edge of the precipitation shield preceding the IGW axis of minimum sea level pressure for the period 1110–1530 UTC 4 January 1994.

  • View in gallery

    Time series of smoothed hourly winds (m s−1, plotted according to the convention given in Fig. 4) from the Bloomfield (BMF), Connecticut, wind profiler.

  • View in gallery

    As in Fig. 13 except for 1550 UTC 4 January 1994.

  • View in gallery

    As in Fig. 13 except for 1750 UTC 4 January 1994.

  • View in gallery

    Observations of winds (m s−1) from the Mount Washington Observatory in North Conway, New Hampshire, for 4 January 1994. Spiral radial arms denote hourly times (UTC) increasing counterclockwise. Mean hourly wind speed and direction (octas) as shown.

  • View in gallery

    Maps of 400-hPa geopotential heights (thin solid lines every 6 dam), Lagrangian Rossby number (see text) computed at 400-hPa (thick solid lines at intervals of 0.5, 0.75, 1.0, 1.5, 2.0, 2.5 and 3.0 units), 250-hPa isotachs (shaded every 10 m s−1 beginning at 60 m s−1), and 300-hPa ageostrophic wind barbs plotted every other grid point in the format of Fig. 4 for (a) 0000, (b) 0600, (c) 1200, and (d) 1800 UTC 4 January 1994. (e)–(g) As in (a)–(d) except for the 400-hPa parcel divergence tendency (positive values shaded at contour intervals of 0.5, 1.0, 1.5, 2.0, 3.0, 4.0 × 10−8 s−2), 400-hPa geopotential heights (thin solid lines every 6 dam), and unbalanced divergence (zero and positive values only contoured every 2 × 10−5 s−1) averaged over the 350–250-hPa layer. All computed quantities are derived from the gridded Eta Model fields. IGW pressure minimum axis is denoted by a bold dotted line at 1200 and 1800 UTC 4 January 1994.

  • View in gallery

    (Continued)

  • View in gallery

    Left: cross sections (orientation shown in Fig. 2) of potential temperature (medium solid every 8 K) and potential vorticity (thin solid every 0.5 PV and shaded according to the scale below the top panel for PV > 1.5 PVU where 1.0 PVU is as in Fig. 5) for (a) 0000 and (d) 1200 UTC 4 January 1994. Middle: Cross sections of equivalent potential temperature (medium solid every 5 K) and vertical motion (shaded according to the scale along the top panel every 3 hPa s−1 beginning at ±3 hPa s−1) with ascent and descent regions outlined by thin dashed and solid lines, respectively, for 0000 (b) and 1200 (e) UTC 4 January 1994. Right: Cross sections of parcel divergence tendency (zero and positive values only; contoured bold dashed every 0.5 × 10−8 s−2), critical level (heavy solid line), based upon a wave phase velocity of 20 and 30 m s−1, respectively, at 0000 and 1200 UTC 4 January 1977, Richardson number (shaded according to the scale below the top panel for values of 2 and less), and unbalanced divergence given by thin solid (dashed) contours for positive (negative) values every 2 × 10−5 s−1 for 0000 (c) and 1200 (f) UTC 4 January 1994. Top and bottom cross sections extend about 2850 and 3000 km, respectively, with selected station locations (Fig. 2) shown along top of each panel.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 48 48 10
PDF Downloads 32 32 5

A Study of Cyclone Mesoscale Structure with Emphasis on a Large-Amplitude Inertia–Gravity Wave

View More View Less
  • 1 Department of Earth and Atmospheric Sciences, The University at Albany, State University of New York, Albany, New York
© Get Permissions
Full access

Abstract

An analysis is presented of prominent mesoscale structure in a moderately intense cyclone with emphasis on a long-lived, large-amplitude inertia–gravity wave (IGW) that moved through the northeastern United States on 4 January 1994. Available National Weather Service WSR-88D Doppler radar and wind profiler observations are employed to illustrate the rich, time-dependent, three-dimensional structure of the IGW. As the IGW amplified [peak crest-to-trough pressure falls exceeded 13 hPa (30 min)−1], it also accelerated away from the cyclone, reaching a peak forward speed of 35–40 m s−1 across eastern New England. The IGW was one of three prominent mesoscale features associated with the cyclone, the others being a weak offshore precursor warm-frontal wave and an onshore band of heavy snow (“snow bomb”) in which peak hourly snowfalls of 10–15 cm were observed. None of these three prominent mesoscale features were well forecast by existing operational prediction models, particularly with regard to precipitation amount, onset, and duration. The observed precipitation discrepancies illustrate the subtle but important effects of subsynoptic-scale disturbances embedded within the larger-scale cyclonic circulation. The precursor offshore warm-frontal wave was instrumental in reinforcing the wave duct preceding the IGW. The snow bomb was an indication of vigorous ascent, large upper- (lower-) level divergence (convergence), unbalanced flow, and associated large parcel accelerations, environmental conditions known to be favorable for IGW formation.

Small-amplitude IGWs (<1 hPa) are first detected over the southeastern United States from surface microbarogram records and are confirmed independently by the presence of organized and persistent mesoscale cloud bands oriented approximately along the wave fronts. The area of IGW genesis is situated poleward of a weak surface frontal boundary where there is a weak wave duct (stable layer) present in the lower troposphere. In the upper troposphere the region of IGW genesis is situated on the forward side of a deep trough where there is significant cyclonic vorticity advection by the thermal wind. Diagnostic evidence supports the importance of shearing instability and/or unbalanced flow in IGW genesis.

The large-amplitude IGW originates on the downstream edge of the northeastward-advancing packet of small-amplitude IGWs. Wave amplification occurs near the upshear edge of a high, cold cloud shield that generally marks the warm conveyor belt. Although it is not possible to conclusively state whether the amplifying IGW forms in situ or grows from a predecessor weaker (<1 hPa) disturbance, rapid amplification occurs 1) as the wave encounters an increasingly deeper and stronger wave duct, possibly permitting wave overreflection, in the cold air damming region east of the Appalachians, and 2) downshear of an area of significantly positive unbalanced divergence and parcel divergence tendency. The authors raise the possibility that IGW amplification can be associated with the penetration and perturbation of the wave duct by vigorous subsynoptic-scale vertical motions whose vigor is increased by wave-induced latent heat release.

Corresponding author address: Dr. Lance F. Bosart, Department of Earth and Atmospheric Sciences, The University at Albany/SUNY, 1400 Washington Ave., Albany, NY 12222.

Email: bosart@atmos.albany.edu

Abstract

An analysis is presented of prominent mesoscale structure in a moderately intense cyclone with emphasis on a long-lived, large-amplitude inertia–gravity wave (IGW) that moved through the northeastern United States on 4 January 1994. Available National Weather Service WSR-88D Doppler radar and wind profiler observations are employed to illustrate the rich, time-dependent, three-dimensional structure of the IGW. As the IGW amplified [peak crest-to-trough pressure falls exceeded 13 hPa (30 min)−1], it also accelerated away from the cyclone, reaching a peak forward speed of 35–40 m s−1 across eastern New England. The IGW was one of three prominent mesoscale features associated with the cyclone, the others being a weak offshore precursor warm-frontal wave and an onshore band of heavy snow (“snow bomb”) in which peak hourly snowfalls of 10–15 cm were observed. None of these three prominent mesoscale features were well forecast by existing operational prediction models, particularly with regard to precipitation amount, onset, and duration. The observed precipitation discrepancies illustrate the subtle but important effects of subsynoptic-scale disturbances embedded within the larger-scale cyclonic circulation. The precursor offshore warm-frontal wave was instrumental in reinforcing the wave duct preceding the IGW. The snow bomb was an indication of vigorous ascent, large upper- (lower-) level divergence (convergence), unbalanced flow, and associated large parcel accelerations, environmental conditions known to be favorable for IGW formation.

Small-amplitude IGWs (<1 hPa) are first detected over the southeastern United States from surface microbarogram records and are confirmed independently by the presence of organized and persistent mesoscale cloud bands oriented approximately along the wave fronts. The area of IGW genesis is situated poleward of a weak surface frontal boundary where there is a weak wave duct (stable layer) present in the lower troposphere. In the upper troposphere the region of IGW genesis is situated on the forward side of a deep trough where there is significant cyclonic vorticity advection by the thermal wind. Diagnostic evidence supports the importance of shearing instability and/or unbalanced flow in IGW genesis.

The large-amplitude IGW originates on the downstream edge of the northeastward-advancing packet of small-amplitude IGWs. Wave amplification occurs near the upshear edge of a high, cold cloud shield that generally marks the warm conveyor belt. Although it is not possible to conclusively state whether the amplifying IGW forms in situ or grows from a predecessor weaker (<1 hPa) disturbance, rapid amplification occurs 1) as the wave encounters an increasingly deeper and stronger wave duct, possibly permitting wave overreflection, in the cold air damming region east of the Appalachians, and 2) downshear of an area of significantly positive unbalanced divergence and parcel divergence tendency. The authors raise the possibility that IGW amplification can be associated with the penetration and perturbation of the wave duct by vigorous subsynoptic-scale vertical motions whose vigor is increased by wave-induced latent heat release.

Corresponding author address: Dr. Lance F. Bosart, Department of Earth and Atmospheric Sciences, The University at Albany/SUNY, 1400 Washington Ave., Albany, NY 12222.

Email: bosart@atmos.albany.edu

1. Introduction

On 4 January 1994 multiple inertia–gravity waves (IGW) were observed in association with an extratropical cyclone moving poleward along the Atlantic coast. A dominant IGW that intensified and accelerated as it moved northeastward (forward speed of 35–40 m s−1) over New England and adjacent Canada is the primary focus of this paper. This IGW is associated with exceptionally large observed surface pressure falls (>10 hPa h−1), a significant (and unforecast) mesoscale wind and precipitation perturbation, and a 1–2-m rapid sea level oscillation (seiche) in coastal harbors along Narragansett Bay and the Maine coast. The availability of National Weather Service (NWS) Doppler radar observations (WSR-88D) from several locations and a set of wind profiler observations from Bloomfield, Connecticut, provided incentive to examine the three-dimensional structure of the IGW. A map showing smoothed hourly isochrones of the dominant IGW-induced pressure minimum, as constructed from the available surface station microbarograms using time-to-space conversion analysis techniques, is displayed in Fig. 1 for overview purposes. Multiple small-amplitude (<1 hPa) IGWs were observed over the southeastern United States prior to the appearance of the dominant IGW (Fig. 1).

Although the IGW event was extraordinary, the cyclone was very ordinary in terms of minimum central pressure (988 hPa), modest deepening (<6 hPa in 12 h), and the extent of the overall cyclonic circulation. Noteworthy was the presence of significant mesoscale structure embedded within the cyclone environment including (in addition to the IGW) a concentrated region of exceptionally heavy snow with peak snowfall rates in excess of 10 cm h−1 (snow bomb; Fig. 1) that moved northeastward along the Appalachians and just to the west of the track of a midlevel cyclonic vorticity maximum, and an area of coastal precipitation that moved northeastward in association with a predecessor weak offshore warm front wave. Both of these precipitation areas, and the precipitation-enhanced and suppressed regions ahead and behind the IGW, respectively, proved difficult to predict by forecasters and the operational National Meteorological Center [NMC, now the National Centers for Environmental Prediction, (NCEP)] Regional Analysis and Forecast System (RAFS) and Eta Models.

The purpose of this paper will be to document the three-dimensional structure of the IGW and to examine the IGW life cycle within the context of other mesoscale features and the evolving large-scale flow. We accept a priori, based upon subjective visual inspection, that the NMC RAFS and Eta Model performed admirably in simulating the broad structure of the cyclone-related geopotential height and thermal fields over eastern North America in the short-range 12–24-h forecasts. More difficult to quantify, however, is how well these models depicted the important mesoscale features embedded within the cyclone. The distinction is important because model success in simulating the large-scale mass field is no guarantee of similar success in depicting the prominent mesoscale features that are responsible for the important sensible weather events of greatest concern to the general public.

A number of investigators have documented the existence and surface structure of long-lived, large-amplitude IGWs during the past five decades (e.g., Brunk 1949; Tepper 1954; Ferguson 1967; Bosart and Cussen 1973; Young and Richards 1973; Bosart and Sanders 1986; Bosart and Seimon 1988; Koch and Golus 1988;Schneider 1990). These waves typically have wavelengths of 100–200 km, periods of 2–3 h, amplitudes of 3–6 hPa, and phase velocities of 15–35 m s−1. Investigations of the vertical structure of IGWs have been comparatively rare, owing to the spatial and temporal limitations of conventional radiosonde observations. However, persuasive evidence has accumulated through detailed analysis of satellite observations and more recent wind profiler and Doppler radar observations indicating that these waves span the troposphere and have phase tilts consistent with downward energy propagation (e.g., Pecnick and Young 1984; Ralph et al. 1993;Ramamurthy et al. 1993). The pioneering numerical investigation of the large-amplitude IGW of 15 December 1987 over the Midwest (e.g., Schneider 1990) by Powers and Reed (1993) confirmed the tropospheric deep IGW structure and helped to establish the importance of latent heat release associated with convection to IGW maintenance. A further numerical simulation analysis of this case by Powers (1997) confirms that convection occurring above a frontal inversion can force IGWs in the given model.

As demonstrated by Uccellini and Koch (1987), a common attribute of IGWs is that they tend to occur 1) poleward of surface frontal boundaries, 2) near the inflection point between an upper-level trough and downstream ridge, and 3) immediately downstream of an upper-tropospheric jet streak. Wave genesis mechanisms have focused on shearing instability and/or geostrophic adjustment processes, but definitive evidence for the direct importance of either mechanism has been elusive. Also at issue in many of the above papers is whether the IGWs are triggered by preexisting convection or whether the waves themselves help to initiate convection. As to wave maintenance mechanisms, Uccellini and Koch (1987) also noted that the majority of IGWs in their sample of earlier cases had phase velocities in agreement with the Lindzen and Tung (1976) ducted gravity wave model. We will show that 1) multiple small-amplitude IGW development occurs in an environment favorable for shearing instability and unbalanced flow; 2) significant IGW development occurs on the downstream edge of a packet of a small-amplitude IGWs; and 3) rapid IGW amplification occurs where the low-level stable layer (wave duct) is especially prominent, deep convection is absent, and vigorous ascent penetrates and perturbs the duct.

Following the introduction we present in section 2 a description of our data sources and research methodology. A detailed observational synopsis is given in section 3 and is followed in section 4 by a description of the IGW life cycle. IGW origin and maintenance is discussed in section 5 while section 6 contains a presentation of IGW genesis, organization, and amplification. The concluding discussion appears in section 7. A map locating the key places described in the text and the orientation of cross-section lines appearing in section 6 is presented in Fig. 2 for reference.

2. Data and methodology

Data sources consisted of all available conventional surface and upper-air observations, satellite imagery, and microbarograms. Appeal to various NWS offices in the region uncovered additional digitized pressure, wind, and temperature observations. A poststorm e-mail solicitation produced another batch of digitized observations and microbarogram records from individual observers at stations all along the Atlantic coast. Gridded initialized analysis and forecast fields from NMC’s RAFS and 80-km Eta Model were utilized to diagnose the synoptic-scale forcing associated with the IGW event. The dynamical tropopause (DT) viewpoint of Hoskins et al. (1985), based upon the conservation of potential vorticity (PV), is employed to describe the synoptic environment [e.g., Bosart and Lackmann (1995); Hakim et al. (1995); Hakim et al. (1996); Lefevre and Nielsen-Gammon (1995); Nielsen-Gammon (1995); Nielsen-Gammon and Lefevre (1996); Davis et al. (1996); Bosart et al. (1996)]. gempak (Koch et al. 1983) is employed to process and display analyses and model fields.

3. Synopsis

a. Synoptic overview

Large-scale Eta Model–initialized 1000-hPa heights and 1000–500-hPa thickness analyses for 0000, 0600, 1200, and 1800 UTC 4 January 1994 (hereafter we will use a UTC/day convention to represent times and dates) are shown in Fig. 3 (The Richardson number patterns are discussed in section 6). The corresponding Eta Model–initialized 500-hPa height, absolute vorticity, and vertical motion fields (ascent only) are shown in Fig. 4. For additional synoptic-scale perspective, we present in Fig. 5 maps of Eta Model–initialized potential temperature θT, pressure PT, and winds on the DT [defined by the 1.5 PV unit surface (PVU) where 1 PVU = 10−6 K m2 kg−1 s−1] together with the 850-hPa equivalent potential temperature θe, heights, and winds for 1200/3, 0000/4, and 1200/4.

At 1200/3 a surface cyclone with a central pressure of 1006 hPa (approximately 45-m 1000-hPa height) is situated near the southern Alabama–Mississippi border (not shown), downstream of a θT (PT) minimum (maximum) marking the location of a strong disturbance on the DT (Figs. 5a,b). Advection of lower (higher) θT (PT) is indicated over the low-level cyclone center (see also the 850-hPa map in Fig. 5c), consistent with expected ascent (not shown). A weak predecessor warm front wave cyclone (∼1010 hPa) is also indicated just east of the Georgia–South Carolina border near the western edge of the Gulf Stream (not shown). This frontal wave weakens rapidly upward and is marked by a northwest–southeast axis of cyclonic vorticity (estimated) as seen on the 850-hPa map for 1200/3 (Fig. 5c).

By 0000/4 the primary sea level cyclone, now situated over northern Georgia (Fig. 3a), has deepened to 998 hPa (approximately −15-m 1000-hPa height) while the offshore frontal wave has moved northeastward and is marked by a broad trough that extends poleward offshore to almost 38°N. The observed 8-hPa cyclone deepening is supported by the appreciable lower (higher) θT (PT) advection over the low-level cyclone center in advance of the lowered DT marking the upshear cold (θ sense) trough (Figs. 5d,e). The tightening of the gradient of θT and PT in the southwesterly flow ahead of the upstream trough is manifest in a strengthening of the subtropical jet (STJ) with peak wind speeds of about 75 m s−1. Alternatively, the maximum 500-hPa ascent is found over northern Georgia in a region of cyclonic vorticity advection by the geostrophic wind and the 1000–500-hPa thermal wind (compare Figs. 3a and 4a). Likewise, the offshore frontal wave is well defined by an estimated maximum of warm air advection at 850 hPa (Fig. 5f). Note also the well-defined confluent polar front jet (PFJ) entrance region situated poleward of New England where the θT and PT contours come together (Figs. 5d,e). This overall DT flow configuration is in broad agreement with the interactive PFJ and STJ model described by Uccellini and Kocin (1987) as favorable for heavy snow events over eastern North America.

By 0600/4 (Fig. 3b) the sea level cyclone has deepened further to 994 hPa (approximately −45-m 1000-hPa height) and is now located over northwestern South Carolina. Meanwhile, the cold front has swept offshore equatorward of 34°N (as defined by the surge of westerly geostrophic flow), accompanied by a diminishing line of thunderstorms. Given that sea level pressures have continued to fall offshore as the frontal wave moves northeastward, a strong onshore easterly geostrophic flow is now indicated from North Carolina poleward and extending inland as far as the Ohio Valley. The dominant IGW originates after this time in the cold air poleward of the frontal boundary over southeastern Virginia (see Fig. 1; discussed more fully in section 4). The 500-hPa trough, now negatively tilted, is marked by an absolute vorticity maximum (>30 × 10−5 s−1) centered near the Georgia–South Carolina coastal border (Fig. 4b). The strongest 500-hPa ascent is situated over West Virginia at the leading edge of the area of cyclonic vorticity advection by the geostrophic and thermal winds (compare Figs. 3b and 4b). A weaker 500-hPa ascent center marks the position of the offshore frontal wave.

Further cyclone deepening to about 988 hPa (approximately −90-m 1000-hPa height) is indicated by 1200/4 (Fig. 3c) as the center edges into southeastern Virginia and the Delmarva Peninsula while the frontal wave elongates northeastward toward Long Island and southern New England. The 500-hPa absolute vorticity center is now defined by a remarkable 42 × 10−5 s−1 maximum in southeastern Virginia. Intensification of the 500-hPa trough has occurred in conjunction with a shortening of the distance between the trough axis and the downstream ridge axis (downstream half wavelength) and an increase in the coastal winds to almost 70 m s−1 (Figs. 4c, 5g,h). The maximum 500-hPa ascent (∼−20 × 10−3 hPa s−1) is situated near the common border of Pennsylvania, West Virginia, and Maryland (Fig. 4c), beneath a region of appreciable advection of lower (higher) θT (PT) on the DT (Figs. 5g,h) and at the southwestern end of an area of significant 850-hPa warm-air advection wrapping westward around the cyclone center (Fig. 5i). Very heavy snow is taking place in this region (discussed more fully in section 3c). Weaker, but still appreciable, 500-hPa ascent defines the frontal wave to the southeast of New England.

The shortening of the 500-hPa downstream half wavelength, a process under way at 1200/4 and very evident by 1800/4 (compare Figs. 4c,d), also occurs as the θT (320–350 K) and PT (300–200 hPa) contours are displaced to the left (toward the north and west) relative to the west–southwest background flow in the region from the eastern Great Lakes to New England in the 12 h ending 1200/4 (Figs. 5g,h). This contour displacement occurs above the area of appreciable 850-hPa warm-air advection on the poleward side of the primary surface cyclone and predecessor offshore frontal wave where there is significant precipitation and associated latent heat release (Fig. 10). The resulting increase in the gradient of θT and PT across southeastern Ontario and Quebec helps to define a strengthening PFJ in which wind speeds now exceed 90 m s−1. This occurs in response to nonconservative processes as simple horizontal advection of θT and PT by the observed winds at 0000/4 cannot explain the observed distribution of θT and PT at 1200/4 in this region as judged by the appearance of a closed θ = 350 K contour over eastern New York and Vermont at 1200/4 whereas no such θT contour value existed anywhere upstream at 0000/4 (compare Figs. 5e,h).

Between 1200/4 and 1800/4 the cyclone fills slightly to 989 hPa as it moves toward the New York City region (Fig. 3d) while the IGW amplifies (see section 6) as it accelerates northeastward across New England to a position over northeastern Maine (Fig. 1). The prominent 500-hPa ascent center is now situated from northeastern New York to northwestern Maine along the anticyclonic shear side of the now well-defined PFJ entrance region (Fig. 4d). A secondary 500-hPa ascent center south of Nova Scotia also defines the area of prominent 850-hPa warm-air advection associated with the frontal wave. Comparison of Fig. 4d with Fig. 1 shows that the IGW trough over northeastern Maine at 1800/4 lies close to the intensifying downstream short-wave ridge. Broadly speaking, the configuration and evolution of the synoptic-scale flow satisfies the three criteria mentioned in the introduction as identified by Uccellini and Koch (1987) as common to a number of large-amplitude IGW events reported in the literature.

b. IGW overview

As an example of the remarkable IGW structure associated with this event, we present in Fig. 6 the digitized surface pressure trace obtained from the Air Force Phillips Laboratories at Hanscom Field in Bedford (BED), Massachusetts (A. Jackson 1994, personal communication). The IGW is superimposed on a broad synoptic-scale pressure fall (Fig. 6a). It is manifest as a wave of elevation around 1400/4 and is followed by a pronounced wave of depression concentrated near 1500/4 and additional smaller-scale pressure oscillations that cease after 2000/4. The 1400–1600/4 expansion of the BED surface pressure trace (Fig. 6b) reveals a pressure plunge of 9 hPa in 8 min (impressive by tropical storm standards) beginning prior to 1500/4 with a maximum observed fall of 1.3 hPa min−1 at 1500/4!

The weather and anemometer-level wind sequence from nearby Boston (BOS), Massachusetts, shown in Fig. 7a reveals a classic large-amplitude IGW signature. Heavy precipitation and the onset of very gusty winds occurs at BOS as the wave crest approaches. The high wind speeds continue (peak winds are about 30 m s−1) as the precipitation intensity lessens with the approach of the IGW trough axis. The overall period of strong and gusty wind speeds in excess of 15 m s−1 is approximately 30 min. We also show the anemometer-level wind trace from Providence (PVD), Rhode Island, in Fig. 7b. The peak wind speeds at PVD occur about 30 min earlier than at BOS and are weaker. Likewise the character of the BOS and PVD wind traces differ considerably with BOS sustaining a 30-min period of quasi-constant wind speeds approximately 15–20 m s−1 while PVD reports a weaker and more sharply peaked wind speed maximum.

This continuous evolution of the IGW in time and space as it propagates is also evident in plots of selected microbarogram traces for stations over the northeastern United States shown in Fig. 8. Note the highly variable nature of the IGW-related pressure oscillations superimposed on a general synoptic-scale pressure fall pattern. The largest IGW-induced pressure changes (defined as the wave crest to wave trough pressure change) are generally confined to east of the Appalachians and increase from southwest to northeast poleward of 35°N. In much of southeastern New England (e.g., BOS and PVD; recall Fig. 7) the IGW-induced pressure perturbation is characterized by a steep fall and a rapid recovery (primary wave). It is followed by additional multiple smaller-amplitude pressure oscillations that are difficult to see in Fig. 8 but are readily apparent on the BED trace shown in Fig. 6a.

c. Precipitation signatures

Comparison of Figs. 3–5 shows good temporal and spatial continuity for ascent associated with the offshore warm-frontal wave and the area of rapidly shortening wavelength between the upper-level trough and the developing downstream ridge where differential cyclonic vorticity advection is especially strong. These twin ascent regions, along with the dominant IGW, played an instrumental role in controlling the spatial and temporal mesoscale precipitation distribution. Although the operational NCEP Eta and RAFS Models correctly indicated that a major precipitation event would take place over much of the eastern United States with this storm, neither model captured the spatial and temporal precipitation variation associated with the twin ascent regions.

We present in Fig. 9 the NCEP manually prepared 24-h quantitative precipitation forecasts (QPF) verifying 1200/4 (Fig. 9a) and 1200/5 (Fig. 9b). The 24-h forecast QPF update verifying 1200/4 is shown in Fig. 9c. Forecasters at NCEP have the responsibility of providing national heavy precipitation guidance using all available numerical and statistical guidance and a full suite of observational products. As noted by Olson et al. (1995), forecasters have been able to maintain a competitive advantage over model QPF in the face of increasingly skillful operational models. According to Fig. 9a, forecasters expected the heaviest precipitation to fall from the Appalachians eastward to the middle Atlantic coast in the 24 h ending 1200/4, consistent with climatological expectations for a coastal storm. The 24-h update (Fig. 9c) continued this trend while the 24-h forecast verifying 1200/5 indicated the heavy precipitation would spread northeastward into New York and New England.

For comparison we display the observed precipitation in 6-h increments for the period 0000–1800/4 in Fig. 10. Although there is broad agreement between Figs. 9 and 10, a closer inspection reveals that substantial precipitation spreads more rapidly than forecast into coastal New England in association with the warm-frontal wave (Fig. 3) and northeastward along and to the west of the Appalachians ahead of the vorticity center aloft. Similarly, precipitation is less than forecast east of the Appalachians in the middle Atlantic states. These discrepancies illustrate the subtle but important effects of subsynoptic-scale disturbances embedded within the larger-scale cyclonic circulation.

Neither the Eta nor the RAFS Model forecasts fully captured the intensity of the offshore frontal wave (not shown). Precipitation was enhanced ahead of this disturbance in the area of low-level warm air advection well downstream of the trough aloft (Fig. 3). Similarly, precipitation was suppressed behind and inland from the track of the offshore frontal wave, partly accounting for the unexpected reduction of precipitation in the middle Atlantic region. A second reason for the unexpected precipitation reduction in this region was the pronounced drying behind the organizing IGW as it traversed the region between 0600 and 1000/4 (Fig. 1; section 4b).

However, as is evident from a comparison of Figs. 9 and 10, the axis of heaviest observed precipitation from West Virginia to central New York averages 100–200-km farther west than forecast. Given the inland track of the 500-hPa vorticity center (Fig. 4), and given the rapid decrease of the downstream half wavelength and developing PFJ entrance region across southern Canada (Fig. 5), deep ascent and low-level warm air advection was able to extend well inland to west of the Appalachians (Figs. 3–4). In the Pittsburgh (PIT), Pennsylvania, area upward of 40–50 cm of snow fell between 0600 and 1800/4 with the bulk of the snow falling in a few hours during the morning rush hour. Air carrier US Airways, with a major hub at PIT, was forced to suspend flight operations unexpectedly, leading to lengthy system-wide air travel disruptions.

The highly concentrated nature of the intense snowfall, which we term a “snow bomb,” is demonstrated in Fig. 11 by peak hourly snowfall rates of 10–15 cm at Elkins (EKN), West Virginia, Williamsport (IPT), Pennsylvania, and Syracuse (SYR), New York. Note also the approximately 2-hPa positive pressure perturbation with the heavy snow at EKN and IPT (and to a much lesser extent at SYR). These positive pressure perturbations, superimposed upon a broad synoptic-scale pressure minimum at EKN and IPT, are reminiscent of the positive “bubble high” pressure perturbations resulting from evaporative cooling associated with a warm season mesoscale convective system (MCS). Given the observed widespread large-scale quasi-saturation in the heavy snow region (not shown), it is unlikely that evaporative cooling can account for the observed positive pressure perturbations. The presence of vigorous midlevel ascent over the heavy snow region (Fig. 4c) raises the possibility, however, that the warm frontal inversion aloft (Fig. 3c) may have been perturbed sufficiently to briefly increase the depth of cold air in the boundary layer, allowing for the temporary local pressure increases.

4. IGW life cycles

a. Initiation

Multiple low-amplitude (<1 hPa; not shown) IGWs, oriented northwest–southeast, are first evident over northeastern Georgia and southwestern South Carolina near 0000/4 (recall Fig. 1). Infrared satellite imagery for 0000, 0300, 0600/4 (Figs. 12a–c) shows that the antecedent IGWs can be associated with organized, persistent bands of higher and colder cloud tops that extend from the central and southern Appalachians southeastward toward the coast near the tip of the developing dry slot (clearer imagery can be accessed via http://www.atmos.albany.edu/facstaff/bosart/jan94/images.html). It is not possible to conclusively state whether the dominant IGW forms in situ or grows from a predecessor weaker (<1 hPa) disturbance. This dominant IGW is first detectable as a coherent feature in extreme northeastern North Carolina at 0700/4 (Fig. 1). A regional surface map for 0600/4 is shown in Fig. 13 to better document the antecedent conditions. The dominant IGW organizes in a region of cold air damming (Forbes et al. 1987; Bell and Bosart 1988) west of a nearshore frontal boundary and poleward of an occluded front stretching eastward from the primary cyclone center over South Carolina. Comparison of Fig. 13 with Fig. 4b shows that IGW genesis and/or amplification occurs immediately downstream of an inflection point in the strong 500-hPa flow and along the poleward margin of an area of significant ascent. Passage of a weak“zipper low” (Keshishian and Bosart 1987) frontal wave offshore may have also contributed to the reinforcement of the cold air onshore as judged by a 3.2°C (18.1°C to 14.9°C) temperature decrease at buoy 44014 (36.6°N, 74.8°W) in the 3 h ending 0800/4 as the sea level pressure bottomed out at 990.5 hPa there. Given the 10°C temperature contrast in less than 100 km between buoy 44014 and Cape Henry Light Vessel (CHLV2) and the lower-tropospheric warm-air advection over the region (Fig. 3b), it is clear that a significant frontal inversion, and therefore an IGW duct must be present over southeastern Virginia at 0600/4.

A comparison of the 0000/4 and 0600/4 soundings from Greensboro (GSO), North Carolina (Fig. 14), confirms that GSO is situated on the equatorward margin of a cold air damming region as deduced from the layer of cold, stable air below about 870 hPa containing easterly flow. This surface-based cold-air layer (wave duct) appears to be too thin to support a long-lived large-amplitude ducted IGW through 0600/4 [Lindzen and Tung (1976); section 5] but may still permit short-lived small-amplitude ducted IGWs. Based upon the 20–40 m s−1 southerly winds in the 850–500-hPa layer over GSO at 0600/4, the observed northwest–southeast orientation of the IGW in Fig. 1, and the Eta Model 6-h forecast valid 0600/4 a critical level is present in the 700–500-hPa layer above the duct (not shown), a condition shown by Lindzen and Tung (1976) to be favorable for ducted IGWs. This duct strengthens by 0600/4 in response to 3°–5°C warming near 850 hPa and cooling of the layer below 900 hPa to below 0°C. Given that the 0000/4 soundings for Cape Hatteras (HAT; not shown), North Carolina, and Athens (AHN), Georgia, depict shallow wave ducts (∼75 hPa; Fig. 14), and given that HAT lies very close to the coastal frontal boundary while AHN is near the cyclone center (Fig. 3a), we conclude that favorable wave ducts for IGW propagation are likely to be found poleward of the Carolinas and east of the Appalachians in the cold air damming region indicated in Fig. 13.

A map of cloud-to-ground (CG) lightning flashes obtained from the National Lightning Detection Network for the 0000–1200/4 period is shown in Fig. 15. Comparison of Fig. 15 with Figs. 3a,b and Fig. 13 shows that CG lightning activity near the primary cyclone center over Georgia at 0000/4 rapidly dissipates in northwestern South Carolina by 0600/4. Inspection of the 0600/4 GSO sounding shown in Fig. 14 confirms the existence of a weak conditionally unstable layer between 850 and 500 hPa immediately downstream of the decaying CG activity over South Carolina. A second significant area of CG lightning activity is seen in the offshore waters east of the Carolinas. This CG activity lies near and equatorward of the intersection of the cold front sweeping offshore with the quasi-stationary frontal boundary east of the Carolinas (Figs. 3a,b and Fig. 13) and rapidly weakens after 0600/4. A third concentration of CG lightning activity well east of Florida and extending to near Cuba is associated with an active prefrontal squall line. The squall line is readily apparent as a meridionally oriented band of high cold cloud tops near and west of 75°W in the satellite infrared imagery for 0000/4 shown in Fig. 12a. By 0600/4 this convective cloud band becomes better organized equatorward of 35°N (Fig. 12c). It also broadens poleward and merges with the main high cold cloud shield that extends eastward from New England (Fig. 3b).

Also of interest at 0600/4 (Fig. 12c) is the well-defined western edge of a secondary band of somewhat lower high cold clouds that extends equatorward from extreme south-central Pennsylvania across east-central Virginia and northeastern North Carolina where the band crosses the coast just northeast of Wilmington (ILM). The western edge of this band lies along an Asheville (AVL), North Carolina, to Savannah (SAV), Georgia, line at 0000/4 (Fig. 12a) and a Parkersberg (PKB), West Virginia, to Myrtle Beach (MYR), South Carolina, line at 0300/4 (Fig. 12b). The area of CG lightning activity that decays east of North Carolina by 0600/4 appears to be associated with the back edge of this secondary high cold cloud band. Likewise, the initiation of the dominant IGW over extreme northeastern North Carolina and extreme southeastern Virginia in the 0700–0800/4 period (Fig. 1) is occurring close to the back edge of this same secondary high cold cloud shield and downstream of the developing dry slot that contains multiple low-amplitude IGWs (Figs. 1 and 12a,b). Deep convection is absent near the organizing IGW subsequent to 0600/4 as measured by the near cessation of CG lightning activity and the dearth of convective signatures in the satellite imagery (Figs. 12 and 15).

b. Organization

By 1200/4 the IGW has organized sufficiently to be readily detectable to the northeast of a 988-hPa cyclone now located near Richmond (RIC), Virginia (Fig. 16). The IGW is oriented along a northwest–southeast line from northeastern Pennsylvania to northern New Jersey to the southwest of New York City and then offshore. The downstream environment, as measured by the 1200/4 soundings from Albany (ALB), New York, and Chatham (CHH), Massachusetts, is very favorable for IGW amplification and propagation (Fig. 17a). The ALB and CHH 1200/4 soundings exhibit deeper and stronger ducts and deeper capping quasi-moist adiabatic layers above the duct in comparison to the 0600/4 GSO sounding (Fig. 14). This duct is especially strong inland from the Maine coast [as seen in the 1200/4 sounding from Portland (PWM)] where a strong coastal front in the western Gulf of Maine, as measured by an estimated surface temperature gradient of approximately 10°C (200 km)−1, exists. [Weaker, secondary IGWs with similar orientation lie farther to the southwest of the dominant IGW. An especially long-lived secondary IGW (Fig. 16) can be associated with the narrow band of higher, colder clouds in southeast Virginia and extreme northeastern North Carolina at 0900/4 (Fig. 12d). Its passage is also seen, for example, in the barogram traces (Fig. 8) from RIC, GSO, Rocky Mount (RWI), Cherry Point (NKT), and Wilmington (ILM), North Carolina, and Columbia (CAE), South Carolina.]

Strong and gusty northeasterly winds are apparent immediately in advance of the dominant IGW pressure minimum axis in Fig. 16. Precipitation was absent across portions of southern and central New Jersey and southeastern Pennsylvania immediately behind the dominant IGW pressure minimum axis, another characteristic signature of large-amplitude IGWs. This precipitation “hole” is first readily apparent by 0900/4 across much of northern Virginia, southern Maryland, and the Chesapeake Bay as seen in the WSR-88D base reflectivity map for 0908/4 from Sterling (KWBC), Virginia, shown in Fig. 18. The precipitation hole has a general northwest-southeast orientation, similar to the orientation of the dominant IGW, and is 1–3 h “wide” on the basis of simple time-to-space conversion estimates. Although the precipitation hole contains a few embedded patches of light-to-moderate precipitation, its presence, along with the passage of the offshore frontal wave, contributes to the unexpectedly light precipitation amounts in the middle Atlantic region (Figs. 9 and 10).

Widespread precipitation across much of New England with observed 3-h pressure falls in excess of 6 hPa, occurring at the time of normal diurnal pressure rises, attests to the unexpectedly severe impact of the stronger-than-forecast offshore warm-frontal wave (recall Figs. 9 and 10) at 1200/4 in this area. Sustained 5–10 m s−1 low-level north-northeasterly winds across interior New England to the northwest of the offshore frontal wave continue to be associated with near-surface cold-air advection, further contributing to the duct strengthening in advance of the organizing IGW. This duct is further reinforced by weak offshore coastal frontogenesis where oceanic sensible and latent heat fluxes likely contribute to the generation of a 10°C (100 km)−1 thermal gradient in the coastal region of eastern New England. The effect of the frontogenesis process is to enable air warmed and moistened by contact with the sea surface to be advected westward over the onshore cold dome (wave duct), thereby increasing the strength of the wave duct. The northwest–southeast-oriented mesoscale pressure ridge located behind the dominant IGW pressure minimum axis marks a region of resumed precipitation in Virginia and exceptionally heavy snow in parts of West Virginia and extreme southwestern Pennsylvania where the snow bomb is in progress at 1200/4.

Infrared satellite imagery for 0900, 1000, and 1200/4 shown in Figs. 12d and 19a,b illustrates the dominant IGW organization. The meridionally oriented back edge of the main high cold cloud shield lies across eastern New England at 0900–1000/4 (Figs. 12d, 19a). Comparison with Fig. 12c shows that the back edge of this cloud shield has moved eastward 100–200 km in the 3–4 h ending 1000/4. Noteworthy is the cooling of the high cloud tops across eastern Pennsylvania, eastern New York, and New Jersey at the same time as the back edge of the secondary cold high cloud shield becomes better defined farther west. Comparison with Fig. 1 and the 1200/4 surface map (Fig. 16) shows that the organizing IGW at 1000/4 is situated approximately 100 km downstream (to the northeast) of the back edge of the secondary high, cold cloud shield and near the strengthening cloud-top temperature gradient across southeastern Pennsylvania and extreme southern New Jersey. Cloud-top cooling in the secondary high cloud shield continues and by 1200/4 (Fig. 19b; see arrow) a distinctive ellipse-shaped area of colder cloud tops can be seen in this cloud shield extending from northeast of SYR to just south of New York City (LGA) and offshore.

Further comparison of Figs. 16 and 19b discloses that the dominant IGW is organizing within about 100 km of the upshear back edge of the now very well defined northwest–southeast-oriented region of high cold cloud tops by 1200/4. This new axis of developing cold high cloud tops continues to remain oriented quasi-parallel to, and downstream of, the IGW pressure minimum axis. This new cold cloud top growth is occurring to the west of the primary warm conveyor belt cloudiness and strong 500-hPa ascent associated with the offshore frontal wave (Fig. 4c), approximately 500-km poleward of the cyclone dry slot, and to the east of cold cloud tops appearing in southwestern Pennsylvania just downshear of the vigorous 500-hPa ascent center marking the Appalachian snow bomb. Inspection of Fig. 4c shows that the Eta Model initialization is unable to capture any distinctive mesoscale vertical motion couplet in the organizing IGW wave environment at 1200/4, a failure likely attributable to the model analysis and initialization process.

c. Amplification and intensification

A regional surface mesoanalysis map for 1400/4 appears in Fig. 20. Although Fig. 1 indicates that the dominant IGW pressure minimum axis lies near an ALB–PVD line, Fig. 20 shows ample evidence of nonlinear and nonuniform structure along the IGW axis. Considerable alongwave and crosswave structure exists in the wind and sea level pressure fields across southern New England. Heavy wind-blown snow and/or ice pellets are reported from many locations [for example, Windsor Locks (BDL), Connecticut] immediately prior to the passage of the IGW pressure minimum. By 1400/4 a narrow zone of very cold cloud tops has formed as marked by the lighter band oriented along an approximate Massena (MSS), New York, to New London (GON), Connecticut, line, in the infrared satellite imagery (Fig. 19c). This narrow band of high cold cloud tops lies somewhat downstream (∼25–50 km) of the IGW pressure minimum (Fig. 20), and is interpreted to represent the effect of deep tropospheric ascent associated with the advancing IGW. Meanwhile, the precipitation hole has expanded into southwestern Connecticut and across Long Island underneath the trailing back edge of the now prominent secondary cold high cloud deck. The snow bomb region, marked by a mesoscale pressure ridge over southwestern Pennsylvania, is manifest by a secondary area of expanding high cold cloud tops downstream (Fig. 19c).

The relationship of the IGW-modified pressure and precipitation patterns is explored in more detail in Figs. 21 and 22. Presented are base reflectivity composites (beginning contour is 15 dBZ) derived from the WSR-88D radar sites in KENX, KTAW, and KOKX along with 20-min pressure change maps calculated using time-to-space conversion techniques applied to the microbarograms (and others) displayed in Fig. 8. The precipitation hole noted at 0908/4 in Fig. 18 can be seen passing from southwest to northeast across New Jersey, New York, and New England from 1130/4 to 1450/4. The trailing back edge of the precipitation shield is approaching central New Jersey and eastern Pennsylvania at 1130/4 (Fig. 21a), is crossing the New York City metropolitan region at 1210/4 (Fig. 21b), and reaches into southwestern Connecticut and southern New York at 1250/4 (Fig. 21c). The large base reflectivity gradient marking the trailing edge of the primary precipitation shield remains closely collocated with the axis of maximum IGW-induced 20-min pressure falls at these times (Figs. 21d–f). Note also the continuing IGW amplification as a northwest–southeast-oriented band of 20-min pressure falls in excess of 8 hPa becomes evident by 1250/4 (Fig. 21f). Although a linear structure is readily apparent in the pressure and base reflectivity perturbations to first order in Fig. 21, it is also clear that important alongband pressure and reflectivity variations exist.

Relative pressure rises in advance of the IGW pressure minimum are generally less than 1.5 hPa and are best defined at 1250/4 (Fig. 21f), 1330/4 (Fig. 22d), and 1410/4 (Fig. 22e). At 1330/4 (Fig. 22d) two such pressure rise centers are located in west-central Massachusetts and southern Rhode Island immediately to the northeast of the two prominent fall centers (greater than 8 hPa in 20 min) associated with the dominant IGW. Pressure rises behind the dominant IGW pressure minimum are more robust and average about 4 hPa in 20 min and tend to increase in time as denoted by the area of 20-min pressure rises greater than 8 hPa observed near the Rhode Island–Massachusetts border at 1450/4 (Fig. 22f). The increasingly distorted and two-dimensional nature of the IGW-induced pressure perturbations across southern New England becomes quite apparent by 1410/4 (Fig. 22e) and 1450/4 (Fig. 22f) based upon the orientation of the pressure rise and fall axes. As evidence, note that across southeastern Connecticut, Rhode Island, and southeastern Massachusetts IGW-induced pressure falls less than −8 hPa (20 min)−1 are followed by nearly comparable 20-min pressure rises between 1330 and 1450/4. A similar situation is apparent across northwestern Massachusetts, adjacent New York, and southern Vermont. Over north-central Connecticut and south-central Massachusetts (along a CEF–BDL line), however, a much more modest pressure recovery subsequent to the initial pressure fall is observed.

The WSR-88D base reflectivity patterns at 1330/4 (Fig. 22a), 1410/4 (Fig. 22b), and 1450/4 (Fig. 22c) also illustrate the existence of important two-dimensional IGW structural variations. Noteworthy are the “reflectivity indentations” created by a more rapid northeastward movement of the 15-dBZ contour marking the back edge of the main precipitation shield as seen, for example, across northern Massachusetts and Cape Cod as compared to Rhode Island at 1410/4 (Fig. 22b). This two-dimensional structure is made more explicit in Fig. 23, which maps 20-min isochrones showing the position of the IGW-triggered precipitation shutoff (15-dBZ contour) during the 1110–1530/4 period. A slow counterclockwise rotation of the isochrones is observed, suggestive that the IGW propagates northeastward faster across southern New England than it does in the Hudson Valley region of New York. Small-scale variations along individual isochrones appear to have some time continuity and may indicate evidence for internal IGW structure and/or the effect of IGW–terrain interactions. The tendency for isochrone spacing to increase slightly from southwest to northeast in Fig. 23 is consistent with the broad-brush IGW acceleration indicated in Fig. 1. Although there is no obvious precipitation signature associated with the secondary IGW (compare Figs. 16 and 20), precipitation remains suppressed at what should be the height of the storm until its passage. Spotty and somewhat disorganized precipitation resumes after 1410/4 (Fig. 22b) in association with the northeast movement of a modest region of 20-min pressure rises (∼0.5 hPa) following the secondary IGW.

To close this section we present in Fig. 24 a vertical time series of smoothed hourly wind observations from the wind profiler located in Bloomfield (BMF), Connecticut. Well prior to the passage of the dominant IGW pressure minimum at BMF (about 1330/4 according to Fig. 23) a band of strong easterly winds is observed in the 0.5–1.8-km layer and well within the wave duct region. Easterly wind speeds begin to exceed 20 m s−1 after 0400/4 and reach 30 m s−1 near 1300/4 immediately preceding the passage of the dominant IGW pressure minimum. The duct layer (top near 2500 and 1500 m, respectively, at ALB and CHH at 1200/4 from Fig. 17a) is characterized by veering winds above it, indicative of the deep warm-air advection ahead of the primary cyclone discussed in conjunction with Fig. 3. Note also the strengthened southeasterly flow in the 2–3-km layer centered near 1000–1100/4 that accompanies the gradual lowering of the region of southwesterly flow as the warm front approaches. This strengthened southeasterly flow above the wave duct may be one possible indication of deep flow acceleration toward the approaching IGW trough. A similar BMF time series of signal power (not shown) illustrates the well-defined back edge to the precipitation shield associated with the passage of the dominant IGW and is in excellent agreement with Fig. 23.

d. Maturity

Subsequently, the IGW accelerates northeastward to a speed of about 35 m s−1 and by 1600/4 is situated along a northern Vermont–extreme southwestern Maine line (Fig. 25; absence of observations offshore does not permit an analysis in the Gulf of Maine at this time). Satellite imagery for 1500/4 (not shown) and 1600/4 (Fig. 19d) suggests the existence of a narrow northwest–southeast-oriented band of warmer cloud tops propagating northeastward approximately 120 km upshear of a similarly oriented band of high cold cloud tops. At 1600/4 this warm cloud-top band stretches from extreme southern Vermont across Massachusetts and is situated about 60 km upshear of the dominant IGW axis of minimum sea level pressure. This observed relationship between the warm and cold cloud-top bands and the IGW pressure minimum axis indicates the existence of a tropospheric-deep upshear-tilting wave. Sustained surface northeasterly wind speeds are now commonly in the 15 m s−1 range in the region of rapid pressure falls ahead of the dominant IGW pressure minimum. A trailing bubble high can be found over northeastern Massachusetts. The snow bomb region over central Pennsylvania and New York continues to be associated with a slight surface pressure ridge behind a weak secondary IGW.

By 1800/4 the now mature IGW accelerates to a forward speed of approximately 40 m s−1 over northeastern Maine (Fig. 26). This acceleration is consistent with the observed downstream increase in duct thickness and strength as judged by the 0000/5 sounding for Stephenville, Newfoundland, Canada (YJT), shown in Fig. 17b. The IGW pressure minimum that appears as a broad zone of lower pressure over Maine in Fig. 26 (confined between the two bold lines) is representative of the observed band of quasi-uniform pressure that follows the steep pressure plunge. The IGW is still manifest as a narrow band of high cold cloud tops (not shown). The trailing weaker secondary IGW, now better organized, lies from just south of Montreal (YUL), Quebec, Canada, to Mount Washington (MWN), New Hampshire, and to PWM. The snow bomb region and associated downstream high cold cloud area move northeastward (not shown) with the heaviest snow at 1800/4 falling over northeastern New York, extreme northwestern Vermont, and portions of adjacent Quebec. Comparison of Fig. 4d with Fig. 26 shows that the heavy snow region remains associated with the area of strong 500-hPa ascent situated between the equatorward PFJ entrance region and the poleward STJ exit region.

Finally, we present in Fig. 27 a replotted wind speed trace from the MWN observatory for 0000–2300/4 (the original traces were in English units). MWN summit wind speeds begin to pick up shortly before 1500/4 with many gusts reaching 35–40 m s−1 after 1515/4. A further abrupt wind speed increase is observed at 1615/4 and continues until shortly after 1700/4 with many gusts in the 45–50 m s−1 range (the mean wind speed in the hour ending 1700/4 averages 37.1 m s−1 from the east) with the approach of the dominant IGW (Fig. 25). After a lull in which the wind speeds slacken to 35–40 m s−1 the wind increases again to near 45 m s−1 around 1800/4 at which time a weaker secondary IGW is passing MWN (Fig. 27). Subsequent to 1800/4 the wind rapidly decreases to less than 20 m s−1, where it remains for the duration of the storm. Clearly, the summit of MWN at an elevation of 1800 m experiences IGW wind signatures that are felt at surrounding stations at much lower elevations.

The temperature trace (not shown) is also of interest because of the abrupt 3°–4°C cooling that is observed shortly before 1400/4 at which time the wind is starting to become more gusty. Given that the MWN observatory reports a ceiling and visibility of zero with light to moderate snow throughout the 1300–1500/4 period, the possibility exists that the cooling can be attributed to the uplift of the inversion (deepening of the cold air) with the approach of the dominant IGW. An approximate 1.5°C warming is observed in the 30 min ending about 1700/4 with the passage of the dominant IGW and is followed by a slight cooling and a further 2.5°C warming in a 30-min period ending shortly after 1800/4 as a weaker secondary IGW passes the MWN summit.

e. Dissipation

The dominant IGW moved across southeastern Canada at speeds in excess of 40 m s−1 subsequent to 1800/4. Infrared satellite imagery (not shown) shows that the narrow band of high cold cloud tops marking the approach of the IGW pressure minimum axis strengthens further and elongates northwest–southeast as the IGW reaches Newfoundland and coastal eastern Canada. There is no evidence for an IGW passage in the near coastal waters south of Nova Scotia. Note that while the 0000/5 sounding for Sable Island (WSA), Nova Scotia, Canada, exhibits a very shallow surface-based inversion, the absence of a wave duct in the warmer southwesterly flow aloft precludes significant IGW activity as is observed (Fig. 17b).

5. Origin and maintenance of IGW activity

We have seen (Figs. 1, 3–5, and 12–14) that the large-amplitude IGW event of 4 January 1994 occurs in a large-scale environment conducive to IGW evolution as identified by Uccellini and Koch (1987). Other studies of large-amplitude IGW events [e.g., Lin and Goff (1988); Bosart and Seimon (1988)] have stressed the likely importance of mesoscale circulations associated with deep convection to the wave amplification process. Although a compelling argument cannot be made that deep convection is crucial to IGW amplification in this case, the CG lightning flashes reported in Georgia and South Carolina in the 0000–0600/4 period (Fig. 14a) overlaps the genesis region of the observed small-amplitude IGWs. Subsequent to 0600/4 there are no reported CG lightning flashes within 500 km of the IGW amplification region over southeastern Virginia (Fig. 14b).

The unexpected availability of a 0600/4 sounding from GSO, situated immediately upshear of the region of IGW amplification, allows us to test for the existence of criteria that have previously been shown to be important to IGW organization, amplification, and propagation. Lindzen and Tung (1976) identified three criteria as crucial to IGW maintenance: 1) a duct (stable layer) of sufficient thickness in the lower troposphere; 2) a near-neutral layer in the middle and upper troposphere above the duct; and 3) a critical layer in the near-neutral layer of atmosphere above the duct. Inspection of Fig. 14 reveals that there is a modest stable layer below 820 hPa at GSO. According to Lindzen and Tung (1976) the thickness of the stable layer determines the phase velocity of the dominant wave that can be supported in the duct according to the following relationship:
i1520-0493-126-6-1497-e1
where D is the duct thickness, θ is potential temperature, and g is gravity. The subscripts t and b refer to the value of θ at the top and bottom of the duct, respectively, while the overbar denotes an average value of θ over D. Given that Lindzen and Tung (1976) showed that the duct must be sufficiently thick to accommodate one-quarter of the vertical wavelength of the wave traveling with the phase velocity supported by (1), an additional requirement is that
i1520-0493-126-6-1497-e2
In (2) N is the Brunt–Väisälä frequency, C is the observed phase speed of the wave, and U∗ is the component of the observed wind in the direction of the observed wave propagation. Note also that the values of N and U∗ that are used in (2) are computed as layer averages over D.

In order to evaluate (1) and (2) we adopt the parameter values given in Table 1 from the GSO 0600/4 soundings shown in Fig. 14. These estimates suggest that the observed duct is too thin over GSO at 0600/4 to support the existence of a long-lived large-amplitude ducted IGW. Given that the IGW is first detected approximately 250 km east-northeast of GSO near the time of the nominal sounding launch (∼0515/4), it is likely that the GSO sounding is representative of the IGW environment. Note, however, that a comparison of the 0600/4 and 1200/4 surface maps shown in Figs. 13 and 16, respectively, and the 1200/4 soundings from Sterling (IAD), Virginia, and Atlantic City (ACY), New Jersey (not shown), reveals that the duct thickness increases rapidly to the northeast in the cold-air damming region east of the Appalachians. This rapid increase in duct thickness and strength northeast of GSO should support a ducted IGW.

6. IGW genesis, organization, and amplification

In this section we discuss a number of issues related to the genesis, organization, and amplification of the dominant IGW. At 1200/4 the rapidly intensifying IGW lies over central New Jersey and vicinity (Figs. 8, 16, 19c) and is near the inflection point in the 300-hPa flow between a sharp upstream trough located over Virginia and North Carolina and a prominent downstream ridge over northeastern New York and northwestern New England. The relatively short wavelength between the trough and downstream ridge (∼500–600 km) implies the existence of large horizontal divergence (and associated vertical motion) in this region as air parcels accelerate from subgeostrophic to supergeostrophic speeds in a few hours. Although some authors [e.g., Zack and Kaplan (1987); Kaplan et al. (1997); Koch and O’Handley (1997)] define unbalanced flow based upon the magnitude of the Lagrangian Rossby number,
i1520-0493-126-6-1497-e3
it is recognized that balanced motions associated with flow curvature are contained in (3). The Rossby number RoL as given by (3) will be used to identify regions where large flow accelerations and associated vertical motions are possible. If these vertical motions are vigorous enough to perturb the lower-tropospheric stable layer then large surface pressure oscillations and IGW genesis may result. Van Tuyl and Young (1982) showed that IGW generation near jet streaks is possible when RoL > 0.5. Likewise, the magnitude of the right-hand side of the divergence equation,
i1520-0493-126-6-1497-e4
where D ≡ ∂u/∂x + ∂υ/∂y, φgz, Jxy(u, y) is the Jacobian of u and υ, ζk· × V, and β ≡ ∂f/∂y, will be used to assess the potential for unbalanced flow. Large parcel accelerations are likely when the right-hand side of (4) differs significantly (∼10−8 s−2) from 0.

Equations (3) and (4) are evaluated using the NCEP Eta Model–initialized fields for 0000, 0600, 1200 and 1800/4 at 50-hPa intervals from 1000 to 100 hPa. As an independent check on the validity of (3) and (4), and at the suggestion of a reviewer, we computed the unbalanced divergence defined to be the difference between the full Eta Model–initialized divergence and the balanced divergence computed from a quasigeostrophic (QG) model that includes diabatic heating from large-scale stratiform precipitation. (As noted by an anonymous reviewer this divergence difference would likely be smaller in magnitude and areal coverage if a higher-order balanced model had been used, but phase differences are expected to be much smaller.) Large parcel accelerations associated with geostrophic adjustment processes might be expected to occur in regions where the unbalanced divergence is large. We computed QG vertical motions following Tsou et al. (1987) with a stable latent heat release scheme based upon Krishnamurti and Moxim (1971) as modified by Vincent et al. (1977). In this scheme grid-scale precipitation is proportional to the vertical advection of saturation specific humidity and latent heat release is assumed to occur whenever there is ascent, horizontal moisture convergence, and the relative humidity exceeds 80%. The static stability is taken to be a function of pressure only and zero vertical motion is used on all boundaries. The balanced divergence is calculated by differentiating the QG omega solution with respect to pressure (using centered differences) and substituting this result into the isobaric form of the continuity equation.

Since the 80-km Eta grids were only available over North America, the NCEP–NCAR reanalysis gridded fields (Kalnay et al. 1996), interpolated to a 1° latitude–longitude grid using the Cressman (1959) analysis scheme, were used to define the hemispheric mass field. The 80-km Eta grids were interpolated from a polar stereographic projection to a 1° latitude–longitude grid using a bilinear interpolation scheme. These interpolated Eta grids, available only over North America and vicinity, were then inserted into the larger-domain NCEP–NCAR reanalysis prior to making the QG calculations. Prior to performing the analysis meld the Eta Model geopotential height and temperature fields were smoothed using the gempak nine-point smoother (Koch et al. 1983) to remove unrepresentative small-scale features.

Shown in Figs. 28a–d are maps of 400-hPa RoL and geopotential heights, 300-hPa ageostrophic winds, and 250-hPa isotachs valid for 0000–1800/4. Similarly, shown in Figs. 28e–h are maps of 400-hPa parcel divergence tendency (dD/dt) and geopotential heights and the unbalanced divergence (averaged over the 350–250-hPa layer) for the same time periods. The 400-hPa level is chosen because it is representative of conditions near the top of the duct downshear of the intense trough. The 350–250-hPa layer-averaged unbalanced divergence is representative of conditions below the jet. (Caution: The parcel divergence tendency signal at 0600/4 and 1800/4 is less robust than 0000/4 and 1200/4 since these times are strongly influenced by the 6-h model forecast.) In general larger values of RoL (>0.5) are confined to the deep trough, selected regions along the jet, and the developing shortwave ridge upshear of the IGW domain.

At 0000/4 the parcel divergence tendency (Fig. 28e) is significantly positive (∼10−8 s−2) over southern Georgia and northern Florida. This area of positive parcel divergence tendency is located upshear of the region where small-amplitude IGWs are observed to develop prior to 0600/4 (Figs. 1, 12a–c). The significant overlap of the unbalanced divergence and positive parcel divergence tendency regions confirms the likely existence of unbalanced flow in this area at 0000/4. By 0600/4 the region of positive parcel divergence tendency moves northeastward into South Carolina and weakens (Fig. 28f). Despite this apparent weakening (which may be an artifact of the 6-h forecast), the maximum positive parcel divergence tendency region remains situated upshear of where the dominant IGW will organize. This interpretation is reinforced by the meridionally oriented maximum in unbalanced divergence whose equatorward edge significantly overlaps the parcel divergence tendency maximum in South Carolina at 0600/4 (Fig. 28e). Note also that the maximum in unbalanced divergence at this time extends poleward across the southern Appalachians and roughly corresponds to the axis of heavy precipitation (snow bomb) (Fig. 10a). Finally, the local maximum in positive parcel divergence tendency in the eastern Ohio Valley at 0600/4 coincides with the developing shortwave ridge axis (Fig. 28f).

By 1200/4 (Fig. 28g) the now rapidly amplifying dominant IGW is located immediately downshear of a prominent region of positive parcel divergence tendency. This area of positive parcel divergence tendency remains significantly overlapped by an unbalanced divergence maximum. These regions of independently diagnosed unbalanced flow are located in an area where the aforementioned prominent short-wave ridge is developing, where the wavelength between the upshear trough and this ridge is shortening rapidly, and where RoL > 0.5, all of which are indicative of (implied) large parcel accelerations. It is also significant that the dominant IGW is located in a region where RoL > 1.0 between two strong jet streaks (Fig. 28c). As the dominant IGW accelerates across northern New England through 1800/4 (Fig. 28h) it continues to lie downshear of a region of diagnosed unbalanced flow. Note also that the snow bomb, a prominent mesoscale circulation itself, remains closely associated with the region of diagnosed unbalanced divergence at 1200–1800/4 (Figs. 28g,h).

Application of (1) and (2) to the ALB 1200/4 sounding (Fig. 17a) yields the duct parameter values given in Table 1 (values in parenthesis define a thinner and more conservative estimate of duct thickness based upon the most stable part of the sounding). Even with the use of more conservative estimates of duct thickness and strength and allowing for theoretical limitations, it seems reasonable to conclude that the atmosphere over inland portions of New York and New England at 1200/4 is supportive of IGW amplification and propagation. By 0000/5 this duct is even better defined as measured by the YJT sounding (Fig. 17b) and the derived duct parameters shown in Table 1 (note that a more conservative estimate of duct thickness yields a better agreement between the computed and observed phase velocities, as at ALB). Clearly, as the IGW amplifies and accelerates across New England subsequent to 1800/4 it encounters a large-scale environment inland from the coast even more conducive to the existence of long-lived, large-amplitude IGWs.

To help assess the possibility that shearing instability is acting as an important physical mechanism to IGW development we computed the bulk Richardson number (Ri) from the NMC Eta Model gridded fields (vertical resolution: 50 hPa) for the period 0000–1800/4. The Ri is computed as follows (e.g., Schneider 1990):
i1520-0493-126-6-1497-e5
where u and υ refer to the horizontal wind components in the (positive) eastward and northward directions, respectively, and the other symbols are as in (1) and (2). We display the Ri results in map form (Fig. 3) and on selected cross sections (Fig. 29). We shade regions where Ri < 1 in Fig. 3 for the 500–450-hPa (0000 and 0600/4) and 350–300-hPa (1200 and 1800/4) layers, respectively. These pressure levels were chosen based upon the estimated height of the critical level in the observed IGW region.

At 0000/4 (Fig. 3a) an area of Ri < 1 extends from the Tennessee Valley to northern Georgia. This region lies poleward of the cyclone center and is in a region where there is appreciable (estimated) advection of cyclonic vorticity by the thermal wind downshear of the main trough and where there is vigorous ascent (Fig. 4a). It also corresponds to the area where low-amplitude IGWs are observed (Fig. 1) and where the infrared satellite imagery suggests the existence of persistent, organized northwest–southeast-oriented cloud bands (Fig. 12a). By 0600/4 the area where Ri < 1 shrinks considerably (Fig. 3b). This shrinkage is probably an artifact of the Eta Model 0600/4 initialization (6-h forecast) and the reduction of strong wind shear regions. Support for this interpretation comes from a calculation of the Ri based upon the 0000/4 and 0600/4 GSO soundings, which show the existence of a layer of Ri < 0.5 centered near 500 hPa. At 1200/4 values of Ri < 1 are found in the 350–300-hPa layer over much of inland Pennsylvania, New York, and northwestern New England (Fig. 3c) along the northwestern edge of the path of the IGW. Similarly, at 1800/4 the position of the IGW over eastern Maine overlaps appreciably the region where Ri < 1.0 (Fig. 3d) and again the Eta Model 6-h forecast appears to underestimate the areal extent of low-Ri regions.

In general, the Ri patterns in other 50-hPa-thick layers ranging from 500–450 hPa to 300–250 hPa look broadly similar to Fig. 3 (not shown). From an inspection of the GSO (0600/4), ALB, CHH, YJT, and WSA soundings shown in Figs. 14 and 17a,b, we estimate that the observed south to southwesterly flow, when resolved into a component perpendicular to the IGW wave fronts, will match the observed phase velocity of the IGW of 30–40 m s−1 somewhere in the 500–300-hPa layer. Given that this “matching” takes place in the quasi-neutral layer above the observed wave ducts, and given that the bulk Ri values are about 1.0 or less in much of the IGW domain, it is likely that IGW wave energy will be maintained (or even increase) in the reflecting layer above the wave ducts. Accordingly, shearing instability processes cannot be ruled out as important to IGW organization and amplification.

To help summarize the last two sections, we present in Figs. 29a–f cross sections of PV, θ, θe, vertical motion, Ri, critical levels, and parcel divergence tendency at 0000/4 and 1200/4. At 0000/4 (Fig. 29a) the cross section shows a low DT coincident with the strong trough (500-hPa absolute vorticity greater than 30 × 10−5 s−1) over the southeast United States. Ascent (Fig. 29b) is found in the troposphere downshear of this trough and beneath the inclined DT near the equatorward margin of a well-defined stable layer that deepens toward the northeast. Somewhat downshear of the positive PV anomaly and deep trough can be found a thick layer of unbalanced flow as measured by the large positive parcel divergence tendency. This deep region of unbalanced flow has centers near 250, 400, and 600 hPa. The latter center is bisected by a critical level that itself lies just below a layer where Ri < 2. Note also the substantial overlap between the deep layer of positive parcel divergence tendency and the layer of unbalanced divergence (convergence) in the upper (lower) troposphere near the equatorward edge of the wave duct.

At 1200/4 (Fig. 29d) the DT has become more steeply sloped downshear of the now very impressive trough with 500-hPa absolute vorticity values of greater than 40 × 10−5 s−1 (Fig. 4c). Deep ascent (<−12 × 10−3 hPa s−1) is situated immediately downshear of the steeply inclined DT (Fig. 29e). Noteworthy is the downward penetration of this ascent into the duct. The dominant IGW at this time is embedded in the duct and is located northeastward of this ascent core. The region of maximum unbalanced flow as defined by the area of maximum positive parcel divergence tendency has strengthened over the last 12 h (Fig. 29f). Although the area of positive parcel divergence tendency spans the troposphere, it is best defined below the critical level. This deep column of implied unbalanced flow maximum penetrates the strong duct and also intersects a layer of low Ri containing a critical level above. This low-level duct is characterized by great stability in the 900–600-hPa layer and is capped by a layer of relatively weak stability between 600 and 400 hPa (Fig. 29e). We hypothesize that the flow configuration just described allows IGWs generated near the equatorward margin of the duct to propagate into and through the duct and to amplify (possibly) by wave overreflection processes. A second downshear maximum in unbalanced flow may also contribute further to IGW genesis and/or amplification, given the increasing duct depth and strength to the northeast and the continuation of low Ri values and the existence of a critical level above the duct (Fig. 29f).

7. Concluding discussion

We have examined an eastern North America cyclogenesis event from 4 January 1994 in which existing NMC/NCEP operational prediction models and human forecasters performed admirably in identifying the large-scale circulation features associated with storm development but were less successful in distinguishing three important embedded mesoscale features. These mesoscale circulations included a long-lived, large-amplitude IGW (and several other much smaller amplitude disturbances), a weak predecessor warm front wave cyclone that moved northeastward in the coastal waters well downstream of the primary surface cyclone, and a very heavy snow band that moved northeastward along the Appalachians. Urban dislocation was maximized in association with these three mesoscale circulations. Our attention was attracted to this case for three reasons: 1) the extreme nature of the dominant IGW event, 2) the availability of profiler and WSR-88D Doppler radar observations from stations in the path of the wave, and 3) the likelihood that deep convection does not play a significant role in IGW formation and amplification.

Noteworthy features of the dominant IGW include 1) its amplification near 0600/4 over northeastern North Carolina and southeastern Virginia occurs well to the northwest of an offshore baroclinic zone; 2) its association with peak crest-to-trough pressure falls of 10–13 hPa in 20–30 min with falls of 9 hPa (8 min)−1 and 1.3 hPa (1 min)−1, respectively, measured near BED; 3) its observed forward acceleration to a phase velocity of 30–40 m s−1 (wavelength of 200–300 km) as it moves northeastward across New England; 4) and its association with short-lived blizzard conditions (heavy snow and ice pellets, peak surface wind speeds of approximately 30 m s−1) prior to the arrival of the wave-induced surface pressure minimum.

Our analysis suggests that multiple small-amplitude (∼1 hPa) IGWs developed just prior to 0000/4 in a favorable environment for shearing instability and unbalanced flow in the middle troposphere ahead of a vigorous baroclinic trough. The northwest–southeast orientation of these IGWs ensured the existence of a critical level where the Ri was small. The multiple IGWs could be associated with distinctive northwest–southeast-oriented cloud bands in the infrared satellite imagery (best seen in image loops). These cloud bands were typically 25–50 km wide and separated by 100 km. In the infrared satellite imagery loops, these multiple cloud bands can be seen propagating northeastward toward the western edge of a band of higher and colder cloud tops marking the warm conveyor belt near the Atlantic coast. The satellite imagery loops also suggest that the leading members of the advancing cloud band packet tend to“disappear” beneath the deck of higher and colder cloud tops. Although we lack the critical observations necessary to be able to distinguish between in situ IGW amplification versus possible amplification of the downstream IGW in a traveling IGW packet, we believe that the rapid amplification of the dominant IGW occurred as the wave encountered an increasingly deep and well-defined stable layer in an atmosphere very favorable for vigorous ascent on subsynoptic scales.

The WSR-88D and profiler observations illustrate the rich, three-dimensional structure of the IGW including the discontinuous along-wave undulations in base reflectivity gradients, a deep (∼200 hPa) wave duct, and a 100–200-hPa layer of backing winds above the duct that may represent the acceleration of air toward the approaching IGW-induced pressure minimum. Given that rapid surface pressure falls result from either thinning of the cold layer adjacent to the ground and/or subsidence into this cold layer, and rapid pressure rises result from a thickening of this cold layer and/or ascent with roots in the cold layer, and given that along-wave wind and pressure variations also simultaneously exist in association with topographical variations, it is perhaps not surprising that the IGW evolves continuously as a nonlinear entity.

Our analysis points to three possible IGW genesis mechanisms: 1) shearing instability, 2) unbalanced flow and associated geostrophic adjustment, and 3) mechanical perturbation of the wave duct by vigorous vertical motions. In the first process shearing instability associated with large vertical wind shears in the middle troposphere in the presence of a critical level is hypothesized to create a favorable environment for the genesis of small-amplitude IGWs over north Georgia and the interior Carolinas. These small-amplitude IGWs then propagate northeastward into the very strong duct. In the second process unbalanced flow associated with the shortening of the wavelength between the deep trough and the downshear short-wave ridge excites IGW genesis immediately downshear and in an area where there is a very strong duct. The diagnosed unbalanced flow is concentrated below the jet cores and is especially prominent between the poleward exit region of the STJ and equatorward entrance region of the PFJ. In the third process a very intense subsynoptic-scale updraft core that penetrates a surface-based stable layer (duct) acts to produce IGWs by perturbing the top of the inversion and thus varying the depth of the cold air and creating a surface pressure perturbation. A fourth process may involve the generation of IGWs by flow over complex topography. We consider this last mechanism less likely since the southwesterly flow aloft near the mountaintop level is quasi-parallel to the Appalachians.

We hypothesize that there are at least three possible IGW amplification mechanisms. The first mechanism involves small-amplitude IGW amplification through overreflection in the strong wave duct in the cold-air damming region east of the Appalachians. The second mechanism is based on wave-CISK concepts [e.g., Powers and Reed (1993); Powers (1997)] in which there is a positive feedback between ascent, latent heat release, and wave growth. The third mechanism involves IGW growth in response to large unbalanced flow (and divergence) in the upper troposphere and (estimated) large parcel accelerations and vigorous deep subsynoptic-scale ascent associated with the IGW itself that penetrates and perturbs the increasingly strong and deep wave duct.

Evidence for the first two amplification mechanisms exists in the form of a well-defined wave duct overlain by a critical level near the base of a layer of low Ri, and significant latent heat release with heavy precipitation associated with the passage of the IGW as measured by Doppler radar base reflectivity and rain gauge observations. The third mechanism rests on the concept that if the subsynoptic-scale vertical circulations are both vigorous and deep enough to penetrate the increasingly deep and strong wave duct, then significant surface pressure perturbations may arise, persist, and amplify. In this regard these subsynoptic-scale vertical circulations are hypothesized to play the role of “surrogate convection” in that they are strong enough to perturb the surface pressure field in the presence of a strong wave duct.

At 0600/4 there is a close correspondence between regions of large parcel divergence, ascent, and IGW amplification. By 1200/4, however, the dominant IGW has outrun the main synoptic-scale ascent as indicated in the NCEP Eta Model and therefore must be amplifying through mechanisms other than “surrogate convection.” We suggest that the 80-km Eta Model lacked the necessary resolution to depict the dominant IGW signature. However, there is considerable direct and indirect evidence for vigorous subsynoptic-scale ascent in association with the IGW, snow bomb, and offshore predecessor wave cyclone. The evidence is represented by observed hourly snowfall rates of 10–15 cm in the mesoscale snow area along the northern Appalachians near 1200/4 (Fig. 11), 25–35-dBZ base reflectivities in the narrow band of heavy precipitation ahead of the IGW-induced pressure minimum (Figs. 21, 22), the presence of an Eta Model–initialized 500-hPa absolute vorticity center of 42 × 10−5 s−1 at 1200/4 (Fig. 4c), and the associated (estimated) large value of cyclonic vorticity advection upshear of the IGW amplification region (Fig. 4c). In particular, the 25–35-dBZ base reflectivity echoes in the snow/sleet region ahead of the IGW-induced pressure minimum correspond to instantaneous stratiform precipitation rates capable of producing 2–5-mm liquid precipitation amounts on time scales of 20–40 min (equivalent to 10–20 cm s−1 ascent rates). Ascent that must be associated with this precipitation and axis of maximum base reflectivity is suggested to play the role of the “surrogate convection” hypothesis discussed above.

We also suggest that a unique combination of events contributed to the coupling of vigorous deep tropospheric ascent with a very strong lower-tropospheric stable region (wave duct). First and foremost is the observed rapid shortening of the downshear half-wavelength in the 12 h ending 1800/4 (see the 500- and 400-hPa geopotential height fields in Figs. 4 and 28, respectively). This observed wavelength decrease must be associated with increasing cyclonic vorticity advection and ascent as can be confirmed from Figs. 4b–d and by the very large observed hourly snowfall rates (Fig. 11) on the western fringe of the area of vigorous ascent. A Lagrangian interpretation would be that air parcels in the upper troposphere must undergo large values of divergence (confirmed by the positive parcel divergence tendency and unbalanced divergence maxima situated near the inflection point in the region between the trough and downshear ridge; Figs. 28 and 29c,f) as they move rapidly from the trough region, where the flow is subgeostrophic and the absolute vorticity is large, to the downstream ridge region, where the flow is supergeostrophic and the absolute vorticity is small. The required vorticity loss must be associated with divergence aloft (to first approximation at the level of quasigeostrophic theory), convergence below, and deep ascent by mass continuity. A similar shortening of the downshear wavelength was observed in the 15 December 1987 Midwest cyclone that was also associated with a long-lived, large-amplitude IGW (Schneider 1990; Powers and Reed 1993; Pokrandt et al. 1996).

Second, the required deep ascent, occurring between the equatorward entrance region of the strengthening PFJ and the poleward exit region of the STJ along the coast (Figs. 4c and 5g,h), is reinforced by latent heat release associated with widespread stratiform precipitation. The impact of the diabatic heating is felt on the DT in the form of an observed tightening of the PT and θT gradients normal to the accelerating southwesterly flow (compare Figs. 5a,b,d,e,g,h). The resulting implied thermally direct circulation in the entrance region of the PFJ, reinforced by diabatic heating and manifest by the appreciable ageostrophic flow toward lower heights (Figs. 28c,d), must be associated with an acceleration of the observed southwesterly flow in response to Coriolis torques. [Alternatively, the existence of the deep ascent zone sandwiched between the two jets can be interpreted equivalently in terms of interacting jet streak circulations (Uccellini and Kocin 1987) or differential vorticity advection (Bell and Bosart 1989)].

Third, the unexpected development of a predecessor offshore warm-frontal wave favors a strengthening of the warm-frontal inversion over New England (Fig. 16) as warm air aloft is advected westward over the continental cold air, leading to the creation of a strong low-level wave duct. Fourth, the generation of absolute vorticity values greater than 40 × 10−5 s−1 in the 500–400-hPa layer in the prominent trough upstream of the IGW formation and amplification region occurs in association with the development of a large area of negative absolute vorticity in the middle and upper troposphere on the east side of the jet over the ocean (Fig. 4). This development further favors vigorous lateral and vertical parcel accelerations.

IGW organization and amplification after 1000/4 is marked by a distinctive satellite signature in which 1) the IGW-induced pressure minimum axis remains situated approximately 100 km downstream (to the northeast) of the back edge of an increasingly prominent secondary high, cold cloud shield, and 2) the IGW-induced pressure minimum axis remains situated 50–100 km upshear (to the southwest) of an increasingly well-defined narrow band of high, cold overshooting cloud tops that erupts from within the secondary cloud shield and is probably a manifestation of IGW-associated deep ascent. These cloud-top temperature, sea level pressure, and precipitation relationships suggest that the dominant IGW spans the troposphere and tilts upshear, consistent with possible wave origin in the upper troposphere and attendant downward energy propagation (Pecnick and Young 1984). What appears to be crucial in this case is to understand how an environment favorable for the existence of multiple small-amplitude IGWs can be transformed into one favorable for the rapid amplification of an individual IGW.

Acknowledgments

This research would not have been possible without the help of many individuals and organizations who provided us with pressure, wind, and temperature records from various supplementary weather observing locations throughout the northeastern United States. In particular we would like to thank Anthony Lupo of SUNY/Albany for providing computational assistance, Warren Snyder of the Albany, New York, NWSFO for providing data records, Arthur Jackson of the Geophysics Directorate of the Phillips Laboratories at Hanscom AFB in Bedford, Massachusetts, for providing the digitized pressure observations from BED. Chris Davis and Jordan Powers of NCAR provided valuable criticisms of the manuscript. Ken Rancourt of the Mount Washington Observatory in Conway, New Hampshire, provided key wind and temperature observations from the summit observatory. Other important observations were provided by Bob Cunningham (Lincoln, Massachusetts), Stephen Colucci (Cornell), Walter Drag (NWS/Taunton), Evan Gillespie (SUNY/Albany), Phillip Falconer (Scotia, New York), David Fitzjarrald (Atmospheric Sciences Research Center), Tom Galletta (Niagara Mohawk Power Corporation), Paul Gluhosky (Yale), R. Brad Harvey (Yankee Atomic Group), Mike Iacono (Atmospheric Environment and Research), Paul Jacobson and Dominic Scerbo (Northeast Utilities), Scott Lindstrom (University of Rhode Island), Joe Maggio (Lindehurst, New York), Jeff Morrison (Parsippany, New Jersey), Bill Munger (Harvard), Dennis O’Keefe (SUNY/New Paltz), Tony Praino (Poughquag, New York), John Quinlan (NWS/Albany), Glenn Ralph (NOAA/Air Resources Laboratory), Frederick Sanders (MIT/emeritus), John Scala (SUNY/Brockport), Chris Seeber (Charlestown, Rhode Island), Eric Sinsabaugh (Shaftsbury, Vermont), Rick Vaughn (Morris Plains, New Jersey), David Woodenstore (NRG Systems), and Richward Woolley (EA Engineering Science and Technology, Inc.).

Surface and upper-air observations and microbarogram observations were also obtained from the National Climatic Data Center. Gridded analyses from the European Centre for Medium-Range Weather Forecasts were obtained from the National Center for Atmospheric Research. Marilyn Peacock prepared a number of the crucial figures and Celeste Iovinella completed the manuscript and provided important editorial assistance. This research was supported by Grant F496209310002 from the Air Force Office of Scientific Research.

REFERENCES

  • Bell, G. D., and L. F. Bosart, 1988: Appalachian cold-air damming. Mon. Wea. Rev.,116, 137–161.

  • ——, and ——, 1989: The large-scale atmospheric structure accompanying New England coastal frontogenesis and associated North American east coast cyclogenesis. Quart. J. Roy. Meteor. Soc.,115, 1133–1146.

  • Bosart, L. F., and J. P. Cussen Jr., 1973: Gravity wave phenomena accompanying East Coast cyclogenesis. Mon. Wea. Rev.,101, 5445–5454.

  • ——, and F. Sanders, 1986: Mesoscale structure in the megalopolitan snowstorm of 11–12 February 1983. Part III: A large amplitude gravity wave and coastal frontogenesis. J. Atmos. Sci.,43, 924–939.

  • ——, and A. Seimon, 1988: A case study of an unusually intense atmospheric gravity wave. Mon. Wea. Rev.,116, 1857–1886.

  • ——, and G. M. Lackmann, 1995: Post-landfall tropical cyclone reintensification in a weakly-baroclinic environment: A case study of Hurricane David (September 1979). Mon. Wea. Rev.,123, 3268–3291.

  • ——, M. A. Bedrick, W. E. Bracken, M. J. Dickinson, G. J. Hakim, D. M. Schultz, and K. R. Tyle, 1996: Large-scale antecedent conditions associated with the 12–14 March 1993 cyclone (“Superstorm ’93”) over eastern North America. Mon. Wea. Rev.,124, 1865–1891.

  • Brunk, I., 1949: The pressure pulsation of 11 April 1944. J. Meteor.,6, 181–187.

  • Clark, J. D., 1983: The GOES user’s guide. United States Department of Commerce, 7-38–7-39.

  • Cressman, G. P., 1959: An operational objective analysis system. Mon. Wea. Rev.,87, 367–374.

  • Davis, C. A., E. D. Grell, and M. A. Shapiro, 1996: The balanced dynamical nature of a rapidly intensifying oceanic cyclone. Mon. Wea. Rev.,124, 3–26.

  • Ferguson, H. L., 1967: Mathematical and synoptic aspects of a small-scale wave disturbance over the lower Great Lakes area. J. Appl. Meteor.,6, 523–529.

  • Forbes, P. A., G. S., R. A. Anthes, and D. W. Thomson, 1987: Synoptic and mesoscale aspects of an Appalachian ice storm associated with cold-air damming. Mon. Wea. Rev.,115, 564–591.

  • Hakim, G. J., L. F. Bosart, and D. Keyser, 1995: The Ohio Valley wave-merger cyclogenesis event of 25–26 January 1978. Part I:Multiscale case study. Mon. Wea. Rev.,123, 2663–2692.

  • ——, D. Keyser, and L. F. Bosart, 1996: The Ohio Valley wave-merger cyclogenesis event of 25–26 January 1978. Part II: Diagnosis using quasigeostrophic potential vorticity inversion. Mon. Wea. Rev.,124, 2176–2205.

  • Hoskins, B. J., M. E. McIntyre, and A. W. Robertson, 1985: On the use and significance of isentropic potential vorticity maps. Quart. J. Roy. Meteor. Soc.,111, 877–946.

  • Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-year reanalysis project. Bull. Amer. Meteor. Soc.,77, 437–471.

  • Kaplan, M. L., S. E. Koch, Y.-L. Lin, R. P. Weglarz, and R. A. Rozumalski, 1997: Numerical simulations of a gravity wave event over CCOPE. Part I: The role of geostrophic adjustment in mesoscale jetlet formation. Mon. Wea. Rev.,125, 1185–1211.

  • Keshishian, L. G., and L. F. Bosart 1987: A case study of extended East Coast frontogenesis. Mon. Wea. Rev.,115, 100–117.

  • Koch, S. E., and R. E. Golus, 1988: A mesoscale gravity wave event observed during CCOPE. Part I: Multiscale statistical analysis of wave characteristics. Mon. Wea. Rev.,116, 2527–2544.

  • ——, and C. O’Handley, 1997: Operational forecasting and detection of mesoscale gravity waves. Wea. Forecasting,12, 253–281.

  • ——, M. desJardins, and P. J. Kocin, 1983: An interactive Barnes objective map analysis scheme for use with satellite and conventional data. J. Climate Appl. Meteor.,22, 1487–1503.

  • Krishnamurti, T. N., and W. J. Moxim, 1971: On the parameterization of convective and nonconvective latent heat release. J. Appl. Meteor.,10, 3–13.

  • Lefevre, R. J., and J. W. Nielsen-Gammon, 1995: An objective climatology of mobile troughs in the Northern Hemisphere. Tellus,47A, 638–655.

  • Lin, Y. L., and R. C. Goff, 1988: A study of a mesoscale solitary wave in the atmosphere originating near a region of deep convection. J. Atmos. Sci.,45, 194–205.

  • Lindzen, R. S., and K. K. Tung, 1976: Banded convective activity and ducted gravity waves. Mon. Wea. Rev.,104, 1602–1617.

  • Nielsen-Gammon, J. W., 1995: Dynamical conceptual models of upper-level mobile trough formation: Comparison and application. Tellus,47A, 705–721.

  • ——, and R. J. Lefevre, 1996: Piecewise tendency diagnosis of dynamical processes governing the development of an upper-tropospheric mobile trough. J. Atmos. Sci.,53, 3120–3142.

  • Olson, D. A., N. W. Junker, and B. Korty, 1995: Evaluation of 33 years of quantitative precipitation forecasting at the NMC. Wea. Forecasting,10, 498–511.

  • Pecnick, M. J., and J. A. Young, 1984: Mechanics of a strong sub-synoptic gravity wave deduced from satellite and surface observations. J. Atmos. Sci.,41, 1850–1862.

  • Pokrandt, P. J., G. J. Tripoli, and D. D. Houghton, 1996: Processes leading to the formation of mesoscale waves in the Midwest cyclone of 15 December 1987. Mon. Wea. Rev.,124, 2726–2752.

  • Powers, J. G., 1997: Numerical model simulations of a mesoscale gravity wave event: Sensitivity tests and spectral analyses. Mon. Wea. Rev.,125, 1838–1869.

  • ——, and R. J. Reed, 1993: Numerical simulation of the large-amplitude mesoscale gravity-wave event of 15 December 1987 in the central United States. Mon. Wea. Rev.,121, 2285–2308.

  • Ralph, F. M., M. Crochet, and S. V. Venkateswaran, 1993: Observations of a mesoscale ducted gravity wave. J. Atmos. Sci.,50, 3277–3291.

  • Ramamurthy, M. K., R. M. Rauber, B. P. Collins, and N. K. Malhotra, 1993: A comparative study of large-amplitude gravity-wave events. Mon. Wea. Rev.,121, 2951–2974.

  • Schneider, R. S., 1990: Large-amplitude mesoscale wave disturbances within the intense midwest extratropical cyclone of 15 December 1987. Wea. Forecasting,5, 533–558.

  • Tepper, M., 1954: Pressure jump lines in midwestern United States, January–August 1951. Research Paper No. 37, U.S. Weather Bureau, 70 pp.

  • Tsou, C.-H., P. J. Smith, and P. M. Pauley, 1987: A comparison of adiabatic and diabatic forcing in an intense extratropical cyclone system. Mon. Wea. Rev.,115, 763–786.

  • Uccellini, L. W., and S. E. Koch, 1987: The synoptic setting and possible energy sources for mesoscale wave disturbances. Mon. Wea. Rev.,115, 721–729.

  • ——, and P. J. Kocin, 1987: The interaction of jet streak circulations during heavy snow events along the east coast of the United States. Wea. Forecasting,2, 289–308.

  • Van Tuyl, A. H., and J. A. Young, 1982: Numerical simulation of nonlinear jet streak adjustment. Mon. Wea. Rev.,110, 2038–2054.

  • Vincent, D. G., G. B. Pant, and H. J. Edmon, 1977: Generation of available potential energy of an extratropical cyclone system. Mon. Wea. Rev.,105, 1252–1265.

  • Young, J. A., and C. F. Richards, 1973: A climatological study of surface pressure events at Madison, Wisconsin: Preliminary results relating to possible gravity wave activity. Studies of the Atmosphere using Aerospace Proves, Space Science and Engineering Center 1973 Annual Report, University of Wisconsin—Madison, 172–187. [NTIS COM-74-11626/0.].

  • Zack, J. W., and M. L. Kaplan, 1987: Numerical simulation of the subsynoptic features associated with the AVE-SESAME I case. Part I: The preconvective environment. Mon. Wea. Rev.,115, 2367–2393.

Fig. 1.
Fig. 1.

Dominant IGW isochrone analysis for 0700–1900 UTC 4 January 1994. Area affected by “snow bomb” outlined by bold-dashed ellipse. Region of multiple small-amplitude IGW disturbances outlined by bold-dotted ellipse.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 2.
Fig. 2.

Station locator map. Solid dots, triangles, and open triangles denote surface stations, upper-air stations, and WSR-88D sites, respectively. Profiler site at Bloomfield (BMF), Connecticut, is denoted by an open square. Inset shows cross-sectional orientations used in Fig. 29.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 3.
Fig. 3.

National Meteorological Center (NMC, now the National Centers for Environmental Prediction) Eta Model–initialized 1000-hPa heights (solid, every 4 dam) and 1000–500-hPa thickness (dashed, every 6 dam) for (a) 0000, (b) 0600, (c) 1200, and (d) 1800 UTC 4 January 1994. Shading (scaled at left) denotes regions where the Richardson number (Ri) is <1. Richardson number is computed as a layer average over 500–450 (350–300) hPa in (a), (b) [(c), (d)].

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 4.
Fig. 4.

As in Fig. 3 except for 500-hPa heights (solid, every 6 dam), 500-hPa absolute vorticity (dashed, every 6 × 10−5 s−1), and 500-hPa vertical motion (ascent only; every 3 × 10−3 hPa s−1 and shaded according to the gray scale at the left of each panel). Representative gridpoint winds plotted with one pennant, full barb, and half-barb denoting 25, 5, and 2.5 m s−1, respectively.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 5.
Fig. 5.

Left column (a, d, g): pressure (solid, every 50 hPa and shaded beginning at 350 hPa according to the gray scale) on the dynamic tropopause (DT) defined by the 1.5 potential vorticity unit (PVU) surface (1.0 PVU = 10−6 K m2 kg−1 s−1) for 1200 UTC 3 January 1994 (top), 0000 UTC 4 January 1994 (middle), and 1200 UTC 4 January 1994 (bottom). Middle column (b, e, h): as in the left column except for potential temperature (solid, every 10 K) on the DT. Right column (c, f, i): as in the left column except for 850-hPa heights (solid, every 3 dam) and equivalent potential temperature (dashed, every 10 K). Winds as in Fig. 4.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 6.
Fig. 6.

Digitized pressure (hPa) versus time (UTC) for Bedford, Massachusetts: (a) 24 h ending 0000 UTC 5 January 1994, and (b) 1-min resolution data for the 2-h period ending 1600 UTC 4 January 1994. Source: A. Jackson of the Air Force Phillips Laboratories, Hanscom Field, Bedford, Massachusetts.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 7.
Fig. 7.

Reconstructed analog wind speed trace for (a) Boston (BOS), Massachusetts, and (b) Providence (PVD), Rhode Island, centered at 1500 (1430) UTC 4 January 1994 for BOS (PVD). Vertical lines denote 5-min intervals. Wind speeds in knots (multiply by 0.515 to convert to m s−1) with horizontal lines every 10 kt. Precipitation type and intensity indicated at BOS along the abscissa.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 8.
Fig. 8.

Reconstructed microbarogram traces for selected stations in the northeastern United States for 4 January 1994. Pressure (hPa) and time (hours) scale as shown with tick marks denoting every 1 hPa and 1 h, respectively. Solid contours denote peak ridge-to-trough pressure decreases (hPa) associated with IGW passage and are shaded accordingly.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 9.
Fig. 9.

National Meteorological Center (NMC, now the National Centers for Environmental Prediction) manually prepared 24-h quantitative precipitation forecast (QPF) for 24-h periods ending (a) 1200 UTC 4 January 1994, (b) 1200 UTC 5 January 1994, and the 24-h update ending (c) 1200 UTC 4 January 1994. Contour intervals are 6.2, 12.5, and 25.0 mm.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 10.
Fig. 10.

Accumulated precipitation (solid, contoured every 1, 5, 10, and 20 mm) for 6-h periods ending (a) 0600, (b) 1200, and (c) 1800 UTC 4 January 1994.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 11.
Fig. 11.

Reconstructed station pressure (solid, hPa; tick marks every 2 hPa on the ordinate) for Elkins (EKN), West Virginia (top), Williamsport (IPT), Pennsylvania (middle), and Syracuse (SYR), New York (bottom), for 4 January 1994. Times (UTC) shown by two-digit numbers along the abscissa; sloping vertical lines every 1 h. Hourly snowfall rates (cm h−1) are shown stippled. Conventional plotting for present weather with temperature and dewpoint temperature in degrees Celsius and winds in m s−1 according to Fig. 4.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 12.
Fig. 12.

(a) Infrared satellite imagery for (a) 0001 UTC, (b) 0301 UTC, (c) 0601 UTC, and (d) 0901 UTC 4 January 1994. Gray scales correspond to operational MB curve (Clark 1983) except that temperatures −31.2° to 1.8°C are linearly stretched. Arrows mark the position of multiple cloud bands discussed in text.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 13.
Fig. 13.

Manually prepared surface analysis for 0600 UTC 4 January 1994. Mean sea level isobars (solid, every 2 hPa). Plotting format of conventional surface observations is as in Figs. 4 and 11.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 14.
Fig. 14.

Sounding in skew T–logp format for Greensboro (GSO), North Carolina, for 0000 (dotted) and 0600 (solid) UTC 4 January 1994 and for Athens (AHN), Georgia, for 0000 (dashed) UTC 4 January 1994. Winds (m s−1, in format of Fig. 4) for GSO at 0000 UTC (left), AHN (center), and GSO at 0600 UTC (right).

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 15.
Fig. 15.

National Lightning Detection Network cloud-to-ground lightning flash distribution: (a) 0000–0600 UTC 4 January 1994, (b) 0600–1200 UTC 4 January 1994. Triangle, box, and circle symbols denote first 2 h, middle 2 h, and last 2 h of each 6-h period, respectively.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 16.
Fig. 16.

As in Fig. 13 except for 1200 UTC 4 January 1994.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 17.
Fig. 17.

As in Fig. 14 except for (a) Albany (ALB; solid), New York, and Chatham (CHH; dashed), Massachusetts, at 1200 UTC 4 January 1994, and (b) Stephenville (YJT; solid), Newfoundland, and Sable Island (WSA; dashed), Nova Scotia, at 0000 UTC 5 January 1994. Winds:(a) ALB (left) and CHH (right); (b) YJT (left); and WSA (right).

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 18.
Fig. 18.

Base reflectivity from the WSR-88D Doppler radar at Sterling (KWBC), Virginia, for 0908 UTC 4 January 1994. Solid (dashed) contours denote 15 (30) dBZ with reflectivities greater than 30 dBZ shown stippled.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 19.
Fig. 19.

As in Fig. 15 except for (a) 1001, (b) 1201, (c) 1401, and (d) 1601 UTC 4 January 1994. Arrows denote region of high cold cloud tops associated with dominant IGW.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 20.
Fig. 20.

As in Fig. 13 except for 1350 UTC 4 January 1994.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 21.
Fig. 21.

Left column: composite WSR-88D base reflectivities for (a) 1130, (c) 1210, and (e) 1250 UTC 4 January 1994. Contour intervals every 10 dBZ beginning 15 dBZ. Right column: composite 20-min pressure change analysis with solid (dashed) lines denoting pressure falls (rises) contoured every 0.5, 1.0, 2.0, 4.0, and 8.0 hPa with zero contour dashed for (b) 1130, (d) 1210, and (f) 1250 UTC 4 January 1994.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 22.
Fig. 22.

As in Fig. 21 except for 1330, 1410, and 1450 UTC 4 January 1994.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 23.
Fig. 23.

Composite WSR-88D 20-min isochrones of the 15-dBZ base reflectivity contour along the back edge of the precipitation shield preceding the IGW axis of minimum sea level pressure for the period 1110–1530 UTC 4 January 1994.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 24.
Fig. 24.

Time series of smoothed hourly winds (m s−1, plotted according to the convention given in Fig. 4) from the Bloomfield (BMF), Connecticut, wind profiler.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 25.
Fig. 25.

As in Fig. 13 except for 1550 UTC 4 January 1994.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 26.
Fig. 26.

As in Fig. 13 except for 1750 UTC 4 January 1994.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 27.
Fig. 27.

Observations of winds (m s−1) from the Mount Washington Observatory in North Conway, New Hampshire, for 4 January 1994. Spiral radial arms denote hourly times (UTC) increasing counterclockwise. Mean hourly wind speed and direction (octas) as shown.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 28.
Fig. 28.

Maps of 400-hPa geopotential heights (thin solid lines every 6 dam), Lagrangian Rossby number (see text) computed at 400-hPa (thick solid lines at intervals of 0.5, 0.75, 1.0, 1.5, 2.0, 2.5 and 3.0 units), 250-hPa isotachs (shaded every 10 m s−1 beginning at 60 m s−1), and 300-hPa ageostrophic wind barbs plotted every other grid point in the format of Fig. 4 for (a) 0000, (b) 0600, (c) 1200, and (d) 1800 UTC 4 January 1994. (e)–(g) As in (a)–(d) except for the 400-hPa parcel divergence tendency (positive values shaded at contour intervals of 0.5, 1.0, 1.5, 2.0, 3.0, 4.0 × 10−8 s−2), 400-hPa geopotential heights (thin solid lines every 6 dam), and unbalanced divergence (zero and positive values only contoured every 2 × 10−5 s−1) averaged over the 350–250-hPa layer. All computed quantities are derived from the gridded Eta Model fields. IGW pressure minimum axis is denoted by a bold dotted line at 1200 and 1800 UTC 4 January 1994.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Fig. 29.
Fig. 29.

Left: cross sections (orientation shown in Fig. 2) of potential temperature (medium solid every 8 K) and potential vorticity (thin solid every 0.5 PV and shaded according to the scale below the top panel for PV > 1.5 PVU where 1.0 PVU is as in Fig. 5) for (a) 0000 and (d) 1200 UTC 4 January 1994. Middle: Cross sections of equivalent potential temperature (medium solid every 5 K) and vertical motion (shaded according to the scale along the top panel every 3 hPa s−1 beginning at ±3 hPa s−1) with ascent and descent regions outlined by thin dashed and solid lines, respectively, for 0000 (b) and 1200 (e) UTC 4 January 1994. Right: Cross sections of parcel divergence tendency (zero and positive values only; contoured bold dashed every 0.5 × 10−8 s−2), critical level (heavy solid line), based upon a wave phase velocity of 20 and 30 m s−1, respectively, at 0000 and 1200 UTC 4 January 1977, Richardson number (shaded according to the scale below the top panel for values of 2 and less), and unbalanced divergence given by thin solid (dashed) contours for positive (negative) values every 2 × 10−5 s−1 for 0000 (c) and 1200 (f) UTC 4 January 1994. Top and bottom cross sections extend about 2850 and 3000 km, respectively, with selected station locations (Fig. 2) shown along top of each panel.

Citation: Monthly Weather Review 126, 6; 10.1175/1520-0493(1998)126<1497:ASOCMS>2.0.CO;2

Table 1.

Inertia–gravity wave and wave duct parameters. IGW and wave duct parameters from Greensboro (GSO), North Carolina, at 0600 UTC 4 January 1994, Albany (ALB), New York, at 1200 UTC 4 January 1994 (numbers in parentheses are the conservative values discussed in the text), and Stephenville (YJT), Newfoundland, at 1200 UTC 5 January 1994. Parameters computed are (from left to right) pressure at base of duct, potential temperature at base of duct, pressure at top of duct, potential temperature at top of duct, the average potential temperature of the duct, observed duct thickness, computed minimum duct thickness from Eq. (2), computed IGW wave speed from Eq. (1), observed IGW phase speed, observed direction that wave is moving toward, wind direction in the duct, wind speed in the duct, and the component of the observed wind direction in the direction of IGW propagation. Numbers in parentheses refer to values derived for a more conservative duct thickness estimate (see text).

Table 1.
Save