Wake lows, often associated with mature or dissipating mesoscale convective systems (MCS), can provide unique forecast challenges for operational forecasters (Johnson and Hamilton 1988). Strong, post thunderstorm winds associated with a wake low can significantly affect both aviation and public forecasts. In rare instances, damaging winds associated with wake lows have been observed (Voelker and Wetzel 1994). The purpose of this paper is to demonstrate the use of meteorological sensing instruments (e.g., radar and satellite) and other technology for detecting the presence of wake lows given the current knowledge of MCS conceptual models.
WSR-88D radar images and derived products, satellite images, AFOS application graphics and surface observational data are used to show the presence of two wake lows which occurred across North Dakota and Minnesota on June 25, 1996 and July 6, 1996. Both mesoscale convective systems did not produce damaging winds along their respective trailing flanks. However, surface winds exceeding 30 knots were reported near the vicinity of the wake lows. Examples revealing the mesoscale rear inflow jet, critical for wake low development, are shown by or inferred from improved satellite and radar technology available at the time.
Understanding the basic processes for wake low development and utilizing improved meteorological technology such as satellite and radar allow wake lows to be more easily recognized by the operational forecaster. This becomes particularly helpful in data sparse regions where surface mesoanalysis would show little evidence of the low. Even more importantly, under certain conditions, it may be possible to forecast wake low development. This, in turn, may help lead to better forecast and warning products available to the public.
A. Radar Echo Patterns
Conceptual models of the surface pressure, storm relative flow, and precipitation fields associated with the symmetric and asymmetric type MCS commonly resulting from a multicell evolution process are shown in Figures 1a and b. In PRE-STORM studies in Oklahoma and Kansas in the middle 1980s, Loehrer and Johnson (1995) make the argument that these two patterns are not different MCS types, but conceptual models of different stages in the MCS life cycle.
In a symmetric MCS example, the leading convective line is arc shaped with the strongest cells towards the center and weaker cells on the flanks (Figure 1a). The mesohigh is located near the rear of the convective line. The wake low is located along the trailing flank of the MCS behind the stratiform rain region. A notch-like concavity may be present in the rear of the stratiform region. This reflectivity notch is believed to reflect the location of the mesoscale rear inflow jet advecting lower theta-e air downward and eroding the back edge of the stratiform rain region (Loehrer and Johnson 1995).
For the asymmetric type MCS, the northern end of the convective line is weaker, with the strongest cells located on the southern flank (Figure 1a). The intense convective cells reflect the presence of strong low-level convergence along this flank compared to the northern part of the line. The stratiform region is displaced to the left of the convective line. The mesohigh is present well behind the leading stratiform region which helps to increase the pressure gradient between the mesohigh and wake low.
B. Wind Field in an MCS
The wind field of a mature midlatitude squall line (Figure 1b) consists of two major airstreams, front to rear flow and mesoscale rear inflow. The front-to-rear flow slopes upward from low levels ahead of the surface gust front to high levels in the rear (Smull and Houze 1985). This flow transports hydrometeors rearward from the convective line to the stratiform region. These hydrometeors contribute to precipitation in that region by seeding clouds (cells) leftover from previous convective cells which originated from the convective line. Beneath the sloping updraft is a region of relative descending rear to front flow through the stratiform region to the convective line.
Although not fully understood, Houze (1993) theorizes that the rear inflow could be one branch of a mid-level mesoscale convective vortex (MCV), associated with an asymmetric squall line or a wind shift line uniform in the along-line direction of a symmetric-type system (Houze 1993). This descending rear inflow jet, interacting with the trailing stratiform rain region, is believed to be the major factor in the development of the wake low. Subsidence warming associated with the descending rear inflow jet eventually out weighs the effects of diabatic cooling from the interaction with the stratiform rain region. This warming hydrostatically lowers the surface pressure producing a wake low (Johnson and Hamilton 1988).
C. Pressure Field in an MCS
Commonly, the surface pressure field associated with a mature squall line consists of a mesohigh centered under or just to the rear of the leading convective line and a wake low adjacent or just to the rear of the stratiform region (Figure 1a). In addition, a pre-squall mesolow or trough is often observed ahead of the leading convective line.
The pre-squall mesolow is attributed to convectively induced subsidence warming in the middle to upper troposphere ahead of squall lines. Warm air advection may also play a minor role in the development of the pre-squall mesolow (Johnson and Hamilton 1988). The mesohigh is linked primarily to rainfall induced evaporative cooling which hydrostatically produces an area of high pressure (Bluestein 1993). Hydrometeor loading can also contribute to the enhancement of the mesohigh. This feature is usually located just to the rear of the region of the squall line heavy precipitation and convective downdrafts. The mesohigh may shift rearward towards the wake low strengthening the isallobaric gradient (Loehrer and Johnson 1995).
Wake lows are well documented by numerous authors in recent years (Johnson and Hamilton 1988, Stumpf and Johnson 1991, Gallus 1996). These studies show a close link for wake low development between the trailing stratiform region and subsidence associated with the descending rear-inflow jet. Observational studies have shown that the mesoscale rear inflow jet either descends gradually or may exhibit a steep descending slope (Smull and Houze 1987, Przybylinski and Schmocker 1993). Downward motion and adiabatic warming are balanced by diabatic cooling from evaporation preventing any wake low development. However, in wake low development, it is generally accepted that the relatively low theta-e air, associated with the mesoscale rear inflow, interacts with the easterly front-to-rear flow and the back edge of the stratiform rain region. The interaction of these two flow regimes results in a strongly convergent mid-level flow. The combination of the mid-level convergence and lower theta-e air interacting with the back edge of the stratiform rain region results in evaporative cooling and feeding strong localized downdrafts in this region. Stumpf and Johnson (1991) have shown that the location of strong downdrafts have been linked to the greatest pressure falls and location of wake low development. Adiabatic warming within the strong downdrafts outweighs evaporative cooling. Thus, the area of low-level warming near the rear of the stratiform rain region is a reflection of the negatively buoyant downdrafts within the descending mesoscale rear inflow.
Recent modeling studies show that micro-physical processes alone within the stratiform region may induce enough descent for wake low development (Gallus 1996). Modeling simulations of collapsing stratiform precipitation cores, or a tight reflectivity gradient at the rear of the stratiform region, advancing rapidly with time display the best results for wake low development. Rapid dissipation of stratiform rain regions may be caused by either the demise of the convective line or a change in wind shear preventing the rearward transfer of hydrometeors.
A. Case One (25 June 1996)
The first wake low case occurred during the late morning and early afternoon on June 25, 1996. The 1500 UTC surface analysis (Figure 2) showed an area of low pressure centered over Western South Dakota and a well defined east-west oriented warm front that extended from the low along the North and South Dakota border.
A series of GOES-8 3.9 micron water vapor images (Figures. 4a-d) from 1132 UTC to 1332 UTC showed the evolution of an expanding dark area at the back edge of the MCS with time. This expanding dark area reflects the intrusion of lower theta-e air, associated with the mesoscale rear-inflow jet, behind the MCS. The arrows indicate the probable location of the rear-inflow jet. Although not available for this study, wind profilers, WSR-88D velocity azimuthal display (VAD) wind profiles, or velocity cross sections could also help in determining the existence of a rear inflow jet.
Figure 4a. GOES-8 water vapor images for 1132 UTC 25 June 1996.
By 1338 UTC, the WSR-88D reflectivity image (Figure 5) continued to reveal an asymmetric MCS with new moist convection developing along the southern flank. Again, a large stratiform rain region remained over the northern end of the MCS mainly across northeastern North Dakota. The AFOS Data Analysis Program (ADAP) graphic two-hour surface altimeter change (SAC) in hundredths of inches from 1000-1200 UTC (Figure 6) shows a well defined rise-fall couplet over north central and northeastern North Dakota. This couplet correlates well with the recent location of the mesohigh and wake low. In contrast, a well defined rise-fall couplet was absent along the trailing flank of the stronger convective line, south of the enhanced stratiform region across central North Dakota (Figure 6). It will be shown that Jamestown, North Dakota (JMS), closer to the convective line, experienced 20-30 knot post thunderstorm easterly surface winds. However, sustained winds were of a shorter duration and not as strong as the winds at Devils Lake, North Dakota (DVL), which is located in or near the stratiform region.
As the convective system continued to propagate eastward into eastern North Dakota, a 3.4° KMVX base velocity image at 1535 UTC (Figure 7) confirms the wind field pattern one would expect to infer from the presence of a wake low. Note a 30 knot inbound or east-west couplet exists. Also, at approximately 7000 ft above ground level 20 nautical miles east and west of the radar site a 50 knot west-to-east flow which would be an inflow jet. Last, at higher levels, an ascending front to rear flow is indicated. These air flow patterns at different elevations fit well with the conceptual models of MCS wind flow shown in Figure 1b.
Once developed, the area of greatest pressure falls, or wake low, propagated eastward trailing the stratiform region of the mature MCS. The ADAP three-hour total altimeter change chart from 1300-1600 UTC (Figure 8) shows a large area of strong pressure falls over the northern half of Eastern North Dakota with a maximum in the DVL area. At this time, the mesohigh has weakened. However, the wake low remained well established.
Figure 9 shows a time series of surface pressure, wind speed and direction at DVL, JMS, Grand Forks, North Dakota (GFK) and Fargo, North Dakota (FAR), during the wake low event. In viewing the pressure trend, this time series shows the passage of the mesohigh, wake low and return to ambient pressure at DVL and especially GFK. At JMS and FAR, closer to the convective line than the stratiform region, the pressure field shows the mesohigh and return to ambient pressure with no real indications of a wake low passage. With the approach and passage of pressure falls associated with the wake low, winds veered to an easterly direction and increased in speed at DVL and FAR. After the passage of the maximum pressure falls, wind speeds decreased as the atmosphere returned to the ambient pressure. Surface winds at DVL maintained wind speeds of 30 to 35 knots through the period 1335-1600 UTC. A peak gust of 39 knots was recorded at 1355 UTC. By 1900 UTC, sustained winds had dropped below 10 knots. This graphic also shows that at JMS, closer to the convective line, the degree of the pressure rise-fall and associated wind speeds were less that what occurred at locations closer to the trailing stratiform region.
Figure 9. Time series of surface pressure, wind speed and direction at Devils Lake (DVL), Jamestown (JMS), Grand Forks (GFK), and Fargo (FAR), North Dakota.
B. Case II (6 July 1996)
Case II is a good example of the relationship between velocity and reflectivity fields associated with a wake low as it moved across the KMVX radar site on July 6, 1996. Although the time sequence of reflectivity versus velocity differ due to lack of available archived data, many important features can be inferred by examining the images.
A mature MCS and associated trailing stratiform region propagated across eastern North Dakota, northeastern South Dakota and northwestern Minnesota. A composite reflectivity image from 0320 UTC (Figure 10), indicates the northern edge of a convective line over southeastern North Dakota with a trailing stratiform region over portions of extreme eastern North Dakota, northwestern and west central Minnesota. Using storm motion vectors nearest the radar site, the stratiform region is moving from 250 degrees at 25 knots. Using this motion, the back edge of the stratiform region would be near the vicinity of the RDA at the time of the velocity image in Figure 11.
Figure 11 shows, near the RDA, an inbound and corresponding outbound component indicating a surface east to southeast flow of approximately 20 to 30 knots at the KMVX RDA site. Looking east of the RDA at approximately 25 nautical miles at 300 meters above ground level, a relatively strong outbound flow is possibly representing the lower levels of the rear-inflow jet. Comparing this area of directional shear to its projected location at the time of the reflectivity image in Figure 10, one can deduce the possibility of a wake low. Notice this area of strong, inbound flow relative to the RDA is in the rain-free area immediately behind the area of strongest reflectivity gradient in the stratiform region.
Also Figure 10 shows a weaker gradient, rear-inflow notch signature is located in the area of the strong outbound winds east of the RDA lending another indicator of a strong rear-inflow jet. By combining these radar signatures along with other available tools, one can determine the wake low location and the possible time of the onset of strong post-thunderstorm winds and wind direction in the short term.
By understanding the process on how wake lows develop and being able to recognize MCS conceptual models favorable for wake low formation, forecasters will not be surprised by sudden changes in wind speeds and direction associated with wake lows. This process, in turn, will lead to better forecasts and statements to the public. Improved meteorological tools such as GOES high resolution images, WSR-88D products, improved high resolution gridded data and hourly surface analysis from observational data sets and the Rapid Update Cycle (RUC) will aid in determining if wake low development is possible, observing an established wake low and forecasting passage of this feature and a subsequent decrease in wind speed.
A. Forecasting or Anticipating Wake Low Development
Satellite images provide an excellent means of observing MCS development and propagation and can provide clues to possible wake low development. GOES water vapor (WV) and infrared (IR) images, cyclonic circulation patterns associated with a Mesoscale Convective Vortex (MCV) can be identified and may contribute towards enhancing the mesoscale rear-inflow jet. The dark area on WV imagery trailing the back edge of an MCS can be a reflection of the descending mesoscale rear-inflow jet. This subsiding region is comprised of lower theta-e air as compared to its environment.
The WSR-88D has been a valuable tool for identifying possible wake low development as shown by the cases presented. A primary factor was recognizing which type of MCS was involved (symmetric or asymmetric). Additional important clues to identify the favorable position of wake low development include: 1) location of the trailing edge of the stratiform rain region, 2) identification of a relatively large, weak reflectivity notch at the rear of the stratiform rain region signifying the probable location of the mesoscale rear-inflow jet, and 3) detecting the rear-inflow jet on base velocity data. By looping base reflectivity images, MCV cyclonic circulations have also been observed in the stratiform rain region. This has assisted in locating the mesoscale rear-inflow jet. The notch-like concavity at the rear edge of the stratiform region is associated with mesoscale rear inflow comprised of lower theta-e air that evaporates a portion of the stratiform echo. This is not unlike a rear-inflow notch on a convective line associated with a strong descending rear-inflow jet. The WSR-88D VAD Wind Profile, velocity cross section, and storm relative velocity may all provide indicators of a descending rear-inflow jet especially if the RDA is in the favored location near the wake low area.
B. Observing, Tracking and Forecasting Wake Lows
There are several other tools to observe, track and forecast winds associated with wake lows. Mesohighs (or wake lows) may be identified by surface mesoanalysis. However, because of wide spatial surface observation locations, these small scale features may occur between observation sites.
The AFOS ADAP pressure change graphics, (specifically the SAC, SA3, or SA6) have been perhaps the best tools for tracking and forecasting winds associated with wake lows. This is due to an ageostrophic isallobaric acceleration that result in the high wind speeds observed. In the AWIPS era, the Mesoscale Analysis and Prediction System Surface Assimilation System (MSAS) will be used to assume this function. Also the three-hour surface pressure change from the RUC may also help to identify and track a wake low. The area of strongest winds generally occur from the outer pressure fall gradient to the isallobaric fall center. Surface winds in turn gradually diminish with the passage of maximum pressure falls. Tracking the isallobaric fall maximum can give an approximate time frame of duration of the wind event. This product has also proved valuable in determining the strength of a wake low by the magnitude of the rise-fall couplet or isallobaric gradient as well as detecting the dissipation of the wake low.
C. Forecast Implications
An ability to detect the presence of (or anticipate the development of) wake lows can have many operational forecasting advantages. Because of the transient nature of wake lows, short term and zone forecasting can be updated to depict the rapid changes in surface winds which can approach severe limits. Due to the location of wake lows relative to an MCS, these strong winds occur usually after thunder and associated precipitation have ended. This is at a time when one would usually expect improving conditions. In addition, aviation forecasts can be updated to reflect near term changes in the surface wind direction and speed which may greatly affect aviation interests.
Several wake low events occurred during the summers of 1997 and 1998 in the Grand Forks county warning forecast area. Using the above mentioned methodology, wake low development was anticipated, in most cases, either in advance or in the initial stages of the occurrence. Aviation and public forecasts were updated accordingly. This provided the public with advanced notice of changing wind conditions. In several instances, wind advisories and one high wind warning were issued using these techniques.
Thanks to the staff, past and present, at WFO Eastern North Dakota at Grand Forks for data collection. A special thanks to Philip Schumacher, SOO at WSFO Sioux Falls, SD for his guidance, suggestions, and expertise regarding this paper. I also wish to thank Ron Przybylinski, SOO at WSFO St. Louis and Brad Bramer, SOO at WFO Eastern North Dakota for their constructive and thorough review which further strengthened this paper.
Bluestein, H.B., 1993: Synoptic-Dynamic Meteorology in Midlatitudes Vol. II, Oxford University Press. 594pp.
Bothwell, P.D., 1988: Forecasting Convection with the AFOS Data Analysis Programs (ADAP Version 2.0). NOAA Technical Memorandum NWS SR-122. DOC, NOAA, NWS, Southern Region Headquarters, Scientific Services Division, Fort Worth, TX, 92pp.
Gallus, W.A. Jr., 1996: the Influence of Microphysics in the Formation of Intense Wake Lows: a Numerical Modeling Study. Mon. Wea. Rev., 124, 2267-2281.
Houze, R.A., Jr., 1993: Cloud Dynamics. Academic Press, 573pp.
Johnson, R.H., and P.J. Hamilton, 1988: the Relationship of Surface Pressure Features to the Precipitation and Airflow Structure of an Intense Midlatitude Squal Line. Mon. Wea. Rev., 116, 1444-1470.
Loehrer, S., and R.H. Johnson, 1995: Surface Pressure and Precipitation Life Cycle Characteristics of Pre-storm Mesoscale Convective Systems. Mon. Wea. Rev., 123, 600-621.
Przybylinski, R.W., and G.K. Schmocker, 1993: the Evolution of a Widespread Convective Windstorm Event over Central and Eastern Missouri. Preprints 13th Conf. on Weather Analysis and Forecasting, Vienna, AMS (Boston) 461-465.
Smull, B.F., and R.A. Houze, Jr., 1987: Rear Inflow in Squall Lines with Trailing Stratiform Precipitation. Mon. Wea. Rev., 115, 2869-2889.
Stumpf, G.J., and R.H. Johnson, 1991: The Wake Low in a Midlatitude Mesoscale Convective System Having Complex Convective Organization. Mon. Wea. Rev., 119, 134-157.
Voelker, A.M., and H.A. Wetzel, 1994: Application of Adap and Mesoanalysis During an Unusual Post Thunderstorm Wind Event. Central Region Applied Research Paper 12-13. DOC, NOAA, NWS, Central Region Headquarters, Scientific Services Division, Kansas City, MO, 12, 82-92.