Preprints, 14th AMS Conference on Weather Analysis and Forecasting

Dallas, Texas, 15-20 January 1995


Harold E. Brooks
Charles A. Doswell III

NOAA/ERL/National Severe Storms Laboratory
Norman, Oklahoma

Erik N. Rasmussen
Sonia Lasher-Trapp

Cooperative Institute for Mesoscale Meteorological Studies
Norman, Oklahoma


The dryline frequently observed in the Southern Plains is important as a favored zone for convective development in the spring and early summer (Rhea 1966), although the mechanism of convective initiation is not well known (Ziegler and Hane 1993). It represents the boundary separating warm, moist air, extending northward from the Gulf of Mexico, from drier air originating over the high terrain in the southwestern United States and Mexico. The basic characteristics of the dryline have been detailed by Schaefer (1986).

Study of the dryline, as with any mesoscale meteorological phenomenon, has been hampered by the lack of high temporal and spatial resolution data. One solution has been to use observations from aircraft and mobile ballooning facilities to try to capture cross-sections across the dryline (Ziegler and Hane 1993). With the development and deployment of automated meteorological observing systems, it is now possible to observe the surface characteristics of the dryline with higher resolution than in the past. A network of such systems, the Oklahoma Mesonet (Crawford et al. 1994), has put at least one station in every county in Oklahoma recording standard meteorological variables every five minutes. In addition, over a small area of southwestern Oklahoma, a second network of identical sensors, the Agricultural Research Service (ARS), has placed 41 stations over an approximately 1000 km2 area, providing extremely dense observations in time and space. The two networks combine to provide "targets of opportunity" for case studies of meteorological events when they occur in the network any time of the year.

We wish to document one such event in this paper. On 14 April 1994, a small area of very dry, warm air moved northward through the Oklahoma Mesonet and ARS networks, just ahead of the larger-scale dryline. With the standard observational network, this event would have gone undetected. Thus, besides being of scientific interest, this case illustrates the problem of detecting potentially important mesoscale boundaries with the standard observing network. Although no storms formed on this day, owing to the strong capping inversion, the presence of the strong mesoscale gradients of moisture seen here could be significant on other days.

Observations were made on this day by the mobile mesonet of the Verification of the Origins of Rotation in Tornadoes EXperiment (VORTEX) (Rasmussen et al. 1994), but we will focus in this preliminary study on the fixed-site mesonet observations. Blending of these observational data sets with those from the field project will be the focus of future work.


A series of observations over a period of 150 minutes illustrates the movement and basic behavior (Fig. 1). At 1900 UTC, the dryline can be seen in western Oklahoma, where the dewpoints go from 3 C or less in the extreme west up to 20 C in the south- central part of the state. The observation of a 9 C dewpoint at Grandview ("G") at 1900 UTC with south-southwesterly wind provides an interesting challenge for the analyst as to the placement of the dryline in the extreme southern part of the state. As time goes on, the dewpoint drops at Grandview and Walters ("W") and the wind veers around to the southwest (1930 UTC). The dewpoint then begins to increase again at those sites with a backing of the wind. Further to the northeast, the dewpoint rapidly drops at Acme ("A"), reaching 3 C at 2030 UTC, at which time the Grandview dewpoint has risen to 14 C. Dewpoint values equal to the extremely low dewpoint at Acme are still confined to extreme western Oklahoma, almost 200 km to the west, and a case could be made that the main dryline is still west of Grandview. Recovery to dewpoints in the teens at Acme is rapid and by 2100 UTC, the lowest dewpoint in central Oklahoma is at Minco ("M"). Finally, by 2130 UTC, dewpoints across western Oklahoma begin to drop (Acme is back to 10 C) as the main dryline approaches, finally reaching Acme by about 2230 UTC (not shown).

A small area of dry air from southwest to northeast can be tracked using the Oklahoma mesonet and the ARS sites located near Acme. A series of stations lies nearly along the direction of movement of the dry air over a distance of approximately 30 km (Fig. 2). When the dewpoint temperature is plotted versustime for these sites, the continuous progression of the dry air and subsequent moisture return is apparent (Fig. 3). The dewpoint falls by 10-14 C over a period of two hours, but recovers to within 2 C of the original dewpoint value within 30 minutes. The onset of the return of moisture is the most easily tracked feature. It advances at a speed of 10-12 m s-1 from a direction of 200. This is approximately the speed and direction of the low-level winds in the moist sector.

The relationship of meteorological variables to each other during the course of the event is best illustrated by looking at a time series from Acme (Fig. 4). The temperature steadily rises at a rate of about 1 C hr-1 until the return of moisture begins. At the onset of the moistening, there is a 0.9 C drop in the temperature over five minutes. The wind direction shows more dramatic changes, with the wind veering about twenty degrees during the dry episode, but then rapidly returning to its original direction in the moist air. Note that the wind speed remains relatively constant throughout the entire episode. Also of interest, the standard deviation of the wind speed and direction show remarkably small changes throughout the period of drying and moistening. There is no evidence of the gustiness of the wind in the dry air typically associated with dryline passage. Instead, the impression is of a distinct dry air mass, with slightly different momentum.

The existence of the high spatial resolution observations from the ARS network allows us the unique opportunity to see the horizontal extent of the dry pocket. At the time of the lowest dewpoint reading in the network, the dewpoint changes from 11.2 C to -0.4 C, and back up to 4.7 C over a distance of approximately 12 km (Fig. 5). The gradient is particularly strong on the eastern side of anomaly, with the minimum dewpoint of 7.4 C observed at the southeasternmost station in the network, while the station 5 km to the west reached 1.1 C (Fig. 6). The two stations have similar dewpoints until approximately 1940 UTC, when the eastern station quickly returns to the moist air dewpoint value while the dewpoint continues to drop at the western site. The horizontal scale of this anomaly appears to be on the order of 30 km, but it persists for a period of a few hours within the state of Oklahoma. Clearly, the standard observing network would have a difficult time detecting this feature, let alone capturing it in sufficient detail to resolve it. Properly timed, hourly reports would only see one observation at most from this feature. An analyst would be hard-pressed to determine whether it is real or not. Only with the high temporal and spatial resolution data can the existence and extent of this feature be verified. Note that it could not be identified it until it reached the mesonet region.


Clearly, a small region of dry air moved through southwestern Oklahoma on 14 April 1994. Although the boundaries of this region were not associated with deep convection on this day, it seems plausible to believe that phenomena of this kind could have significant impacts on the initiation of convection, perhaps with only slightly different environmental conditions. Davies- Jones and Zacharias (1988) identified waves on the dryline that produced observed changes in dewpoint almost as large as we have seen on a day that produced multiple violent tornadoes. Their waves were of a much larger scale than the feature seen here and moved along the dryline approximately with the speed of the v-component of wind over the lower half of the troposphere. They speculated on the origin of the waves, dismissing gravity waves and suggesting that symmetric instability might be relevant. Sanders and Blanchard (1993) looked at the same case and concluded that horizontal advection of dry air was not responsible for the waves on that day since the maximum of the correlation of the dewpoint and the westerly component of the wind was at zero time lag. If advection was responsible, they reasoned, the dewpoint should lag the winds. In our case, however, there is a maximum correlation of almost the same magnitude at a lag of 50 minutes in addition to the maximum at zero lag (Fig. 7).

The question of the origin of the dry air remains open to speculation. 1800 UTC upper air soundings for the VORTEX project from Norman (OUN) and just west of the Acme site illustrate the strong gradient in moisture above the boundary layer in Oklahoma that day (Fig. 8). To reach the surface values observed in the ARS network the OUN sounding would have to be mixed almost to 500 hPa. This seems unreasonably deep, but the Mobile Cross-chain Loran Atmospheric Sounding System (M-CLASS) (Rust et al. 1990) sounding would only need to be mixed to approximately 800-750 hPa. (Note that the apparently bad data near and above 220 hPa in the M-CLASS sounding does not affect our interest in the lower troposphere.) The strong moisture gradient suggests that it is possible that even drier air existed off to the southwest of the area and that by the time the surface dry pocket reached Acme at 2030 UTC, the environmental conditions might have been drier. We do not have observations at 1800 UTC to support this speculation, but if true, an even shallower layer would have to be mixed. A significant amount of momentum would have had to be lost during any mixing process from the 1800 UTC soundings, since the winds just above 850 hPa in the M-CLASS sounding are 35 kts.

If mixing is responsible for the dry air, a significant question is why the mixing remained confined to such a small area. It is possible that such small areas of mixing are relatively common and that it is only with the advent of very high-density observing networks that they become obvious. A more careful, long-term look at the mesonet observations will be necessary to identify the frequency of such events. Further, the lack of gustiness of the wind in the dry air is interesting if mixing of a significant depth of the atmosphere is invoked as the explanation.

Other mechanisms could be responsible. It is at least conceivable that the inhomogeneities in the moisture field are being advected. During the previous night, convection took place in north and west Texas and it is possible that some of the heating of the processed air could result in the observed temperature and dewpoint. Again, the absence of high-density observations south of the Oklahoma Mesonet makes this unverifiable. In any event, the small horizontal extent of the anomaly makes its persistence remarkable.

There are a number of interesting forecasting implications of this event (see also Doswell et al. 1995). The first is the detection of the anomaly. As mentioned before, lower resolution observing systems would have yielded one observation of the dry air at most. The question of representativeness immediately comes up. It is possible that remote observations, from platforms such as satellite or the WSR88-D operating in clear air mode, might provide indirect evidence of the different air mass.

A second question relates to the relationship of such anomalies to convection, in particular. We do not know what the vertical motion fields are associated with the passage of the dry air, although there is weak convergence in conjunction with the wind shift. If these events are not extremely rare, it is possible that their vertical motion fields could be significant contributors to convective initiation in the vicinity of the dryline. Doswell (1987) discussed the importance of mesoscale (or smaller scale) events in the initiation of convection. While the scale of the anomaly detected here is small in some sense, it is large when compared to an individual thunderstorm. As such, convergence at the boundary of a similar anomaly on a day with a weaker capping inversion (see Fig. 8) might be significant in the development of convection.

We plan to look more closely at this day and include data from a wide variety of sources, including the standard NWS observations and the mobile observations from VORTEX. The blending of data from these various platforms will require significant effort since the instrumentation is different, but it should help provide a more complete picture of the evolution of the atmosphere on this day.


Crawford, K. C., F. D. Brock, R. L. Elliott, G. W. Cuperus, S. J. Stadler, H. L. Johnson, and C. A. Doswell III, 1994: The Oklahoma Mesonet: A twenty-first century project. J. Atmos. Oceanic Technol., 11, submitted.

Davies-Jones, R., and D. Zacharias, 1988: Contributing factors in the 10 May 1985 tornado outbreak in northwest Kansas. Preprints, 15th Conference on Severe Local Storms. Amer. Meteor. Soc., Baltimore, Maryland, 284-287.

Doswell, C A. III, 1987: The distinction between large-scale and mesoscale contribution to severe convection: A case study example. Wea. Forecasting, 2, 3-16.

Doswell, C. A. III, H.E. Brooks, and E. N. Rasmussen, 1995: Forecasting isues and implications from the VORTEX-94 project. Preprints, 14th Conference on Weather Analysis and Forecasting. Amer. Meteor. Soc., Dallas, Texas, this volume.

Rasmussen, E. N., J. M. Straka, R. Davies-Jones, C. A. Doswell III, F. H. Carr, M. D. Eilts, and D. R. MacGorman, 1994: Verification of the Origins of Rotation in Tornadoes EXperiment: VORTEX. Bull. Amer. Meteor. Soc., 75, 995-1006.

Rhea, J. O., 1966: A study of thunderstorm formation along dry lines. J. Appl. Meteor., 5, 58- 83.

Rust, W. D., R. Davies-Jones, D. W. Burgess, R. A. Maddox, L. C. Showell, T. C. Marshall, and D. K. Lauritsen, 1990: Testing a mobile version of a Cross-Chain Loran Atmospheric (M-CLASS) Sounding System. Bull. Amer. Meteor. Soc., 71, 173-180.

Sanders, F., and D. O. Blanchard, 1993: The origin of a severe thunderstorm on 10 June 1985. Mon. Wea. Rev., 121, 133-149.

Schaefer, J. T., 1986: The dryline. Mesoscale Meteorology and Forecasting, P. S. Ray, Ed., Amer. Meteor. Soc., Boston, 549- 572.

Ziegler, C. L., and C. E. Hane, 1993: An observational study of the dryline. Mon. Wea. Rev., 121, 1134-1151.

Return to VORTEX Page