STORM TRACK: November 30, 1982 (Volume 6 Issue 1)

(This article concludes the excellent seven-part series by John Weaver on the History of Thunderstorm Forecasting. ST and its readers very much appreciate the many hours of research and writing that were invested by John on this project. While reading and rereading these articles, ST's Editor has mused on what it would be like if the founding fathers of meteorology could somehow return and see where their early efforts have led.)

A ghostly visit during the late night shift.

The end of the nineteenth century marked the true beginnings of the technological era. It had taken four hundred-fifty years for Mankind to grope his way from the simplest mechanical devices for assisting his previous endeavors, to the increasingly complex machines of the twentieth century. Along the way, these devices had helped the sciences progress at an ever accelerating pace. Current electro/mechanical creations exist which are beyond the ability one person to understand completely with even a lifetime of study. And the sophistication seems to increase daily. Many philosophers claim that we have opened a 'Pandora's box,' which will lead eventually to our own destruction (even though, given our nature, technological development is almost inevitable). However, the purpose of this paper is not to try to solve unsolvable problems but rather to outline how some of this advancement helped develop the young science of meteorology.

In the previous section, it was shown that a great deal was known regarding the nature of the thunderstorm's environment prior to the turn of the century. It was also noted that, little of this knowledge was being applied, for a variety of reasons. As in the case of the general circulation, progress in thunderstorm forecasting awaited the arrival of regular upper air observations. The sequence for availability of these data in the U.S. was as follows: 1) instrumented kite flights were conducted (for research purposes) at sixteen stations, beginning in 1898; 2) the first regular kite flights for weather forecasting began in 1907; and 3) balloons, fitted with continuous recording instrument packages, began to partially replace kites in 1909. New findings followed rapidly. In 1914, W. J. Humphreys, of the U.S. Weather Bureau, noted that strong thunderstorms required air mass instability, which could result from strong surface heating, cold air advection aloft, forced lifting by low-level dense air, or some combination of these. In 1920, Charles Brooks noted, in the conclusion to a tornado case study, that severe storms seemed to be associated with a surface low; a convergence zone with cyclonic shear; moist, southeasterly flow in the low layers; and cooler, southwesterly flow aloft. In 1920, Humphreys (again) found similar pre-severe storm conditions but offered more specific information, which included: 1) the strongest storms seemed to occur near pressure troughs, which are also convergence zones; 2) winds from the surface to about 2 km are strong, moist and southerly; 3) the winds veer above this layer and the flow is cooler and drier; and 4) most often, a temperature inversion separates the two layers. 1926 was also the year that Sir Napier Shaw designed the T-Phi gram for plotting upper air data, and, in 1927, Stuve designed his pseudoadiabatic chart.

On July 1, 1931, aviation interests persuaded the Weather Bureau to replace balloon and kite soundings with aircraft observations. These 'APOBS' were discontinued in 1940 in favor of the newly developed radiosonde (actually invented in 1928 by Moltchanoff). The radiosonde is a balloon-mounted device, which does not require retrieval to obtain the observations, since it is equipped with a small, inexpensive transmitter to telemeter pressure, temperature and humidity readings to radio receivers back at, the surface. With the advent of these efficient transmitters, daily upper air maps could be produced routinely. Furthermore, dissemination of the information had become more efficient due to the establishment of a Weather Bureau teletype network in 1928 and the gradually increasing usage of wire photo transmission of maps. The 'facsimile' machine had been invented in 1926 by Francis Jenkins.

By 1943, enough synoptic data had accumulated to allow Showalter and Fulks, of the U.S. Weather Bureau, to reach the following conclusions regarding the atmospheric environment prior to the formation of tornadic storms:

1) A conditionally unstable layer of dry air is usually found superimposed on a lower layer of moist air. The moist air is at least convectively neutral but may be both conditionally and convectively unstable;

2) the two layers are typically separated by a temperature inversion, which (they felt) serves to prevent deep convection from beginning, until some external lift mechanism (such as a frontal passage) compels a 'violent convective upsurge';

3) the wet bulb potential temperature decreases with height;

4) vertical wind shear is generally present, and normally winds veer with height; and

5) thunderstorms are often occurring elsewhere in the vicinity.

In 1951, Fawbush, Miller and Starrett presented the results of tests and refinements to these conclusions in order to establish an operational set of forecasting criteria for use by the U.S. Air Force. A few more specific criteria were identified. The horizontal distribution of the moist air in a tornado threatened region, for example, was found to exhibit a distinct maximum along a relatively narrow band (i.e. 'a moisture tongue') in most cases. Another example: a mid-level jet (between 3 and 6 km altitude) of 35 kts or more in the overriding dry air was identified in many of the cases studied. The most likely area for severe activity is near the intersection of this 'mid-level' jet with the moist tongue.

Several other additions and refinements to Showalter and Fulks work have been made over the years. A recent listing of many important criteria is given by Miller in a Technical Report for the Air Force, which is over one hundred pages long. In fact, so many 'secondary', or supplementary, severe storm conditions have been defined, that a full listing would be too large to be useful in any practical sense. One of the reasons for this, of course, is that the atmospheric environment does not have to be the same for every severe thunderstorm case. The severe storm forms through a combination of many factors, sometimes one factor dominates, other times it may play a secondary role, and sometimes it is missing completely. Missing, that is, to the degree that we can detect it with our somewhat coarse data gathering network (i.e. sounding sites separated by some 400 km and release times for soundings at 12 hours apart). A great deal of air mass instability and many of the Fawbush et. al. criteria seem to be present in the vast majority of the large outbreaks of severe weather. Additionally, Lee and Galway, in 1956, documented the importance of another feature present in many major episodes of severe weather, i.e. wind maxima in the jet stream. Investigation proved the jet stream to be a regular feature of the high troposphere, and these authors found a strong correlation of jet stream maxima and the western and southern edge of the -60C cold pool at 200 mb to severe weather.

About this time, another device appeared which would result in a major advance to weather forecasting. At the end of World War II, several captured German rockets were brought to White Sands, New Mexico for study. On March 7, 1947, the first photographs of earth from very high altitudes (160 km) were taken. For the next ten years, more photos from even higher altitudes excited meteorologists about the possibility of regular weather observations from space. However, little impetus for weather satellites could be generated until the Russians bested the Americans by launching Sputnik I in October of 1957, four months ahead of America's launching of its experimental satellite, Explorer I. Under great pressure to 'get more U.S. satellites into space', the U.S. launched the first weather satellite (TIROS-I) on April l, 1960. Subsequent satellites have become quite sophisticated. The latest of these (Geostationary Operational Earth Satellite -- GOES) hovers some 22,500 miles above the equator and routinely delivers half-hour interval photos, both in visual wavelengths (as great as one half mile resolution) as well as infrared wavelengths (used to sense radiated temperature to resolutions as great as one mile). Meteorological analysis and prediction has been aided on every scale of action. Prior to the availability of satellite data, the large-scale organization of weather systems (quite apparent from space) was not very well understood. Data are now available over the oceans, polar regions, and uninhabited portions of the globe heretofore unaccessible. Numerical predictions have improved noticeably, as only a part of the result. Tropical storms can no longer form and grow to mature hurricanes without being observed. Precise frontal positioning has been simplified, as has identifying and estimating the strength and location of sub-scale shortwave troughs. The major role played by thunderstorm outflow 'arc olouds' in triggering new convection is, thanks to satellites, just beginning to be understood. A listing of all of the benefits provided by satellite imagery is beyond the scope of this paper. Suffice it to say that the list is long and growing.

The modern thunderstorm forecaster has at hand a myriad of tools unavailable to his nineteenth century counterpart. Data arrives more quickly, in greater variety and in larger quantity than ever before, and prediction techniques are much more highly polished. Yet, the forecaster still finds himself in the middle ground between science and art. When all the maps are analyzed and all the data synthesized, there still remains the need for subjective interpretation and human judgment. This is true in all facets of meteorological forecasting, but particularly so for small-scale, short lived events such as thunderstorms. Perhaps someday, enough will be learned regarding the subsynoptic scales of the atmosphere to provide better tools for this level of prediction. At this writing, several efforts involving research on the smaller scale are underway. Until they succeed, the thunderstorm forecaster will continue to depend, at least in part, on the human talent.