(Reprinted from preprint Volume: Eighth Conference on Weather Forecasting and Analysis, June 10-13, 1980. Denver, Colo. Published by the American Meteorological Society, Boston. Mass.)


Terrain roughness may play an important role in initiating and sustaining convective activity. Convective activity can occur as a result of a) differential heating over southward-sloping topography, b) moist air being forced upward over an escarpment or c) amplification of lee waves beneath a low-level inversion when air is lifted over a small barrier.

Previous investigations have usually involved radar studies with mountainous terrain. Most noted are those by Henz (1974) and Karr (1976) with numerical simulations completed by Gerbier (1961) and Hosler (1962). Also, results from aircraft observations have been presented by Braham (1960). On a smaller scale, Myers (1964) and Jones et al. (1974) have studied the effects of hilly terrain on convection. In general, these studies have demonstrated the evolution of convection resulting from mechanical lifting.

In northwest Texas, an escarpment called the "Caprock", slopes southeastward from Amarillo (AMA) to Lubbock (LBB) (see Figure 1). With a moist southeasterly surface wind, the influence of the Caprock can occur. During the spring months, this can be seen visually as a line of towering cumulus during the late afternoon. Another typical springtime phenomenon is the development of orographic thunderstorms in Eastern New Mexico. These storms migrate eastward off the mountains during the late afternoon and approach the Amarillo vicinity during the late evening.

As an initial attempt to document the influence of the topography, a radar climatology within 100 nm around Amarillo was assembled for the spring convective season (}~y and June) from 1975 through 1978. The objective of this study was to obtain significant evidence supporting the influence of topography in initiating convection in the study region. In conjunction with the radar study, synoptic patterns favorable for convective activity were also identified.


Before data extraction commenced, it was first recognized that with increasing range several radar biases might be present. The most common are beam widening and over or undershooting of the radar beam. Echo detail is lost as the radar beam widens with range. This effect is most significant when echoes are small and appear in clusters. Also, since the radar beam elevation increases with range, echoes which occur close to the radar can be missed due to undershooting of the radar beam. Likewise, overshooting effects become more important at longer ranges. These effects on initial echo counts must be considered in the analysis and it is unjustifiable to assume that the radar detects targets equally at all ranges.

Previous radar climatologies completed by Henz (1974), Bark (1976) and Driscoll (1978) have used a uniform box grid in the data extraction process (see Figure 2a). The grid necessitates employing a correction scheme to account for range bias effects in order to arrive at an approximation of the true echo distribution.

It is believed that a more accurate echo distribution can result by using a concentric grid (Figure 2b) which does not employ corrections on the original data. The grids consists of three annular rings, 25 nm apart, divided into 45 sectors. The data in each ring is treated independently with the moan and departures from the mean computed.

The magnitude of the departures is then tested statistically using a non-parametrical procedure described by Conover (1971). The results locate those sectors which had relatively high or low echo occurrences during the four--year study period. For a more detailed explanation refer to Marshall (1980).


The data used in this study consisted of 34 microfilm reels of the plan position indicator (PPI) scope from the WSR-57, 10-cm radar, located at the National Weather Service in Amarillo. Data was extracted from observations made on an hourly basis. During the four-year study period, there were a total of 14,216 observations made from 2,186 hours of available radar data. A total of 3,573 initial echoes developed within the study region.


The long-term precipitation climatology supports the hypothesis that topography enhances convection. Haragan (1978) analyzed the 30-year pattern for 63 stations in the Texas Panhandle (Figure 3). In May, the general west-to-east precipitation gradient rapidly increases just east of Miami (A). In June, the same max-imam can be distinguished however, a new maximum appears over Palo Duro Canyon (B), with a ridge extending southwestward to Littlefield (C). It should be mentioned that the four-year precipitation pattern closely resembles the long-term pattern.

Surface reports from Amarillo and Lubbock were analyzed for May and June 1975-1978, to examine the relationship between convective activity and wind direction. The results are based on observations between 18 and 00 GMT (Table 1). The primary wind direction on convective days was from the south or southeast. From these directions, moisture can be advected upslope into the region from the Gulf of Mexico. Parmenter (1976) has shown that low-level moisture intrusions into West Texas are crucial for the development of convection. Rarely does convection take place when the surface wind direction is from a northerly component.

Initial echo values were stratified by month for comparison with the long-term precipitation pattern. In an attempt to sample diurnal convection only, departures from the mean were computed for each ring for those echoes which appeared between 18 and 00 GMT (Figures 4a and b). Using a 95% confidence interval, departures greater than 30% are significant.

In May, several isolated echo maxima occur. The largest and most extensive area is located southeast of Amarillo, parallel to the Caprock. In this region, maximum uplifting would be expected under a prevailing southeasterly surface wind. This maximum is located upwind from the precipitation maximum near (A). Another echo maximum occurs just west of Lubbock and appears to be the result of convective disturbances which propagate northeastward with mid-level winds, not directly influenced by the topography. The other echo maximum which occurs near Tucumcari (TCC), is possibly the result of orographic activity which migrates into the study region from Eastern New Mexico. This effect is reflected in the west-to-east long-term precipitation gradient. Notice that a broad minimum of echo activity occurs northeast of Amarillo. This deficit, which appears both in May and June, could be the result of a "shadow effect". Under the conditions of a southeast surface wind, moisture could be uplifted and condensed out prior to reaching the Canadian River Basin.

During June, the large two-fold maximum is still positioned southeast of the Caprock. Moreover, the orientation has shifted more towards an east-west direction, possibly due to the prevailing southerly surface winds during June. The orographic influence near Tucumcari is more pronounced, shifting more to the south than in May.


Two synoptic patterns were found to be associated with enhancing convection in the study area. The first, consists of a 500 mb trough or cut-off low developing over Southern California. As this wave progresses inland, strong southeasterly surface winds develop, accompanied by increasing southwesterly winds aloft. Surface moisture is advected northward, upslope over the Caprock. The resulting air mass becomes unstable.


During the four-year study period, there were a total of 20 cases involving 998 initial echoes that occurred with this pattern. The echoes were plotted and only the maximum departures from the mean ring value were calculated (Figure 5a). Departures greater than 40% were significant using a 95% confidence interval. The results indicate that the primary source for convection is located east of Amarillo, off the Caprock. It is theorized that the strongest lifting occurs near Palo Duro Canyon. The initial echoes appear just northeast of the Canyon where they grow to maturity. The upper windflow steers the storms to the northeast where they decay while new echoes continue to propagate the system eastward. Another echo maximum occurs near the White River Canyon (D), Just east of Lubbock. Here, lifting occurs on a smaller scale, due to a southeastward extension of the Caprock.

In the second pattern, a large ridge of high pressure builds towards the east, with light southeasterly surface winds and stronger northwesterly winds aloft. As a short wave passes, orographic thunderstorms develop and migrate towards the southeast along with the upper windflow. New echoes develop and propagate eastward.

During the four-year study period, there were a total of 12 cases involving 458 initial echoes that occurred with this pattern. The percent departure from the mean ring value was tested and is plotted in Figure 5b.


A concentric grid was devised to account, in part, for the large range bias effects associated with conventional radar. The procedure involves computing the departures from the mean ring value and applying a nonparametric statistical test to determine regions of unusually high or low echo occurrences.

This climatology has provided supporting evidence that topography indeed influences the local convection in the Texas Panhandle. It is determined that the most pronounced effect of the Caprock occurs under two synoptic conditions. Under these two conditions the most favorable regions for convection are Palo Duro Canyon, White River Canyon, the Caprock, and Eastern New Mexico foothills.


This research was supported in part by the U.S. Nuclear Regulatory Commission under Contract N1~.C-04-77-016.


Bark, D., 1975: A survey of the radar echo population over the western Kansas High Plains. Kansas Water Resources Board, Project 5-369, 68 pp.

Braham, R., and M. Draginis, 1960: Roots of orographic cumuli. J. Meteor., 17, 214-226.

Conover, W., 1971: Practical Nonparametric Statistics. John Wiley Press, p. 111.

Driscoll, D., 1978: A radar echo climatology for the southern High Plains. Texas Dept. of Water Resources, Project report LP-64, 89 pp.

Gerbier, N., and M. Berenger, 1961: Experimental studies of lee waves in the French Alps. Quart. J. R. Meteor. Soc., 87, 13-23.

Haragan, D., 1978: Precipitation climatology for the Texas High Plains. Texas Journal of Science, 30, 107-123.

Henz, J., 1974: Colorado High Plains thunderstorm systems - a radar synoptic climatology. Colorado St. Univ., M.S. thesis, 82 pp.

Hosler, C., L. Davis, and D. Booker, 1962: The effect of mountainous terrain on convection. Final report, G73~3, Penn Univ. St.

Jones, D., F. Huff, and S. Changnon, Jr., 1974: Causes for precipitation increases in the hills of southern Illinois. Illinois State Water Survey, Report 75, 36 pp.

Kart T., and R. Wooten, 1976: Summer radar echo distribution around Limon, Colorado. Mon. Wea. Rev., 104, 728-734.

Marshall, T., 1980: Topographic influences on radar echo climatology. Texas Tech Univ., M.S. thesis.

Myers, J., 1964: Preliminary radar climatology of central Pennsylvania. J. Appl. Meteor., 3, 421-429.

Parmenter, F., 1976: Low-level moisture intrusion from infrared imagery. Mon. Wea. Rev., 104, 100-104.