National Severe Storms Laboratory and Cooperative Institute for Mesoscale Meteorological Studies, Norman, Oklahoma

In the western US, mountainous regions act as focal points for convective initiation (e.g., Braham 1958; Karr and Wooten 1976; Banta and Schaff 1987). The earliest studies on this topic use single-radar data (e.g., Braham 1958; Ackerman 1959) or photogrammetric data (Orville 1965) to study primarily: 1) where clouds first form, 2) cloud or radar echo characteristics, and 3) differences between clouds that form over mountains versus those that form over relatively flat terrain. More recent studies tend to use precipitation (Brazel and Balling 1987), lightning (Watson et al. 1994), or satellite data (Banta and Schaaf 1987) to investigate the frequency of convection in areas of mountainous terrain. Besides identifying convective initiation "hot spots" in Colorado and New Mexico, Banta and Schaaf (1987) show that prevailing flow directions strongly influence where convection forms. Since a study examining the influence of flow direction on convective initiation location in central Arizona is lacking, completing such an analysis is a long-term goal of the current study.

This paper focuses on the role terrain plays in producing convection over central Arizona during the monsoon season (roughly July-mid-September). Arizona receives 40- 60% of its annual rainfall (Jurwitz 1953) during the monsoon season, and these storms generally produce flooding, severe weather, and lightning. Hence, a better knowledge of areas favorable and unfavorable for storm development would be useful for forecasters and weather-sensitive operations (e.g., utilities, fire managers, farmers, etc.). Balling and Brazel (1987) and Watson et al. (1994) document Arizona’s diurnal-precipitation-cycle during the monsoon season. An important difference between these studies is that Balling and Brazel use rain gauge data only, whereas Watson et al. use lightning data. Since rain totals are measured by a discrete set of gauges, using lightning data improves data coverage considerably. The result is a more detailed climatology (compare Figs. 5- 11 in Watson et al. to Fig. 5 in Balling and Brazel). However, in Watson et al., lightning-location errors may be up to 15- 20 km.

Even though Watson et al.’s results are more detailed than Balling and Brazel’s, both studies show that mountainous regions tend to receive precipitation during the afternoon and early evening hours, whereas desert regions tend to receive precipitation during the late evening and early morning hours. Knowing this diurnal-precipitation-cycle helps forecasters anticipate the time of day precipitation is most likely over the desert and mountain regions. Although the above studies identify areas favorable and unfavorable for storm development, the relatively poor data resolution limits the utility of the results. In order to improve the resolution of Arizona’s convective climatology, this study uses high-resolution (1 km) reflectivity mosaics to analyze the diurnal cycle of convective in central Arizona during the 1999 Monsoon.

2. METHOD

For this study, radar reflectivity mosaics, gridded at 1 km resolution, are created using level II data collected from the Phoenix (KIWA) and Flagstaff (KFSX) Weather Surveillance Radar-1988 Dopplers (WSR-88D) during July and August 1999. The mosaic technique maps reflectivity data from multiple-radars onto a common Cartesian grid (440 km x 440 km) using objective analysis techniques (Zhang et al. 2000). The domain (Fig. 1) includes Phoenix, the most populated city in Arizona, and mountainous terrain that affects the area’s precipitation variability.

Diurnal reflectivity frequencies are created for each grid box by calculating the number of reflectivity observations greater than or equal to 40 dBZ for each hour (6 obs/hr) over the two-month period. This reflectivity threshold is used to highlight areas conducive to convection. Since the radar sample volume increases with range, the maximum possible reflectivity magnitude decreases with range. Thus, in this study, the relative frequency of reflectivity values greater than 40 dBZ is underestimated at ranges far from the radar. To address this sampling issue, range-dependent-thresholds may be used in the future.

3. RESULTS

The diurnal cycle of convection is distinguished clearly in this analysis (Fig. 2). A complete loop of the diurnal reflectivity frequencies and more information concerning this study are available at the web address: http://www.nssl.noaa.gov/~mackeen/radar-climo.html. The first areas to initiate convection after sunrise (18 UTC – 20 UTC) include the Mongollon Rim, San Francisco Mountain, Prescott vicinity, and Southeast Highlands (Fig. 2a). Note in Fig. 2 that convection initiates northwest of the White Mountains, rather than over the mountains. It is likely the lack of signal over the White Mountains is due to the radar sampling degradation discussed earlier.

During the afternoon (Fig. 2b; 20 UTC – 00 UTC), areas favored for convection progress from the mountain ridges and peaks toward lower elevations, including parts of the Sonoran and Painted Deserts. Finally, convective activity is limited mostly to lower elevations and its occurrence decreases through the evening and into the early morning hours (Fig. 2c; 01 UTC – 09 UTC).

4. DISCUSSION AND FUTURE WORK

Using high-resolution reflectivity mosaics to improve understanding of the diurnal convective cycle in central Arizona exemplifies a new analysis technique that is beneficial to areas where multiple WSR-88D’s improve radar estimation of rainfall. This is especially true in areas like central Arizona, where terrain blockage significantly affects radar sampling. Using 1 km resolution reflectivity mosaics and high-resolution terrain data provides an excellent opportunity to identify areas of repeated convection and the role of terrain forcing in central Arizona. Hence, the current study will be expanded to cover the 1995- 2000 monsoon seasons. In the future, we will be stratifying the convective events according to synoptic regime. This stratification will help us test the hypothesis that discernibly different synoptic patterns are associated with convective development in specific regions of Arizona.

ACKNOWLEDGEMENTS

REFERENCES

Ackerman, B., 1959: Characteristics of summer radar echoes in Arizona, 1956. The University of Arizona Institute of Atmospheric Physics Scientific Report, No. 11, 66 pp.

Balling, R. C., Jr., and S. W. Brazel, 1987: Diurnal variations in Arizona monsoon precipitation frequencies. Mon. Wea. Rev., 115, 342- 346.

Banta, R. M., and C. B. Schaaf, 1987: Thunderstorm genesis zones in the Colorado Rocky Mountains as determined by traceback of geosynchronous satellite images. Mon. Wea. Rev., 115, 463- 476.

Braham, R. R., 1958: Cumulus cloud precipitation as revealed by radar- Arizona 1955. J. Meteor, 15, 75-83.

Jurwitz, L. R., 1953: Arizona’s two season rainfall pattern. Weatherwise, 6, 96- 99.

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

Orville, H. D., 1965: A photogrammetric study of the initiation of cumulus clouds over mountainous terrain. J. Atmos. Sci., 22, 700- 709.

Watson, A. I., R. E. Lopez, and R. L. Holle, 1994: Diurnal cloud-to-ground lightning patterns in Arizona during the southwest monsoon. Mon. Wea. Rev., 122, 1716- 1725.

Zhang, J., 2000: Three-dimensional mosiac displays of data from multiple WSR-88D Radars. In these proceedings.