WMOD00535 - Laserfiche WebLink

Home Browse Search

3. MESOSCALE CONTROLS OF HEAVY PRECIPITATION. TWO EXAMPLES The general circulation in the Northern Hemisphere from January to May 1983 was much more intense than normal with a large number of heavy winter storms affecting the Western United States. These storms were associated with the polar jet stream, which was stronger than normal and displaced several degrees south of its normal position in the U. S.. Early in the winter, several severe storms pounded the west coast resulting in precipitation exceeding 200% of normal (Willhite et aI., 1987). In late winter, the storm track continued to penetrate the Western U.S. producing a series of winter storms that caused heavy [ate season snowfalls in the central Rocky Mountains. This precipitation combined with an unusually cool spring, followed by sudden warming throughout the Rockies and locally heavy showers in June, to produce record streamflows on the Colorado River that resulted in widespread flooding and property damage (Rhodes et al., 1984). Analvsis of the 700- and 300-mb level National Meteorological Center (NMC) synoptic charts from January to June clearly shows a persistent strong extratropical jet that passed across southern California and northern Mexico from late February through mid-April and slowly shifted northward into Arizona and Colorado in late April and May 1983. This pattern placed the mean long-wave trough over the central Rockies from March to May. MM4 simulations for this period accurately reproduced the mesosynoptic scale features controlling wmt~r .storms and precipitation. Comparisons of MM4 predicuons and NMC analyses indicated that the location of the jet stream core, wind speeds and directions all verified reasonably well as shortwave perturbations passed through the long-wave trough. Model-simulated 24-hr precipitation patterns generally matched the locations of observed precipitation and relative amounts; however, the model tended to have more precipitation than observed in the synoptic-scale data. Model quantities matched the observations from high elevation stations better than the NMC data, which generally agreed with the amounts observed at lower elevations. Giorgi et al. (1992) found that the correlation between observed and model-simulated monthly mean precipitation for the West was quite good, nearly 0.8 for the 40 months analyzed. Conditions from 21-30 April produced a series of storms that resulted in heavy precipitation in the mountains of Utah and Colorado. This period illustrates the typical structure of 1983 spring storms. The long-wave trough was located over the Southwest with a very intense jet stream pattern south of the central Rockies. Figure 2 shows the 300- mb wind vectors at 0000 UTC on April 21, 1991 during the period of heaviest mountain precipitation. Note the 64 m s-1 jet maximum over the Arizona-Mexico border. This moist southwesterly flow from the Pacific Ocean supplied large amounts of water for precipitation from the surface to 300 mb across the southern and central Rockies. ~imil~ patterns were c?rrectly diagnosed by the MM4 si.mulauons on 40 occaSlOns from March to April 1983. Flgure 3 shows the 24-hr precipitation ending at 1200 UTC 21 April 1983. Note the 1-2 cm of precipitation associated with this storm in Utah and southern Colorado. High elevation observations in southwest Colorado indicated that from 0.5 to 3.8 cm of precipitation fell during this storm. The extratropical jet axis remained south of Colorado throughout March and all but 2 days in April. During May, the jet gradually moved northward across Colorado as the long-wave flow became more zonal. However, during this period, precipitation continued to occur nearly every day at " .>\ ~ )"-: ....... -..-......+...... -. .................h......_...._.......n...... .... Fig. 1. To]POgraphy used in the MM4 simulations over the Western United States. high elevaltion stations, with daily accumulations often exceeding LO rom each day. From 14-18 May the long-wave trough reestablished over the Southwest producing heavy precipitation in the central Rockies. The Gunnison Basin received from 0.3 to 3.8 crn of precipitation each day during this period. Figure 4 shows the 300-mb airflow at 0000 UTC on 14 May, when the initial perturbation moved into the Colorado Rockies. This produced from 0.8 to 2.5 cm of precipitation on May 14 and 15th. Figure 5 shows the 24-hour precipitation simulated by MM4 ending at 1200 UTC 14 May. These quantities generally matched the precipitation observed at high-elevalion sites in the Gunnison Basin. By 17 May the storm produced the heaviest precipitation as the jet stream moved northward over Colorado and the storm center moved onto the High Plains. Figure 6 shows the transition of flow as ridging occurred in the Pacific Northwest and the jet moved northward over Colorado at 0000 UTC 17 May. Precipitation was locally heavy with over 3.8 em measured at high elc:vations and 2.5 cm at lower elevations in the mountains. Figure 7 shows the 24-hr precipitation simulated by MM4 elllding at 1200 UTC 17 May 1983. Four weaker short-wave perturbations passed through the Rockies from 20 May - 14 June 1983. These systems also produced heavy precipitation on 21 and 29-30 May and 6-7 and 11-14 June 1983. MM4 simulated these systems allld predicted locally heavy precipitation in the central Rockies. Cumulative precipitation plots from 1 October 1982 to 30 June 1983 clearly indicated the unusually heavy late winter and spring precipitation in the Colorado Rockies. Figure 8 shows the time series plot of cumulative rnean precipitati.on observed at 31 stations in southwestern Colorado including 10 high elevation (9000-11440 ft msl) sites. The: cumulative precipitation in 1982-83 after mid- March (dashed line) was significantly larger than that in the drier 1988-89 (solid line) period. The two examples of storm conditions discussed above are marked by arrows labeled 1 (21 April) and 2 (14-17 May). Note the 40 rom increase in precipitation contributed by the May storm. Model-simulated cumulative precipitation averaged over 8. grid points in southwestern Colorado are marked by m for 2