BACKGROUND: The statistics of updrafts and downdrafts were substantially different between tropical/subtropical and midlatitude continental cumulus convection. The Thunderstorm Project (Byers and Braham 1949) provided the only statistics for midlatitude continental convection.
Recent aircraft observations also suggested that the averaged thermal buoyancy in downdrafts was positive and similar to that of updrafts. Updrafts with negative buoyancies were also frequently observed. These observations revealed that decelerating drafts commonly occurs within convective systems.
MOTIVATION: There were a few observational studies of the draft statistics of tropical and midlatitude cumulus convection. All aircraft observations were restricted to the lower and middle troposphere (Table 1). The lack of complete draft statistics has diminished the values of these studies for improving cumulus parameterizations.
OBJECTIVES:
1) to provide a complete set of the draft statistics for any height; and
2) to compare the similarities and to contrast the differences in the draft statistics between tropical and midlatitude cumulus convection.
APPROACH: Use of outputs generated from cloud-resolving simulations of tropical and midlatitude cumulus convection, imposed with observed large-scale advective forcings during GATE and July 1995 IOP of ARM program (Xu and Randall 1996, 1999).
METHODOLOGY: Strong convective cores with absolute vertical velocities over 1 m/s and total condensate mixing ratios over 0.1 g/kg are sampled for the statistical analysis. An average is taken if two or more adjacent grid points satisfy the criteria.
SAMPLE SIZES of this study are much greater than any aircraft measurement (Table 1). The sample size is greatly dependent upon the height; bimodal for GATE and nearly unimodal for ARM (Fig. 1).
COMPARISON with OBSERVATIONS: Draft statistics are comparable to observations from the recent tropical experiments (Fig. 2), but much smaller than those of the Thunderstorm Project (not shown), which only sampled the convective towers during their growth phases. Figure 2 shows the 90 percentiles of updraft and downdraft velocities for both ARM and GATE simulations, as well as the GATE observations by LeMone and Zipser (1980).
THERMAL BUOYANCY distributions (10 percentile, medium, and 90 percentile) are rather similar between updrafts and downdrafts except for smaller shifts in values.
TOTAL CONDENSATE mixing ratio distributions are rather different between updrafts and downdrafts. Figure 4 shows the 10 percentile (blue), medium (green), and 90 percentile (red) of total condensate mixing ratios for ARM simulations only.
LOADING EFFECTS have the largest impact on the strongest updrafts. The convective available potential energy (CAPE) reduction due to the loading effects are shown in Table 2. The CAPE values for updrafts are shown in parenthesis of Table 2.
THE STRONGEST 1% AND 10% DRAFTS shows the largest difference between tropical and midlatitude cumulus convection. Selected results are shown in Fig. 5, which corresponds to the averages of the strongest 1 and 10% updrafts, respectively.
The CAPE and convective inhibition (CIN) values for the strongest 1% and 10% updrafts are shown in Table 3, which also provides the CAPE reduction due to the loading effects. The difference between tropical and midlatitude convection is substantial.
NARROW VS. WIDE DRAFTS: The comparison between drafts with diameters of 2 km and greater than 2 km shows the importance of less diluted, larger convective cores. Figure 6 shows the 90 percentiles of vertical velocities and 10%, medium and 90% of updraft buoyancies for ARM simulations.
1. The medium drafts are rather similar between tropical and midlatitude cumulus convection.
2. The strongest 10% of midlatitude drafts are much stronger than their counterparts in the Tropics.
3. Decelerating drafts are abundant in both the tropical and midlatitude convective systems.
4. Wider drafts are much stronger than narrow drafts.
5. It is time for the ARM program to conduct aircraft measurements of convective drafts to replace the only outdated, midlatitude draft statistics by Byers and Braham (1949).
Acknowledgments: This research has been supported by DOE grant DF-FG03-95ER61968 as a part of the ARM program.