Reanalysis of Validation Data for the 2009 Convective Season

Pryor (2009) presented validation results for the GOES Microburst Windspeed Potential Index (MWPI) for the 2009 convective season over the U.S. southern Great Plains. Further statistical analysis of a dataset built by comparing wind gust speeds recorded by Oklahoma Mesonet stations to adjacent MWPI values for 35 downburst events has yielded some favorable results. Correlation has been computed between key parameters in downburst process, including temperature lapse rate (LR) and dewpoint depression difference (DDD) between two levels (670mb/850mb), CAPE, and radar reflectivity (Z). The first important finding is a statistically significant negative correlation (r=-.34) between lapse rate and radar reflectivity, as shown in Figure 1. Similar to the findings of Srivastava (1985), for lapse rates greater than 8 K/km, downburst occurrence is nearly independent of radar reflectivity. For lapse rates less than 8 K/km, downburst occurrence was associated with high reflectivity (> 50 dBZ) storms. The majority of downbursts occurred in sub-cloud environments with lapse rates greater than 8.5 K/km. Adding the dewpoint depression difference to lapse rate yielded an even greater negative correlation (r=- .42) when compared to radar reflectivity, as demonstrated in Figure 2. Finally, comparing the sum of LR and DDD (the former hybrid microburst index (HMI)) to CAPE resulted in the strongest negative correlation (r=-.82), with a confidence level above 99%. This emphasizes the complementary nature of the HMI and CAPE in generating a robust and physically meaningful MWPI value. This result also shows that CAPE can serve as an adequate proxy variable for precipitation loading (expressed as radar reflectivity) in the MWPI . The strong negative correlation, or negative functional relationship, between HMI and CAPE terms in the MWPI algorithm indicates that the MWPI should be effective in capturing both negative buoyancy generation and precipitation loading as downburst forcing mechanisms.

Figure 1. Scatterplot of lapse rate versus radar reflectivity.

Figure 2.Scatterplot of the sum of lapse rate and DDD versus radar reflectivity.

Figure 3. Scatterplot of the sum of lapse rate and DDD versus CAPE.

References

Pryor, K.L., 2009:Microburst windspeed potential assessment: progress and recent developments.arXiv:0910.5166v1 [physics.ao-ph]

Srivastava, R.C., 1985: A simple model of evaporatively driven downdraft: Application to microburst downdraft. J. Atmos. Sci., 42, 1004-1023.

A New Look at the 2004 Baltimore Water Taxi Accident

During the afternoon of 6 March 2004, the "Lady D", a water taxi servicing the Baltimore Harbor, was capsized in a strong convective windstorm, resulting in five deaths. A cluster of convective storms developed over north-central Maryland during the afternoon, and then tracked rapidly eastward through the Baltimore Harbor and upper Chesapeake Bay between 2050 and 2120 UTC. Radar wind velocity measurements from Dover AFB NEXRAD between 35 and 45 knots were associated with a cluster of downbursts as the convective storm complex approached the Baltimore Harbor. About two hours prior to the water taxi accident, the GOES imager microburst product, derived from 1847 UTC sounder data, indicated output BTD 35 to 36K over western Baltimore City in proximity to the location of the generation of the downbursts. Previous validation has identified that BTD greater than 35K corresponds to wind gust potential greater than 35 knots. According to news reports, the Lady D capsized between 2050 and 2100 UTC (see Baltimore Sun article).


Figure 1. GOES imager microburst product derived from 1847 UTC 6 March 2004 sounder data. Radar reflectivity and velocity imagery from DOVER AFB NEXRAD at 2048 UTC are overlying the microburst product image.

Figure 1 shows the 1847 UTC GOES microburst product with overlying radar imagery from Dover AFB NEXRAD. Figure 1 displays several features associated with high downburst wind gust potential: Output BTD greater than 35K (orange shading) over western Baltimore City, and radar wind velocity of 35 to 45 knots (green shading) surrounding a high reflectivity (red shading) convective storm over downtown Baltimore near the inner harbor. Between 2050 and 2100 UTC, the Lady D capsized in the Baltimore Harbor due to the strong convective winds. The cluster of convective storms then continued to track rapidly southeastward over the Chesapeake Bay, moving over the Eastern Shore at Tolchester Beach, Maryland. Although the high winds were not verified in surface observations in proximity to the Baltimore Harbor, an NOS PORTS station recorded a wind gust of 48 knots at Tolchester Beach about 20 minutes later at 2118 UTC. This high wind report was the result of a downburst, clearly marked in a wind histogram from Tolchester Beach PORTS station in Figure 2.



Figure 2. Wind histogram from Tolchester Beach PORTS station on 6 March 2004 (top) and GOES imager microburst product at 1946 UTC with overlying radar reflectivity from Dover AFB NEXRAD at 2117 UTC.

The 1946 UTC microburst product indicated high risk values downstream of the convective storm over Tolchester at 2117 UTC. Maximum output BTD over the Delmarva Peninsula of 35 to 37K indicated wind gust potential of 35 to 37 knots. The measured wind gust of 48 knots at Tolchester signifies that the rapid forward motion of the storm as well as precipitation loading, with radar reflectivity greater than 55 dBZ, were also factors in the magnitude of the wind gust associated with the downburst at Tolchester Beach. Based on radar velocity and the measured wind gust at Tolchester, wind gusts of 35 to 45 knots were likely with the convective storm complex as it moved over the Baltimore Harbor. The GOES imager microburst product two hours prior to the event would have been useful in assessing convective wind gust potential and, perhaps, may have provided guidance in issuing more timely warnings.

An Early November Downburst on the Chesapeake Bay

During the afternoon of 5 November 2009, an upper-level disturbance triggered the development of scattered convective storms over the mid-Atlantic region. Solar heating of the lower atmosphere during the early afternoon fostered conditions for strong convective storm downdrafts that were more typical of the Great Plains region. A convective storm that developed west of Baltimore, Maryland during the late afternoon produced a strong downburst as it tracked from the Chesapeake Bay eastward into the Delmarva Peninsula. An associated wind gust of 38 knots was recorded at Tolchester Beach, Maryland at 2242 UTC. The 1800 UTC Geostationary Operational Environmental Satellite (GOES) imager microburst product, derived from brightness temperature differences obtained from the GOES-11 imager (in full disk mode), indicated high downburst risk in proximity to Tolchester Beach.

Figure 1. GOES-11 imager microburst product at 1800 UTC 5 November 2009 with radar reflectivity from Dover Air Force Base NEXRAD at 2237 UTC overlying the image (top) and RUC model analysis sounding over Tolchester Beach, Maryland at 1800 UTC (bottom).

Figure 1 compares the Geostationary Operational Environmental Satellite (GOES)-11 imager microburst product to a corresponding Rapid Update Cycle (RUC) model sounding over Tolchester Beach, Maryland at 1800 UTC, 5 November 2009. At 1800 UTC, the GOES-11 image product showed a large area of high downburst risk (red shading) over southern New Jersey and the Delmarva Peninsula where strong solar heating was destabilizing the lower atmosphere. Output brightness temperature difference (BTD) in this region was greater than 40°K, indicating wind gust potential greater than 40 knots based on a previously established statistical relationship. Overlying radar reflectivity imagery from Dover Air Force Base NEXRAD at 2237 UTC displayed the downburst-producing convective storm as a spearhead echo over Tolchester Beach. At 2242 UTC, the Tolchester Beach Physical Oceanographic Real-Time System (PORTS) station recorded a wind gust of 38 knots. The corresponding RUC sounding profile echoed favorable conditions for downbursts in the Tolchester Beach area with the presence of a 5000-foot deep mixed layer and steep temperature lapse rate in the lower atmosphere. These boundary layer conditions promoted strong downdraft generation due to the effects of precipitation loading, evaporational cooling, and subsequent generation of negative buoyancy.

October 2009 Chesapeake Bay Downburst

A week after the passage of an intense coastal storm and associated winter-like conditions, a more spring-like regime over the Chesapeake Bay region resulted in a significant downburst event. During the afternoon of 24 October 2009, a rain band with embedded convective storms developed ahead of a strong cold front over the Blue Ridge Mountains. Similar to the 15 November 2008 event, temperatures were above normal ahead of the cold front, especially over the Delmarva Peninsula. A convective storm on the eastern flank of the rain band produced a strong downburst at Tolchester Beach, Maryland. The GOES imager microburst product, derived from from both the GOES-12 sounder and GOES-11 imager (full disk mode), indicated favorable conditions for downbursts three hours prior to the event.

Figure 1 compares the GOES-12 sounder and GOES-11 imager products. At 1746 UTC, the sounder derived product image displayed a small break in the large frontal cloud band over the upper Chesapeake Bay and northeastern Maryland. Brightness temperature difference (BTD) between 30 and 35K indicated wind gust potential of 30 to 35 knots. By 1800 UTC, the GOES-11 image product showed a small break in the frontal cloud band over the Chesapeake Bay near Tolchester. Again, output BTD in this region ranged from 30 to 35K, indicating wind gust potential of 30 to 35 knots. Overlying radar reflectivity imagery from Dover Air Force Base NEXRAD at 2059 UTC displayed the downburst-producing convective storm as a spearhead echo over Tolchester Beach, with reflectivities greater than 50 dBZ. At 2100 UTC, the Tolchester Beach PORTS station recorded a wind gust of 39 knots, well-marked in the wind histogram in Figure 2. In a similar manner to the 15 November 2008 event, steep low to mid-level temperature lapse rates as inferred by elevated BTDs and precipitation loading as inferred by high radar reflectivity in the parent storm favored strong downdraft instability.The RUC sounding profile displayed in Figure 2, over Tolchester Beach at 1800 UTC, suggests that downward momentum transport was also a factor in downburst generation with winds near 40 knots near the top of the mixed layer at 955 meters above the surface.

Figure 1. GOES-12 sounder microburst product at 1746 UTC 24 October 2009 (top) and GOES-11 imager microburst product at 1800 UTC (bottom). Radar reflectivity from Dover AFB NEXRAD at 2059 UTC is overlying both images.

Figure 2. Wind histogram for Tolchester Beach PORTS station (top) and RUC sounding profile over Tochester at 1800 UTC 24 October 2009 (bottom).

Comparing product images in Figure 1 reveals that the GOES-11 imager product in full disk mode provides a higher spatial resolution and a more precise display of output BTD than that produced by the GOES-12 sounder. Thus, the GOES-11 imager product may provide forecasters with more useful guidance pertaining to downburst risk over the Chesapeake Bay region. More information pertaining to downburst activity over the Chesapeake Bay region is available in this paper.