2. Pre-Storm Link to Cyclonic Intensification
A recent climatological study conducted by Cione et al. (1993) investigated cold season (November-April) coastal cyclonic episodes from the period 1982 through 1990. The major goal of this research was to investigate the potential effects Gulf Stream-induced low-level baroclinicity had on subsequent cyclonic intensification within the US mid-Atlantic coastal zone. To determine the effects of marine boundary-layer baroclinicity directly associated with the presence of the Gulf Stream off the Carolinas and Virginia, the region between 38° N and 32° N and 79° W and 72° W was chosen as the storm domain for the Cione et al. (1993) study (see Figure 1). The region highlighted in Figure 1 was selected so that storms entering this region would be subjected to the highly variable baroclinic zone associated with the lateral meanders of the Gulf Stream (Pietrafesa and Janowitz, 1980). It should be noted that all cyclonically oriented, surface low pressure systems analyzed with a closed isobar entirely contained within the study domain were considered 'storm events'. All east coast winter storms between 1982 and 1990 that remained within this region for a period exceeding 6 h were analyzed. Past storm track information was retrieved from the National Centers for Environmental Prediction (NCEP) North American surface weather maps via the National Climatic Data Center (NCDC). A 13 year, digitized (1978 through 1990) once-to-twice weekly data set was used to depict regional SST conditions within the study domain. Since detailed SSTs were only available in nine of these thirteen years dating back to January 1982, only storms occurring after 1982 were included in this study.
In order to accomplish the research objective linking pre-storm low-level baroclinicity with cyclogenic intensification, low-level baroclinic indicators were established. These indicators were devised as a means to ascertain quantitatively the average near-surface thermal contrasts present between the SSTs near the GSF and the cold continental air advecting off the NC coastline during the wintertime pre-storm period. A measure of the low-level pre-storm baroclinicity was obtained using the near-surface air temperatures at coastal locations and the corresponding SST of the GSF. Cape Hatteras and Wilmington, NC were the coastal observation sites used in the Cione et al. (1993) study. Near-surface air temperatures were obtained at hourly intervals at both locations. The air temperatures were averaged during the offshore cold advective period. After combining the averaged surface air temperatures with the SST at the western boundary of the GSF, the average pre-storm air-sea temperature contrast was calculated. In addition to the average pre-storm air-sea temperature contrast, an average low-level pre-storm thermal gradient was calculated by using the distance between the GSF and the coastal stations of Cape Hatteras and Wilmington, NC. The gradient was included so as to incorporate the potential relative impact GSF position has on the low-level thermodynamic modification of the offshore advecting airmass.
It should be noted that SSTs were assumed to be constant throughout the period over which coastal
surface air temperatures were averaged. Storm intensification was taken to be the observed surface
central pressure decrease experienced by a cyclone located in the study area. The normalized units for storm intensification are taken to be mb 12 h-1.
Results from this winter storm climatology revealed that the pre-storm baroclinic nature of the coastal mid-Atlantic region was strongly linked to the subsequent development of regional coastal cyclones. Figure 2 illustrates a linear regression of the total pressure change of the surface cyclone dependent (normalized to a 12 h period) on the pre-storm baroclinic index (PSBI). The results from this 116 storm study illustrate a statistically significant geometric mean of the regression coefficients (r) of 0.562 (at the 0.01% level) between the pre-storm baroclinicity and future (i.e., 24-48 h) storm
development that explains 31.6% of the total variance. Much of the scatter, or unexplained variance (i.e.,1 - r2) observed in Figure 2, is a result of other important contributing cyclogenic processes that were not included in this research.
Also evident in Figure 2 is a solid relationship between the PSBI and dp/dt for weaker intensifying cyclonic events (i.e., <12 mb (12 h)-1). In fact, stratification of storms exhibiting surface intensification rates less than 12 mb (12 h)-1 illustrates a statistically significant r value of .698 (at the 0.01% level) between the PSBI and dp/dt, accounting for 48.7% of the total variance (not shown). In addition, it can be shown from Figure 2 that for the 38 storms that experienced average PSBI values less than - 1.0 °C (10 km)-1 only one (i.e., 2.6%) was observed to rapidly intensify (i.e., dp/dt at least 12 mb (12 h)-1). Of the remaining 37 cyclones, 36 exhibited rates of cyclonic intensification no greater than 8 mb (12 h)-1. From these results, it appears that the likelihood for rapid cyclonic intensification (i.e., 'bombs') when the PSBI is less than approximately 1 °C (10 km)-1 (or more conservatively 0.9 °C (10 km)-1) is highly unlikely even when other contributing factors potentially impacting upon surface cyclonic intensification are not explicitly taken into account. Also from Figure 2 we see that for a PSBI greater than 1.7 °C (10 km)-1, 12 of the 16 storms (75%) were observed to 'bomb' (i.e., dp/dt at least 12 mb (12 h) -1).