Click here for
catalog of sonde drops.
Click here for zipped post-processed sonde data.
1. Synoptic situation
Tropical Storm Alex developed from an easterly wave just west of the Cape Verde Islands on 27 July, and moved westward, strengthening slowly. Due to the potential threat to the Virgin Islands and Puerto Rico, a synoptic surveillance mission was tasked for nominal time 1 August 1998 0000 UTC. At that time, Alex was embedded in the westerlies to the south of the subtropical ridge about 1000 km east of Martinique, and a strong upper-level cold low was located to the northwest, just to the northeast of the Leward Islands (Fig. 1). The upper-level feature had started to shear Alex, leading to dissipating by the next afternoon.
2. General Assessment of dropwindsonde impact
A. GFDL model
Figure 2 shows the GFDL forecast tracks for Tropical Storm Alex, and Table 1 shows the errors and impact of the synoptic surveillance mission. The control forecast was very good, and the dropwindsonde data helped the forecast only at 36 h. The upper-tropospheric data degraded the forecast at 12 h and improved it by 36 h.
Figure 3 shows the VICBAR forecast tracks for Tropical Storm Alex, and Table 2 shows the errors and impact of the synoptic surveillance mission. The dropwindsondes had only slight impact on the VICBAR forecast track, with a small improvement at 12 h, and a slight degradation at 36 h. The upper-tropospheric data degraded the forecast at both 12 and 24 h. This last result can be expected in cases in which the tropical cyclone is weakening due to strong vertical shear since the tropical storm likely is advected by the mid- or lower-layer mean flow than the deep-layer-mean of the VICBAR model.
Figure 4 shows the GSM forecast tracks for Tropical Storm Alex, and Table 3 shows the errors and impact of the synoptic surveillance mission. The dropwindsondes had a largely negative impact on the GSM forecast track. The upper-tropospheric data degraded the forecast at 12 and 36 h. Burpee et al. (1996) showed that the GSM responded favorably to the relatively uniform data distribution in the original HRD synoptic flow experiments. However, in Alex, the dropwindsonde observations improved the GSM forecasts only slightly at 24 h. Like in the 1997 synoptic surveillance missions (Aberson and Franklin 1999), the asymmetric sampling logistically required for the Alex mission is probably not the optimal sampling strategy to improve GSM forecasts, and is the most likely explanation for the discrepancy between this result and those of Burpee et al. (1997). On the other hand, asymmetric sampling was able to produce somewhat better mixed results in the GFDL and VICBAR models.
Table 4 shows the GFDL intensity forecast errors and impact of the synoptic surveillance mission. The dropwindsondes had a positive impact after 12 h, and the upper-tropospheric data seems to have only helped through 24 h. Tuleya and Lord (1997) also showed modest improvements to GFDL model intensity forecasts in the HRD synoptic flow cases.
The atmosphere has long been recognized as a chaotic system (Lorenz 1963), e.g. very small perturbations to initial conditions result in increasingly large differences in the evolution of the atmosphere with time. Since the exact state of the atmosphere can never be measured, all analyses contain errors whose magnitudes can only be estimated. An indeterminate number of initial conditions consistent with the observational data can therefore be used in numerical weather prediction, and single model runs at any synoptic time only give one possible solution to the evolution of the atmosphere. Many operational forecast centers around the world, therefore, now employ ensemble forecasting as a means of quantifying the uncertainty in the evolution of the atmospheric system. Small perturbations from a best "control" state are calculated and added to and subtracted from this control to allow for different integrations starting from theoretically equally likely initial states. These perturbations are designed so as to mimic the fastest growing modes in the model and to create the largest envelope of possibilities in the forecast. Therefore, they generally correspond to locations where large analysis errors will most impact the forecast. Those features corresponding to perturbations which most impact the tracks of tropical cyclones must be found, properly sampled, and thoroughly tested, to prove the efficacy of targeting techniques.
Figure 5 shows the variance of the size of the perturbations in the National Centers for Environmental Prediction (NCEP) global model ensemble forecasting system (Toth and Kalnay 1993). The large perturbation to the north of Alex corresponds to the upper-level cold low. The slightly larger perturbation to the west of the cold low is related to the strength of the western extent of the deep-layer subtropical ridge. Due to logistical constraints, only the feature closer to Alex, the cold low, was adequately sampled during the mission.
Two sets of model runs have been performed. The first, the TG run, includes only the dropwindsonde data taken within and around the cold low (the ten dropwindsondes extending from near 20N 60W to 27N 58W, or about one-third of the entire mission), and the other, the NT run, includes the complement of the first. Results are shown in Tables 1-4 and Figs. 2-4. The TG run provides better forecasts at almost all forecast times in all three models than the runs with either all the dropwindsonde data or none of the dropwindsonde data included.
Figure 6 shows the difference in the 850 - 200 hPa averaged winds between the runs in which all the dropwindsonde data and none of the dropwindsonde data are included. The largest difference is to the south of Alex, in an area of small ensemble perturbation (Fig. 5). However, by 24 h into the forecast (Fig.7), this large difference has decayed. The largest difference at 24 h corresponds to the cold low to the north of Alex despite the initial condition difference being very small there. This confirms that the largest (smallest) ensemble perturbations correspond to amplifying (decaying) modes in the model. Therefore, the dropwindsondes surrounding the cold low are expected to make the largest positive impact on the model forecast. The large initial condition difference to the south of Alex, despite decaying, seems to have pushed the forecast slightly to the south, making the forecast slightly worse than the forecast without those dropwindsonde data. The differences, however, are small.
The dropwindsonde data obtained during the synoptic surveillance mission for Tropical Storm Alex at nominal time 1 August 1998 0000 UTC has provided mixed results. In general, the forecasts without the dropwindsonde data were quite good, and therefore difficult to improve. The MRF ensemble forecasting system suggested that data obtained in and around the cold low to the northwest of Alex would be the most important in the subsequent forecast. Forecasts using only these data mainly showed more substantial improvements than those using all the dropwindsonde data.
Return to Table of Contents
|GFAL Error (km)|
|GFP3 Error (km)|
|GFTG Error (km)|
|12||46.1||49.2||( -7%)||39.5||( 14%)||39.5||( 14%)|
|36||89.4||69.5||( 22%)||76.2||( 15%)||44.4||( 50%)|
|VBAL Error (km)|
|VBP3 Error (km)|
|VBTG Error (km)|
|12||77.6||77.6||( 0%)||71.9||( 7%)||71.9||( 7%)|
|24||77.2||84.0||( -9%)||77.2||( 0%)||77.2||( 0%)|
|GSAL Error (km)|
|GSP3 Error (km)|
|GSTG Error (km)|
|12||67.4||77.7||(-15%)||67.4||( 0%)||69.9||( -4%)|
|24||123.9||122.9||( 1%)||129.2||( -4%)||91.3||( 26%)|
|36||167.6||191.4||(-14%)||172.4||( -3%)||152.1||( 9%)|
|GFAL Error (kn)|
|GFP3 Error (kn)|
|GFTG Error (kn)|
|12||11||10||( 9%)||10||( 9%)||9||( 9%)|
|24||15||0||(100%)||5||( 67%)||14||( 7%)|
Return to 1998 assessment page
Return to HRD home page