The time and location of the dropsonde missions will be selected adaptively in order to minimize the error
in selected forecasts associated with possibly large societal impact and large forecast uncertainty. The
weather events for which forecasts are to be improved through enhanced initial conditions provided by the
dropsonde data will be selected by the Hydrometeorological Prediction Center of NCEP, after
consultations with NWS field offices and other NCEP service centers.  The sensitive areas where extra
data can reduce most the expected forecast error for the selected weather events at a given forecast lead
time will be determined by the Environmental Modeling Center of NCEP, collaborating with PSU and MIT

Because of constraints due to Air Traffic Control  requirements, the flight plans need to be prepared 24-36 hrs in advance.  So typically, forecasters identify potentially threatening and uncertain weather features in 72-96 hours lead time forecasts. The sensitive areas are then identified at 36-48 hrs lead time. The extra data collected at that lead time should then lead to an enhanced forecast of the weather event of interest in the 24-48 hrs time range.

As an example, let us assume that today is 990111, and that the latest forecast runs available are from the 00Z cycle  (that was the case when the first reconnaissance flight was planned). HPC identified low pressure waves reaching the northwest coast with potentially large amount of precipitation, with a center verification location of 47.5 N, 125W, at an estimated verification time of 99011412 (84 hrs lead time). Based on this information, EMC, EMC ran the ensemble transform sensitivity calculations based on the NCEP 99011100 ensemble. Assuming that the flight takes place at 99011300 (48 hrs lead time), the most sensitive area is centered around 35N, 165W.  Based on the sensitivity chart, a predesigned flight track (North Pacific #16) was selected, with planned dropsonde locations indicated on the sensitivity chart. The flight request had to be filed with CARCAH, as usual, before 18:30 UTC time.

A day later, in the morning of 990112, the sensitivity calculations were repeated, now with the 99011200 ensemble (with a 24-60 hrs time window between targeted observations and verification times. The updated sensitivity pattern has not changed substantially so no adjustments were made to the original flight request.

The plans call for ten flights with USAF Reserve C-130 planes out of Anchorage, Alaska, and
approximately nine flights with the NOAA G-Vl aircraft, operated by NOAA AOC, out of Honolulu, Hawaii.
The first and last possible flight days for the C-130s are January 12, and February 11, while those for the
G-Vl are January 15 and February 14, respectively. The dropsonde data will be transmitted through regular channels (GTS) in real time.