Below an example is given for demonstrating the procedures used for taking adaptive observations over the northeast Pacific with the goal of improving downstream forecasts over the continental US. The example pertains to a possible development of a low pressure system over the northeastern US, verifying around 021200. The 5-day ensemble mean mean sea level pressure  forecast chart indicated an elongated low pressure system extending northeast, with maximum normalized spread values up to 2.5 times larger than their usual values at this lead time near the southern tip of Lake Michigan (orange colors). The ensemble mean forecast over the same area called for 1016 hPa pressure level. The MRF control forecast gave a prediction of 996 hPa; the ensemble spread around the ensemble mean in the area was around 10 hPa (green colors). The verifying analysis shows a value around 1024 hPa. The large ensemble spread (10 hPa) correctly indicated that much lower or higher values than the 1016 hPa ensemble mean were plausible. Fast growing atmospheric instabilities, given the current hemispheric observing network, did not allow for more accurate predictions for this particular area at 5-day lead time.

Given that the 5-day forecast initialized at 020700 gave around 50% probability of more than half inch of precipitation associated with this storm, HPC requested that adaptive observations be taken in order to reduce the large uncertainty associated with the above forecast feature (see attached text below). Due to the long advance notice required by Air Traffic Control before dropsondes can be released from any aircraft, flight planning for a 020900 flight had to be prepared on 0207. Given the verification times and locations in the HPC request, sensitivity calcuations were carried out to determine the area from where data taken around 020900 can most inmfluence the selected forecast feature. The sensitivity chart indicates that on this day a large area extending from 175E to 120W, and from 35 to 50N is highly relevant for downstream forecasts verifying around 40N, 87W, 3 days later. The 2-day forecast initialized at 020700 indicated that the two most sensitive areas are associated with a very deep low pressure system west of the dateline, and a small, shallow low pressure wave around 140W. Of the two maxima reachable from Anchorage with the available NOAA G-lV plane, the eastern one (around 37N, 140W) was estimated more important. The sensitivity technique also provides an estimate how the data can be expected to influence the forecasts. The estimated data impact  chart should be evaluated only in a qualitative way. It indicates the impact of the data would travel slowly to the east, reaching the west coast within 36 hrs; and larger impact over the eastern half of the continent can be expected around 72-84 hrs lead times. We can compare the predicted data impact chart to the actual impact of the data, determined as the difference between two independently run analysis/forecast  cycles. The actual signal evolved similarly to what was predicted and the data had a large impact in the preselected verification areas/times. Much of this impact was positive. The red contours on the comparative verification chart indicate that forecast errors were reduced due to the use of the adaptive data, while the blue contours over smaller areas indicate an increase in forecast errors. Overall, the 72-hour forecast initialized using all operationally available - including targeted - data,  provided a much improved guidance compared to the 120-hour forecast, indicating that there would be two low pressure systems following each other along similar paths, and that the verification area around 021200, around 40N, 87W would be dominated by a high pressure system between the two low pressure systems.


       Reconaissance Areas from 0000 UTC Feb 7 2000
       Mon, 07 Feb 2000 09:08:23 -0500
       Stephen Flood <>


Here are todays's key areas. They may look a bit redundant, but you can
vary by as much as a day a flight with the same weather system to
maximize logistics.

Priority            VT             Area               Remarks

High             021112         40N/95W       Central US  system highly
medium range models.

High             021212        45N/75W       Same system as above...24
hrs later.

Medium       021300        40N/120W     Broad area of  spread  NOAM W

Medium       021300        50N/135W     Second max within same broad

021200    42N    125W    west coast precip event     (low)
021400    42N    97W    possible second low pressure wave affecting eastern US     (low)