Forecasts for 6-7 January 2002

Operational forecasts

In advance of the 6-7 January 2002 snowfall over the interior northern Mid Atlantic States and New England, both the operational numerical weather prediction (NWP) models and the newly developed NCEP short-range ensemble forecasts (SREFs) were indicating that:

• The storm would track south of PA, NY, and New England.
• There would not be much precipitation where the air was cold enough for snow.

The 00 UTC 6 January 2002 operational AVN and Eta-12 precipitation forecasts versus verification for the 24 hours ending 12 UTC 7 January 2002 are shown in this graphic. Note that the AVN shows more precipitation further inland than the Eta-12 (which shows a relative minimum of less than 5-mm precipitation where there was an observed maximum. This run of the operational AVN indicated that most, if not all, of this precipitation in central PA would fall as snow (not shown), resulting in low-end winter storm warning criteria accumulations. However, the AVN was an outlier relative to the ensemble members, and as such would typically be treated as an unlikely forecast outcome, under normal circumstances.

The 06 and 12 UTC 6 January 2002 forecasts for both operational models can be seen by clicking on the graphics buttons below.

Eta-12 fcsts from 06 UTC 6 Jan. 2002	AVN fcsts from 06 UTC 6 Jan. 2002
Eta-12 fcsts from 12 UTC 6 Jan. 2002	AVN fcsts from 12 UTC 6 Jan. 2002

The 06 UTC 6 January 2002 Eta-12 forecast showed a relative precipitation minimum moving through central PA during the period from 12 UTC 6 January to 12 UTC 7 January 2002, with precipitation type indicated to be snow. Total precipitation predicted over the 24 hour period is less than 0.3" . Meanwhile, the AVN from this forecast cycle shows more precipitation over central PA (0.5-0.7") for the same period. Note should be made that the Eta-12 forecast to the northeast of PA for the night of 6-7 January was much better, with a maximum of precipitation in excess of 0.75" over 6 hours from 00-06 UTC on the 7^th, remarkably close to the verification. Equally impressive is the structure of the precipitation; the Eta-12 indicated a heavy mesoscale snowband would develop by 00 UTC on the 7^th, and move northeastward from the Poconos through the Catskills and into adjacent central New England.

The 12 UTC 6 January 2002 Eta-12 forecast developed the heavy mesoscale precipitation band 6 hours earlier, and further north and west, than the 06 UTC run, resulting in a model forecast of heavy snow over central PA (with heavy freezing rain diagnosed to the southeast from the Baldwin/Schictel algorithm). While the precipitation type was a problem in the southern parts of central PA, the location was almost perfect, with its movement to the northeast very well-predicted in time and space. Meanwhile, the 12 UTC AVN forecast for the same period showed a shift further eastward with its precipitation, giving central PA less precipitation than the previous 06 UTC run. It turns out that the initial conditions provided by the EDAS and GDAS, respectively, were at least a partial explanation for this difference. These initial states will be examined in the next page.

Ensemble forecasts

As to the ensembles, the table below links to animated graphics that show how successive SREF forecasts (09 and 21 UTC January 5, 09 UTC January 6) handled the sea level pressure, 3-hour accumulated precipitation, and precipitation type from 12 UTC January 6, 2002 through 12 UTC January 7, 2002. Note that all ensemble members and means are identifiable by the color coding in the legend at the top of each graphic.

Ensemble forecasts from 09 UTC 5 January 2002 cycle

Ensemble forecasts from 21 UTC 5 January 2002 cycle

Ensemble forecasts from 09 UTC 6 January 2002 cycle

The following trends can be noted in the SREF ensemble members:

• Initially, the SREF Eta had more of a tendency to develop a double low pressure center, with an inland center moving northeastward along the spine of the Appalachians as far north as southwestern VA by 21 UTC 6 January 2002. In later runs, the RSM also hinted at two low pressure centers.

• Both SREF models had a primary low tracking from AL to somewhere off the Mid-Atlantic coast from 12 UTC 6 January to 06 UTC 7 January 2002. There was a tendency for the low to be:
  1. Further north and west
  2. Stronger, and
  3. A bit faster

  in the later SREF runs, with an accompanying shift in the precipitation to the north and west as well.

• When the storm track and heavy precipitation was shifted to the north and west, the rain-snow line also shifted north and west, so that heavy precipitation was not collocated with snow. The ensemble means and individual ensemble members all indicate this trend.
  • For example, the 21 UTC 5 January 2002 RSM n1 and n2 members show the 1004 hPa contour encompasing a larger area, and extending further west, at 21 UTC 6 January 2002 than the other ensemble members, with the 0.2"/3-hr precipitation contour as far north and west as south-central and southeast PA. At the same time, the eastern edge of the categorical snow extends from the southwest corner of PA to the PA line well west of the "heavy" precipitation line as defined above. At 00 UTC on the 7th, these members show the eastern edge of the rain-snow line and the western edge of the heavy precipitation line for the RSM n2 barely touching.
  • The 09 UTC 6 January 2002 SREF ensemble mean and ensemble members, however, tended to show a merging of heavier precipitation with the southern edge of the snow area from northeastern PA through the central Hudson Valley of NY to VT.

None of the ensemble members or ensemble means from any of the ensemble runs indicate heavy snowfall for central PA, where amounts ultimately ranged from 8-14".