Spring is typically a time of greater error in NWP forecasts, as the atmosphere/land/ocean system transitions from winter to summer. Ensemble forecast systems usually show this error in terms of increased uncertainty in the forecast, particularly beyond the first 3 days. This particular case involves two forecast verification times, 12 UTC April 11, 2002 and 12 UTC April 16, 2002, from two MRF runs separated by one day (from 00 UTC 8 and 9 April).
The verification for 12 UTC April 11 2002 is shown in the graphic below. The verification will be compared to model forecasts for the same valid time from 00 UTC 8 April 2002 through 00 UTC 9 April 2002, at 6-hour intervals.
The 00 UTC 8 April through 00 UTC 9 April 2002 forecasts for days 3 and 4 (48-96 hours) were for significant deepening of a short-wave trough in the western US. The operational forecast for 84 hours can be seen by clicking on the first "display graphic" button below.
Operational
forecasts valid 12 UTC 11 April 2002
Consecutive MRF forecasts are represented by the first and last frames of theanimation. After the animation has gone through one cycle, click on the "STOP" button (the middle button in the "loop advance" column of buttons). The first frame shows the forecast from 00 UTC 8 April 2002 indicating a deepening 500-hPa short-wave trough in the intermountain west, appearing to be ready to cut off within 6-12 hours. A lee-side low is developing over the Central Great Plains. Inflow from the Gulf of Mexico is beginning, with strong south winds inferable from the strong pressure gradient existing from OK north-northeastward into south-central Canada. One could project a potential severe weather situation developing for the next day over the central and southern Plains.
Clicking the top button in the "Loop Advance" group will bring you to frame 5, from the MRF forecast initiated 24 hours later. The 500 hPa trough, instead of deepening and preparing to cut off as in the earlier forecast, is now positively sheared, weaker, and further east than previously forecast. Surface inflow as inferred from the MSLP gradient is weak and directed toward old Mexico, rather than the US Great Plains. A much weaker surface low and trough are found further east than before, centered over MN rather than eastern CO and the NE panhandle. This situation seems much less volatile with respect to severe weather than the previous MRF scenario.
If you click on the bottom button in the "Loop Advance" group, you'll step through the consecutive AVN/MRF forecasts at 6-hour intervals. As is often the case with western US trough development, the sharpness of the upstream short-wave ridge on or just off of the west coast of North America is a key determinant for downstream development. At 00 UTC on the 8th, the upstream ridge is amplified, probably as a result of the forecast development of a strong trough in the east-central Pacific.
As we step forward in time, the next forecast valid at 12 UTC 11 April weakens
the intermountain West short-wave and upstream ridge near the west coast, and
begins the process of deepening the central Pacific trough further west. The 12
UTC forecast continues this process, developing a strong Gulf of Alaska storm, but
weakening the features downstream over the western US even more. By 18 UTC 8 April,
the Pacific trough is broken into two separate features, with emphasis on the
further west of the two, while the lead system is further flattening the west coast
ridge and downstream trough, now found over the western US Great Plains with a weak
surface reflection. So rather than a sudden change in the forecast, we note that
there was a transition from a strong cut-off western trough to a weak and
progressive system that occurred over the course of consecutive forecasts separated
by 6 hours. Note that each forecast followed a data assimilation system cycle, and
that the consecutive MRFs were separated by four assimilation cycles! So,
there was plenty of opportunity for new data to get both ingested into the model
and influence the model forecast over consecutive data cycles.