Page History

...

Usually the different evolution can be traced day-by-day through the forecast sequence back to the the source area.  Ideally the reason for these differences can be identified in the source area and some assessment can be made of the reason (e.g. incorrect observations, active convection and MCSs etc).  In this way the more likely evolution may be tentatively accepted.   However, sometime this is not possible.

Image RemovedImage Added

Fig4.2-10: The plot shows the difference in 500hPa ensemble mean height initialised DT 00UTC 26 October 2024 (black) and DT 00UTC 28 October 2024 (red).  In the 26 October forecast run, VT T+192 01 Nov the forecast has a trough over Scandinavia.  In the 28 October forecast run, VT 28 October, VT T+144h 01 Nov the forecast has something like a weak ridge.  This is highlighted by a large area of negative (purple) difference.  The error can be traced back to north-eastern U.S.A and eastern Canada at DT 00UTC 26 October, T+48h where compared with DT 00UTC 28 October, T+0.  In this case it is difficult to identify severe weather in the model or in reality, or any significant analysis increments over that part of U.S.A.

The jump between the forecasts has a major impact on the forecast surface weather in Scandinavia.  The forecast of DT 00UTC 26 October develops an intense low over Finland at T+192 VT Fri 1 Nov bringing strong winds and cold air with a cold snap over southern Scandinavia.  The forecast of DT00UTC 28 October develops a broader low and more zonal winds over southern Scandinavia with more neutral conditions.

Image RemovedImage Added

Fig4.2-11: Comparison between the forecast from DT 00UTC 26 October 2024 (black) and the corresponding analyses (red).  The forecast from DT 00UTC 26 October 2024 (black) failed to capture the weak ridge over Scandinavia shown by a large negative (purple) difference from the analysis.  The forecast of DT 00UTC 26 October 2024 was effectively a bust.

...