|
|
The ensemble
information at an individual grid-point location may be displayed
through a probabilistic meteogram, which indicates the time
evolution of a given parameter for all ensemble members. The spread
is indicated by the range of forecast values. 50% of the members
are distributed evenly around the median to define a vertical
rectangle. The remaining members define the extreme 25% "spikes".
The epsgram thus provides a discrete probability information in the
intervals 0-25%, 25-50%, 50-75% and 75-100% which is sufficient for
many applications.
The deterministic TL511
and TL255 forecasts may be included as a reference, as could the
ensemble mean. To help the assessment of useful predictability the
typical climatological variance might also be included. It is often
useful to know when the spread might cover most of the
climatological range.
As stressed in chapter
6.7 users should be careful when interpreting direct model output.
Only the temperature forecasts in the epsgrams can be corrected for
differences between true height above the sea level and the model
height (according to the Standard atmosphere). Locations along seas
borders often suffer from large systematic errors, in particular
with respect to temperature. The reason is that the strong
temperature differences between land and sea may be poorly resolved
by the TL255 resolution.
- The forecast
uncertainty, as indicated by the EPS spread, is changing with the
parameter, it is usually larger for cloud cover, and it usually
increases with the forecast step. However uncertainty is also flow
or feature dependent and there might be cases, as this EPSgrams
example shows, where forecast uncertainty is larger at shorter
forecast steps. In this case uncertainty in the positioning of a
surface low has resulted in a large spread in wind and
precipitation.
|
