Geoscience Reference
In-Depth Information
a
b
(mm)
(mm)
(m/s)
(m/s)
c
d
(mm)
(mm)
(m/s)
(m/s)
Fig. 20.7
(
a
) Ensemble mean and (
b
) spread of 1 h rainfall (
shaded regions
), horizontal wind at the
height of 20 m (
vectors
) and sea level pressure (
contours
) at 15 JST 5th September 2008 obtained
by performing the assimilation without the Inner LETKF. (
c
,
d
) Enlarged distribution of spreads
that are shown in Fig.
20.7
b (without the Inner LETKF) and
20.5
b (with the Inner LETKF).
Arrows
in (
c
)and(
d
) indicate the regions in which spreads of sea level pressure and rainfall were larger,
respectively
wind distributions of the ensemble member #005 that were obtained by deploying
4 Inner LETKFs side by side. In this experiment, the configuration of the Inner
LETKFs was the same as the aforementioned one except that their horizontal grid
number was changed to
. In the reproduced distributions, the rainfall
regions and horizontal wind varied smoothly near the boundary of the Inner
LETKFs and the unrealistic convection cells were not generated there. It is deduced
that the blending of the analyzed values of the Outer and Inner LETKFs at the
boundaries of Inner LETKFs reduced the generation of unrealistic convection cells.
121
121
20.4.3
Impact of No-Cost Smoother
Next, other improvements of the nested system are explained in this section. The
aforementioned experiment of the nested system will be called “CNTL” hereinafter.
In CNTL, the initial seeds and boundary conditions of the Inner LETKF were
Search WWH ::
Custom Search