Geoscience Reference
In-Depth Information
a
4
Precip Squared Bias
24h
48 h
3
2
1
0
A
B
C
D
E
F
G
H
I
b
60
50
Precip MSE
24h
48h
40
30
20
10
0
A
B
C
D
E
F
G
H
I
c
60
24h
48h
50
Precip SD
40
30
20
10
0
A
B
C
D
E
F
G
H
I
d
40
35
24h
48h
Precip Ratio of Squared Bias Vs MSE
30
25
20
15
10
5
0
A
B
C
D
E
F
G
H
I
Fig. 25.4
Same as Fig.
25.2
but for precipitation (mm/day) forecasts. Unit: square of precipitation
amounts (mm/day) in (
a
), (
b
), (
c
) (Adapted from
Xu et al.
(
2009
))
Over SWA domain, the MSEs of 10-m wind speed in 24-h forecasts correspond
fairly well to the low SD errors (Fig.
25.7
b, c). The forecast errors from systematic
error in the western mountains of Iran are relatively large values (Fig.
25.7
a). The
10-m wind speed statistical fields are quite different from the 2-m temperature fields,
and the nonsystematic model errors compose a much larger portion of the total
forecast errors for 2-m temperature forecasts (Fig.
25.7
d). The 48-h forecast errors
are similar to the 24-h forecast errors in most of the areas (not shown).
The above results suggest that the MSEs near the surface contain a substantial
spatial heterogeneity, as seen by the relatively larger errors in higher mountainous
areas. However, the source of the errors indicates a significant difference among
temperature, precipitation, and wind speed. The inaccuracies in 2-m temperature
forecasts are mainly from systematic errors, which are controlled largely by the
physical representation in the model. In contrast, the inaccuracies in precipitation
and 10-m wind speed forecasts are dominated more by nonsystematic errors,
Search WWH ::
Custom Search