The Advances in Targeted Observations for Tropical Cyclone Prediction Based on Conditional Nonlinear Optimal Perturbation (CNOP) Method - Data Assimilation for Atmospheric, Oceanic and Hydrologic Applications - page 573

a b c

Matsa Matsa Matsa

30km 60km 120km

39N 39N 39N

1.2
1
0.8
0.6
0.4
0.2

1.2
1
0.8
0.6
0.4
0.2

1.2
1
0.8
0.6
0.4
0.2

33N 33N

27N

33N
27N

27N
21N 21N 21N

15N 15N 15N

111E 120E 129E 138E 111E 120E 129E 138E 111E 120E 129E 138E

d e f

Meari Meari Meari

37N 37N

30km 60km 37N 120km

1.2
1
0.8
0.6
0.4
0.2

1.2
1
0.8
0.6
0.4
0.2

1.2
1
0.8
0.6
0.4
0.2

31N 31N 31N

25N 25N 25N

19N

19N 19N

13N 13N

13N

111E 120E 129E 138E 111E 120E 129E 138E 111E 120E 129E 138E

g h i

Mindulle Mindulle Mindulle

30km 60km 120km

28N
24N
20N
16N
12N

28N
24N
20N
16N
12N

28N
24N
20N
16N
12N

1.2
1
0.8
0.6
0.4
0.2

1.2
1
0.8
0.6
0.4
0.2

1.2
1
0.8
0.6
0.4
0.2

114E 121E 128E 135E 114E 121E 128E 135E 114E 121E 128E 135E

Fig. 24.1 The vertically-integrated energies of CNOP: ( a - c )TCMatsa;( d - f ) TC Meari; ( g - i )TC
Mindulle. ( a , d , g ) at a resolution of 30 km, ( b , e , h ) at a resolution of 60 km, and ( c , f , i )at
a resolution of 120 km. The boxes indicate the verification areas. The “ ˚ ” symbol indicates the

initial position of the cyclone (From Zhou and Mu 2012a)

For each case, the sensitive areas identified using different resolutions are
different from each other (Fig. 24.1 ); however, common sensitive areas occur at
the three resolutions, and the sizes of the common areas are different from case
to case. In general, the sizes of common areas are bigger between sensitive areas
at the lower resolution. This can be deduced from the similarities of the energy
distributions between each resolution for the three cases (Table 24.1 ). Tor the three
cases, the similarities between the lower resolutions (60 and 120 km) are greater
than those between the finer resolutions (30 and 60 km); moreover, this illustrates
that more small-scale activity would be resolved at higher resolutions.
From the analysis of the similarities, it can be induced that the sensitive areas
identified at lower resolutions are also helpful for improving the forecast at finer
resolution. However, to get the largest improvement at a high resolution, it is better
to use the sensitive areas identified at the same resolution. A resolution at which the
nonlinearity could be explored has been suggested to be used in the identification
of the sensitive areas. Generally speaking, because usually the forecasts at higher
resolutions are better than those at lower resolutions, and the high resolutions are