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 580

a b c

CNOP CNOP CNOP

37N

12h

37N

24h

37N

36h

31N 31N 31N 10

8
6

10
8
6

10
8
6 25N 25N 25N

4 4 4

19N

2

19N

2

19N

2

13N 13N 13N

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

Fig. 24.9 TC Meari (2004). The vertically integrated energies of CNOP ( shaded , units: J/kg) and
the 500 hPa stream fields for the second approach. ( a ) forecast starts at 12 h ahead, (b) forecast
starts at 24 h ahead, and ( c ) forecast starts at 36 h ahead (From Zhou and Mu 2012b )

a b c

CNOP CNOP CNOP

24h 36h

12h

39N 39N 39N

10
8
6
4
2

10
8
6
4
2

10
8
6
4
2

33N 33N 33N

27N 27N 27N

21N 21N 21N

15N 15N 15N

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

Fig. 24.10 Same as Fig. 24.9 , but for TC Matsa (2005) (From Zhou and Mu 2012b )

24.4 Examination of the CNOP Sensitive Areas

In Sect. 24.3 , we have studied the properties of the CNOP sensitive areas. In this
section, we will examine the efficiency of the CNOP sensitive areas with a lot of
cases by carrying out the sensitivity tests, OSSEs, and OSEs.

24.4.1 Sensitivity Tests

In this part, random initial errors are introduced into the CNOP sensitive areas, and
their impacts on the TC forecasts are explored ( Chen and Mu 2012 ). Two tropical
cyclones, Longwang (2005) and Sinlaku (2008) are studied. For comparison, the
roles of the random initial errors in the areas identified by FSV and CSV, and the
randomly selected areas are considered. Based on these four areas, experiments are
designed to determine which area have the greatest impact on TC forecasts.