311
Radar Polarimetry for Rain Estimation
estimator apparently overestimates the rainfall rate for convective cores, because the
Z h value has been significantly enlarged by a few large drops and/or the contamination of melting hail. Because an increase in drop size also causes an increase in Z dr ,
the R(Z h ,Z dr ) estimator effectively reduces the overestimation of R(Z h ). The R(K DP )
estimator, less sensitive to the hail contamination, also has a smaller but more reasonable estimation than R(Z h ) in this region. However, the uncertainty of the K DP
measurement is much larger than Z h or Z dr . It is shown in Figure 13.2c that in a lightrain region, R(K DP ) has a worse rain estimation than the other estimators.
One inconvenience of an empirical approach is that empirical relations are not
identical for different radar platforms. There have been many polarimetric radar–
rain relations for S-, C-, and X-band applications (Zhang et al. 2001; Brandes et al.
2002; Bringi and Chandrasekar 2001; Matrosov et al. 2002; Cao 2009). Due to the
diversity of radar frequency, these relations cannot match the frequency of every
radar platform exactly. Characterization error of radar variables could be introduced
in the polarimetric estimator and would consequently increase the rain estimation
error. Another inconvenience of the empirical approach is that a set of relations needs
to be derived to estimate other rain variables of interest. Similarly, those relations are
100
75
50
30
20
15
10
5
3
1
0.1
100
75
50
30
20
15
10
5
3
1
0.1
100
75
50
30
20
15
10
5
3
1
0.1
100
75
50
30
20
15
10
5
3
1
0.1
150
100
50
–50
–100
–150 –100
0
Zonal distance (km)
Meridional distance (km)
100
0
(a) R(Z h ), mm h –1
–100
0
Zonal distance (km)
Meridional distance (km)
100
(b) R(Z h , Z dr ) mm h –1
150
100
50
–50
–100
–150 –100
0
Zonal distance (km)
Meridional distance (km)
100
0
150
100
50
–50
–100
–150
0
150
100
50
–50
–100
–150
0
(c) R(K DP ), mm h –1
–100
0
Zonal distance (km)
Meridional distance (km)
100
(d) DSD retrieval, mm h –1
FIGURE 13.2 Rain retrievals using different methods: (a) R Z
Z
( )
.
h
h
0.714
= 0 017
; (b) R Z
( ,
h
Z
Z Z
dr
h
dr
)
.
.
.
= 0 0142
0 77
1 67
−
; (c) R K
K
(
)
DP
DP
0.822
= 44
; and (d) DSD retrieval.
Radar Polarimetry for Rain Estimation
estimator apparently overestimates the rainfall rate for convective cores, because the
Z h value has been significantly enlarged by a few large drops and/or the contamination of melting hail. Because an increase in drop size also causes an increase in Z dr ,
the R(Z h ,Z dr ) estimator effectively reduces the overestimation of R(Z h ). The R(K DP )
estimator, less sensitive to the hail contamination, also has a smaller but more reasonable estimation than R(Z h ) in this region. However, the uncertainty of the K DP
measurement is much larger than Z h or Z dr . It is shown in Figure 13.2c that in a lightrain region, R(K DP ) has a worse rain estimation than the other estimators.
One inconvenience of an empirical approach is that empirical relations are not
identical for different radar platforms. There have been many polarimetric radar–
rain relations for S-, C-, and X-band applications (Zhang et al. 2001; Brandes et al.
2002; Bringi and Chandrasekar 2001; Matrosov et al. 2002; Cao 2009). Due to the
diversity of radar frequency, these relations cannot match the frequency of every
radar platform exactly. Characterization error of radar variables could be introduced
in the polarimetric estimator and would consequently increase the rain estimation
error. Another inconvenience of the empirical approach is that a set of relations needs
to be derived to estimate other rain variables of interest. Similarly, those relations are
100
75
50
30
20
15
10
5
3
1
0.1
100
75
50
30
20
15
10
5
3
1
0.1
100
75
50
30
20
15
10
5
3
1
0.1
100
75
50
30
20
15
10
5
3
1
0.1
150
100
50
–50
–100
–150 –100
0
Zonal distance (km)
Meridional distance (km)
100
0
(a) R(Z h ), mm h –1
–100
0
Zonal distance (km)
Meridional distance (km)
100
(b) R(Z h , Z dr ) mm h –1
150
100
50
–50
–100
–150 –100
0
Zonal distance (km)
Meridional distance (km)
100
0
150
100
50
–50
–100
–150
0
150
100
50
–50
–100
–150
0
(c) R(K DP ), mm h –1
–100
0
Zonal distance (km)
Meridional distance (km)
100
(d) DSD retrieval, mm h –1
FIGURE 13.2 Rain retrievals using different methods: (a) R Z
Z
( )
.
h
h
0.714
= 0 017
; (b) R Z
( ,
h
Z
Z Z
dr
h
dr
)
.
.
.
= 0 0142
0 77
1 67
−
; (c) R K
K
(
)
DP
DP
0.822
= 44
; and (d) DSD retrieval.