Geoscience Reference
In-Depth Information
4.4.1 Wet versus dry hours discrimination
From Table 4, we can see that on the one hand the number of Missing Detections (22) is
relatively high: hence POD is relatively small 0.65; FAR is relatively small: 0.089. Finally,
HSS
results to be
0.587
, which is by chance exactly the same as HK .
G 0.4 mm/h G < 0.4 mm/h
R 0.4 mm/h 41 4 45
R < 0.4 mm/h 22 59 81
63 63 126
Table 4. Contingency table for the Torino radar and 3 gauges (42-hour observation period).
2000
2500
Terrain Profile
Radar
Gauge
Terrain Profile
Radar
Gauge
1800
1600
2000
1400
1200
1500
1000
1000
800
600
400
500
200
0
0
0
0.5
1
1.5
2
2.5
x 10
4
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
x 10
4
Fig. 15. As in Fig. 6, but for the Torino radar. Left: Castagneto Po-Torino Radar profile; right:
Ciriè-Torino Radar profile.
Distance [m]
Distance [m]
4.4.2 Quantitative agreement between gauge and radar-derived hourly rain rates
During the 63 “Wet-Gauge hours”, the total rainfall amount measured by the three gauges
was 119.7 mm. The corresponding radar-derived amount was 108.0 mm. Based on these
“
Wet-Gauge
” 63-hourly amounts the Bias is 0.4 dB. As can be clearly seen from Fig. 16, the
agreement between the radar and the 3 gauges is quite poor. This fact is obviously reflected
in the amazingly large value of the Scatter, which is as bad as 5.38! dB.
Fig. 17 shows the 63 “Wet-Gauge hours” hourly amounts as measured by the gauges at the
ground and as derived from radar echoes aloft: the large scatter between such different
devices and their different sampling modes is again evident.
Using only the 41 hours where BOTH the Radar AND the Gauge amounts were larger or
equal to 0.4 mm/h, the total rainfall amount measured by the three gauges is reduced to
94.5 mm. The corresponding radar-derived amount remains instead almost the same: 106.2
mm. Consequently, the “wet-wet” Mean Field Bias increases and becomes even positive:
+0.5 dB (radar overestimation). Also the “wet-wet” Scatter improves remarkably: based on
such 41 “wet-wet” hours, it results to be 3.73 dB, which is still a figure much worse than the
one obtained for the 2 radars in Sicily. Such huge Scatter value decrease (high sensitivity to
different radar thresholds) is again a clue of the poor QPE agreement for the Torino radar.
