Geoscience Reference
In-Depth Information
Table 6 summarizes the dispersion of the radar-gauge errors using the hydrological-oriented
score called
Scatter
(see Sec. 4.1.2). Since, as discussed previously, in the radar detection
process (see for instance the “multiplicative” nature of the meteorological radar equation
derived by
Probert-Jones
[1962]), the multiplicative nature of error prevails, the Scatter is
defined as a ratio between the Radar (the device under test) and the Gauge (the reference).
Hence, dry hours cannot be considered in evaluating the Scatter (unless using some trick,
like for instance adding a negligible amount…) Consequently, only “Wet-Gauge” hours or
“wet-wet” hours are considered in Table 6.
It can be concluded that the three-presented X-band radars are less reliable at low rain rates.
By limiting the observations to hours with both Radar and Gauge amounts larger than 0.4
mm/h, the agreement improves not only in terms of Bias, but most of all in terms of Scatter.
Finally, in the interpretation of these values of Bias and Scatter, it is important to bear in
mind the large intrinsically different sampling modes as well as mismatches in time of the
radar and gauge devices (e.g.
Zawadzki
[1975]).
Site “Wet-Gauge”
Scatter
“Wet-Wet”
Scatter
Palermo radar
2.38 dB
(48 h)
1.97 dB
(38 h)
Bisacquino radar
1.47 dB
(37 h)
1.31 dB
(29 h)
Torino radar
5.38 dB
(63 h)
3.73 dB
(41 h)
Table 6. QPE evaluation summary in terms of Scatter for three mini-radar sites. The number
of hours of each data set are given in parentheses.
5. Open issues and limitations
Short-wavelength (X-band) radar has the benefit of attaining high spatial resolution with a
smaller antenna. However, there is a clear disadvantage compared to longer wavelengths: an
increased attenuation in the presence of precipitation. Imagine three 2-km convective cells with
instantaneous rain rate of 20, 40 and 100 mm/h respectively: at X-band frequencies these cells
would cause two-way attenuation in radar reflectivity values of approximately 1.5, 3.6! and 11!!
dB. Such figures preclude not only the use of X-band radar for long-range monitoring but also
an accurate QPE, even at short-range. A partial remedy could be the use of polarimetric
information, but this would remarkably increase the cost of the system: in the interesting work
by
Mc Laughlin et. al
[2009], the cost of a Doppler, fully Polarimetric advanced X-band system
developed within the framework of the CASA project is estimated in approximately 180 kEuro,
which is almost 8 dB more expensive than our low-cost, semi-quantitative approach.
6. Summary: Filling the gap, which is observing the lowest part of the
troposphere at short-range with portable, low-cost radars
Radar sampling volume increases with the square of the range (beam broadening) therefore,
at longer ranges, small but intense features of the precipitation system are blurred (non-
homogeneous beam filling). Furthermore, it is more likely to include different types of
hydrometeors (e.g. snow, ice, rain drops), especially in the vertical dimension. At long-
range, because of the decreased vertical resolution, the lower part of the sampling volume
can be in rain whereas the upper part can be even characterized by an echo weaker than the
radar sensitivity itself (beam overshooting).
