118
C.G. Collier
rather that its presence and the fraction of enhanced echoes close to the O°C or brightband altitude be known.
Analytic methods based on using radar data alone. Procedures described by
Harrold and Kitchingham (1975), Koistinen (1991), Gray (1991), Andrieu and
Creutin (1995) and Andrieu et al. (1995), derive an average reflectivity profile by
analysing data from several radar beam elevations at ranges up to a few tens of
kilometres from the radar. The average profile is then used to correct data from longer
ranges. Assumptions of spatial homogeneity are necessary, both in the derivation of
the profile and in its application to data at a longer ranges.
Smith (1986) devised an analytic technique for reducing errors due to the bright
band which proved very promising in tests. In routine operation, however, it was
found to be susceptible to large errors when significant variations in the freezing level
occurred over the area covered by a radar. Such variations are common in frontal
precipitation in the UK. (Colour Plate 6.C)
Physically-based methods using independent meteorological data. Austin (1987)
(see also Dalezois and Kouwen, 1990, Fabry et aI., 1992) recommended physicallybased techniques. One such has recently been described by Kitchen et al. (1994). An
idealised reflectivity factor profile was constructed from analysis of radar data. The
heights of significant turning points in the profile are diagnosed from relevant
meteorological data (for example surface temperatures) at each radar pixel. The
parameterised profile is weighted by the radar-beam power profile and the surface
precipitation rate found by an iterative method in real-time.
Since this technique can exploit a wide range of meteorological information from
conventional observations and forecast models to derive the reflectivity factor profile
Kitchen et al. (1994) felt that the technique is better than those using analyses of radar
data along. This is undoubtedly true, although the strong dependence on other
meteorological observations and model data makes the procedure vulnerable to the
availability and accuracy of these data.
A somewhat different approach has been described by Hardaker et al. (1995). A
microphysical model is used to calculate the reflectivity profile using as input lapse
rate information from either radio-sondes or a mesoscale numerical model. Simple
relationships between bright-band intensity and surface rainfall rate were derived
which may be useful in the absence, in real-time, of extensive computing facilities or
independent meteorological data. This approach enables improvements in the stylised
bright band profile to be made, and tests using numerical model temperature profiles
as input to a correction procedure showed encouraging results.
6.3.2 Measurement of snowfall and hail
Snowflakes make deviate significantly from the spherical shape assumed in the
estimation of Eq. 6.1. It may be shown (Smith, 1984) that
C.G. Collier
rather that its presence and the fraction of enhanced echoes close to the O°C or brightband altitude be known.
Analytic methods based on using radar data alone. Procedures described by
Harrold and Kitchingham (1975), Koistinen (1991), Gray (1991), Andrieu and
Creutin (1995) and Andrieu et al. (1995), derive an average reflectivity profile by
analysing data from several radar beam elevations at ranges up to a few tens of
kilometres from the radar. The average profile is then used to correct data from longer
ranges. Assumptions of spatial homogeneity are necessary, both in the derivation of
the profile and in its application to data at a longer ranges.
Smith (1986) devised an analytic technique for reducing errors due to the bright
band which proved very promising in tests. In routine operation, however, it was
found to be susceptible to large errors when significant variations in the freezing level
occurred over the area covered by a radar. Such variations are common in frontal
precipitation in the UK. (Colour Plate 6.C)
Physically-based methods using independent meteorological data. Austin (1987)
(see also Dalezois and Kouwen, 1990, Fabry et aI., 1992) recommended physicallybased techniques. One such has recently been described by Kitchen et al. (1994). An
idealised reflectivity factor profile was constructed from analysis of radar data. The
heights of significant turning points in the profile are diagnosed from relevant
meteorological data (for example surface temperatures) at each radar pixel. The
parameterised profile is weighted by the radar-beam power profile and the surface
precipitation rate found by an iterative method in real-time.
Since this technique can exploit a wide range of meteorological information from
conventional observations and forecast models to derive the reflectivity factor profile
Kitchen et al. (1994) felt that the technique is better than those using analyses of radar
data along. This is undoubtedly true, although the strong dependence on other
meteorological observations and model data makes the procedure vulnerable to the
availability and accuracy of these data.
A somewhat different approach has been described by Hardaker et al. (1995). A
microphysical model is used to calculate the reflectivity profile using as input lapse
rate information from either radio-sondes or a mesoscale numerical model. Simple
relationships between bright-band intensity and surface rainfall rate were derived
which may be useful in the absence, in real-time, of extensive computing facilities or
independent meteorological data. This approach enables improvements in the stylised
bright band profile to be made, and tests using numerical model temperature profiles
as input to a correction procedure showed encouraging results.
6.3.2 Measurement of snowfall and hail
Snowflakes make deviate significantly from the spherical shape assumed in the
estimation of Eq. 6.1. It may be shown (Smith, 1984) that