IMPROVED RAINFALL RATES FROM POLARISATION RADAR

1
IMPROVED RAINFALL RATES FROM POLARISATION
RADAR IS THE EXTRA INFORMATION FROM ZDR REAL?
Anthony Illingworth, IAC, Exeter, 7 April 2004
ZDR
Z
2
Simple Z-R relationship

• Z 200R1.6
• This is equivalent to assuming constant drop
  concentration, but as rain gets heavier the mean
  size of the drops increase.
• In fact drop concentrations do change so the
  sinple Z-R relationship is in error.
• Z ND 6 R N D 3.67

3
Z and Zdr give us N and D

• Zdr 10 log(Zh/Zv) gives us drop shape and hence
  size, D.
• Suppose Z is constant.
• N higher than normal, so drops smaller, Zdr below
  normal, so R higher than normal for this Z
• N lower than normal, so drops larger, Zdr higher
  than normal and so R lower than normal for this Z.

4
PLOT OF Z AND ZDR AT EACH GATE FOR PPI
Black line const N Z 200 R1.6 Is spread
around line due to changes in drop spectra? Red
line is for 1mm/hr, What is R for each Z/Zdr? Z
and R both scale with N, So if Z is xdB gtred
line Then R is xdB gt1mm/hr.
High N
x
x
Low N
5
FOR A GIVEN Z WE OBSERVE A SPREAD OF ZDR IS
THIS REAL INFORMATION ON DROP SPECTRA?

• NATURAL CHANGES IN ZdR DUE TO CHANGES IN DROP
  SPECTRA ARE QUITE SMALL
• Zdr IS QUITE DIFFICULT TO MEASURE.
• Any imperfections in our system (e.g. antenna)
  could lead to an artificial spread in values of
  Zdr for a given Z.
• If this is so, then using Z and Zdr may give
  worse values of of R than from Z alone.

6
FOR EACH 5dB SPREAD IN Z, CALC R FOR ALL THE
Z/ZDR POINT, FIND MEAN SPREAD OF R
e.g. Z from 35 to 40dBZ 14,000 data
points Average R 10.8mm/hr Average fractional
change for each point ? 38. i.e average R error
from Z alone is 38, use of Zdr reduces this.
High N
Low N
7
IS THIS ZdR SPREAD REAL OR INSTRUMENTAL??
Lee and Zawadski calculate Z and R from use one
minute observed spectra. For given Z individual
spectra have an R 32 different from average R
for that Z. For Chilbolton data the average
spead or R for a given Z for this scane is Z
from 20-25dBZ 1mm/hr 35 to 40dBZ
25-30dBZ 2.5mm/hr ?36
30-35dBZ 5.1mm/hr ?36 35-40dBZ
10.8mm/hr ?38 40-45dBZ 26.6mm/hr ?29
45-50dBZ 50.9mm/hr ?
26 CONSISTENT WITH DISDROMETER
Quality check for data is Zdr spread correct?
8
Conclusion

• A study of the observed spread of Z and ZDR at
  each gate from the new Kent radar would provide a
  useful check on the data quality.
• It would confirm if the values of ZDR being
  measured reflect the true variability in mean
  raindrop size or are artificially broadened by
  some radar instrumental artefact.
• This should be done as soon as data in rainfall
  is available.