WMOD00399

• Title: WMO Training Workshop on Weather Modification for Meteorologists - Lecture Notes
• Date: 12/1/1979
• Weather Modification - Doc Type: Report

-~-

inversely with the time interval for the estimation or with the area for which the
comparison is made. Thus, for a period of six hours over an area of 50;km2 or for
two hours over 500 km2 this discrepancy is 13%.

I
2.25 Data obtained from a meteorological observatory in the Fe~eral Republic
of Germany where regular measurements of precipitation have been carried out since
1971 using a 3-cm radar, gives a RMS error in determining the amount of precipitation
of 22% for all cases when attenuation between the radar and the precipitation area
did not exceed 10 dB.

2.26 Research carried out over the last few years in Canada ha~ shown that a
10-20% accuracy can be achieved for radar measurements of precipitation under the
following conditions: resolution over an area of 100 km2; accumulation: time ~ 3 hours;
measurement distance < 100 km; raingauge density for calibration no less than one
instrument per 3000 km2, measurement frequency for radar data no less t~an six times
per hour and amount of precipitation no less than 1 mm/h.

2.27 During 1973 - 1975 scientists in the Soviet Union worked 6n the possible
accuracy of radar precipitation measurements using the constant coefficients of the
Z-R relationship. Estimates made on the basis of spectr~l measurements:show that the
discrepancies between radar and raingauge measurements which were causeq by variations
in the raindrop spectra for the precipitation amount collected during o~e rain
occurrence, should be characterized by an RMS value of no more than 25%. Moreover,
many researchers concerned with such data comparisons have come to the qonclusion that
the Z-R relationship on which the method is based has an extremely high :variability.
I

2.28 This substantial discrepancy between the estimates has necessitated
reconsideration of the question of the relative contribution of methodo~ogical and
instrumental errors to the radar measurement error. There is substantial evidence
that the instrumental errors in most cases give a much bigger contribution to the total
measurement error than was previously supposed. In the Soviet Union work carried out
during 1973-1975 an instrument, for continuous recording of the potential of a meteo-
rological radar with an accuracy of up to 0.7 dB, was attached to the radar instrument-
ation. This enabled the potential accuracy of radar measurements of precipitation to
be verified experimentally. The measurements were carried out in the M~scow district
within a radius of 100 km and in cells of 3x3 km2. After processing and comparison,
the results were averaged over the resolution elements of 10xlO km2. Cpmparisons
were then carried out with surface precipitation data at nine points of :the polygon.
The total number of measurements were: 1973 - 35 rain occurrences with ;a total
duration of 90 hours; 1974 - 50 rain occurrences with a total duration of 115 hours;
and 1975 - 49 rain occurrences with a total duration of 100 hours. According to
surface measurements, the precipitation amount accumulated over a rain dccurrence
varied from 0.5 to 15 mm. Precipitation intensity varied from the minimum recorded
level of 0.1 mm/hour to 67 mm/hour. Comparison of the two.types of measurement~ over
the given period shows that the discrepancy between ra~ar and surface-base~ est~mates
of the amount of precipitation over a rain occurrence ~s not of a systemat~c nature and
fluctuates between -60 and 140%. In 75% of cases the discrepancy does not exceed 40%

~