WMOD00583

• Contract/Permit #: 1-07-81-V0175
• Title: Some Laboratory Tests of the Accuracy of the Belfort Universal Recording Raingage - Technical Report
• Date: 3/1/1985
• Weather Modification - Doc Type: Report

IV. Summary and Conclusions

The results of the laboratory tests that have been presented
appear to support the following conclusions concerning the
Belfort weighing precipitation gages used.

,)

OVer periods of several hours, chart readings generally
indicated less precipitation than the measured simulated amount
dripped into the gages. The mean departure was approximately
0.02 inches with the largest departure 0.04 inches. These
departures were somewhat reduced by passage of time after
simulated precipitation, and markedly reduced by tapping of the
mechanism. Apparently some hysteresis existed in the mechanical
weighing mechanism. However, it was usually limited to 0.01 to
0.03 inches precipitation equivalent.

OVer periods of 1- to 3-hours, chart readings generally
under-estimated the simulated precipitation amounts, which ranged
from 0.001 to 0.12 inches. Almost 40 percent of the cases were
within -0.007 to -0.003 inches of the estimated amounts entering
the gages, while 81 percent were within -0.012 to +0.002 inches.
On a percentage basis, 54 percent of the cases had a chart
response within ~ 10 percent of the estimated simulated
precipitation amount, and 91 percent of the cases were within ~
30 percent.

This is considered remarkable performance in view of the
various errors possible. These include errors in reading the
charts (usually within 0.005 inch with the considerable care
taken), long-term drift of the mechanism (ranging over as much as
0.015 inch in the two 20-hour tests), mechanical hysteresis, and
probably other sources of error not considered such as
expansion/contraction of the chart paper due to changes in
temperature and humidity. In addition, errors were possible in
weighing the total simulated precipitation for the few- to
several-hour tests, and in estimating the test flow amounts in
the 1- to 3-hour tests, particularly where the flow rate varied
with time.

~

In summary, the results of the tests presented suggest that
the Belfort weighing gages, when carefully calibrated and with
care given to manually reading the charts, can provide accurate
precipitation data with low precipitation rates of 0.01 to 0.10
inch per hour. Errors should be limited to approximately 0.01 to
0.03 inch in most cases, and this could be reduced with an
orifice large.r than the standard 8- inch-diameter orifice (as is
often needed to minimize gage capping when measuring snowfall).

A note of caution is in order; it was not practical to
conduct many tests at below freezing[ temperatures because of
unusually warm weather during the tests. It would be worth
repea ting the tests at colder ambien.t temperatures to determine
whether accuracy is further degraded (e.g. increased hysteresis).

15