[Weather] Weather measurement algorithms (Was: Tipping BucketRainfall Rate Algorithm)

Karl Uppiano Karl.Uppiano at verizon.net
Fri Dec 22 15:10:31 EST 2006

I think I used exactly the same documents you located, and cobbled together 
an algorithm for reporting wind behavior. A while back, I stumbled on a 
document that stated that the standards are changing. The resolution used to 
be three seconds for polling wind speed. Now with faster electronics and 
data channels, the trend seems to be going to one second. There is an 
averaging window implied by these polling resolutions, so the historical 
data must be corrected to reflect this change.

I cannot remember where I saw this... BTW, I tried the links below, and 
could not connect. Maybe it's me...

----- Original Message ----- 
From: "Paul Davis" <pdavis at mrv.com>
To: "List for 1 Wire Weather Stations and devices" <weather at buoy.com>
Sent: Friday, December 22, 2006 8:21 AM
Subject: [Weather] Weather measurement algorithms (Was: Tipping 
BucketRainfall Rate Algorithm)

I just spent the bulk of yesterday afternoon searching for similar
information. I was looking for the exact specification for reporting a
wind gust, what constitutes a gust, and how it's measured. There is a
document "Federal Standard Algorithms for Automated Weather Observing
Systems" which presumably has the algorithms used for various purposes
(at least in a METAR context). I could not find it anywhere. The best I
could come up with was two other documents. The 'Federal Meteorological
Handbook' <http://www.ofcm.gov/fmh-1/pdf/fmh1.pdf.> and the 'ASOS User
Manual' <http://www.nws.noaa.gov/asos/pdfs/aum-toc.pdf>. Between these
two I found what I wanted regarding wind measurement. I just reviewed
them and see no reference to rainfall rate. You may find them useful

While on the subject of rainfall, I found it interesting to read in the
ASOS User Manual in section 3.4.2 - Precipitation Accumulation
Algorithm, that there is a correction factor applied to their tipping
bucket measurements. C = A(1 + .60A) where C is the computed rainfall
and A is the amount measured from the tipping bucket. Their rain gauges
apparently under report with as little as 1.8 inches/hr. Since under
reporting seems to be a characteristic of tipping bucket rain gauges, I
expect that a correction factor should also be applied to our rain
gauges as well. Given the variations in design and the dynamics of the
bucket mechanisms, I think we'd have to establish a correction factor
for each manufacturer.

If someone comes across that "Federal Standard Algorithms for Automated
Weather Observing Systems" please post a link.


-----Original Message-----
From: weather-bounces+pdavis=mrv.com at buoy.com
[mailto:weather-bounces+pdavis=mrv.com at buoy.com] On Behalf Of Karl
Sent: Friday, December 22, 2006 2:23 AM
To: weather at buoy.com
Subject: [Weather] Tipping Bucket Rainfall Rate Algorithm

The worldwide web seems to be completely devoid of information about how
calculate rainfall rate from a tipping bucket rain gauge (total
is a piece of cake). Furthermore, neither agency that I report to, CWOP
Weather Underground, have any specifications about to calculate rainfall

rate. So I made my own specification.

The rainfall task records the number of bucket tips since the previous
sample. The current rate is calculated by the following formula:

    rate = (msec / hr) * (btcount / imsec)

where btcount is the number of bucket tips, and imsec is the interval in

milliseconds since the last sample. The constant msec / hr = 3600000.
Milliseconds cancel, giving an instantaneous rate in units of bucket
per hour. This is similar to the way instantaneous speed in miles per
is calculated without having to wait a full hour to determine the speed.

Each bucket tip represents 0.01 inch, but rate can be converted into any

units desired.

For example, if btcount is 1 and imsec is 60000 (one minute), the
rate becomes 3600000 / 60000 = 60 bucket tips = 0.6 inches per hour.
This is
the rate resolution of the rain gauge.

Since I report to the weather services every 10 minutes, I decided to
average the results using a 10 minute sliding window. For example, if I
record one bucket tip every minute for ten minutes, the average rainfall

rate will be 0.6 inches per hour.

Without any standards and practices or examples used by various
agencies, I have no way to know whether this is a valid algorithm or
not. If
anybody knows a better way, please let me know.

Weather mailing list
Weather at buoy.com

More information about the Weather mailing list