Snow fall = percipitation?

[dalecoy]

I agree on all points.

xylotik asked "who can reverse engineer the Davis console to use that little icon?" and wanted to spark a discussion because it is "fun to talk about possibilities".

I think it's impractical to automate snow (anything) measurements in this context, but I tried asking some questions to illustrate the difficulties.  So far, nobody seems to see any "possibilities".

[Bushman]

Most ski areas have automated snow depth systems.  Usually ultrasonic.  If one could make a unit cheaply there is no reason why snowfall and snow depth could not be captured.  No different than daily rain and yearly total.  So I see a possibility. 

[dalecoy]

Most ski areas have automated snow depth systems.  Usually ultrasonic.  If one could make a unit cheaply there is no reason why snowfall and snow depth could not be captured.  No different than daily rain and yearly total.  So I see a possibility. 

Assume you have an automated snow depth sensor (for example: http://www.rickly.com/MI/SnowCover.htm )

You mention "daily rain and yearly total", so assume you read this sensor, do some calculations, and post the results.

1.  How would you calculate "daily snow" from the snow depth?

2.  How would you calculate yearly total? [Alternatively, you would you measure "yearly total snow" using a snow depth sensor?]

[Bushman]

Start and zero and count up.    So Day 1, 10cms of snow fall.  Sensor records 10 cms.  Zero the DAILY total and then measure at the start of the next snow day (due to settling etc.)  Say start of day 2 depth  is at  8 cms.  Snow falls - may 5 cms more.  Daily snow for day 2 is 5 cms; total YTD is 15cms.  CURRENT DEPTH (due to aforementioned settling/sublimation whatever) is 13 cms.  Rinse and repeat for the next 363 days.  Or less if it gets warm.

This in fact is how ski resorts  do it with their automated systems.  EG. http://www.kickinghorseresort.com/mountain/information/advanced-weather-data.aspx  Note that KHMR also does a manually measure snow stake (use for avalanche control)

[dalecoy]

Start and zero and count up.    So Day 1, 10cms of snow fall.  Sensor records 10 cms.  Zero the DAILY total and then measure at the start of the next snow day (due to settling etc.)  Say start of day 2 depth  is at  8 cms.  Snow falls - may 5 cms more.  Daily snow for day 2 is 5 cms; total YTD is 15cms.

How do you determine that it's a "snow day"?
What about days with mixed rain and snow? 
Blowing snow? 
Snowing in the morning, followed by a sunny day melting some of the snow?
How to account for settling/melting/sublimation/etc. that occurs during the entire 24-hour period, while precipitation is (or is not) occurring?

"Say start of day 2 depth  is at  8 cms.  Snow falls "... end of day, snow depth is 7 cm.  What was the daily snow fall?

A piece of the selection you referenced:

                                 Hrly  24hr       Total           
                     Air DewPt    RH   New   New Storm  Snow  Hrly  24hr
      Date  Time    Temp  Temp        Snow  Snow  Snow Stake    WE    WE
    --------------------------------------------------------------------
   
     11 13   100    -7.8   0.0    93   0.0  18.7  15.8    56   0.0   0.0
     11 13   200    -7.6   0.0    93   0.0  15.7  15.8    54   0.0   0.0

Note that the "Hrly Stake" went down from 56 to 54, although there was 15.7 of snow in the last 24 hours.  (centimeters?)

It's not quite as simple as "start at zero and count up".

[Bushman]

You are a bright person.  I gave an overly simple example.  CLearly if you monitor hourly (or minutely ?)  you can get a picture of what is going on WRT snow.  All over the continent automated sowfall is recorded.  Please don't descend to sophistry.

[C5250]

I don't he's doing that, just pointing out some of the complications of trying to automate a snowfall measurement.

I find it hard to believe it could even be done accurately from snow depth.

[Bushman]

I don't he's doing that, just pointing out some of the complications of trying to automate a snowfall measurement.

I find it hard to believe it could even be done accurately from snow depth.

Well tell that to the hundreds of ski areas across the continent.  The device he listed is really a fancy tipping bucket.  But for solids rather than liquids.

[dalecoy]

I don't he's doing that, just pointing out some of the complications of trying to automate a snowfall measurement.

I find it hard to believe it could even be done accurately from snow depth.

Well tell that to the hundreds of ski areas across the continent.  The device he listed is really a fancy tipping bucket.  But for solids rather than liquids.

No, I was referring to is the first one in the page.  It's an ultrasonic instrument.  
http://www.rickly.com/MI/SnowCover.htm#Ultrasonic%20Snow%20Depth%20Sensor

It measures snow depth.  No tipping bucket involved.  

I also do not believe that snow fall can be measured using only an automatic snow depth sensor.  However, I would be happy to be referred to one (or several) of the ski areas that do so.

[Bushman]

Clearly, analogies escape you.  A tipping bucket measures, well, tips.  An ultrasonic snow sensor measure pulses (very fast tips).  The point is that they operate  in an analogous manner.

No matter - go check out any of the hundreds of ski areas that measure snowfall automatically.  They use ultrasonic guages (like Kicking Horse - BTDT).  Some use SWE  via snow pillows and some use/add a manual snow stake (a la Kicking Horse).

I get the feeling you think measuring snow is impossible.

[DaculaWeather]

So, who can reverse engineer the Davis console to use that little icon? 

Why?  [What input device would be used?]

1.  Because it is fun to talk about possibilities.  That's why there are silly little icons available. 

2.  Anything is better than my daughter-with-ruler gauge, which hasn't provided a measurement since the first time I used it.

I think she's doing a great job! 

[dalecoy]

Clearly, analogies escape you.  A tipping bucket measures, well, tips.  An ultrasonic snow sensor measure pulses (very fast tips).  The point is that they operate  in an analogous manner.

Thanks.

[SLOweather]

Clearly, analogies escape you.  A tipping bucket measures, well, tips.  An ultrasonic snow sensor measure pulses (very fast tips).  The point is that they operate  in an analogous manner.

Your explanation is a bit overly-simplistic. Strictly speaking, an ultrasonic sensor measures time, the time it takes for the pulse to travel from the transducer face and back after being reflected off a surface, in this case snow.

[dalecoy]

I get the feeling you think measuring snow is impossible.

Of course I don't think that.  It is clearly possible, and the NWS gives a set of ways for doing so.  And ultrasonic gauges provide an example of good automated ways to measure snow depth.  We also have good automated ways to measure and report the precipitation (disregarding whether it is melted snow).

I think that accurately measuring snow fall is quite a bit more difficult, and that automating that measurement is even more difficult.

[Subject for another discussion:  I'm not even sure weather experts have a useful/practical definition of snow fall.]

No matter - go check out any of the hundreds of ski areas that measure snowfall automatically.  They use ultrasonic guages (like Kicking Horse - BTDT).  Some use SWE  via snow pillows and some use/add a manual snow stake (a la Kicking Horse).

I have checked seven ski areas, and have not found any that claim to accurately measure snowfall automatically.  If you actually know of one, please let me know and I'll check it out.  

Your reference Kicking Horse is interesting:  At the time I'm posting this:

1.  "HN 24 Board   Depth of new snow in last 24 hours in centimetres. Cleared each morning by the Avalanche Forecaster."

2.  The above is a depth, measured by ultrasonic gauge.  Generally increasing since each morning, although sometimes decreasing a bit (i.e., 20.1 to 19.6 cm over a 1-hour interval).  That's reasonable, since they call it a depth, not a fall.

3.  "Hrly New Snow   Hourly New Snow depth in centimetres"  It's not clear how this is measured and reported, but the report has been 0.0 every hour for the past two days, while the 24-hour report has been 24.1 + 1.0.  Again, this is reported as a depth, not a fall.  

[In summary, from the total of hourly reports, the total depth over the past 48 hours has been 0.0.  But the 24-hour reports imply about 25 cm]

4.  "Total Snow Stake   Depth of snow at the snow stake in centimetres." (Ultrasonic gauge) started 48 hours ago at 57 and is now 58, having been as high as 61 and as low as 52.

Water content reports are continuously zero, presumably an instrumentation problem.

Please note that the point here is that Kicking Horse is measuring and reporting snow depths, not snow fall.  The two things are not identical.

[Bushman]

Holy crap.  Talk about pendantry.  If snow falls, the amount that lands becomes a depth (measure).  Which may or may not be the same until snow falls again due to settling, sublimation, etc.  When it does, its next snow  "fall" adds to total snow "depth".  I'm not sure how it could be more  clear.  And speaking of clear, you need to get out more - automated snowfall measurement is pretty much standard at modern ski resorts.

Maybe this will enlighten you on moderen snowFALL and snowDEPTH technology.  http://ams.confex.com/ams/91Annual/webprogram/Paper178201.html

[dalecoy]

I really don't appreciate the abusive ad hominem attack. 

Having said that, I must say that I really appreciate the reference you supplied, which so thoroughly illustrates my points.

It was particularly gratifying to hear Mt. Bill Scott of the Meteorological Service of Canada state that there is no accurate automated way to determine rate of snowfall.  Manual methods had to be used to satisfy a requirement to determine if the rate exceeded 2 cm of snowfall per hour.

[Bushman]

Now we are talking RATE of snowfall?  Nice misdirection from the original dicussion. 

Here's a quick test - if the current snow depth is 20 cms and 30 cms is recorded 5 hours later (as measured by ultrasound); (a) How much snow has snow has fallen; and (b) what is the snowfall rate?

And there was no personal attack - merely pointing out  sophistry.

[dalecoy]

Now we are talking RATE of snowfall?  Nice misdirection from the original dicussion. 

No misdirection. By analogy, Davis displays rain rate by using the incremental rain amount measurements.  If you had accurate incremental snowfall ammounts, it would be easy to do rate of snowfall. 

Here's a quick test - if the current snow depth is 20 cms and 30 cms is recorded 5 hours later (as measured by ultrasound); (a) How much snow has snow has fallen; and (b) what is the snowfall rate?

a.  I do not know accurately, but clearly more than 10 cms.
b.  I do not know.

Now, since that was a test, what are your answers?

[C5250]

Here's a quick test - if the current snow depth is 20 cms and 30 cms is recorded 5 hours later (as measured by ultrasound); (a) How much snow has snow has fallen; ...

You didn't provide a key piece of data to enable an answer to that question. How much did the snow on the ground compact from the weight of newly fallen snow?

[dalecoy]

Here's a quick test - if the current snow depth is 20 cms and 30 cms is recorded 5 hours later (as measured by ultrasound); (a) How much snow has snow has fallen; ...

You didn't provide a key piece of data to enable an answer to that question. How much did the snow on the ground compact from the weight of newly fallen snow?

...and what is the compaction of the new snow itself during the 5 hours.  (Which depends on several factors). 

...and how much would the level of the snow on the ground have changed, without the snowfall (internal compaction, sublimation, etc.)

[C5250]

...and what is the compaction of the new snow itself during the 5 hours.  (Which depends on several factors). 

I left that out because NWS best practices state that a snow board should not be cleared more often than every six hours (and at least once a day).

[Weather Display]

there is a 1 wire snow detector people are using
(uses averages and reference points (that reset) to allow for snow melt)

[dalecoy]

...and what is the compaction of the new snow itself during the 5 hours.  (Which depends on several factors). 

I left that out because NWS best practices state that a snow board should not be cleared more often than every six hours (and at least once a day).

That's quite correct if we are trying to make measurements that agree with the NWS snowfall-total-over-xxx -hours measurements (so, maximum of 4 reports per day).  And, of course, no earthly idea of snow rate at any time.

there is a 1 wire snow detector people are using
(uses averages and reference points (that reset) to allow for snow melt)

I've seen the literature on the 1-wire snow depth sensors, and associated calculations, etc.  If there is a 1-wire snow detector (directly measuring snowfall at short intervals), I would be interested in a pointer to information.

My big "heartburn" about the use of snow depth measurements to say "snowfall" is that the measurement is of the effect of a sizeable number of variables, and then must be analyzed to accurately isolate a single variable (by thoroughly understanding and removing the effects of all of the variables).  Automatically and quickly (If we wait until tomorrow to state the snowfall between 9 and 10 today, who would care).

Add to that the fact that the best available automatic snow depth instruments have a resolution only 1/5 as good as a Davis rain gauge.

I have no question that ski areas (and others) can easily and automatically measure current snow depth.  That's very useful. 

[Sigdigit]

I'm still laughing at xykotik's "daugher-with-a-ruler gauge" which hasn't provided a measurement since the first time he used it.

[Weather Display]

sensor then

alot of work has been done on getting it calibrated and initial testing looks very promising