Rainfall accuracy

[serowe]

The resolution of a tipping bucket gauge is dorectly related to the capture area or funnel. There have been numerous articves/drawings produced over the years on how to increase the sensitivity of the WM/WMR Oregon series from 1.0 to 0.5 mm resolution. As long as the software reading the 'tips' knows what is to be expected, then you could, theoretically inclrease the collection area even further to obtain 0.1mm accuracy if you wanted - BUT the area MUST be highly accurate otherwise your readings are all over the place.

But, again, even if you got down to 0.1mm accuracy, it still can't read 'trace' (if you think about it, a few dozen spots of rain would be classified as a 'trace' even if nothing was collected in the gauge or bucket.

The other type of rain that isn't catered for is that which falls but doesn't hit the ground. Is this then not classified as rain by some?? (think before answering though )

[iceberg]

The resolution of a tipping bucket gauge is dorectly related to the capture area or funnel. There have been numerous articves/drawings produced over the years on how to increase the sensitivity of the WM/WMR Oregon series from 1.0 to 0.5 mm resolution. As long as the software reading the 'tips' knows what is to be expected, then you could, theoretically inclrease the collection area even further to obtain 0.1mm accuracy if you wanted - BUT the area MUST be highly accurate otherwise your readings are all over the place.

But, again, even if you got down to 0.1mm accuracy, it still can't read 'trace' (if you think about it, a few dozen spots of rain would be classified as a 'trace' even if nothing was collected in the gauge or bucket.

The other type of rain that isn't catered for is that which falls but doesn't hit the ground. Is this then not classified as rain by some?? (think before answering though )

THIS IS CLASSIFIED AS DISTANT PRECIPITATION......
I THINK YOU ARE GOING A BIT OVER BOARD. WHAT HAPPENS IF A BIRD PEES ON YOUR RAIN GAUGE. YOU WILL HAVE TO CALCULATE IT AS A BIRD TRACE!!!!

[serowe]

Call it what you like 9you don't have to SHOUT either btw...)

I gave that example as it is directly related to a new Cumulus web tag of 'Days since rain'. Some people couldn't give a rats if rain isn;t recorded - they just record it as a no-rain day.

I collect, and submit, rainfall records here (and have for some years) and these records are supposed to include all days precipitation falls, whether it is recorded in any of my gauges or not.

Basically - if oyu have a piece of data that states something, then it should be as accurate as possible. As I pointed out before, there are users here who have complained about temperatures being 0.1C out of whack compared to others; pressure being slightly too high or low - so why, when it comes to rain fall, do some users say 'it doesn't matter'?

[serowe]

THIS IS CLASSIFIED AS DISTANT PRECIPITATION......

Oh and this is incorrect btw - 'distant precipitation' is observable rainfall or precipitation away from the observing site.

The term you were looking for is 'virga' - rain fall which evaporates before it hits the ground. A common enough phenomenom especially in western USA and the Canadian Prairies.

Both terms are easily found by searching for them..

[Matt.j5b]

It isn't (as Tau Bootis states). But what a recording of 'Trace' in the data does do is show that those 10 or 20 days your electronic records show as being rain free are not, in fact, totally free of precipitation - just that on some day or days, there was precipitation BUT it was too small to be recorded.

I make my personal proper rain recording manually because my station only records in a resolution of 0.5mm so if I get a light shower less than 0.5 I don't get a recording, that but would affect the number of rain days if just use the weather station, anyway that how I done it a while now with the manual gauge. I would never expect any device to measure down to the 0.1mm or a trace , it is my thought the tipping bucket is not designed to do that, only my thought.

If you have a piece of data that states something, then it should be as accurate as possible

Yes, as my station only records 0.5, 1.0, 1.5 etc and nothing in between I use the weather station reading only as a guide as to how much rain falls over a peroid of time rather than use for the correct figure down to the 0.1mm readings that my manual gauge does.

[serowe]

Matt - believe it not we *do* agree on this. What I don't agree with is discounting those days (for the 'Days since rain' data in particular) where the rainfall is less than the resolution of the gauge. Just because it didn't record doesn't mean it didn't rain! (And, as an aside, over the past 48 hours I have had the opposite effect between manual and electronic gauges - 13.0mm in the electronic gauge and 11.5 in the cylinder...sono matter which way we choose, thee will always be some discrepency about the readings).

[serowe]

I make my personal proper rain recording manually because my station only records in a resolution of 0.5mm so if I get a light shower less than 0.5 I don't get a recording, that but would affect the number of rain days if just use the weather station, anyway that how I done it a while now with the manual gauge. I would never expect any device to measure down to the 0.1mm or a trace , it is my thought the tipping bucket is not designed to do that, only my thought.

You can adjust any electronic gauge to record down to 0.1m if you want though. A 1mm resolution gauge might have, say, a collection area of 5 square cm. (figures pulled out of the air btw). To get 0.5mm accuracy you double the collection area - so it becomes 10 square cm. Obviously you use mathematics to determine the new diameter of the circular collection area. To get 0.1 mm accuracy use mathematics to determine the area needed but remembering that this resolution could literally flood the tipping bucket as it wasn't designed to measure thus you end up missing (because of 'floods) much of the water coming into it..

Problem is - the only software I know of that allows you to change (on the WM/WMR series at least) the recorded accuary is Weather Display - Brian has a flag for 0.5mm resolution (good idea but the conversion MUST be accurate) as well as 'ignore first tip of the day' to allow for any remaining water at the roll over time.

If anyone is interested (probably not given some responses...) - 2 links on how to increase the resolution of a rain gauge:

http://www.weather-watch.com/smf/index. ... 309.0.html
and
http://www.weather-watch.com/smf/index. ... 441.0.html

But on the other side of the coin NO software can automatically detect a trace - that's why it is called a 'trace' - but the software SHOULD allow for the manual intervention and recording of this - so instead of recording what would be a zero value, a manual edit would be required. Not rocket science (and I don't believe I ever mentioned the gauge being able to detect such a small amount anyway). Do it this way (as I suggested - a storage value of -999 would be a plausible and workable value to use) and the software takes care of this when counting the number of days since rain.

There have been plenty of forum discussions on http://weather-watch.com about these topics though.

[Matt.j5b]

Well I was saying that I don't expect a "electronic station" to read as accurate and/or precise as a manual, thats why I go for the manual gauge for the real reading, and of course you count a rain day for rain that falls that day, regardless of whether it recorded or not. Because why would you say it did rain, but it not counted as a rain day. I am not just agreeig , because I count a drop as rain for my records and also a daily paper log of the general weather and if rain fell etc....

You can adjust any electronic gauge to record down to 0.1m if you want though

Of course you can change it to read down to what resolution you want, I just don't see the point for me because even when it is modified there can be a difference between the gauge and the manual, when I have to use the manual for the real reading.

[serowe]

Well I was saying that I don't expect a "electronic station" to read as accurate and/or precise as a manual, thats why I go for the manual gauge for the real reading, and of course you count a rain day for rain that falls that day, regardless of whether it recorded or not. Because why would you say it did rain, but it not counted as a rain day. I am not just agreeig , because I count a drop as rain for my records and also a daily paper log of the general weather and if rain fell etc....

Think you missed what I have been saying here - the Cumulus web tag of 'days since last rain' (or whatever the correct tag name is) can't do this UNLESS you have a way of adding your own informaiton to say 'rain fell'. You can't put in, say, 0.1mm because that alters tha actual data. But the addition of a marker to indicate 'trace' or similar would then adjust this web tag. But this is moot because steve has already decided it shouldn;t be done because he doesn;t agree with it...

Of course you can change it to read down to what resolution you want, I just don't see the point for me because even when it is modified there can be a difference between the gauge and the manual, when I have to use the manual for the real reading.

But the way you and I do it is probably unique. How many others do you thnk will actually run both type of gauges? There wouldn't be many judging by omments here and in other forums Again it comes down to individuals expectations - the expectation of a number of this forum's users is that their weather stationshould be highly accurate (see the recent sea level pressure and temperature threads) even when we know they aren't that accurate. And then this thread - 'Rainfall accuracy'...

[Matt.j5b]

Think you missed what I have been saying here - the Cumulus web tag of 'days since last rain' (or whatever the correct tag name is) can't do this UNLESS you have a way of adding your own informaiton to say 'rain fell'.

Of Course if the data from the weather station gauge doesn't capture of all the data(ie less than the lowest figure possible (0.5mm for me) , that would create a wrong figure for the days since last rain since it missed days that it did rain but not enough for it to register. I agree with on that one, ( I agreeing again ). It would bother those who want to know the last rain on a day with 0.1mm, but if it is thought as days with a least x mm (0.5mm for me) it is no problem.

But the way you and I do it is probably unique. How many others do you thnk will actually run both type of gauges?

Well, before my La Crosse, I did it manually and is a shame to stop that, it does it well, the station really is only a extra, but it helps with ploting rain over time, rain rate etc. But most people would not bother but it does annoy me if I get 0.4mm and therefore nothing on the station, thats why I do it. But the weather station gauge is only so accurate, and never expect it to be perfect.

[serowe]

It would bother those who want to know the last rain on a day with 0.1mm, but if it is thought as days with a least x mm (0.5mm for me) it is no problem.

http://en.wikipedia.org/wiki/Atacama_Desert

[Eric]

Think you missed what I have been saying here - the Cumulus web tag of 'days since last rain' (or whatever the correct tag name is) can't do this UNLESS you have a way of adding your own informaiton to say 'rain fell'.

I can't claim thirty years as a professional meteorologist, or whatever it is, but it seems to me that those items are dealing with measurable rain, i.e 'rain days', and of course a 'trace' is not a rain day, by definition. So they are perfectly valid, with or without your 'suggestion' of some kind of 'trace' indicator in the data logs. Manually recording a 'trace' wouldn't affect them anyway.

I've looked at quite a few software packages for hobbyist weather stations, and as far as I can tell, none of them offer the ability to log a 'trace', so I don't know why you're constantly whining on to the good people of this forum about it, and nowhere else, apparently. It seems to me that it would be of use to only a small minority of Cumulus users, as very few people with home weather stations have a 'manual' rain gauge that they go out and read every day at 9 am (and in support of this assertion, no-one seems to have created an enhancement request for it). As the very first line of the Cumulus description says, it's "software for retrieving, storing and displaying data from an electronic Automatic Weather Station". Note: Automatic. If you have manual equipment, by all means make your own manual records.

Eric

[b.e.wilson]

I guess the question of whether "trace" means something or not is if trace rain makes any discernible difference having fallen. Do plants grow differently because of trace rain? Does the soil moisture change when trace rain falls? Is the trace of rain even detectable an hour after it falls? If the answer is no, then trace can safely be ignored.

Here in the desert western US, trace means nothing. But then, 0.01 inches also means nothing. Small insects need to avoid getting stuck in a droplet, but otherwise, trace rain is safely and comfortably ignored. Well, not completely ignored: trace rain will bring down the dust from the air and really mess up windows. So yes, window washers care. They love it.

[steve]

I think Ian's thread has been hijacked sufficiently now (and I'm not without part of the blame myself). Please take discussion of the recording of 'trace' elsewhere and restrict discussion here to the original question. If anyone feels strongly that Cumulus should permit manual recording of 'trace' amounts of rain, then as I've said in the past, feel free to create an enhancement request. Thank you for your co-operation.

[iceberg]

Please keep your comments nice, like captzero said, it is boring and not useful, Cumulus is great piece of work by Steve , it is not the fault of Cumulus that it can't measure a trace. Your comments are very helpful, no need to knock that, just keep the comments clean.

Not right - Tr. is an exceptional amount and could be catered for with proper coding. As this thread *is* about 'rainfall accuracy' it is a point that does need to be made because accumulated trace amounts which, in a cylindrical gauge can easily top more than 1 or 2 mm, can be quite easily be ignored by an automated tipping bucket. As I mentioned - I already have a discrepency of 1.75mm between the cylinder and the auto bucket here since it started raining last night around midnight - primarily because the falls have been in hot conditions and each fall has been very light (most likely evaoprated between showers during the day today).

Well serowe why dont you just purchase a quality weather station like a Capricorn Orion station......

http://www.columbiaweather.com/Orion_We ... ation.html

Rainfall Measurement
The rainfall measurement is based on an impact sensor, which detects the impact of individual rain drops. The signals exerting from the impacts are proportional to the volume of the drops. Hence, the signal from each drop can be converted directly to the accumulated rainfall. Due to the measurement method, flooding, clogging as well as wetting and evaporation losses in rain measurement are eliminated.

and under current you can add (TRACE OF RAIN) to your web site