4 stations sited in the water in same area different rain readings

So there are 3 weatherflow stations near me, the 3 are located at a rowing lake, 3 locations along the main lake, within 200m of each other and about 5km line of site from me, they show different rain totals to each other and different to mine yet we all experienced the same rain event today…

This clearly shows a problem with the way the weather flow stations detect and measure rain! Amazing… I’ll raise it with weatherflow and see what they say.

The first 3 pics are the lake the last location pic is me then the last pic shows where we are.

I know those stations and their purpose. Ignore the rain values. They are used strictly for wind.


the rain sensor shows accumulated values with an accuracy specified by weatherflow of +- 20%. These station have an average value of 66mm. The 20% gives a range of 53-80mm. I would say that is just within specs. (the 85mm is outside that range, but there is no reason to assume that it rains the same amount in each spot)

Did you really mean TWENTY percent ? That’s not good. Where’s that documented ?

1 Like

Scroll to the bottom of https://weatherflow.com/smart-home-weather-stations/

Just for the Stealer fans here, it’s +/- 20% of actual total rain. You would need a very accurate gage to get actual.

If the actual is 66mm the high would be 79.2mm (66 + 13.2) and the low would be 52.8mm (66 - 13.2).

I think that the bigger issue here is not knowing how these devices are mounted. They are obviously being used for measuring wind speeds at various points on a regatta course. We don’t know if they are mounted to floating docks, buoys, or even just sitting on the ground, all of which will throw rain measurement accuracy out the door…

Again, +/- 20% is just horrible.

Just to mention that famous label brand, Davis claims “For rain rates up to 4"/hr (100 mm/hr): ±4% of total or ± one tip of the bucket (0.01"/0.2mm), whichever is greater.” for the VP2 and the same numbers for the Vue.

My recollection of the CoCoRAHS gauge accuracy is it’s in the 2-3% kind of number, but I can’t find the reference quickly.

1 Like

but the sky reacts with the first drop! That’s different from the bucket systems. To get to the rain total, the sky basically has to add up every single drop it detects. It has to estimate the size of the drop based on the sound it makes. It is amazing that the sum of those estimated sizes add up to something that is 20% accurate.
Initially the unit was advertised to have a capacitance sensor as well. At that time I just assumed they used the change in capacitance to determine the drop size. Apparently they decided to drop the capacitance sensor and only use the sound.

There are many reasons these stations would not be yielding the exact same rain accumulation:
device to device variation, mounting variation, environmental variation (namely turbulence), rain fall spread variation. Also, as Gary mentions, these SKYs are sited ideally for wind readings. Any rain gauge, either from an all-in-one station or collection style should be sheltered from winds for more accurate accumulation readings.


just wondering how much the influence of wind is on the rain readings…

I’d agree, +/-20% for a meteorological parameter is not very good. Hopefully, as they refine their algorithms, things will improve. However it’s for this reason I’d love to see them integrate a conventional tipper into their system.


Overiaying graphs from two sky and one VP2 - sky2 is low at 1m elevation, sky95 (serial number ends in that) is at 7m higher, the VP2 is at about 5m high.

Truth is the red VP2 line which matches my CoCoRAHS gauge exactly.

Look for the light blue for when the Davis tipper tipped ? :slight_smile:

1 Like

now it would be really interesting if you would interchange the two sky units (assuming there is no real obstruction for the rain at 1 m).

Beats the average accuracy of the evening weather report! :wink:

hey a day when the two sky, the VP2, and manual gauge all agreed !!!


1 Like