Nearcast rain result

Just an interesting observation.

The last 3 days we have had some rain with the most of it yesterday. Actual total rain of 2.24" (cocorahs).

Without NC Rain option on… the Tempest showed a total of 1.68".

When I turned NC Rain on… the total went to 1.92" probably using some data from my Davis station, at the same location… and other factors.

Once I entered my cocorahs totals of 2.24" and turned NC rain on… about an hour or so later the tempest NR totals changed to 2.19".

So between my Davis station data and COCORAHS reports… the Tempest NC rain option comes very close to the COCO reports. I think, when possible, their NC routine heavily weighs the COCO reports over other data.

Again… just an observation that I have seen repeat several times with past rainfalls.

1 Like

The problem there is timing of the CoCoRaHS data entry. Tempest does midnight-to-midnight totals. CoCoRaHS wants 7am-7am. Actual report times of course vary site by site and day by day for manual reports.

I understand that.

With the use of radar and other station data… their algorithm can certainly determine “when” the rain fell and portion out the COCO report accordingly.

i.e. X% from 7AM to midnight (the day before) and Y% (100%-X%) from midnight to 7 AM (current day).

We’ll have to disagree there.

I just find the whole automagical ‘fixing’ of readings just wrong. You can’t fix bad hardware with software. The sensors should read accurately without these kludges.

2 Likes

Vince,

We actually don’t disagree.

It is “wrong” to try to accommodate the inaccuracy of the Tempest rain “gauge” by using other stations data and COCOrahs reports.

I was just pointing out how it calculates… not condoning the fact that it has to adjust.

George

but that isn’t a real problem. Tempest has minute by minute rainfall and could calculate the accumulated rain in any period that it wants and adjust the calibration accordingly. Please note I’m saying “adjust the calibration” not adjust its reported amount (as nearcast does). The different timings on the report, is only a problem for humans that want to compare the results with other measurements. The amount of rain fallen in your garden doesn’t change and if you are going to ignore the tempest values anyway, why is it a problem.

The Tempest raw rainfall results being unreliable is not a problem personally for me… because I do not use the Tempest rainfall values for anything important.

It could very well be a problem for Tempest users who really expect/want to have it “measure” rainfall
accurately at their site.