Operator error rainfall

[vetgmc]

My system just logged a high rainfall in the last hour or so. There was no rain, but I did tap the post in a bit which has the pluviometer on it. How do I know which numbers to edit in which log file to correct my stupidity please?

[steve]

I'm going to write a document on how to fix rainfall when I get chance, but in the meanwhile...

First thing to do is stop Cumulus and look at Nov09log.txt to see what the 'total rainfall counter' is now. Then edit today.ini and set the Start= value so that today's rain is correct. Today's rain is calculated from 'total rainfall counter' minus the Start= value, so you will need to add an appropriate amount to the Start= value. the Start= value is the actual value of the total rainfall counter at start of day; this is why it needs fudging when the counter jumps abnormally. Then edit the 'total rainfall today' values in Nov09log.txt to correct them.

That should get most of your data back on track, you may also need/want to edit the 'current rainfall rate' figures in Nov09log.txt as they have probably also been affected. Today's high rate is in today.ini (the High= value) so you may also need to edit that.

The format of the Nov09log.txt file is in the help and the wiki.

[vetgmc]

Thanks Steve

[vetgmc]

Steve

I am looking at the nov09log.txt & today.ini as suggested, but where you say:

Today's rain is calculated from 'total rainfall counter' minus the Start= value, so you will need to add an appropriate amount to the Start= value

THe Start=value is bigger than the total rainfal counter

- not quite sure what to do.

I think yesterdays rainfall is also suffering from the same disease (nut behind the wheel).

I have attached the two files so you see what I mean.

I am leaving Cumulus turned off for now so the files don't change - is that the right thing to do?

Greg


Later - oops I see I looked at the wrong place what I said above is not correct.

Do I now correct each line that is incorrect separately?
Is it possible to do this in Excel then save it as a CSV?

[steve]

Your Start= value is 230.699996948242 (round to 230.7), and your latest total counter is 240.3:

04/11/09,20:00,10.1,85,7.7,8.7,14.8,167,0.0,9.6,1011.3,240.3,18.4,50,3.6,9.0

So Start= isn't bigger than the counter?

Your problem apparently occurred here, with a jump of 9.6 mm:

04/11/09,13:15,15.4,59,7.2,11.0,15.8,237,0.0,0.0,1010.2,230.7,18.7,51,5.0,15.0
04/11/09,13:30,17.6,52,7.2,8.4,12.2,248,0.0,9.6,1010.1,240.3,18.7,51,5.0,17.6

So you need to add 9.6 to your Start= value to make today's rain zero (assuming all of that 9.6mm is bogus).

You problem from yesterday apparently occurred here, with a 6.3 mm jump:

03/11/09,18:15,14.5,77,10.5,11.3,18.4,250,0.0,0.3,1006.5,219.3,21.4,49,18.4,13.8
03/11/09,18:30,13.4,82,10.4,13.0,25.6,222,0.0,0.3,1006.8,219.3,21.4,50,11.2,12.3
03/11/09,18:45,13.5,78,9.7,14.0,20.9,230,0.0,6.6,1006.8,225.6,21.3,50,20.9,12.3
03/11/09,19:00,13.6,78,9.8,13.8,17.3,199,0.0,6.6,1006.9,225.6,21.1,49,8.6,12.5

So you need to correct the 'total rainfall today' figures from that point on (the 6.6 etc).

You can also change that high rate in yesterday.ini if you want, but that is purely cosmetic (it won't get stored anywhere). You probably have a corresponding bogus 'high rainfall rate' in yesterday's entry in dayfile.txt that you may want to modify.

[vetgmc]

Steve - I am sorry to be a really painful novice...

I have checked the records of the local BOM station again.

The calendar year to 31 October 09 rainfall recorded by them is 464.4mm
They also recorded 1.2mm on 3/11/09.

All my data for October are really iffy, so I wa wondering what you thought of wiping all the October data and starting november with their data as YTD then adding in the 1.2 for 3/11. Alternatively since I have stuffed up fairly comprehensively just wipe all my data and start fresh tomorrow?

I have uploded the nov09log.txt to Excel, then saved it as a .csv file, opened it with Notepad and saved as .txt file. the data don't LOOK damaged - does that mean it is OK to edit this way?

Greg

[steve]

You can't really compare your rainfall with the local station, unless it is really very close to you (or rainfall patterns are different in Australia). Your data doesn't look too bad to me and should be fairly easily corrected.

I think the only worry about using Excel is that it may change the date format (and possibly also the time format) and then Cumulus won't recognise the formats. If you're sure it hasn't changed those then you should be OK. I usually recommend the use of something like CSVed -http://csved.sjfrancke.nl/

[vetgmc]

Thanks for all that Steve.

I got my YTD rainfall from the other station in the first place, which is why I can't understand why the Total Rainfall Counter is only in the 200s. Should the Total Rainfall Counter be the same as the Calendar Year to date rainfall?

I am not sure exactly where the other station is, but it can't be far away - Kyneton is not a very big town!

for all my mistakes

Greg

The only format change I can see is the leading 0 on the time is dropped as in 09:00 becomes 9:00. Would that make any difference to Cumulus?

[steve]

I got my YTD rainfall from the other station in the first place, which is why I can't understand why the Total Rainfall Counter is only in the 200s. Should the Total Rainfall Counter be the same as the Calendar Year to date rainfall?

The 'total rainfall counter' is read from the station, so the YTD rain figure in Cumulus has no effect on it. Cumulus simply adds the supplied YTD figure on to its calculated annual total.

The only format change I can see is the leading 0 on the time is dropped as in 09:00 becomes 9:00. Would that make any difference to Cumulus?

Hard to say; it depends on how the system routines react to being given a time in that format; things like Select-A-Graph may not work. I converted some data (with a program) for someone a while ago and accidentally wrote the times in that format. I seem to remember it did cause a problem so I had to correct it.

[vetgmc]

I think I have done everything you suggested (using the CSV editor recommended), except I don't get the adding on to Start=value bit. If the error occured during the day shouldn't the Start=value be correct without adjustment? If it needs correcting for yesterday's error, shouldn't I be subtracting yesterday's error from it?

The today's rain figures in Cumulus haven't changed since I edited the files.

Looking at the nov09log.txt file my editing doesn't appear to have been saved. I'll have to try again.

Sorry to be so thick.

Greg

I edited again - I think I saved the edited files in the wrong place or something. Cumulus is still ging weird figures with TYD & this month all wrong. I am getting tempted to just reset everything and start from tomorrow. What do you think?

Greg

[steve]

The Start= value is the snapshot of the total counter read from the station at midnight (assuming a midnight-midnight day). In your case 230.7.

From then on, today's total so far is calculated by subtracting that value from the current counter. So, at 13:15 your current counter is still 230.7, so today's rainfall is (correctly) 230.7 - 230.7 = zero. By 13:30 your station had increased its total counter to 240.3; let's assume all of that is bogus. So your daily total is now calculating as an incorrect 240.3 - 230.7 = 9.6.

To correct this, you set Start= to 240.3, so that your daily total so far is now 240.3 - 240.3 = zero, and for the rest of the day (assuming no other problems occur) the total is calculated correctly.

Basically, you are changing Start= so that it is correct relative to the new value of the counter.

[beteljuice]

Slightly sideways (as ever)

... The 'total rainfall counter' is read from the station, so the YTD rain figure in Cumulus has no effect on it. Cumulus simply adds the supplied YTD figure on to its calculated annual total.

Presumeably, At 00:00:01 of Jan 1st people who have 'added' YTD need to zero it for the New Year, or is it clever enough to self-destruct ?

[steve]

Presumeably, At 00:00:01 of Jan 1st people who have 'added' YTD need to zero it for the New Year, or is it clever enough to self-destruct ?

On the station settings screen, when you specify a YTD value, you have to supply the year to which it applies, which would normally be the year you started using Cumulus.

[beteljuice]

I really must read what's on the screen more often ! thanks Steve

[vetgmc]

Steve

I have done it all again and Cumulus is still telling me the month's rain is 9.6 and today's rain is 9.6. Should I just start again?

The station seem to be editing the log test file.