Rainfall different after restarting cumulus.

[Alcedema]

Just restarted my PC and now the graphs here: http://weather.a3n.co.uk/images/raint.png?1260455037401 is somehow showing that roughly 90mm of rain landed in no time at all. The stats on the front page are wrong too.. http://weather.a3n.co.uk/ .. the yearly rainfall is off, the monthly is off too. I think this has happened before aswell, I thought we had more rain than it said, so reverting to an old data set may not be the complete answer as I would lose a lot of new info.

Any ideas?

[Alcedema]

Below appears to be where it occurs..

Code: Select all

08/12/09,15:38,8.7,88,6.8,2.0,3.1,173,0.0,0.0,1012.9,448.2,20.0,53,1.7,7.6,8.7
08/12/09,15:39,8.8,88,6.9,2.0,3.1,179,0.0,0.0,1012.9,448.2,20.0,53,1.4,7.8,8.8
08/12/09,15:40,8.8,88,6.9,1.9,2.7,186,0.0,0.0,1013.0,448.2,20.0,53,1.4,7.9,8.8
08/12/09,15:41,8.8,88,6.9,2.0,5.4,187,0.0,89.7,1012.9,448.2,20.0,53,1.4,7.8,8.8
08/12/09,15:42,8.8,88,6.9,1.9,5.4,194,0.0,89.7,1012.9,448.2,20.0,53,1.4,7.9,8.8
08/12/09,15:43,8.8,88,6.9,1.9,5.4,195,0.0,89.7,1013.0,448.2,20.0,53,1.7,7.9,8.8
08/12/09,15:44,8.8,88,6.9,1.8,5.4,199,0.0,89.7,1013.0,448.2,20.0,53,1.7,7.9,8.8
08/12/09,15:45,8.7,89,7.0,1.8,5.4,198,0.0,89.7,1012.9,448.2,20.0,53,1.0,7.9,8.7

And it is in the backups too..

[steve]

Could you zip up your diags folder and attach it, and also your dec09log.txt and dayfile.txt, please?

[Alcedema]

Done.

[steve]

The total rainfall counter that Cumulus read from the station apparently went down:

2009-12-08 15:38:00 : Writing todayfile, LastUpdateTime = 08/12/2009 15:38:00 raindaystart = 448.200012207031
2009-12-08 15:39:00 : Writing todayfile, LastUpdateTime = 08/12/2009 15:39:00 raindaystart = 448.200012207031
2009-12-08 15:40:00 : Writing todayfile, LastUpdateTime = 08/12/2009 15:40:00 raindaystart = 448.200012207031
2009-12-08 15:40:20 : Writing todayfile, LastUpdateTime = 08/12/2009 15:40:00 raindaystart = 358.5
2009-12-08 15:41:00 : Writing todayfile, LastUpdateTime = 08/12/2009 15:41:00 raindaystart = 358.5

But it must have gone straight back up again, hence the (448.2-358.5) = 89.7 mm increase in your rain. I'd expected to see a message logged in there which would confirm this for certain, but you're still running the beta, from before I changed it to log here rather than the debug.log. When the counter apparently goes down, Cumulus assumes it's been reset and starts counting from the new value; there's not much else it can do, really. What I could probably do is give it a 'second chance'; ignore the first time in case it's spurious, as it clearly is in this case, and only reset if the next reading says the same. I'll try to put something in to the next beta.

You'll need to edit the log to subtract 89.7 from the entries from that day, and also edit dayfile.txt to subtract the same amount for that day's total.

I also noticed this in the log:

2009-11-25 09:23:52 : Error reading Dayfile: List index out of bounds (0)
2009-11-25 09:23:52 : Rainthismonth from dayfile.txt: 0

Which is why your annual total is wrong (quite apart from this 89.7 mm error).

Code: Select all

18/04/09,9.9,338,14:48,5.3,05:35,14.3,15:22,1012.6,00:00,1021.4,23:54,0.0,00:00,0.0,9.9,346.3
19/04/09,8.2,135,18:25,5.2,05:50,15.5,14:06,1020.8,03:39,1025.7,22:12,0.0,00:00,0.0,9.8,281.8

20/04/09,2.0,90,23:34,0.0,00:00,8.5,22:24,0.0,00:00,1027.3,23:44,0.0,00:00,0.0,8.2,1.4
19/04/09,4.1,180,16:30,4.3,05:19,16.4,16:05,1024.8,17:14,1027.9,08:49,0.0,00:04,0.0,9.4,54.8
20/04/09,5.1,360,08:14,6.0,04:49,17.9,12:59,1026.0,01:49,1027.8,09:04,0.0,00:04,0.0,10.8,69.7

It's the blank line causing the problem, but you also have duplicate entries for 19/04 and 20/04, which you'll need to resolve somehow.

[Alcedema]

That's great and a nice catch. Stats are back to normal again.

Thanks.