[rimprona] RIMpro anomalies

[Vincent Philion]

Hi!

I’ve never used Meteoblue. It was tested last year, but it’s still fairly “new”. Automatically correcting the “near past” is something we are considering for virtual weather. The idea would be:

Currently = “Forecast” is updated with “Nowcast” and we keep the latest nowcast as “reality”.

Depending on how much difference we see between “nowcast” and station data, in a near future we might substitute the latest “Nowcast” with gridded data. i/e a clever interpolation of station and radar data.

Your stations are different in the database. The software won’t change the file unless you force it. So Hudson NEWA, MB, and HRDPS are completely independent.

Now I’m curious: did the sudden change make matters “better” (closer to your station) or worse?
Was the forecast for April 7-8-9 bad with every source?

Vincent


> Le 11 avr. 2016 à 09:58, David A. Rosenberger <dar22 at cornell.edu> a écrit :
> 
> I suspect that it is a bug with MeteoBlue:  It is definitely changing the history going several days back from the present.  I noticed a small change last week, but thought it might be a fluke.  The change this morning was from a non-event on Friday-Saturday to a major infection event even though there was no indication of that occurring on either Friday, Saturday, or Sunday.  It only showed up this morning.  I have not changing any settings or forcing a re-read from the beginning of the season unless that happens when I switch from one station (Highland-Newa to the Highland-MB station).  
> 
>> On Apr 11, 2016, at 9:50 AM, Vincent Philion <vincent.philion at irda.qc.ca> wrote:
>> 
>> Hi! not sure I understand. I don’t have a link to “your” RIMpro station to have a look.
>> 
>> As far as I know, RIMpro never “corrects the past” (!)
>> 
>>> morning.)  What I am finding, however, is that (at least with MeteoBlue) the history also changes over time.  For example, when I checked the Venturia table yesterday at  5 pm, the MeteoBlue table said that the RIM value for Sat, 9 April, was 10. When I checked this morning, the history for Sat, 9 April was showing a RIM value of 1166 and the value for Friday, 8 April (which had previously been 1) was now shoing 1079.
>> 
>> Sounds like a bug to me (?) (Or something is missing in my understanding)
>> 
>> Unless:
>> 
>> 1) the files used to generate the simulation were erased from RIMpro (by clicking, “force read data from beginning of season”) AND
>> 2) that file changed
>> 
>> OR
>> 
>> biofix was changed?
>> 
>> 
>> For the Hudson/HRDPS station, I see an infection that finished on April 7-8 (RIM about 411), and a new ejection and infection that started on 7th PM, that is now at RIM = 22. 
>> 
>> Makes sense?
>> _______________________________________________
>> RIMproNA mailing list
>> RIMproNA at virtualorchard.com
>> http://virtualorchard.com/mailman/listinfo/rimprona
> 
> _______________________________________________
> RIMproNA mailing list
> RIMproNA at virtualorchard.com
> http://virtualorchard.com/mailman/listinfo/rimprona