Page 1 of 1

Stationplots and temperature analysis / 1.0m

Posted: Tue Jun 29, 2004 7:17 pm
by hoevenm
Hi,

at first, apologies for the big picture. But this makes my question more clear.

When I plot stations at a 80% density, and afterwards plot temperature iso-lines (with Analysis menu -> Temperature -> Celsius), I notice the red iso-lines do not really reflect reality. For example, see Northwest Russia (east of Finland) where temps are at the 30-s. Yet the red temperature line indicate a 22/24 degrees. Same in Ukraine and southern Russia: stations plot of lower twenties. Yet a red line of 30 degrees is plotted almost through those stations.

I'm suspecting that the averaging of the red line is far to loose. It's all levelled out in the map, although the real station plots vary a lot regarding temperatures.

My question is: can I set some parameter(s) so the red line will far more closely follow the real temperatures?

PS. Analysis type is set to "Nearest neighbor"

Thanks for responses.

(edited: picture removed. Picture was on my server, but that one just crashed last nigth :()

Posted: Thu Jul 01, 2004 6:34 am
by hoevenm
well, guess I was the only one. It is indeed a setting.

Posted: Thu Jul 01, 2004 9:52 pm
by Weasel
In file, preferences, analysis tab, reduce the smoothing number. This will more closely match actual observations...the isolines look more pointy, however.

Posted: Fri Jul 02, 2004 7:37 am
by hoevenm
I was a bit confused, because when you select "Nearest Neighbor" there is nothing to set on smoothing (these options are grayed out, not accessible). I first have to change the analysis type to something else than Nearest neighbor, then set the smoothing, and then (if I wish) re-set it to Nearest Neighbor.

Posted: Sun Jul 04, 2004 4:05 am
by Weasel
Hoevenm,

I'm not having this trouble with the latest version, nothing is grayed out regarding nearest neighbour..at least nothing out of the ordinary. I did notice some isoline cubes using a smoothing setting of zero over the North Sea; however, with the latest synoptic obs. This seems a bit strange.

Don't know what to say except my m patch has no limitations regarding smoothing reduction with nearest neighbour, excluding the aforementioned cubes/squares/rectangles. Sorry my system isn't acting as yours:(

Pat

Posted: Sun Jul 04, 2004 3:00 pm
by hoevenm
Okay, thanks. There is a tiny work around for me: switch to another analysis type, adjust smoothing to the desired value, and set analysis type to nearest neighbor again.

So, not a big problem, but it took me some time to find out I have to change the analysis type to open the smooting settings (at least in my case).


Maarten

Posted: Sun Jul 04, 2004 3:48 pm
by Tim Vasquez
hoevenm - it would help if you could post what version of the software you have. I am guessing you have something before V1.0m, and a lot of analysis functions were fixed in V1.0m.

Tim

Posted: Sun Jul 04, 2004 9:30 pm
by hoevenm
pfff my deepest apologies :shock:

I am running 1.0m, and I see it is not grayed out at all. Perfectly working! Dont know what I was thinking... Duh

Sorry about this noise

Posted: Sun Jul 04, 2004 11:56 pm
by Weasel
sorry I typed m patch:(