The Weather Forecasting Forum

Strange map behaviour
Page 2 of 2

Author:  Fred [ Fri Nov 11, 2011 9:51 am ]
Post subject:  Re: Strange map behaviour

I compared both maps and the differences are indeed rather small. But there are (e.g. the low pressure area) and there should'nt be. Again, the analysis is based on the plots and not on the dimensions of the map.

I used a scale of 2500. I am pretty curious what will happen in your area with such a scale.

Author:  geophi [ Fri Nov 11, 2011 3:46 pm ]
Post subject:  Re: Strange map behaviour

The analysis is based on the plots, but also the grid spacing. Since the number of grid points in both maps is the same, the grid spacing (used for the analysis and therefore the drawing of the contours) will be larger in the north/south direction for map that encompasses a larger north/south distance. Therefore, I wouldn't expect the two analyses to be exactly the same, even across the middle of the map. Larger geographic spacing between grid points will result in a smoothing of the contours relative to the one with a smaller geographic spacing between grid points. You can see where the grid points are that are used in the analysis by choosing the wind speed (barbs) analysis menu item.

Below are maps similar to yours from your area with a 2500 nm scale. This was from the 06Z 11 November SYNOP obs downloaded from Albany. They have much more detail than the analyzed sea level pressure in your maps. The two analyses at different scales are not exactly the same and I wouldn't expect them to be because of the differences in geographic distance between the grid points in the north/south direction. the biggest differences should be on the edges of the map, and in areas where the spacing between observations is greatest (over the oceans for example). The top map is a plot with a 99% Data Plot Crowding value showing all observations used in the analysis.

My guess is that for whatever reason, you didn't have nearly that many obs that were used in the sea level pressure analyses in your original attachments. Fewer obs will result in a smoother, less accurate analysis. The other possibility is that the analysis routine you used had different settings than mine. I'm using the Barnes analysis with default settings.
Automatic smoothing checked
0.50 smoothing coefficient
0.30 Gamma

Edit...More information on the objective analysis used in Digital Atmosphere can be seen at

680by960VietnamAllObs.gif [ 111.08 KiB | Viewed 1767 times ]
680by960Vietnam.gif [ 56.72 KiB | Viewed 1767 times ]
960by960Vietnam.gif [ 63.52 KiB | Viewed 1767 times ]

Author:  Fred [ Sat Nov 12, 2011 2:07 am ]
Post subject:  Re: Strange map behaviour

Thanks for your comprehensive answer. I'm away all day so I think I can look to it earlist tomorrow.

Author:  Fred [ Sun Nov 13, 2011 1:53 pm ]
Post subject:  Re: Strange map behaviour

I have chanced my parameters to yours to have the same. I also use Albany, so I guess I have the same dataset. I only set crowding on ~80%.

Also thanks for the link. I did not know this one. I only read the less detailed chapter in the (old?) manual. Due to a lack of time this weekend (a lot of troubles with the water well) I only read the parts that has a link with my problem.

“Since the pixel size of a window rarely exceeds 1200 x 1200 pixels…”.
Well, at the moment many people has much larger screens. But I have my doubts if this will be changed, so I think I stay with a maximum of 1280*960 to be more or less close to the 1200.

“The floating domain simply moves to whatever geographic area is displayed and is not used until an analysis is requested.”
Here I have my doubts, based on my maps.

After writing this I looked to my maps and then I saw something strange. Every time I used the same parameters for the map: lat 12, long 109, scale 2500, and height 960. I only changed the width. So you would expect a somehow distorted map. But although I changed the width, the part shown on the map from E to W is always the same! On the top and the bottom of the map parts are missing! Compare the maps 680*960 and 1280*960.

My guess is that the grid is still based on the original height of the 680*960. That also explains the fact that the isobars are shifted mainly in N-S direction. .

If an analysis (see above) says I’m in the 24-25 deg. C. zone in Nha Trang and on the other map with the same data in the 25-26 zone then in my opinion there is something wrong.

Perhaps Tim can shine his light on this?

Page 2 of 2 All times are UTC
Powered by phpBB® Forum Software © phpBB Group