Jump to content

Recommended Posts

  • Replies 67
  • Created
  • Last Reply

Top Posters In This Topic

Not clear, is it. The new BB display is truncating the eastern margin of the upper image, so you can't see what I wrote about. The graph is a N-S temperature profile for the Pacific, east of Japan, but you have to right click and View Image to see it in full.

 

This makes little difference to my amateur climatology, but next winter's snow photos are going to look a little weak at either 500-odd pixels wide, or half a shot.

Link to post
Share on other sites

Not trying to be confusing, just one of my gifts.

 

With three columns turned on, the right hand side of the image was truncated in my display and the graph (N-S profile) didn't relate to what was on the screen. I've turned off the left hand column and now I can see it.

 

The graph shows a profile of the actual temperature difference this year compared with the 2003-2006 average. The dark blue areas on the image are about 1-2C cooler than average, and the yellow areas 1-2C warmer. The sea of Japan and Tohoku Pacific are pretty warm.

Link to post
Share on other sites
  • SnowJapan Admin

With 3 columns turned on for me, the image appears without problems. Things like this are going to depend on the size of your display and the settings of your display.

Link to post
Share on other sites

Thanks for the efforts soubriquet, interesting. I'd like to know how you did it but it would probably confuse me right from the word go! Does that cooler than average water east of Chiba have any say on what might happen in typhoon season?

Link to post
Share on other sites
 Originally Posted By: 2pints,mate
Interesting that last sentence.

The people we met in Nagano said it was by far the least snow they had ever had in a season. (And one of these people was 79!!) Regional variations?


As a long term amateur weather observer (and I actually majored in meteorology at uni) I can tell you that local people are usually not the best resource for anything weather related. Over the years I've been told so many things by locals about the current season or past seasons only to find most of what they are saying doesn't really add up when you look at the actual meteorological data.

The fact is most people take little more than a passing interest in weather and generally have pretty short term memories of seasons of old. Believe me study the meteorological data for an area you live in and you'll have a much more informed view of averages and trends than most locals who've lived there all their lives.
Link to post
Share on other sites
 Originally Posted By: Go Native
.. I actually majored in meteorology at uni..


Where was that Go Native? I did some meteorology as part of my Environmental Sciences degree at University of East Anglia, but I'm coming at this from a remote sensing background. Definitely not making any predictions though. No no.

I totally agree with your point. It's the data that count. We are lucky to have so much stuff available online now, especially the satellite imagery.

I've written a proper algorithm now, so I can easily process the images. We will have to see where the monthly snapshots take us.
Link to post
Share on other sites
 Originally Posted By: happyhappy
Thanks for the efforts soubriquet, interesting. I'd like to know how you did it but it would probably confuse me right from the word go! Does that cooler than average water east of Chiba have any say on what might happen in typhoon season?


Sorry, missed this one. I'm using ENVI which is the industry standard software for dealing with remotely sensed data. A digital image is essentially a raster (a 2-D array), that is a number of rows and columns where every value represents the brightness of an individual pixel. A colour image has three arrays, one each for R, G and B.

Unlike something like Photoshop, ENVI doesn't care how many arrays you have, as long as they are the same size. I can use the Band Math routine and type in something like:

Band 1 = Band 2 - ((Band 3 + Band 4)/2)

Then all I have to do is tell it which bands are which and it finds the average of Bands 3 & 4, subtracts them from Band 2, and gives me Band 1, the difference. That's the easy part.

It's a lot more work to filter out all of the null values in the data, such as where the sensor was pointing at land, or no data exist. For that I make up a Binary Mask, an array where the values are either 0 (null) or 1 (data). Multiplying Band 1 with the Mask leaves the data unaltered (1 x 1 = 1), and multiplying by zero sets all the rubbish to 0. Then I apply a colour table, and transform a single Band grey scale image into funky RGB.

There is some other stuff to do too, but ENVI is just brilliant at reading in unknown file formats and dealing with multiple monster arrays \:\)
Link to post
Share on other sites

I did my degree at The University of Melbourne. Didn't go on to do any post-grad stuff which is probably why I've never had a job actually using my degree (plus I've made a hell of a lot more money over the years than I ever would have working as a meteorologist).

Link to post
Share on other sites
  • 7 months later...

I've worked up some of the sea surface temperatures for 24th December 2003-2007.

 

soubriquet_183.jpg

 

This is the average sea surface temperature around Japan for late December 2003-2006. The plot shows the temperature profile from north to south.

 

soubriquet_184.jpg

 

This is the sea surface temperature around Japan for 24th December 2006. It is clearly visible that the warm water masses of the southern Sea of Japan, and Pacific are much further north than average.

 

soubriquet_185.jpg

 

This is the sea surface temperature around Japan for 24th December 2007. In this you can see that overall temperatures, and the warm and cold water masses are much closer to their average positions than last December.

 

I'm not predicting anything, but at least one of the major controls on the weather is set to more "normal" conditions than last winter.

Link to post
Share on other sites

×
×
  • Create New...