Windy Friday

It has been blowing quite a bit on Friday Jan the 4th 2008 in San Francisco... See the weather map, it was not exactly unpredictable.

Dustlets

Based on the Dustlet concept developed by Nick Thompson, we can now visualize the wind of the GRIB files this way. It is rendered through a Java applet. It's a great tool to see at a glance what the current situation is like.

GoogleMap & wind Improvements

Some improvements have been added to the GRIB visualization in Google Map

Dynamic Patterns

We introduced a while back this concept of pattern. Once you have reshaped all several faxes to fit your needs, you can "replay" the re-shaping step by creating a pattern out of it, so you can apply it again on other similar faxes, stored on your file system. That is pretty convenient, when you use the GetFax utility from SailMail, which stores the downloaded files on your disk.
We have also added some features, allowing you, when you are planning a trip, to download from the UI faxes from the Internet. In that case, you download the fax(es) from the net, save them on your disk, and then apply a pattern on them. Mmmh..., there is room for improvement here!
This is where we've introduced the concept of dynamic pattern, containing the URLs of the faxes to download before applying a pattern on it.
As a result, the picture below is just two clicks away. One to invoke the dynamic pattern, another one to show the result in Google Earth... Google Map behaves the same. Click on it to enlarge.It is made out of three faxes:

1. North Pacific Surface Analysis
2. North Pacific 500mb analysis
3. Central Pacific Streamlines

Notice that this last one has to be rotated 90 degrees clockwise, as it comes from the net on its side. This rotation is now part of the pattern as well.

Contour detection in the GRIB

Some areas, like south Pacific, south Atlantic, are apparently not covered by weather faxes, but it is possible to retrieve GRIB files for those regions, as they are computed using a model. In order to have on those binary data a view equivalent to what the weather charts would be showing, it is necessary to detect the contour lines from the GRIB, about the PRMSL, 500mb, waves heights, temperature, or whatever data the GRIB would contain. This is now done, and here is the rendering it makes. I'm not sure the algorithm wouldn't deserve some improvements (actually, I'm sure it will), but it seems to render the data I'm expecting. Click the image (top-right) to enlarge it.
Just for fun, you can also superimpose the contour lines with their equivalent faxes, to see if it matches...
That should be very useful for the areas not covered by the faxes.

More Google stuff

Still playing with projections and images...
Google Earth and Google Maps can superimpose images on the charts. The problem is that they recalculate the coordinates of the document - an image - based on an Anaximandre (square) projection. The NOAA faxes are available as Mercator projection; so I had this interesting Java exercise to turn an image representing a Mercator projection into an image representing the same data, but using an Anaximandre one. Interesting, indeed. The same operation is probably available for the other projections, we'll see that later.
The Google Earth one is pretty interesting, as you can use it in conjunction with other overlays, like clouds, winds, etc, from different origins. It's quite snappy.

If you have Google Earth installed on your computer, download this kmz file, and try for yourself, it's for the North Pacific. And if you don't have it, you should get it, it's free.
The cloud layer in the picture above can be reached through http://www.barnabu.co.uk/live-global-clouds-for-google-earth.

GRIB and Satellite Picture

A first version of the software, allowing to superimpose a GRIB file on top of a satellite picture is now available. Calculating the projection was quite a challenge, but it seems to work.
It really relates the wind and the clouds. GOES-12 (Geostationary Operational Environmental Satellites, West) provides very suitable pictures for this kind of exercise. The ITCZ (Inter Tropical Convergence Zone) becomes really obvious, very interesting to see that.
The same kind of superimposition can be done with the faxes received from SailMail, I will very soon give it a try.
The image on the right can be clicked, to see more details.
Enjoy!

GRIB Files in Google Map

This is a bit of a gadget, but it looks good. A given GRIB frame can be rendered in Google Maps.

500mb in 3D

The GRIB files can also contain data related to the 500mb charts. A 500mb fax would typically look like this:

It is really supposed to represent the topography of the atmosphere...
Other rendering(s) of the same data might be helpful, like this one:

Plus it is really animated, you can orient it the way you want, just with the mouse.
The four lows are really easy to spot, notice that the one in the Gulf of Alaska is not that impressive, and look - it is not on the fax, it's beyond its edge - how much fun they're going to have in the south Japan...
Sea temperature and PRMSL (Pressure at Mean Sea Level) can also be represented the same way. I want to check if wave heights can be represented the same way...

Smoothing the GRIB

The smoothing algorhythm mentioned in an earlier post can be adapted to the GRIB Data.
It allows you to turn this:

into that:

This smoothing is spatial, in other words, it fills the gap between the points.
This can reduce substantially the size of the GRIB data to download, as it gives the possibility to rebuild the data between the actual points. This will need some validation, but that sounds promising. SailMail users might be concerned; on the example above, the size of the GRIB file is ~15 times smaller.
The next step will be to add a time dimension, smoothing the wind across several frames.
Keep posted!