I almost forgot I was collecting this data!  Here is a time-lapse animation of the map at waterwatch.usgs.gov, which is described as “real-time streamflow compared to historical streamflow for the day of the year.”  You can see where drought persists, and track the sweep of storms that bring streams out of their banks.  Also interesting how entire states flicker in and out, so they must be collecting data by state.  This animation covers the period from January 1 to May 29, 2010.

YouTube Preview Image
Video licensed Creative Commons Attribution license.   Music “Black Rainbow” by Pitx at ccMixter.org, licensed under Creative Commons Sampling Plus license.

This is a 12-month animation of color-enhanced images from NOAA’s GOES East geostationary weather satellite. It covers all of 2009 except for brief periods when my internet connection or my computer were down.  The date and time of each image is at the upper left, if it survives YouTube’s reencoding of the video.  This is a bit long, but here it is for the record:

YouTube Preview Image

This video is made from infrared images, which measure heat radiation, so what you see here is basically temperature. The color is artificial (obviously!) and is used to enhance the contrast between different temperatures. The temperature is significant because severe weather is usually associated with convective systems which generate tall clouds, and the tops of tall clouds are colder than the tops of low clouds. But that daily pulsing you see isn’t necessarily clouds, as discussed below.

Note that the credits at the end of the video have an error in the URL of the original images, which should be http://www.goes.noaa.gov/GIFS/ECI7.JPG.

Read the rest of this entry »

What are those flashing, oscillating spots or blobs on radar animations?  Short answer: noise.  NOAA has the long answer.  In part:

Echoes from surface targets appear in almost all radar reflectivity images. In the immediate area of the radar, “ground clutter” generally appears within a radius of 20 nm. This appears as a roughly circular region with echoes that show little spatial continuity. It results from radio energy reflected back to the radar from outside the central radar beam, from the earth’s surface or buildings.

Under highly stable atmospheric conditions (typically on calm, clear nights), the radar beam can be refracted almost directly into the ground at some distance from the radar, resulting in an area of intense-looking echoes. This “anomalous propagation” phenomenon (commonly known as AP) is much less common than ground clutter. Certain sites situated at low elevations on coastlines regularly detect “sea return”, a phenomenon similar to ground clutter except that the echoes come from ocean waves.

Radar on Blue Marble image

Radar on Blue Marble image

In many national radar animations you can see these blobs of ground clutter reflections sweep across the US from east to west, and the timestamp shows that happens mostly at night.  I guess that must be what they talk about in the second paragraph quoted above.  Follow the link for more information about NOAA’s radars and what you see.

This noise is annoying when displayed on the white map of the US.  It is very annoying when displayed on the Blue Marble map because the noise contrasts so much more with that background.  It’s really too bad that NOAA does not provide radar images with the noise removed.

The ground clutter reflections can be effectively removed, a fact demonstrated by many (most? all?) commercial weather services.  I have made some really amazing looking time-lapse videos with radar images from Weather Underground.  Unfortunately, those images are copyrighted, so I cannot in good faith redistribute the videos made from them.

Could I remove the clutter?  Playing with an image in GIMP makes me think that eliminating all pixels at or close to white would help a lot.  I don’t think I can do that with the command line tools I use (mostly ImageMagick).  Writing a program to do so might happen some day.  It would probably look  for pixels with a high value and very low saturation, i.e., very close to white.  That algorithm wouldn’t get the blue that shows up close to the radar, but just removing the surges of white would help a lot.

Update 1-Sept-09:

Actually, it looks like it might be pretty easy to remove the noise by looking at hue and saturation.  The problem is getting the right headers and libraries on my main development systems.  Right now I’m in dependency hell, it would appear.

Update 8-Sept-09:

See this post about the solution.

Radar on Blue Marble image

Radar on Blue Marble image

This is a test run of the nation-wide (lower 48) weather radar animated on a Blue Marble image, with state outlines.  The YouTube link is here, and it’s embedded below, along with a short discussion of the necessary steps to make it, and what could make it better.

Read the rest of this entry »

Here is all of 2008 in infrared from NOAA’s eastern GOES satellite. This was the wettest year on record here in the mid-Mississippi valley, and the 2008 Atlantic hurricane season was a fairly busy one.

YouTube Preview Image

Note that the images used here do not show clouds per se — it is infrared sensing. That generally correlates to cloud cover, but it’s not identical. There are often low clouds when this image would make you think the sky is clear. High clouds and thick clouds with high tops — the weather makers — are what these images show most clearly. And those are the interesting ones, of course.

The whorl and play of weather systems sweeping across the continent is always fascinating. In certain seasons you can see sea breeze convergence creating diurnal pulses of cumulus clouds over Florida and the Greater Antilles.  The rain shadow of mountain ranges is evident.

Read the rest of this entry »