Lately I’ve been looking at the (FREE!) MODIS data available. It comes in HDF format, and there are lots of HDF viewers out there that seem to be pretty decent. The one I’ve been using is okay, but by browsing today I think I’ve found at least 5 that would work better - although I haven’t tested them out.

What is MODIS?
MODIS is the Moderate Resolution Imaging Spectroradiometer, as if that helped. You can see that it has lots of bands. A band is a portion of the electromagnetic spectrum over which the sensor can detect photons. For instance, 550 nanometers is about in the green portion of the spectrum, thus band 4 of MODIS would detect “green photons”. Most of the bands are in non-visible portions of the spectrum. From all of these bands, it’s possible to combine them to tell you interesting things about the surface and the atmosphere beneath the sensor. For instance, I’m interested in using their aerosol optical depth product (AOD). A downloaded a sample image from their website and displayed the AOD below.

But what is optical depth? Taken from Wikipedia because I’m too lazy to attempt to describe it in my own words, and theirs are pretty good, “Optical depth is a measure of transparency, and is defined as the fraction of radiation (or light) that is scattered or absorbed on a path. One way of visualizing optical depth is to think of a fog. The fog between you and an object that is immediately in front of you has an optical depth of zero. As the object moves away, the optical depth increases until it reaches a large value and the object is no longer visible.”

There is no data in all the white areas. The more red the color the higher the optical depth. The numbers are wrong. An aerosol optical depth of 1887 would be huge! But I don’t know the scaling factor they used, but I bet it’s either in the header information or on the website. In order to retrieve the aerosol optical depth, there cannot be any clouds. If there were no clouds in this image, the coloured area would be approximately a rectangle. Below, I’ve plotted the cloud fraction.

Cloud fraction is just what it sounds like - the fraction of the pixel that is a cloud. Because each pixel is 1km by 1km (directly beneath the satellite, more at off angles), it’s possible to have portions of the pixel be clouds and portions non-clouded. How they determine cloud fraction is beyond the scope of this post. But you can see that the red area is where there is 100% cloud fraction (designated as 1000s in the image for some reason), and there is a huge cloud covering the Western and Southern portions of the United States. So that’s why there isn’t any aerosol information in those areas.

Another interesting thing to look at is the surface reflectance. Different things on the surface reflect differently at different wavelengths. I’m not sure what wavelenth(s) this reflectance is integrated over, but it still shows some interesting things.

Notice the upper-right part of the colored area, that the Washington/Oregon area of the U.S. It’s colored purples and blues. Directly to the East of that there is a large area of oranges and reds. This corresponds nicely with the Rocky Mountain range.

My ultimate goal is to stitch these fragmented aerosol images together to get a time-resolved picture of the aerosol optical depth.

(No Ratings Yet)

 Loading ...

Related Posts:

Trackback URI |