This and the following posts in this series are excerpts from a Landsat tutorial I wrote for a GIS course that I recently taught. Although the procedures are simple, they are far from obvious for many potential users and I thought that I could well put the tutorial here where it can be beneficial to many more people. In this first part, as I did with the MODIS tutorial, we focus on selecting and downloading the data, and leave processing and correction for later posts.
The Landsat program, a partnership between NASA and USGS, has imaged the Earth surface with purpose-built instruments since 1972 (see http://landsat.gsfc.nasa.gov/), offering a unique glimpse into more than 40 years of global environmental change. Landsat 8, launched on 13 February 2013, is the latest addition to this successful program and includes many interesting refinements over its predecessors, like 12-bit radiometric quantization (coded into 16-bit images) and improved signal-to-noise ratio. Perhaps, the most interesting feature of the Landsat program is that all data are free to use for everyone.
We start by opening the USGS Global Visualization viewer (Glovis):
On the left side you will see a tiny map of the world, that you can use to navigate to your spot of interest. Landsat images can also be located using a system known as WRS-2, in which the locations are specified in terms of two coordinates: path and row. All scenes with the same path and row depict the same region of the world. It is a convenient practice to become familiar with the path/row coordinates for your area of interest.
You can choose to see several scenes at the same time, or just one, by choosing the appropriate scale in the resolution tab. You can specify a date, and a maximum cloud cover. Note that cloud cover is estimated automatically and might not be accurate. Always check visually. By clicking on ‘Prev scene’ or ‘Next Scene’ you go forward or back in time through the image database. By default Glovis opens in the ‘Landsat’ collection, that includes all landsat satellites. If you need only Landsat 8 images, just select that specific data set on the tab ‘Collection’.
In this tutorial we will use an image of the beautiful Bay of Naples, Italy, hotspot of history and summer home to a few Roman emperors, corresponding to path/row 189/32. We see that the Landsat 8 image of 29 July 2013 is completely cloud free:
Once we locate the image we desire, we check that it is available for download by looking if a red sign (‘Downloadable’) is shown at the upper left corner. If yes, add it to you list by pressing the ‘Add’ button (left, below). If it is not available for download you may still add it, but instead of downloading you will proceed to request the image for processing, and will be available soon, normally after a day or two.
We then proceed to the download page by selecting the ‘Send to cart’ button. This can seem confusing, because the data are said to be free. What we do in reality is to purchase the images at a cost of zero dollars (the site is also used for other images that one has to pay for, i.e. ASTER). After sending to cart, another tab opens with the EarthExplorer site. Here you might need to register, otherwise sign in. You are guided to a list with your selected data. If you choose ‘Bulk download’ you will get an immense dataset, where not everything must be useful. I advise to click on the download button to the right, that with an icon depicting a green arrow on a hard drive. Select then the last option ‘Level 1 GeoTIFF Data Product’. As you see, files sizes can be very big, and will be even bigger in your computer because these are compressed files. Click ‘Download’:
You will get a big compressed ‘tar.gz’ file, in our case
LC81890322013210LGN01.tar.gz. It takes a while to download. If you are in windows, you might be able to access this format using the program 7-zip. In Linux, any packager will do, or just execute the command:
tar -xvf LC81890322013210LGN01.tar.gz
the output will look like this:
Here we have all the individual band image files (*.TIF), the quality control image (*_BQA.TIF), and the metadata text file, with extension *_MTL.txt.
The details of the individual bands can be found here. Pay special attention that Landsat 8 band names do not correspond with the previous satellites, which can be confusing. A useful comparison with landsat 7 can be found here.
In the next post, we will go through the steps to convert these separate files into colour composites, and transform them to radiance values at top-of-atmosphere. Happy downloading!