Skip to content

Lab 4: Downloading Remote Sensing Data

For this lab we will go through the steps of downloading Landsat 8 data from Earth Explorer.

  1. Go to the website [http://earthexplorer.usgs.gov]{.underline}

{width="6.0in" height="3.225in"}

  1. If you do not have an Earth Explorer account, you can register for a new account by clicking Register at the top right of the web page.

  2. Once logged into the system, you can navigate to a location of interest using either:

    • the zoom in/out and pan functions provided at the top left of the page (in the red rectangle shown above), or

    • By clicking under the Coordinates tab, selecting Decimal or Degree/Minute/Second for coordinate entering and then clicking Add Coordinate to enter one or more pairs of coordinates.

  3. For this lab, use the second method to enter the coordinates of Lafayette, IN as shown in the box below. Note that directions to the North and East are positive, while those to the South and West are negative.

{width="4.1252416885389325in" height="1.4025in"}

  1. You can also add multiple coordinates to generate a bounding area, or define a region of interest by uploading an ArcGIS shapefile or Google Earth KML file.

  2. For this lab, zoom in to Lafayette, IN using the coordinates you marked in Step 4. When you have the city (West Lafayette and Lafayette) centered and mostly filling in the window, click on the Use Map button. This will set your region of interest to the extent of the window, so everything will turn red. Zoom out a bit, and you will see that you now have a rectangular region of interest containing both cities, as shown below.

{width="6.319444444444445in" height="3.9974639107611547in"}

  1. Now set the Data Range to search for images over the last 6 months.

  2. Near the bottom of the screen, click on Data Sets >>

  3. Find Landsat -> Landsat Collection 1 Level 1, and select Landsat 8 OLI/TIRS C1 Level-1, then click Results >> to get the list of all images that capture the region of interest during the specified date range.

{width="2.86082895888014in" height="2.3833333333333333in"}

  1. You will now have a list of pre-view images, which are very small but can be used to survey images for clouds or other problems. You will also have the image ID (a very long name that is unique to each image), the date of acquisition, the path and row.

Path and Row

Path and Row are used by Landsat to "grid" all imagery as it is collected.

  • The path indicates the specific orbit,

  • The row indicates the slice of the orbit (so along path distance) used to create the rectangular image you see in the preview.

Depending on how you defined you are of interest, you may find that your search returned images in different paths and/or rows. The Paths increase going from East to West, you should have a selection of Paths 21-23 for the Lafayette area (I am displaying the "footprint" of images on Path 21 (red) and Path 22 (blue) on the image below). Row number increases from the northeast to the southwest across the United States. Landsat is designed to collect a new image of each point on Earth every 16 days, so note in the same image that the first image from Path 22, Row 32 was acquired on December 10, 2016, while the next image of the same location was collected on December 26, 2016. Images cover the same spatial area, and are acquired at the same time of day, but are collected only every 16 days. Parts of Lafayette were acquired more frequently, since Path 21 and 22 overlap, and Path 21 was acquired on December 19, 2016.

{width="5.801234689413823in" height="3.632622484689414in"}

Image Tools

Each preview image has a menu of action icons.

  • Click on the footprint button to see the extent of the full landsat image (as seen above)

  • Click on the image button (shown below, highlighted with green background), and a jpg image will be drawn over the map background (and footprint, if you had it turned on).

{width="3.792075678040245in" height="1.2168744531933509in"}

  1. Now use the footprint and image tools to look through the images that your search identified as covering your area of interest, and being available during the period of time your requested. You want to identify an image that is mostly cloud free (you should be able to identify the Wabash River and major tributaries in the preview image), and covers all of the Lafayette -- West Lafayette area. The image below covers most of the selected area, but is affected by clouds especially in the southwestern part of the image.

{width="5.415228565179352in" height="2.910685695538058in"}

  1. Once you identify a suitable image, download the image data by clicking the download icon (indicated below).

{width="3.547825896762905in" height="0.8521741032370954in"}

  1. Choose to download the Level 1 GeoTIFF Data Product. This is the "raw" imagery that requires special tools to view and analyze[^1].

{width="3.257824803149606in" height="1.0301706036745406in"}

  1. Downloading may take several minutes. While it is downloading, review the ESRI blog page https://blogs.esri.com/esri/arcgis/2013/07/24/band-combinations-for-landsat-8/ to view standard Landsat 8 band combinations. That page also provides information on the spectral and spatial resolution of all bands on Landsat 7 and 8[^2].

  2. Once the image file has been downloaded, it will need to be extracted twice as it is a gzipped-tar file. You can right-click on the file, select 7-Zip -> Extract Here to get the process started.

  3. When it is completely extracted, the directory will have multiple single band raster files, similar to the directory shown below

{width="4.863229440069992in" height="3.890582895888014in"}

  1. Start ArcMap with a blank project document.

  2. Use the Add Data tool to import Landsat 8 band files (LC08_...T1_B#.TIF) and add them to the current project. There are 11 bands in the downloaded data file, for this assignment only import the three bands used for the False Color (urban) visualization, so bands in the list are 7, 6 and 4 (keep them in that order, or you will have trouble remembering which is which after the merge, since they will be named for the order they are added in the composite image). This will take a little time since ArcMap will create "pyramid" layers for each to facilitate rapid visualization at a variety of scales, so as you pan and zoom around the image later.

  3. In order to make use of the band combinations in the ESRI document, you will need to merge the individual raster images into a single multi-band composite raster. Open ArcToolbox and go to Data Management Tools -> Raster -> Raster Processing -> Composite Bands

  4. Import the three bands needed for the False Color (urban) image, and create a new raster image file.

  5. Make sure that the composite raster is visible, and that the display is set so that R = Band 7, G = Band 6, and B = Band 4. You may want to turn the other layers off so that they do not continue to redraw.

  6. Zoom in on the image so that Lafayette is centered and clearly visible.

  7. Switch ArcMap to Layout view, and position Lafayette in the middle of the standard US Letter page that is the default layout.

  8. Export the image to an image file (geoTiff preferred).

  9. Submit the image file to Blackboard, and you are done with the final lab assignment.

[^1]: I put "raw" in quotes, because the USGS completes geo-referencing for all Landsat imagery, so none of it is truly "raw".

[^2]: The thermal bands are resampled from their raw 100 meter resolution to a resolution of 30 meters in the data product that you downloaded. That simplifies analysis in tools like ArcGIS, but is another example of how the Landsat 8 data is not "raw" imagery.