ES 771 Lab 9
Landsat 8 OLI

Cheyenne Bottoms, Kansas
James S. Aber


Landsat 8 OLI

Landsat 8 was launched and began routine operation in 2013. It continues a four-decade legacy of continuous Earth observation by the Landsat series of satellites and instruments. The Operational Land Imager (OLI) has substantial improvements over previous MSS, TM and ETM+ instruments. It is a push-broom scanner in contrast to older whisk-broom scanners—see OLI instrument.

Landsat OLI and ETM+ scanner comparison
Landsat 8
OLI band (΅m)
GSD
(m)
Landsat 7
ETM+ band (΅m)
Color Primary use
1. 0.433–0.453
30
none
Violet-blue Shallow water, aerosols
2. 0.450–0.515
30
1. 0.450–0.515
Blue-green Shallow water, land cover
3. 0.525–0.600
30
2. 0.525–0.605
Green Vegetation, land cover
4. 0.630–0.680
30
3. 0.630–0.690
Red Minerals, soils, land cover
5. 0.845–0.885
30
4. 0.775–0.900
Near-infrared Vegetation, NDVI
6. 1.560–1.660
30
5. 1.550–1.750
Mid-infrared Soil moisture, minerals
7. 2.100–2.300
30
7. 2.090–2.350
Mid-infrared Soil moisture, minerals
8. 0.500–0.680
15
8. 0.520–0.900
Panchromatic Image sharpening
9. 1.360–1.390
30
none
Mid-infrared Cirrus clouds

Note differences in band numbers and spectral limits.
GSD = ground sample distance (spatial resolution).
Thermal-infrared bands are not included.
Based on OLI bands and Landsat 8.

Cheyenne Bottoms

We will utilize Cheyenne Bottoms to explore the potential for Landsat 8 OLI imagery. During 2013, Cheyenne Bottoms underwent a remarkable transformation. The first half of the year culminated a two-year drought. By July 2013, nearly all portions of Cheyenne Bottoms were dry including the state wildlife area and Nature Conservany marshes. The 2011-13 drought came to an abrupt end with heavy rains and regional flooding in late July-early August. Streams overflowed, marshes and wet meadows were inundated, and high water recharged soil moisture and filled the pools.

Nature Conservancy marshes
Panoramic overview, May 2013 in drought conditions. Small puddles are scattered in the marsh, and green emergent vegetation is sparse.

Early September 2013, following extensive flooding. Left: view toward the northwest with Hoisington on the left horizon. Right: looking toward the northeast with many pools in the distance. Kite aerial photos © J.S. Aber.

Remote sensing research at Cheyenne Bottoms.

A complete Landsat OLI dataset is quite large. It amounts to nearly one gigabyte in compressed format; upon decompression it expands to ~1.7 gigabytes. From the full OLI scene, a small window was extracted that corresponds approximately to the title image. The following examples display Cheyenne Bottoms on July 12, 2013 under ultimate drought conditions.

Landsat OLI standard false-color composite; bands 3 (green), 4 (red), and 5 (near-infrared) displayed as blue, green and red. Active vegetation appears bright red, and fallow fields are green. Few water bodies (black) are apparent.
Landsat OLI infrared false-color composite; bands 5 (near-infrared), 6 (mid-infrared), and 7 (mid-infrared) displayed as blue, green and red. Active vegetation appears bright blue, and fallow fields are yellow. Few water bodies (black) are apparent.

Exercise

Download all Idrisi files cb2013_aug* into your student working folder. Note: these are large files. This includes OLI bands 1-8; band 9 is for special purposes that do not apply here and is not included. This OLI dataset was acquired on August 29, 2013 approximately two weeks after peak flooding took place. Excess high water had drained away leaving pools and marshes full of water.

Examine the metadata for details about spatial resolution, reference system, etc. Bring up the visible-light bands (1-4) using the greyscale palette.

1. Of the four bands in visible light, briefly describe their appearances. Which seem darker and lighter overall? Explain based on spectral conditions.

Now display the three infrared bands (5-7), again using the greyscale palette.

2. How do these compare with the visible-light bands? Explain based on spectral conditions.

To complete your review of individual bands, display band 8 using the greyscale palette. Check the metadata and file size for band 8.

3. What is the spectral range of this band, and what other bands does it most closely resemble?

4. How do spatial resolution, image dimensions, and raster file size for band 8 compare with other bands?


Continue by making two false-color composite images, as above, using OLI bands 345 and 567 color coded as blue, green, and red. Examime the results and compare them with the drought images from July 2013.

5. What differences are obvious in terms of land cover between July and August?

6. What features are relatively constant between July and August?

7. In all images, a bright white/cyan feature appears south of Hoisington in the left middle of the scene. Identify what this feature is and explain why it is so bright.

Band 1 (violet-blue) is designed specifically for looking at shallow coastal or wetland water bodies, as this wavelength penetrates clear water several meters. Now make a false-color composite using bands 1, 4 and 7 color coded as blue, green, and red.

Landsat OLI infrared false-color composite; bands 1 (violet-blue), 4 (red), and 7 (mid-infrared) displayed as blue, green and red. Post-flood conditions, 29 August 2013. Your image should appear similar.

8. Briefly describe how different types of land cover (vegetation, water bodies, human structures) appear in this unusual false-color composite.

9. What variations can you discern between water bodies? What conditions within the water bodies give rise to these visual differences?


Now turn your attention to the panchromatic band (8). Upon initial display (in greyscale palette) this image is rather dark and lacking in details. Make a stretched version (Image Processing, Enhancement) using linear stretch with 1% saturation. Display the image using the greyscale palette. Zoom in and move around to examine features in detail.

10. What are the smallest features you can identify with confidence, and how are they visible?

On your OLI band 8 image, zoom in to the Nature Conservancy marsh area of the Ikonos image you used for the photo-layer display in lab 3. Enlarge the image to see details, as illustrated below.

Landsat OLI band 8 (panchromatic) stretched image. Post-flood conditions, 29 August 2013. Enlarged and cropped for the Nature Conservancy marsh area.

11. Examine the spatial details that are visible in both OLI and Ikonos panchromatic images. Compare the spatial resolution of these two images and the kinds of features that are visible. Note: ignore the differences in land cover between years.

12. Examine the spectral details that are visible in both OLI and Ikonos panchromatic images. Compare the spectral resolution of these two images and the kinds of features that are visible. Note: ignore the differences in land cover between years.


For many people, red or blue vegetation is most confusing. For best visual impact, naturalistic coloration is desirable. This means that active vegetation should appear green, water bodies are blue to black, and so on. The previous false-color composites in this exercise may be attractive and technically useful, but they certainly do not display naturalistic coloration. Achieving such naturalistic colors in a false-color composite image is not as easy as it sounds, especially with so many bands available.

Note: the panchromatic band (8) cannot be combined directly with the other bands because of its different spatial resolution. Resampling other bands is necessary for pan sharpening in combination with band 8. This can be done using "Expand" (under Reformat). Expand OLI bands 5 and 7 by a factor of two, which doubles the number of rows and columns to match band 8. Now make a false-color composite based on OLI bands 8, 5 and 7 color coded as blue, green and red.

Landsat OLI false-color image based on bands 8, 5 and 7 color coded as blue, green and red. Bands 5 and 7 were expanded by a factor of two in order to match the spatial resolution of band 8. Active vegetation is green, fallow fields are pink/maroon, and water ranges from blue to black. Although vividly distinct, the bright pink-and-green combination is a bit too garish, however, to call this naturalistic coloration.

As your final task, experiment with different false-color composite images for the Landsat 8 OLI dataset. You may include any of bands 1-7, band 8, filtered or ratioed bands, NDVI, etc. Your goal is to produce a naturalistic color composite that incorporates both visible and infrared bands. Furthermore your results should highlight visually as many different aspects of land cover as possible. Prepare a map composition that has an appropriate title, subtitle with your name and date, and scale bar. Also include a text box with brief explanation of the bands, colors, and what they show. Save a digital image file to turn in with your answers.

Turn in


Return to course schedule.
ES 771 © J.S. Aber (2017).