Part I - To be done with a partner or group of 3

Discrimination of natural features on remotely sensed images

Step 1: Familiarize yourself with Landsat Thematic Mapper data

• This is where the Landsat TM bands lie on the EM spectrum (compared to another available US sensor).
  
  http://www.geo.unizh.ch/~kaeaeb/glims/bands.jpg
  
• Here's a succinct description of TM bands, what they "see" and what RGB combinations best show land features (biased toward vegetation).
  • http://gif.berkeley.edu/documents/Landsat%20Band%20Information.pdf
• And a much longer, more thorough one.
• http://web.pdx.edu/~emch/ip1/bandcombinations.html
• and even more if you want to go berserk
  • 3 page landsat flyer, good overall description http://www.geoimage.com.au/geoimage/uploads/Landsat_summary.pdf
• http://www.casde.unl.edu/tutorial/rs-intro/bands_sensors/landsat.php
• http://gdsc.nlr.nl/gdsc/en/information/earth_observation/band_combinations
• http://www.geoimage.com.au/geoweb/pdfs/flyers/LANDSAT_flyer.pdf
• identifying growing vegetation (and why) http://rangeview.arizona.edu/Tutorials/intro.asp
• urban (http://www.ldeo.columbia.edu/~small/Urban/UrbRef.html )

Step 2: Gather some information about "spectral signature"of common materials at the earth's surface (here are some examples. use image search for others)

http://gdsc.nlr.nl/gdsc/sites/default/files/images/spectrum4.gif

modified from http://www.geoimage.com.au/geoimage/uploads/Landsat_summary.pdf

Step 3: For the this Virginia scene (R:\GEOL260\\sat_data\TM16_34_1999_1-7.ers)

I want you to learn to be able to see different types of features (different kinds of vegetation, bare ground, buildings, waters....etc)

1. from the websites above and course notes, determine apriori what band or bands best describe the unique spectral signatures of the following features for a 7 band TM image.
   • bare ground
   • deciduous vegetation
   • pine trees
   • grass
   • urban
   • water
     
2. Determine how you're going to be able to detect these features using RGB images and ratios of two bands.
   
3. On your web document, write down your hypotheses of how you're going to distinguish these features based on their spectra before you go any further. Email me me the same text before you get rolling.
4. using the cell profiles box (with the averaging checked), determine the spectral signature of the features and screen capture the results of the cell profiles box and some of your example locations (alt - print_screen) . Explain the differences.
5. make a few greyscale and/or RGB images using any of the landsat bands that best highlight your features and the difference between them. Make a screen capture of them and explain why it works.
6. make a ratio of any two bands (or more, and use in RGB images with other ratios or simple bands) to highlight the differences between one or more of your features and capture the image.  Exlpain your ratio using the spectral profiles. 

Part II - Either Option 1: a Geological features in the Puna Plateau or Option 2: an ecology and morphology of Barrier Islands

Option 1: Using remote sensing to map geology in the Puna
The images we're going to use are from an area where Prof. Connors did some of his PhD research in the Argentinian Andes, on the southeastern margin of the Altiplano. It is extremely arid, so it will give us a nice contrast with Lexington. The dataset is called is located in R:\satellite\argentina\231_076_21jun00_TM7. Open the scene with the R:\GEOL260\sat_data\RGB_321_argentina.alg algorithm and save the algorithm to your own folder. (you don't need to save the data there)
The data we have are from an EarthSat EarthCover dataset, which means they've been contrast stretched. As a result, you'd not be able to do some things numerically, but they'll be fine for viewing. If you want to see where you are in Google Earth, open this kmz file .

To the West are closed basins with salt playas surrounded by alluvial fans.  To the east is the Eastern Cordillera underalin predominantly by Precambrian/Cambrian basement.  Underlying them is a complexly folded and faulted section that includes Ordovician, Cretaceous, and Tertiary marine and non-marine sedimentary units.  This entire portion of the range was uplifted in the Miocene, which ended the depostion of large scale units. I'll show you some of the highlights of this area, focussed on the two regions shown above.

Your assignment for Part II-Geology is the following.

A.  For approximately following area ("area 1") in the northern part of the image, locate and briefly decribe the following features:

1. an angular unconformity (try infrared bands)
2. map as many geological structures as you can find (folds, faults, etc)
3. an incised pediment (Miocene) and show me where it is being eroded.

 

B. From the eastern Cordillera (location map), find evidence of glaciation by pointing out one or more

1. cirques
2. tarns
3. moraines.(If you don't know what glacial features look like, try here, or better yet, run a search on google, etc for images of glacial features.)

What's different about the location of the glacial features in these valleys when compared with Colorado, where we find similar glacial features.

C.  Find these features

1. aluuvial fan
2. eolian deposits (dunes and thin sand sheets)

and relate the spectral nature of the transported material to its source rock. Puna photo gallery

We cannot all annotate the ERMapper Image, so you'll have to Alt-Printscreen them into Paint, Word, Photoshop or Canvas (or any other program with drawing capability), and show me what you find. Like this syncline in the Cianzo Basin in the Eastern Cordillera

and then paste the results into your web page.

Option 2: an investigation of Barrier Island Morphology

The ERMapper dataset (.ers) for this option is located in the folder R:\satellite\delmarva_TM7\TM_7_earthsat_rectified. Open the ER Mapper file R:\GEOL260\sat_data\RGB321_delmarva.alg and then save the algorithm to your folder in geodata.

1. For Parramore Island (open or save this Google earth KMZ file to the desktop and drop it on Google Earth) and surrouding areas, produce RGB images that enhance the following features, allowing you to identify them.
   1. The habitats as given in this page (http://www.sms.si.edu/irlspec/BarrierIslnd.htm) or this one (http://geologycafe.com/nyc/shoreline/beaches.htm), including
      1. Salt Marsh (can you further refine that by subdividing the salt marsh based on its spectral character? see http://coastgis.marsci.uga.edu/summit/marshes2.htm)
      2. Beach
      3. Dunes
      4. Maritime forests
      5. Former beach ridge/dune lines (like this one)
      6. Washovers (on photos here)
   2. Ebb Tidal Delta
   3. Evidence for spit growth, barrier/beach recssion, barriet island thickening, and other changes of shape of Barrier Islands through time. Which way is the sand moving along the coast of the Delmarva Penninsula.
      
2. Okay, let's up the scale some. We know that everytime glaciers have covered and uncovered the polar landmasses (down to Chicago, for example) sea level has dropped then risen by 100 m. Can you find lines of former barrier islands from former interglacial high sea level periods (see this page for some help http://coastgis.marsci.uga.edu/summit/islandformation.htm
   
3. Lastly, can you use the imagery to show evidence of the influence of agriculture on water quality (hint, look at the estuaries on the east side of the penninsula) and changes in water quality/depth in the bay.

We cannot all annotate the ERMapper Image, so you'll have to Alt-Printscreen them into Word, photoshop or Paint (or any other program with drawing capability), and show me what you find. Like this

Put your findings for both parts I & II.