Spectral band analysis with satellite imagery of the drying of Lake Mead (UCLA Master's Coursework)

Since September 2021, this writer has been enrolled in the all-online Master of Applied Geospatial Information Systems and Technologies (MAGIST) program at UCLA. Geospatial information systems (GIS), essentially mapping and analysis of all data with a physical location attached, is a key field for understanding and influencing the changing world of the Anthropocene, with potential applications ranging from siting new renewable energy projects to researching temperature, moisture, and biodiversity shifts in ecosystems. As my coursework continues, one project in particular showcased key concepts from satellite imaging: spectral bands and spectral indices. Earth observation satellites (in this example, the latest in America's venerable Landsat program) host instruments that observe, record, and measure light from Earth in specific regions or "bands" of the electromagnetic spectrum. The schematic above  shows several commonly observed bands: red, green, and blue in the visible spectrum.

 The map above is a false-color satellite image of Lake Mead, formed by the Hoover Dam and the largest reservoir in the United States. The map above looks weird because it's showing light from the Near-IR Band, which human eyes can't see, as red, while showing light from the SWIR1 (Short Wave InfraRed 1) band, one of the two "Thermal IR Bands" on the spectrum schematic above, as green, and finally showing red light as blue. This sort of thing is helpful because different wavelengths of light show different things: for example, due to the physical properties of the chlorophyll molecule, living vegetation is bright in green wavelengths of light (as we see) but is really bright in the near-infrared, so living plants pop out unmistakably if the NIR band is viewed as red on a map. In this case, the band combination was chosen to emphasize the difference between land and water. 

The maps above and below takes it a step further, by using a spectral index, where some math is done to the results of information from different spectral bands to show new and interesting things. In this case, the project used the Modified Normalized Difference Water Index (MNDWI), the equation for which uses the green and thermal infrared bands to calculate a light pattern that enhances the distinctions between water and land even more. In this case, I looked at Landsat images from 2000 and 2021, applied MNDWI to both of them, and did an approximate tracing of the lake's coastlines in both years. (Measurements of water/land boundaries, no matter how good, are only ever better or worse degrees of "approximate," never exact, due to the coastline paradox). 

These maps rather clearly show a disturbing consequence of the climate crisis, made worse by mismanagement (particularly overuse by profligate local agriculture) of the Colorado River watershed. Lake Mead is shrinking, profoundly and rapidly. It's now at only 36% of its historic capacity, and rationing, the first ever, will go into effect (for industry/society sectors as a whole, not individual consumers) in 2022. NASA's explainer uses a very similar visualization to the map above. This example is fairly simple, but it's a good illustration of how satellite imaging, and playing around with different sorts of light, can be a source of valuable insight in the Anthropocene-likely a recurring theme in future research and reporting!