Section 1: The Electromagnetic Spectrum 1. The wavelength range that has the highest reflectance for broadleaf vegetation and needle leaf vegetation is 0.75µm to 1.05µm. 2. Dry soil can be distinguished from vegetation at wavelengths when the dry soil percent reflectance value is different from the vegetation percent reflectance value. Dry soil can be distinguished from vegetation between the wavelength ranges of 0.5µm to 0.6µm, 0.65µm to 0.75µm, 0.8µm to 1.05µm, and 1.2µm to 1.4µm. 3. There is little separation between vegetation types at 0.6µm because this is the wavelength of green visible light that is reflected most by vegetation due the specific chlorophyll pigment content in chloroplasts within plant cells. Since both these vegetation types contain chloroplasts in plant cells with mostly chlorophyll pigments that absorbs all visible light but green, there is not a great difference in the reflectance at the green light wavelength (0.6µm) between broadleaf and needle leaf vegetation. 4. At 0.85µm broadleaf vegetation, snow and ice, and clouds look almost identical because each of their reflectance is 80% (+/- 10%). At 0.4µm snow and ice, and clouds look almost identical because each of their reflectance is 80% (+/- 10%). Also at 0.4µm clear water, turbid water, needle-leaf vegetation, broadleaf vegetation, dry soil, and wet soil look almost identical because each of their reflectance is 30% (+/- 10%). 5. I added the 5 new spectra from table 1 onto the graph labeled as figure 1. Spectra 1: The feature associated with spectra 1 has similar spectra to the clear water spectra already represented on the graph (black line). However, the spectral range of reflectance for feature 1 is much narrower (reaches 0% reflectance at 0.6 µm) than the clear water feature (reaches 0% reflectance at 0.85 µm), so it has a higher absorbance of visible light than clear water. The higher absorption of visible light will make this water feature appear darker in colour. From this information I think feature 1 is clear tropical ocean water with very little phytoplankton. Spectra 2: The feature associated with spectra 2 follows a trajectory in between the turbid water spectra (grey line) and the wet soil spectra (brown line). I think this feature is a bog or wetland based on its similarity to bot turbid water and wet soil. However, since spectra 2 does not follow any vegetation spectra pattern, feature 2 does not include any vegetation. Spectra 3: This feature s spectra follows the vegetation spectra, but with a lower reflectance at the green light wavelength and a larger reflectance in the near infrared range. Lower green light reflectance than other vegetation spectra indicates feature 3 is less green in colour. Also, this feature reflects a lot more energy in the near infrared range (0.7µm to 1.4µm) and is similar to the dry soil spectra (brown line). I think this feature is a mature agricultural crop, since it reflects more in the near infrared wavelengths than regular vegetation (broadleaf and needle leaf) but not as much as dry soil. Spectra 4: This feature s spectra very closely follows the vegetation spectra, however it has a much lower overall reflectance than the other vegetation features. Particularly it reflects much less in the green light wavelength and near infrared range than other vegetation, so feature 4 must be less green in colour than
vegetation. In the near infrared wavelength range this feature s spectra is between the turbid water spectra (grey line) and the vegetation spectra (green line), indicating higher water content than in vegetation. From this information I think feature 4 is wet decomposing plant material. Spectra 5: This feature s spectra is at a constant reflectance of about 80% across the 0.4µm to 1.4µm range. This feature must be quite light in colour and shiny since most visible light is reflected, like the snow and ice feature (red line) and the clouds feature (blue line). From this information I think feature 5 is smog. Figure 1: Percent reflectance for various features, including the five spectra from Table 1, at different wavelengths from 0.4µm to 1.4µm. Section 2: Landsat 7 Bands, the Electromagnetic Spectrum & ENVI Software 6. The Landsat s thermal band is recorded with coarser spatial resolution than the other bands because of thermal radiation s larger wavelengths relative to other bands wavelengths (ex. visible light s). Since the thermal band has larger wavelengths it must also have lower frequency. As per the equation Q=h(1/λ) (derived from Q = hv and v=1/λ), the longer the wavelength, the lower the frequency, and the less energy per quanta. However, the LANDSAT detector still requires a certain amount of energy per quanta to make a reliable measurement. In order to make a reliable measurement Landsat coarsens the spatial resolution by increasing the instantaneous field of view (IFOV), which allows more energy to reach the detector at a time. 7. Vegetation appears red in a standard false colour composite because the Landsat image s green band (band 2), which is highly reflected by vegetation, is displayed through the computer monitor s red gun.
Contrastingly, in a true colour composite the vegetation appears green since the Landsat image s green band (band 2) is displayed through the computer monitor s green gun. 8. I think there are many possible best colour composites that can clearly map vegetation, but the best one that I chose uses RGB: band 2 (green), band 4 (NIR), band 1 (blue). The screenshot of this colour composition is shown as figure 2. This composite takes advantage of colour contrasts and makes the green coloured terrestrial vegetation more pronounced against the surrounding purple water and pink urban development. Vegetation has high reflection in the NIR band, making areas of vegetation starkly green and clearly distinguishable. In the true colour composite the water surrounding Vancouver is quite light green in colour and makes the surrounding green vegetation less obvious and harder to distinguish. Using the red gun for green reflected light reduces this difficulty because the greenish coastal water appears pink in the false colour composite. Figure 2: Screenshot of Landsat 5 Thematic Mapper image of Vancouver and its surroundings (taken at 22 September 1999) in RGB: band 2 (green), band 4 (NIR), band 1 (blue). 9. To look at water quality I would use the band combination of RGB: band 4 (NIR), band 2 (green), band 1 (blue). The image produced (figure 3) shows clear water with dark blue, turbid/ silty water with green and turquoise, and aquatic vegetation with red. The red and light blue patchy coloured land starkly contrasts the water features. I chose band 2 (green) since turbid/silty coastal water reflects more green light than clear water, and band 4 (NIR) since phytoplankton/algae contain chlorophyll that reflects NIR wavelengths.
Figure 3: Screenshot of Landsat 5 Thematic Mapper image of Vancouver and its surroundings (taken at 22 September 1999) in RGB: band 4 (NIR), band 2 (green), band 1 (blue). To look at agriculture I would use the band combination of RGB: band 4 (NIR), band 5 (SWIR), band 1 (blue). The image produced (figure 4) shows a stark contrast between vegetation (yellow, orange, brown) and other features, like water (blue), dirt/silt (green), and buildings (blue). Agricultural land (containing vegetation or bare dry soil) is more yellow and orange in colour, where as naturally vegetated areas are brown. This is because in September the crops become dry and ready to harvest, so they have lower water and chlorophyll content, therefore reflecting more into the SWIR bands (like dry soil) than natural vegetation. Figure 4: Screenshot of Landsat 5 Thematic Mapper image of Vancouver and its surroundings (taken at 22 September 1999) in RGB: band 4 (NIR), band 5 (SWIR), band 1 (blue).
To look at an urban area like Greater Vancouver I would use the band combinations of RGB: band 5 (SWIR), band 4 (NIR), band 1 (blue). The image produced (figure 5) shows developed areas (buildings, cement, and asphalt) as purple and pink, and undeveloped areas (agriculture, vegetation) as red, yellow and green. This is because soil and vegetation reflect a lot in the NIR to SWIR range (indicated as colours from red to green) where as buildings, concrete and asphalt reflect more in the blue visible light to NIR range (indicated as colours purple and pink). Figure 5: Screenshot of Landsat 5 Thematic Mapper image of Vancouver and its surroundings (taken at 22 September 1999) in RGB: band 5 (SWIR), band 4 (NIR), band 1 (blue).