The scene shows the Los Angeles basin, the Transverse Ranges (and especially, the San Andreas fault) and adjacent segments of California, broadly the same "real estate" as you saw in the first image brought on line in Section 4. The black and white photomosaic, in which the interior parts of ~ 8000 individual photos are included, when reduced to a 23 x 28 cm (9 x 11 inch) print has moderately better resolution than the TM color composite but the differences are not great. Because of the efforts expended, this mosaic shows little blockiness or variations in tone within each image component nor any discontinuities at the margin, although some parallel dark banding in the Mojave Desert (see sketch map for Los Angeles image in Section 4) may be a processing artifact. In this sense, it approaches in internal "purity" the even lighting that exists in a Landsat image that is scanned over a period of about 28 seconds. By varying contrast and other factors during image production, a single Landsat image can be made to look rather like a photo-mosaic of this quality but the latter is panchromatic (film sensitive to all visible colors) and thus has a different mix of brightnesses (tonal patterns) than that of the more narrow spectral region represented in a black and white image from a particular Landsat spectral band. One can approximate a panchromatic film image by calculating the first principal component made from the multiple bands making up a Landsat scene and then writing this out on a film recorder.

Space mosaics can easily be made from Landsat (or SPOT or other imaging satellite systems) inputs. A sequence of Landsat images downtrack is continuous and has no overlap. Generally, the next orbital line, obtained at roughly the same time the next day, will give rise to about 10% sidelap at the equator, increasing to about 40% near the poles owing to longitudinal convergence. The obvious problem in constructing a mosaic that is several to many Landsat orbital paths wide is that, unlike aerial flight patterns in which the lines are flown minutes to a few hours apart, the space paths are days apart so that the likelihood of constant weather conditions (especially minimal cloud cover) is increasingly lowered as the size of the mosaic enlarges. This can be lessened by using cloud-few images taken at different times of year but then sun angle, vegetation state, and other factors enter the picture (in their own way) so that good scene matches over disparate dates are unlikely. But, since Landsat has operated for many years, the probability of getting cloudfree images at generally the same time of year over the large areas to be mosaicked goes up. However, unlike film-based primary imagery (as are aerial photos), the digital format for the Landsat (SPOT, etc) images permits a number of cosmetic enhancements to be made to the "raw" data such that the variations in natural lighting conditions inherent to Landsat imagery acquired on different dates can be largely eliminated by computer-processing.

(this process is described in Section 11) at a scale of approximately 1:7,600,000 and resolution of 305 m (1000 ft), in which the entire states of California, Nevada, Utah and Arizona are depicted. The DEM image, with its controlled artificial shadowing, graphically singles out the structural expressions of mountain-scale landforms in the physiographic provinces you examined in Section 6. These include elements of the western Rocky Mountains, Basin and Range, Colorado Plateau, Sierra Nevada and the California Coast Ranges. Note also the smoothed out flatness of valley floors corresponding to the Basins and to the Central Valley; this digital rendition also brings out the linear character of the San Andreas fault. A sense of relief is less obvious in the Landsat mosaic but the structural features emphasized in the DEM image generally have visual counterparts in the mosaic largely because elevated terrains tend to be vegetated (hence dark in this red band version) so that tonal contrast substitutes for the shadowing effect in the DEM version. Areas given to agricultural (e.g., the Central Valley and Imperial Valley) show a distinctive light-dark mottled pattern; all such land use patterns are absent in the DEM version which purports only to show "raw" topography. At the scales and resolutions of both Landsat and DEM image mosaics, it is difficult to pick up any tonal or geometric patterns that would disclose the presence of large concentrations of population but an alien observer from outer space presented with the Landsat image might guess at the existence of intelligent beings from the field patterns of growing crops.

---

[an error occurred while processing this directive] [an error occurred while processing this directive]