Some Basic Principles and Examples of Urban and Land Use Applications - Lecture Note - Completely Remote Sensing tutorial, GPS, and GIS - facegis.com
Some Basic Principles and Examples of Urban and Land Use Applications

In this Section we will consider the category Urban or Built-up Land. Since more than half of the American people (and likewise in many countries) now live in cities or metropolitan areas, one might expect, in land cover percentage, a signficant part of the continental land surface falls within this category. The U.S. Census Bureau places urban areas in their category "Developed Land". To qualify in this category, the Bureau requires that the land holds a minimum of 30 people as residents per square mile. It is surprising then to learn that as of the year 2000 the percentage of developed land in the continental U.S. is just 5.4%. The actual metropolitan area perentagewise is even less. But, since most people live in developed land, monitoring its distribution, its growth, and its characteristice becomes an important task amenable to vital input from earth-observing satellites.

As strongly alluded to in the first three Sections, space imagery (and aerial photography as well) is a powerful medium for mapping what's at the Earth's surface. We can identify and categorize the various natural and man-developed features present on the surface in terms of land cover. The term land useis almost a synonym, but refers specifically to how the land is used for human activities. Land cover and use maps are an essential input to Geographic Information Systems, described in detail in Section 15. Most of these maps follow some classification system. One in common use is that developed by James Anderson and his colleagues at the U.S. Geological Survey. This hierarchical system has four levels. The two higher ones (Levels I and II) are categories that we can usually identify and hence, map, using space images, whereas, aerial photos (for higher resolution) or IKONOS-type space images are needed for Levels III and IV. This classification in Levels I and II is shown here:

Level I and Level II Classifications, US Geological Survey.

To illustrate the subdivisions at Level III (not shown), we pick Level I = Urban - and Level II = Residential, then the subdivisions under Level III are: units = Single-Family; Multi-Family; Mobile Homes; Hotels/Motels; and Other. Generally, the finer the breakdown, the more on-site ground truth information we need. Maps for Levels I and II can be colorized, and we can add numerical symbols (e.g., 115 = Residential Hotels) for individual features on maps drawn at a large enough scale to fit them.

As you saw in Section 1, a common map product now almost routinely made from satellite remote sensing data is the scene classification, done through computer processing. An example of a typical land cover and use map (some Level II categories) is this supervised classificaton of the city of Ottawa, Canada (the country's capital):

Classified map of Ottawa, Canada; the blue pattern is the Ottawa River.

For comparison, here is a multi-theme classification of land use/cover derived from Landsat imagery that extends over an area in and surrounding Harrisburg, PA (you saw other Harrisburg examples in the Fiirst Exam, Section 1). The city is located next to the prominent yellow pattern assigned to urban. Smaller yellow blotches to the east and west are indications of downtowns, such as Carlisle, Hershey, and Lebanon. This metropolitan area, with a population exceeding 300,000, is generally spread out beyond the Harrisburg city limits. The legend category called "deforestation" describes the same Gypsy Moth defoliation touched upon on page 3-6. Land use/cover mapping with space images such as Landsat, SPOT, JERS, IRS is capable of displaying with good (not excellent) accuracy some of the Level II categories. IKONOS and similar high resolution satellite imagery can pick up many of the Level III units.

Example of a typical land cover classification (map) for an area around Harrisburg, PA.

Students at the University of Arizona have produced an excellent Web site illustrating in a series of well-documented steps the approach they took to classifying land use in a rural part of Cochise County, AZ. They compared Landsat subscenes for two dates, in 1973 and 1992, working towards a change detection map that picks out new features or different uses for some earlier features.

For the next three pages, we will concentrate on urban land use by looking at space imagery of a group of U.S. and foreign cities:

The Urban Southwest: Los Angeles, San Diego, Tucson, Las Vegas, San Jose, and Mexico City

Because much of the U.S. and world populations are concentrated in and near metropolitan centers, we concentrate in this Section on urban land uses.

Major metropolitan areas are sprawling, often occupying a significant fraction of a full Landsat scene (although on a world scale, these areas comprise less than 10% of the land surface). This is certainly the case in one of the largest (areawise) urban regions in the world - the Los Angeles megalopolis as seen in a resampled (lowered resolution) MSS image that covers this part of southern California.

Landsat MSS full scene of the Los Angeles metropolitan area, with the Tranverse Ranges and Mojave Desert to the north.

This Landsat image graces the cover of the NASA publication "Mission to Earth", which the writer conceived in 1974. This scene covers only a part of the total megalopolis, which extends well to the east and to the south. A MODIS image broadens the coverage to incorporate the full megalopolis:

Quasi-natural color image made by MODIS that covers the entire megalopolis around Los Angeles.

Major geographic features in this image include the western Mohave Desert (containing Edwards Air Force Base), the southern tip of the Great (San Joaquin) Valley, the Tehachapi mountains on the north (extending from the southern tip of the Sierra Nevada), and the Transverse Ranges north of L.A. (bounded sharply on the north by the infamous San Andreas fault and on their south by the Santa Monica and San Gabriel Mountains). Los Angeles (lower right) lies within a structural basin, forming a lowlands that restricts air circulation and is a natural trap for pollutants (the famed L.A. smog).

A more detailed look is given in this Landsat-7 subset:

Subscene of much of greater Los Angeles, imaged by Landsat 7.

The urban signatures in a false color composite are clearly evident in the segment of the MSS image containing a part of the numerous cities and suburbs that makes up the greater Los Angeles regional sphere of influence. The areas of higher population density show two dominant color themes: thin, usually blue to almost black, linear criss-crossing patterns representing streets and roadways; and areas in between that are also usually some shade of blue. The latter are the spectral expression of buildings, which tend to reflect brightest in the blue and green bands. Interspersed within these areas are patches of red, which correspond to city and town parks, cemeteries, golf courses, and pockets of intra-urban agricultural fields. Residential areas often have color signatures ranging from brown to pink to mild red, stemming from the mix of lawns and trees with houses and streets. In the Los Angeles hills and neighboring mountains, signs of homes and neighborhood centers are hard to recognize because the trees and brush swamp those areas with reds. The Palos Verdes Estates section on the ocean is a good example. However, bands and patches of blue in the valleys, such as in the Santa Clara valley east of Ventura, indicate commercial concentrations along highway strips.

The Transverse Range is quite an impressive backdrop for the citizens of the Greater Los Angeles area to look at from various spots in the basin. This next pair of images are perspective views of these mountains and parts of the Los Angeles basin with its many metropolitan areas made by combining data from NASA JPL's SRTM radar mission (see Section 11) with an enhanced image acquired by Landsat-5.

A perspective 3-D view of the Transverse Ranges and part of the Los Angeles Basin, made by combining Landsat-5 with SRTM radar data.
Another SRTM-Landsat view of the Transverse Ranges; in this image the San Andreas fault bounding the Mojave Desert basin is strongly brought out.

It also helps to show a ground photo of some part of a city or metropolitan area to tie in with the space imagery. We will do this for many of the cities (especially in the U.S.) in Section 4. Here is a view of downtown Los Angeles:

Panorama of downtown Los Angeles.

So, suppose you want to know what's where in downtown L.A. Google Earth, as we have seen before, provides a myriad of products online at their Internet site. This is an example of a high resolution Quickbird image labeled to show the main buildings in part of that downtown:

Google Earth image map of a section in downtown Los Angeles.

Urban areas are quite distinctive in radar imagery, as is clearly evidenced in this 1978 Seasat image (see Section 8 for a description of this satellite and a review of radar interpretive principles) that covers the west side of the Los Angeles metropolitan area. We include a segment of the 1989 Rand McNally atlas map that locates roadways, towns, and other features in the urban infrastructure (its writing is not legible; it is shown mainly to aid you in picking out major roadways in the radar image).

Seasat radar image of the western Los Angeles metropolitan area, 1978.
Rand McNally map of the western Los Angeles metropolitan area, 1989.

Many of the dark linear features in this radar image correspond to interstate highways and other major roadways. The two dark, narrow, east-west strips just on shore at Santa Monica Bay are runways at Los Angeles International Airport. These dark patterns come from specular (smooth) reflecting surfaces that deflect most of the radar beam away from the receiving platform. Generally, buildings return much of the radar signal, giving rise to an intermediate gray tonal signature (uniform here because most individual buildings are not resolved). Very light patterns are associated with high levels of backscatter caused, in some instances, by stands of vegetation. The light pattern along the west coast of the Palos Verdes peninsula, a luxuriant residential area, may come from dense vegetation. However, the very bright, squarish pattern at the east end of the Santa Monica Mountains that extends into the San Fernando Valley around Burbank-Glendale is rather mysterious. It approximately coincides on the map with Griffith Park, north of Hollywood, but its position relative to roadways that are discernible in the radar image, indicates an inexact fit (possibly an effect of non-rectification).

The layout of metropolitan Los Angeles is made more evident when this region is viewed at night, as in this astronaut-taken photo from the International Space Station:

Los Angeles at night.

When "Los Angeles" is mentioned, people often think almost at once of "Hollywood". The IKONOS satellite succeeded quite well in finding Hollywood. In fact, it imaged the famous large white letters (look near top) that spell out this name against the backdrop of a steep slope on the south side of the Santa Monica Mountains near this most recognized city in the Los Angeles area:

The famed Hollywood sign (near top) as imaged by IKONOS.

Now that we have led you through this tour of the Los Angeles region, and prompted you with questions designed to familiarize you with its landmarks, we want to make sure you found the major cities and suburbs in an north of Los Angeles. This will be done using first this subscene made from a Landsat ETM+ image, and then with this Rand McNally map that should be legible enough to spell out localities:

Landsat-7 subscene of part of Los Angeles
Rand-McNally map of the above scene.

A provocative question: what towns in the scene do you associate with the movies stars and other celebrities?

Let's leave the glitter of the L.A. world and journey south to what many believe to be the most livable city in the United States. Even the lower resolution subscenes created from MSS images are effective ways of manifesting the major features of urban areas, such as street patterns, large buildings, airports, recreational parks, and some industrial facilities. This effect is certainly evident in this March 30, 1975 MSS view of the central parts of San Diego in southernmost California. It shows the Bay harbors, military and civilian airports, the downtown, Balboa Park, Mission Bay, and Cabrillo Point.

Downtown San Diego, a Landsat-2 subscene

But, the value of much higher resolution to discerning metropolitan features down to the individual building scale is convincingly revealed by this November 22, 1988, SPOT image of inner San Diego, in which the three multispectral HRV band images, at 20 m (66 ft) resolution, are combined and registered with the 10 m (33 ft) panchromatic image taken simultaneously. The details in this image are persuasive proof that civilian space imagery can match, at least, the smaller scale aerial photos in clarity and information content.

SPOT subscene of approximately the same part of downtown San Diego as shown above.

Many who live or visit San Diego consider it the ideal metropolitan area in the United States in which to live. Its climate is famed for its moderate temperatures and high frequency of sunny days. Its downtown waterfront skyline, shown below, defines its modern fast growth.

Looking east across the harbor at downtown San Diego.

America's Southwest and West Coast are among the fastest growing in population within this country. This is brought home by the next image, a change detection depiction of part of San Jose, California, a city that has more than doubled in size since 1973.

Population change in San Jose, California

A dramatic example of building a city from "scratch" is found near the southern tip of the next state to the northwest of Arizona. Here in the midst of a flat desert basin surrounded by mountains lies the gambling heart of North America, Las Vegas, Nevada. This rapidly growing urban area is now approaching two and a half million in population. Water to nourish this "oasis" is drawn from Lake Mead, behind the Hoover Dam.

Landsat subscene showing the dry desert area within which Las Vegas, NV is now expanding; Spring Mountains to the west (left).

Las Vegas is built around a long avenue known as the "Strip", seen here during its prime time at night. You should be able to pick out that narrow area in the Landsat image.

The Las Vegas Strip at night.

Here it is in daylight as seen by a satellite:

The Las Vegas Strip as imaged by a satellite.
To see a very large but detailed image (Terra's ASTER) of Las Vegas, press here (this avoids an initial download that would have lengthened the time to complete this page; if you access the image, press BACK to return to this page).

Here is an IKONOS image of most of the Strip (near top of image); McCarron Field is in the lower left. We have deliberately left this oversized (for some screens) and have put a picturesque map below, so that you can scroll back and forth to try to locate some of the famous strip hotels in the IKONOS image. A cheap way to visit but not nearly as good as the real thing!

IKONOS image of part of the Las Vegas Strip.
Map of the strip.

IKONOS can produce images with even higher resolution. But the one available online is made by Digital Globe's Quickbird-2, achieving a resolution of 0.68 m. Shown here is the 28-story twin towers of the Excalibur Hotel, on the Strip, and opened in June of 1990.

Quickbird 2-ft resolution image of the Excalibur Hotel on the Las Vegas Strip.

In 1973, Las Vegas had about 358,000 residents; by 1992 the city had grown to more than 937,000. Growth of the city is evident in comparing a 1973 Landsat subscene to one acquired in 1992.

Las Vegas, 1973 LasVegas, 1992

This next illustration gives a more detailed account of Las Vegas' growth by showing 6 different years over a span of 25 during which Landsat-5 has been in operation:

Las Vegas growth from 1984 through early 2009/

In more detail, here is the Las Vegas strip area in 1984 and again in 2009:

The Las Vegas strip in 1984
The strip in 2009.

Growth continued unabated through 2006, when the population topped 2,013,000. The city has been spreading south and east. But water availability problems have now slowed growth.

The main competition for Las Vegas as a gambling center is Reno, NV (once also known as the Divorce Capital of the U.S.). Here it is in a space image and from the ground;

ASTER image of Reno, NV.
Ground view of Reno, with the Sierra Nevada Mountains in the background.

We now look at another major western urban area in a somewhat different setting, Tucson, Arizona, nestled between mountains on three sides:

Central Tucson with the Tucson Mountains in the background

Much of the city, displayed here in a Landsat TM False Color Composite subscene extending about 58 km (36 mi) east-to-west, sits on nearly flat terrain, wedged in by several large alluvial fans from the southeast and north. The latter fan actually is becoming a major residential suburb, noted for fine homes amidst dissected canyons. The entire city area was built up from stream deposits coming off the surrounding mountains. The tip of the Rincon Mountains touches the eastern edge of the city, and the scattered hills of the Tucson Mountains extend to the west. Copper mining is a major regional industry, as evidenced by the dark fan-shaped lake and the bluish-white open-extraction pits of the Pima Mission Mines to the south.

Landsat TM false color composite showing a broad valley framed by the Santa Catalina Mountains to the north of Tucson, AZ (checkered area).

The city itself is enlarged in this next view to bring out some of the details of the town; note the fan topography on the top.

Enlargement of the Landsat TM image with the Tucson area extracted.

Many prominent, wide streets criss-cross the town, being laid out on the surveyor's section (one square mile) lines in the Township System. From the east, the Rio Rillito joins the Santa Cruz river (north-south) just north of the city's center. Tucson's International Airport sits to the south, about five miles from Davis-Monthan Air Force Base to its northeast. North of the city are the Santa Catalina Mountains, rising to 2800 m (9184 ft) at Mt. Lemmon, where citizens of this desert town in a basin around 825 m (2706 ft) can ski in the winter.

The writer (NMS) spent nearly all of his army career (1946-47) in El Paso, at Fort Bliss. In the next image, El Paso is the dark bluish-gray area near the bottom left in the upper right quadrant. Across from it is the Mexican town of Juarez (very popular nightlife for the soldiers). The Rio Grande, dividing New Mexico-Texas from our southern neighbor, is marked by the patchwork of farms using irrigation from that river. Volcanic flows are evident in the almost uninhabited desert to the west. The blue patches are normally dry (playa) lakes temporarily covered with a few inches of monsoon rain water. El Paso is sited at the bottom end of the Franklin Mountains (right of center); next up to the north is the Organ Mountains (top edge), near which was the firing range for the 90 mm Anti-aircraft guns the writer's unit was responsible for. The area covered by El Paso-Juarez has grown by at least 50% since 1947:

Landsat image containing El Paso, TX, the Rio Grande, and surrounding desert.

This image shows El Paso and Juarez in their regional context. Since the writer' extended stay there, El Paso has spread out in two directions - a broad area west of the Franklin Mountains, and downstream along the Rio Grande River, which appears as a narrow black ribbon in this satellite view.

El Paso and Juarez.

As it so happened, the 100,000th photograph taken by astronauts aboard the International Space Station shows El Paso and Juarez in much more detail. In the scene below, Fort Bliss is outlined by a faint square, somewhat below the lettering for "El Paso". Note that the Rio Grande is lined with dikes on both sides.

Photo from the ISS showing El Paso and Juarez.

Before we leave the western North American continent, lets look a one more example of pronounced growth. Mexico City has been forecast to enlarge to well above 20 million people by the mid-21st century (by then Tokyo will have gone above 30 million as the most populated urban area on Earth). Here is Mexico City in two satellite images, the left (or top) one taken in 1973 (population then at 9 million) and the right in 2000 (14+ million)

Let's divert this urban odyssey to other parts of the United States.

Source: http://rst.gsfc.nasa.gov