Geog 380 Course Supplement, Continued. Back to last section.
A. Projections
1. BackgroundYou can't squash a grapefruit peel flat without breaking it into many pieces (try it sometime). In the same way, we cannot transfer the spherical surface of Earth to a flat surface without distortion. We can create a logical way to transfer coordinates from the sphere onto a flat map. Such ways of transferring coordinates are known as projections, after the original method of transferring by literally projecting a light through a globe onto a surface. But no projection can accomplish its task without some distortion. Fortunately, we can choose to preserve certain qualities that a globe possesses. But in the process we sacrifice other qualities.
Before we project the actual Earth onto a surface, we usually simplify it. Earth isn't a perfect sphere. It's somewhat flattened at the Poles, so an ellipsoid represents Earth better (an ellipsoid is formed by rotating an ellipse around one of its axes). Actually, the diameter at the equator is only about 42.8 km more than the polar axis. But it is enough to throw off exacting measurements, like property lines.
A particular ellipsoid, with particular values for equatorial and polar diameters, is often used in projecting and measuring on Earth, particularly for highly accurate measuring such as in surveying. A dozen ellipsoids are in common use around the world, including:
An ellipsoid is still not perfect; the geoid is an irregular, but even closer, representation of Earth. It's the equivalent of mean sea level all over the globe. Since gravity and Earth's surface are irregular, the geoid is not a smooth surface, and can't be represented with equations easily, so it's rarely used for mapping.
2. Methods of Projection
Many projections can be visualized as literally projecting a light source through a transparent globe onto a surface. The light source can be any number of places -- at the center of the globe, at the opposite side of Earth, or out in space, for instance (see figure below for examples). The map surface onto which the projection is made can be various shapes, and can also be at various places. In all projections, the map surface touches the globe at at least one point. This is because any map is most accurate where it touches the globe; there is no distortion here. The contact point between globe and map is called the point of tangency; if it is a line, it's a line of tangency, standard line, or (if it is a parallel) standard parallel. Away from the tangent locations, the map surface gets further from the globe, and hence more distorted. Most all projections nowadays are done by computer using equations that relate lat/long to x/y coordinates on the map.
Projections may be made onto three basic shapes, with three types of projections resulting:
a. Planar |
|
b. Cylindrical
Here a cylinder is wrapped around the globe, usually with the map surface touching the globe at a circle (a great circle, to be exact--a circle whose center coincides with the center of Earth). Cylindrical projections are the only one of the three main types that can show the entire globe, and so most world maps are cylindrical.
The most famous cylindrical projection is the one named for Gerhardus Mercator, who developed it in 1569. It was valuable for early navigators, since straight lines on a Mercator map are also compass headings. Unfortunately, it greatly distorts the sizes of areas near the Poles (see section 3 below), so it should not be used as a general-purpose world map!
c. Conical
The third type of projection is made onto a cone. Usually this means contacting the globe along one of the parallels (lines of latitude), i.e., a small circle. Although we cannot use conic projections for a world map, they are excellent for continent-sized areas in the mid-latitudes. Most maps of the United States are made with conic projections.

Conic projections are usually made more accurate by "sinking" the cone part way into the globe (remember, this is all done with computers, not literally!). Then we have two lines of tangency, or two standard parallels, along which the map is extremely accurate. This two-line approach is also called the secant case, as opposed to the simple tangent case. You may see maps of the US with a statement on the bottom like: "Lambert Conformal Conic Projection, 48° and 33° Standard Parallels." The projection was developed by J.H. Lambert (1728-1777), an important figure in cartography.
d. Other
Some projections are not based on any of the above three shapes, and cannot be visualized as literally being projected. Instead, they simply have equations that tell where to plot each latitude/longitude coordinate from the globe. Some examples are the Sinusoidal and van der Grinten projections. World maps are often made with these projections, since they may have less distortion than cylindrical projections.
3. Qualities of Projections
The other major factor you need to know about a projection is the qualities about the globe that the projection either preserves or distorts. Most projections can preserve one or more of the following qualities, but none can retain all of them. Note that the projection method (planar, cylindrical, or conical) does not necessarily mean any of these qualities below are preserved or distorted. It all depends on how the projection is done.
a. Equal-Area. Some projections show all areas in true proportion to their real areas on the globe. For example, a dime placed on the map would cover the same area regardless of where placed. To show areas truly, a map must distort most of the other qualities below, at least subtlely. But if you need a general-purpose map of the world or continental area, an equal-area map, or a map that is very close to equal-area, is your best bet. Some examples: Sinusoidal, Albers Conic Equal-Area, and Lambert Azimuthal Equal-Area.
b. True Shape, or Conformality. Another important characteristic of the globe that can be distorted on a map is the shape of areas. This distortion problem is obvious on a cylindrical map that is equal-area, because the higher latitudes near the Poles have to be distorted to preserve areas. The Mercator projection is conformal, but at the expense of area. The Mercator shows Greenland almost as large as South America, when in reality it is about 1/8 the continent's size. Some people have accused developed nations (which are mostly in the higher northern latitudes) of intentionally portraying their lands as larger than developing countries (which are mostly in lower, tropical latitudes). One projection, known as the Peters projection, has been promoted as the "true" world map, since it shows countries with true areas. Peters is indeed equal-area, but does a number on shape--as one person put it, it makes the world look like it was hung on a laundry line. Many other equal-area projections are available that do a better job with shape.
c. True Scale. In no map can you use one scale accurately for the whole map. Some distortion occurs, although it is slight in many maps. Some projections can preserve true scale and distance along one or more lines. These are may be called equidistant projections. A popular planar projection for polar areas is known as the Azimuthal Equidistant, which has true scale from the central tangent point--the Pole--to any other point on the map. You could also use an Azimuthal Equidistant map centered on your location to measure distances accurately to any other place on the globe. Some map software can draw such a map for you.
d. True Direction. The last major quality of maps is direction. Maps that preserve it are called azimuthal. Most planar projections preserve true direction away from the center of the map (usually the Pole) and so azimuthal is nearly synonymous with planar projection.
e. Other Qualities. Some projections are designed to have specialized qualities. The Mercator projection is one: all constant compass headings (rhumb lines, or loxodromes) are straight lines. The Gnomonic projection is another: all great circle routes are straight lines. As you may know, great circle routes are the shortest distances between points on the globe. For example, when you fly from San Francisco to London, you don't fly along a parallel of latitude, but over the polar route; this is along a great circle. If you're flying or sailing, then, you can combine the gnomonic and Mercator maps for navigating. First you draw your route on the gnomonic map (a straight line connecting the two places), then transfer the route to the Mercator map as a series of straight segments that approximate the gnomonic line. This way, you can follow the straight segments on the Mercator map with a compass, and turn only when you need to follow the next segment. You may notice this when you're flying and the pilot periodically turns to follow these segments.
B. Coordinate Systems
How can we describe locations on Earth? If someone asks you, "where is Hawaii?", what do you tell them? You can give them directions relative to your position ("swim 2000 miles south-southwest"). Other ways are also possible, but what if you needed to pinpoint a location for people coming from many directions? Or if you wanted to record a location for later reference? Or if you had no landmarks to guide you? This is the purpose of coordinate systems. They are ways of describing locations on Earth in reference to an established grid. You have probably been exposed to the most common method, latitude/longitude, but there are many other methods in use.
1. Latitude & Longitude
The latitude and longitude system is also called the geographical grid. This grid exploits the fact that Earth is nearly a sphere, and that it spins on an axis. Looking down on the globe from above the North Pole, we can fit a circle to the rotating Earth. We could assign each location along any circle that surrounds the Pole a measurement in degrees. A circle has 360 degrees. We could use this range of numbers, going from 0° to 360°. Alas, early map-makers didn't do this, exactly. They wanted low numbers on both sides of the Prime Meridian (the 0 line). As a result, the globe is divided into hemispheres, each assigned longitude between 0° and 180°, with the addition of East or West to differentiate the halves. The lines of longitude are meridians.
|
To complement the east-west measurement, a north-south measurement is necessary, so that we may pinpoint locations. Since we only need to measure along one meridian, we only need to assign measures to a half-circle, or 180 degrees. Once again, it's more complicated than necessary. Rather than go from 0° at the North Pole to 180° at the South Pole (or vice-versa), the system starts with 0° at the halfway point (the Equator), and measures north and south to 90° at the Poles. Each line of latitude is a circle; these lines are called parallels (sensible, since they are parallel to one another). |
|
Lat/long is cumbersome to use for at least two reasons. First, notice that meridians converge at the Poles. A degree of longitude decreases from about 111 km at the Equator to 0 at the Poles; 1° is about 88 km in Sonoma County. Convergence makes lat/long poor for use as a rectangular grid, where we want simple x,y coordinates for locations. Second, lat/long is not a decimal system. How far is it from 114° 34' 54" to 116° 14' 33"? Not very far, but you'd have trouble giving me the distance even in terms of degrees/minutes/seconds. For these reasons, lat/long is usually replaced by other coordinate systems, especially at the local level, for most descriptions of location. Most of these systems, including those below, use a projection of the globe onto a flat surface, onto which we can then draw an x/y grid.
2. State Plane Coordinates (SPC)
The National Geodetic Survey developed the SPC system beginning in 1933. Eventually every state was covered, with coordinates identified both on maps and on the ground, so that surveyors and cartographers could accurately identify and measure locations. The key to this system is that rather than having one coordinate system for the entire US, separate systems were assigned to smaller zones. Each zone used its very own projection and coordinate center and system. 120 zones cover the US. Within each zone, you are never far from the standard line. This way, the coordinates would be extremely accurate within each zone (less than 1 foot per 10,000 feet of measurement, in fact). The problem, of course, is that coordinates between zones don't match up, so the SPC system is not useful for small-scale (large-area) maps that include more than one zone.
Nearly all states have multiple zones, but zones never cross county lines. California has 7 zones, most extending as east-west bands; Sonoma County's zone extends to Lake Tahoe. Los Angeles County has its own zone (naturally). Each state uses either the Lambert Conformal Conic or the Transverse Mercator projection (California uses the first).
Within each zone, locations are identified by x,y coordinates in feet. Any x,y coordinate system needs an origin, that is, where the coordinates are (0,0). In order to keep all SPC numbers positive, the origin for each zone is placed off to the southwest of the actual zone covered. This origin is not the actual center of the projection (that is, where the globe "touches" the sheet projected onto). That actual center is in the middle of each SPC zone, so that coordinates are most accurate there. In short, the actual center is assigned an arbitrarily large coordinate (such as 2,000,000 feet East, 400,000 feet North), and all other coordinates are measured from there. This puts the "false origin" off to the southwest.
SPC coordinates are shown on all USGS topographic maps. Usually tick marks on the margins of the map show regular spacing of the grid, and selected marks have the actual coordinates in feet. By examining the topographic map for Cotati, we can find that the SPCs for SSU are 1,806,500' E, 246,200' N. As mentioned above, the SPC system is used widely in conducting local land surveying and public works. It can be used by the cartographer and geographer not only to identify coordinates of places, but to calculate distances between locations by use of the Pythagorean Theorem, as described in the next section.
3. Universal Transverse Mercator (UTM)
The UTM grid is similar to the SPC system, at least regarding how you use it at the local level and in being marked on all USGS topographic maps. The principal differences are that the coordinates are given in meters, not feet, and that the zones are much larger. UTM zones extend north-south, practically from Pole to Pole. The UTM grid system covers the entire globe (well, almost -- except for very near the Poles).
You encountered the Mercator projection before. In the standard Mercator, the cylinder is "wrapped" around the Equator, and areas become very distorted toward the Poles. A transverse Mercator projection turns the cylinder, so that the circle of contact with the globe is around a pair of meridians. This way, the projection is very accurate on a north-south zone near the standard line. Of course, once again it distorts severely at large distances away from the meridian.
The Universal Transverse Mercator grid gets around the distortion problem by the same method as the SPC system. The UTM has many zones, each with its own projection centered on a meridian. There are 60 zones to be exact, each 6° wide (which covers Earth, 60 x 6° = 360° around). Within each zone, then, the grid is very accurate in matching true Earth distance and direction. As with the SPC system, going across zones is difficult, so the UTM is meant primarily for local and regional measurement.
The UTM was adopted and thus popularized by the Army in 1947. The Army included the UTM grid on its topographic maps; later the USGS added UTM coordinates to most of its maps and photoquads. The Army numbered each 6°-wide zone around the globe from 1 to 60, starting at 180° W and going east; northern California is in zone 10. They also lettered north-south segments of each zone from A (south) to Z (north). The north-south segments aren't necessary, and so are rarely used outside the military.

Within each UTM zone, x,y coordinates can be given in meters. Like SPCs, an origin is needed, and is placed outside the zone off the southwest corner. The north-south center of the zone is arbitrarily designated as 500,000 meters east (that is, east of a false origin off to the west). "Eastings" (x-coordinates) for locations east of the center are higher than this, up to about 850,000 m E; westward the coordinates decrease, down to about 150,000 m E; the zone doesn't extend all the way to the false origin. The "northing," or north-south (y) coordinate, depends on which hemisphere you're in. For the Northern Hemisphere part of each zone, the measurement starts at the Equator with 0 and measures the number of meters north (up to about 8,800,000 m N at 80° N). In the Southern Hemisphere, the Equator is designated arbitrarily as 10,000,000 m N, and coordinates decrease as you go south toward the South Pole.

Examples: A location with coordinates 334,400 m E, 4,203,600 m N would be 334,400 meters east of the false origin, or (500,000 - 334,400 =) 165,600 meters west of the central line. It would be 4,203,600 meters, or 4,203.6 km, north of the Equator. The UTM coordinates for SSU are: 4,243,540 m E, 528,390 m N (these are actually close to the coordinates for Stevenson 3065).
The UTM grid is shown on all recent USGS topographic maps. The latest topographic maps draw in the grid as thin black lines. All topos with the UTM have tick marks along the margin, along with values for eastings or northings next to most ticks. Except for a few values near the corners, the easting or northing value is abbreviated . For example, instead of printing "3,445,000 m N", the tick would be labeled 3445, with the thousands and meters-north assumed from the context.
The UTM grid, even if drawn in on the map, may not give us the exact coordinates for a given location. Even on 7 1/2-minute quads, the grid is only every 1,000 meters (1 km). How can you determine coordinates more precisely? The answer is called a roamer. This is simply a sheet of paper, plastic or other material that has finer distance intervals marked off that match the scale of the map. Starting from the nearest grid lines, you can measure over to the location and come close (at least within 100 m) to the actual easting and northing coordinates.
Another big advantage of the UTM (or SPC) grid is that once you have coordinates for two locations within the same zone, calculating the distance between them is simple. Just apply the Pythagorean Theorem. If the two locations are at the coordinates (x1, y1) and (x2, y2), then the distance (D) between them is:
For example, say you find coordinates for two cities: Springfield at 294,100 m E, 3,428,900 m N, and Garden City at 292,400 m E, 3,428,100 m N. The distance between them is then:
that is, the distance is 1880 meters, or 1.88 kilometers.
As you can see, the UTM grid is a very useful system for tracking Earth locations. It is used extensively in remote sensing, computerized mapping, and geographic information systems. It is worth your while to familiarize yourself with it.