Handbook for Predicting Stream Meander Migration and Supporting Software (2004)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

17 4.1 INTRODUCTION Accurate calculation and prediction of bend migration rates from historical sequences of aerial photographs re- quires the application of the principles of photogrammetry. The American Society of Photogrammetry and Remote Sens- ing defines photogrammetry as the art, science, and technol- ogy of obtaining reliable information about physical objects and the environment through processes of recording, mea- suring, and interpreting photographic images and patterns of recorded radiant electromagnetic energy and other phenom- ena (Wolf and Dewitt, 2000). Originally, photogrammetry consisted of analyzing photographs, and aerial photos remain the principal source of information today. However, photo- grammetry now includes the analysis of other records, such as digital imagery. This chapter describes the basic principles of photogram- metry, the application of these principles to meander migration analysis and prediction, the map and photographic require- ments necessary for a detailed study of meander migration, and the major sources of maps and aerial photographs. 4.2 BASIC PRINCIPLES OF PHOTOGRAMMETRY 4.2.1 Types of Photogrammetry There are two fields of photogrammetry: metric and inter- pretive. Metric photogrammetry consists of making various types of precise measurements on photos and other informa- tion sources, with the most common application being the preparation of planimetric and topographic maps and digital orthophotos. Determining the significance of recognized and identified objects through careful analysis is the field of inter- pretive photogrammetry. Both fields are used in the analysis of channel migration. The principal types of records available for photogrammet- ric analysis are photographic and digital. Photographic expo- sures are taken from the ground (terrestrial) or from the air (aerial) with different types of cameras using a wide variety of lenses and filters. Negatives and positive contact prints are then made from the exposed film under darkroom processing procedures. A digital image is a computer-compatible pictorial rendi- tion in which the image is composed of a fine grid of picture elements (pixels). Digital images are produced through the process of discreet sampling, whereby a small image area (pixel) is “sensed” to determine the amount of electromag- netic energy emitted by the corresponding patch of surface on the object. Each pixel in a black and white image is defined by an array of integers (digital numbers) that quantifies its gray level, or degree of darkness. The output image, which appears as a continuous-tone picture, consists of many thou- sands to millions of these pixels. A digital color image is also composed of pixels; the color of the pixels is represented by an ordered triplet consisting of a blue, green, and red value. The shades of color in digital color images are represented by varying the levels of brightness of the three primary colors or channels individually within each pixel, similar to the princi- ples used in a color television screen. 4.2.2 Aerial Photography Aerial photography, which is the principal type of photo- graphy used in meander migration analysis, is obtained via an airborne platform, such as an airplane or satellite. Aerial pho- tos are classified as either vertical or oblique. Oblique photos are taken with the camera axis intentionally tilted away from vertical and produce a panoramic view. Vertical aerial photos are taken with the camera axis as nearly vertical as possible. Unavoidable aircraft tilt creates some minor photographic tilt; this is accounted for in metric photogrammetry through the use of precise photogrammetric instruments and procedures. The most common and versatile type of vertical aerial photo is the standard 9 in. × 9 in. (23 cm × 23 cm) black and white panchromatic print using a metric format camera. Other types of aerial photos, including color and color infrared (CIR or false color), are available and are used in meander migration analysis as well. Metric format means that the camera and lens have been calibrated in a laboratory. The calibration includes determining lens distortions, lens focal length, and film flat- ness and making precise measurements of the camera fidu- cials. Fiducials are the crosshair marks one sees in the corners and sides of an aerial photograph. They are used to establish a photo coordinate system when highly accurate measurements from the photograph are desired. Although they vary, the most common flight altitudes from which aerial photos are taken range from 5,000 to 30,000 ft (1,524 to 9,144 m). The aircraft’s height above the ground and the camera’s focal length determine the scale of the aerial photo. The most common aerial photo obtained today is taken with a 6-in. (152-mm) focal length camera at an altitude of about 20,000 ft (6,096 m). This would produce CHAPTER 4 PHOTOGRAMMETRY

a photo that covers an area of about 33 mi2 (85 km2) at a scale of 1:40,000. The scale 1:40,000 means 1 in. on the photo is equal to 40,000 in. or about 3,333 ft (1 cm = 400 m) on the ground. Vertical aerial photography is usually flown along a series of parallel passes, called flight strips. The photos on a given strip are exposed such that there is overlap of the coverage of the previous photo (known as end lap), and parallel flight strips are flown such that there is lateral overlap (side lap) between adjacent flight strips. Aerial photography for the purposes of mapping is usually acquired with a 60 percent overlap to allow photogrammetrists to view the mapping area in three dimensions for collection of elevation data. The common area of coverage in adjacent pairs of photos that are end lapped produces stereoscopic coverage, and the pair of overlapping photos is called a stereopair. Although aerial photos are flat surfaces, they record pro- nounced relief that can be seen in three dimensions, or “stereo,” using a stereoscope. A simple pocket stereoscope, one of several types of stereoscopes, can be used to view the relief recorded by stereopairs. The use of stereopairs allows the user to identify features, such as an obscured stream bank, that may not be readily visible in an individual photograph. In addition, stereoscopic views can be used to estimate relative heights of objects, especially if an object of unknown height 18 is compared with one of known height. For example, a stream bank height may be estimated by comparing its height with that of a bridge or structure of known height nearby. Because an aerial photograph is a perspective view, rather than an orthogonal view as provided by most maps, an effect called relief displacement must be understood before mea- surements are attempted. Relief displacement is the shift, or displacement, in the photographic position of an object caused by the height of an object. Objects above the average terrain elevation in a photo appear to displace outward from the cen- ter of the photo. Objects below the average terrain elevation of a photo appear to displace inward toward the center of the photo. The effect is most clearly seen in urban areas with tall structures, where buildings and smokestacks tilt away from the center of the photo (see Figure 4.1). Relief displacement can affect the accuracy of measurements made from an aerial photo, and its effect can be eliminated. How it is eliminated depends on what technique is used to make measurements. This effect is not significant with regard to low relief features such as rivers, but it can be an important factor when using tall or high elevation features to make distance measurements or when registering photos to each other. Users need to know the average scale of an aerial photo before accurate measurements can be made. The scale of an aerial photograph is the relationship between a measurement Figure 4.1. Vertical aerial photograph illustrating relief displacements.

19 on the photo and a measurement on the ground. Scale can be expressed as a representative fraction or a unit equivalent. For example, the representative fraction of 1:5,000 means 1 unit on the photo is equal to 5,000 units on the ground, whereas the unit equivalent of 1 in. = 2,000 ft means 1 in. on the photo is equal to 2,000 ft on the ground. The scale of the photo is determined by the elevation of the airplane and the focal length of the camera lens at the time the photo is taken (i.e., scale is the ratio of focal length to flying height). The scale determines the accuracy of measurements taken from an aerial photo—the greater the flight altitude, the lower the accuracy of measure- ments taken from the photo. Scale also varies across an aerial photo because of terrain height differences and tilt in the camera when the photo was taken. If the camera is not pointing straight down at the time of exposure, the scale of the photo will vary, depending on location. Also, as the terrain changes in elevation, so does the scale; a mountaintop in a photo has different scale than a val- ley bottom on the same photo. The photogrammetrist can account for both camera tilt and terrain height differences before measurements are taken through processes that are described below. Most aerial photography has a scale printed on it. How- ever, if the scale is unknown, one method of scale determi- nation is to measure between two points on the photo that have a known distance on the ground. For example, if there is a football field in the photo, measure the distance between the goal lines in the photo. If the distance is 1 in. in length, then the scale of the photo is expressed as 1 in. = 300 ft or 1:3,600 (1 in. = 3,600 in.). Another simple method of determining scale is to identify two points on the photo with the same two points on a map having a published scale and measuring the distances between them. For example, measure the distance between two road intersections in the photo and the same two road intersections on a United States Geological Survey (USGS) 7.5-minute topographic quadrangle map (scale: 1 in. = 2,000 ft). If the distance on the photo is 3 in. and the distance on the map is 6 in., then the photo scale can be calculated by using the map distance to determine the ground distance. In this case, 6 in. on the map is equal to 12,000 ft on the ground. That same 12,000-ft distance is represented by 3 in. in the photo, so the scale of the photo is 3 in. = 12,000 ft, or 1 in. = 4,000 ft. There are three methods of making horizontal position measurements off a single aerial photo. The least accurate method of making a horizontal measurement involves using a scale or ruler. Once the scale of the photo is known, one sim- ply measures between objects and converts the measured photo distance to a distance on the ground by using the photo scale. This only yields a distance, however, and most mea- surements need to be related to a known ground coordinate system to be useful. The photo can be referenced to a ground coordinate system using a scale and a map, but the method is very cumbersome and rarely done. This method assumes the scale across the photo is constant, which is never the case, and it does nothing to mitigate the effects of relief displacement. The second method involves a simple rectification of the photo. The rectification process references the photo to a ground coordinate system first so that all subsequent mea- surements are made in a ground coordinate system. This is accomplished using a scanned digital image of the photo, preferably the negative; a map of the same area, usually a USGS 7.5-minute topographic quadrangle map; and a com- puter program for rectification. After picking at least six points on the photo that can be clearly seen on the map, the ground coordinates of these points are established by reading their coordinates off the map. The ground coordinates of the points are used by the computer program along with the point locations on the digital image, which are measured in an estab- lished photo coordinate system (often the Universal Trans- verse Mercator [UTM] coordinate system). With this infor- mation, the computer program performs a simple rectification and georeferences the photo to the ground coordinate system. The rectification process eliminates the effect of scale differ- ences across the photo, but it does not eliminate the effect of relief displacement. (Relief displacement is the inward or out- ward shift in the photographic position of an object because its elevation is above or below a selected datum.) Measurements can then be made off the digital image using a computer and a mouse or other pointing device. The accuracy of this method depends on the number and accuracy of the points used to georeference the photo. For example, the accuracy of a point interpolated from a 7.5-minute topographic map is ±40 ft, whereas the accuracy that can be attained using modern sur- vey methods is easily within inches. The last and most accurate method is to make horizontal measurements from a digital orthophoto. The creation of a digital orthophoto requires a scanned digital image of the photo, a way of georeferencing the image, a digital terrain model (DTM) or digital elevation model (DEM), and associ- ated computer software. Users can georeference the image for an orthophoto using the simple rectification process, or they can use a more accurate technique called aerotriangula- tion. Aerotriangulation requires stereo photo coverage of the area and usually involves surveying high-accuracy ground points. Regardless of the method used for georeferencing, the image is laid over a DTM or DEM surface. The DTM or DEM provides the necessary elevation information to elimi- nate the effect of relief displacement. This, along with the elimination of scale differences through rectification, pro- duces a perspective-free or orthogonal digital image called a digital orthophoto. Measurements can now be made quickly using a mouse and computer software. All other input data being equal, the digital orthophoto is inherently more accu- rate than a simple rectified image because the effect of relief displacement has been eliminated as a source of measure- ment error. 4.2.3 Photogrammetric Products Photogrammetry has been used to produce topographic maps for many years and is still the principal means of

producing maps. The U.S. Geological Survey (the agency tasked with mapping the United States) and most state depart- ments of transportation compile nearly 100 percent of their maps photogrammetrically. Many of the maps and aerial photos that are presently available can be obtained as digital image files. Digital image files are most commonly available in TIFF (Tagged Image File format) and GIF (Graphics Interchange Format), as well as in JPEG (Joint Photographic Experts Group) file format. TIFF, which is recommended, contains the maximum number of colors possible and performs very limited compression in a manner that does not affect image quality (lossless). TIFF is a tag-based format compatible with a wide range of software applications and can be used across platforms such as Macintosh, Windows, and UNIX. TIFF is complex, so TIFF files are generally larger than GIF or JPEG files. Although TIFF supports lossless compression, com- pressed TIFFs take longer to open. A JPEG image also contains the maximum amount of color information possible, but it has much stronger compression (the image can be up to 30 times smaller than a TIFF). A JPEG file supports 24 bits of color information and is most com- monly used for photographs and similar continuous-tone bitmap images. The JPEG file format stores all of the color information in a red, green, and blue (RGB) image and then reduces the file size by compressing it, or saving only the color information that is essential to the image. However, depend- ing on how much an image is compressed, JPEG compression can compromise image quality. GIF saves space by maintaining the sharpness of an image but drastically reducing the amount of color information con- tained. This makes GIF a poor choice for most photos. It is most commonly used for bitmap images composed of line drawings or blocks of a few distinct colors. GIF supports only eight bits of color information or less. Orthophotos and DEMs are two relatively new photogram- metric products that are now often used together to replace tra- ditional topographic maps. Orthophotos are produced from aerial photos that have been modified so that their scale is uni- form throughout. This uniformity is accomplished through the process of differential rectification, which eliminates image displacements caused by photographic tilt, terrain relief, and photograph distortion. Orthophotos are equivalent to plani- metric maps; but, unlike planimetric maps, which show fea- tures by means of lines and symbols, orthophotos show the actual images of features. Orthophotomaps offer significant advantages over aerial photos and line and symbol maps because they combine the features of both. As described earlier, a DEM is a discrete representation of a topographic surface. The DEM, which is an area composed of an array of points that have had their X, Y, and Z coordi- nates determined, provides a numerical representation of the topography of the area such that contours, cross sections, and profiles can be computed from it. In some cases, a DEM is also known as a DTM or a digital terrain elevation model (DTEM). Unlike DEMs, digital line graphs (DLGs) are dig- ital planimentric representations of the points, lines, and areas that define natural and cultural features, but DLGs pro- 20 vide no elevation information. A digital raster graphic (DRG) is a digitally scanned image of a topographic map. Because orthophotos, DRGs, DLGs, and DEMs are in dig- ital form, they are commonly applied in connection with GISs. These photogrammetric products provide a major con- tribution to GISs through their use in generating spatially accurate layers of information for GIS databases, which can ultimately be used in problem solving. 4.3 APPLICATION OF PHOTOGRAMMETRY TO MEANDER MIGRATION ANALYSIS The most accurate means of measuring changes in channel geometry and lateral position is through repetitive surveys of channel cross sections referenced to fixed monuments. How- ever, these data are rarely available. A relatively simple and accurate method of determining migration rate and direction is through the comparison of sequential historical aerial pho- tography (photos), maps, and topographic surveys. The first major use of photogrammetry in the evaluation of fluvial systems was conducted on the Mississippi River Val- ley. Fisk (1944) used maps, aerial photographs, and ground investigations to document historic and prehistoric Mississippi River channel patterns and valley train positions in the lower Mississippi River Valley. He also examined the effects of fine- grained alluvial deposits on the activity of the contemporary river (Fisk, 1947). Brice (1975) developed a classification system of alluvial rivers by analyzing the planform properties of 200 river reaches from topographic maps and aerial photos in order to correlate aspects of river behavior, such as rate of lateral ero- sion and depth of scour, with river type. From this, he devel- oped a comprehensive methodology for conducting a stream stability and meander migration assessment using a compar- ative analysis of aerial photos, maps, and channel surveys (Brice, 1982). Since Fisk’s work, numerous researchers have used pho- togrammetry to document channel planform changes, erosion and sedimentation patterns, and meander migration rates on a wide variety of streams in geographically diverse regions. For example, Hooke (1977, 1984) used historic aerial photos and maps to document the lateral mobility of rivers in Devon, England, over a 50-year period. Williams (1978) used photos taken of the Platte River in Nebraska to document the spec- tacular reductions in channel width that have resulted from river regulation since 1900. Burkham (1972) used surveys, maps, and photographs to document channel changes on the Gila River in Arizona, and Ruhe (1975) used maps covering the period from 1852 to 1970 and aerial photos from 1925 to 1966 to document changes of the Otoe bend on the Missouri River. Using historic maps and aerial photographs, WET (1988, 1990) conducted a geomorphic analysis of more than 100 miles of the Sacramento River in California. Migration rates for specific bends, a bend evolution model, and a bend cutoff index were developed to identify critical sites requiring bank protection and sites where the potential for cutoffs was high (see Section 2.5).

21 4.4 MAP AND AERIAL PHOTOGRAPHY REQUIREMENTS AND SOURCES Historical and contemporary aerial photos and maps can be obtained inexpensively from a number of federal, state, and local agencies. Table 4.1 lists some of the main sources. The Internet provides a wealth of sites with links to data resources and sites having searchable databases pertaining to maps and aerial photography. Often, just typing a few key- words relative to aerial photos or maps for a particular site into a search engine will generate a large number of links to related Web sites, which can then be evaluated by the user. Extensive topographic map coverage of the United States at a variety of scales can be obtained from the local or regional offices of the U.S. Geological Survey (USGS). In general, both aerial photos and maps are required to perform a com- prehensive and relatively accurate meander migration assess- ment. Because the scale of aerial photography is often approx- imate, contemporary maps are usually needed to accurately determine the true scale of unrectified aerial photos. Georeferenced and rectified maps and aerial photos are the most desirable for analysis of meander migration, but they can often be expensive to obtain. Presently, aerial photos for the 1990s for most areas of the United States can be obtained from three major sources, the Microsoft TerraServer Web site, the United States Department of Agriculture (USDA) Farm Ser- vice Agency, and the USGS (see Table 4.1). A major public source of aerial photos from the 1990s is the TerraServer Web site operated by Microsoft Corporation. TerraServer, in partnership with the USGS and Compaq, pro- vides free public access to a vast data store of maps and aerial photographs of the United States. Aerial photos and topo- graphic maps at a wide variety of resolutions can be down- loaded free of charge from the TerraServer Web site. The advantages of the TerraServer images are that they are rectified, georeferenced, and in digital format so that they are easily TABLE 4.1 Sources of contemporary and historical aerial photographs and maps Source Internet Address Comments Microsoft TerraServer—USA www.terraserver.microsoft.com Free downloads of contemporary digital topographic and aerial photo files. Operated by Microsoft in conjunction with Compaq and USGS. USGS EROS Data Center Sioux Falls, South Dakota For photos go to Photo Finder at edcwww.cr.usgs.gov/Webglis/ glisbin/finder_main.pl? dataset_name=NAPP For maps, go to the USGS store at http://store.usgs.gov/ Operated by the USGS. Interactive database search for historic and contemporary topographic maps and aerial photos. USDA Farm Service Agency Aerial Photography Field Office (FSA—APFO) Salt Lake City, Utah www.apfo.usda.gov/filmcatalog.html Operated by the Farm Service Agency. Catalog of historic and contemporary aerial photos for much of the United States. Sources include Soil Conservation Service (SCS) (now called the Natural Resources Conservation Service [NRCS]), the Forest Service, the Bureau of Land Management (BLM), the National Park Service, and other government agencies. USGS Special Collections Library Denver, Colorado Reston, Virginia library.usgs.gov/specoll.html The Field Records Collection is an archive of historic records including maps and aerial photography collected by USGS scientists dating back to 1879. The map collection includes topographic maps for all states dating back to early 1880s. National Archives and Records Administration (NARA)—Cartographic and Architectural Records Washington D.C. www.archives.gov Archive of historic maps and pre- 1941 aerial photos. Western Association of Map Libraries (WAML) San Diego, California www.waml.org/wmlpubs.html References to information on obscure historic maps and where they can be found for reproduction. U.S. Army Corps of Engineers www.usace.army.mil Corps districts often have a wealth of historic photos, maps, and survey data.

manipulated by a wide variety of software and can be used in GIS applications. General instructions on downloading TerraServer images can be found in Appendix A. It may be necessary to download a number of higher-resolution map or photo images and splice them together to obtain the amount of detail desired. For sites where TerraServer photographic coverage from the 1990s is unavailable, aerial photos can be ordered from the USGS Earth Resources Observation Systems (EROS) Data Center in Sioux Falls, South Dakota, or from the USDA Farm Service Agency Aerial Photography Field Office (APFO) in Salt Lake City, Utah. Both agencies have World Wide Web sites (see Table 4.1) with searchable catalogs of available con- temporary and historic aerial photography. Aerial photographs from the EROS Data Center that were flown in the 1980s and 1990s are usually part of the National Aerial Photography Program (NAPP) or the National High Altitude Photography Program (NHAP) and are at scales of 1:40,000 (1 in. = 3,333 ft) or 1:60,000 (1 in. = 5,000 ft). Because of the scale of these photos, small objects may be dif- ficult to see, the resolution of enlarged portions may be poor, 22 and measurements made from the photos may be inaccurate. Historic aerial photos ordered from EROS or APFO range in scale from 1:5,000 to 1:40,000, with most flights having opti- mal scales of 1:20,000 or 1:24,000. Although both agencies have the ability to enlarge any photo to specification, some resolution is lost with increasing enlargement. Topographic maps, in paper or electronic format, can be obtained from a variety of sources. Paper copies of topo- graphic maps can be obtained from the USGS or any com- mercial map supplier. Digital maps (DRGs, DEMs) can be downloaded free from the EROS Web site or purchased from commercial suppliers. Most digital maps are georefer- enced and can be loaded directly into GIS-based applica- tions. Portions of georeferenced topographic maps can be downloaded free from the TerraServer Web site and pieced together to form a complete map of a given area or used to fill in gaps. Care should be taken when using digital maps and photos because the georeferenced coordinates and dimen- sions are usually in metric (SI) units, whereas the contours and spot elevations shown on the maps may be in U.S. cus- tomary units.