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.

47 8.1 INTRODUCTION This chapter provides the user with illustrated examples of the application of the comparison techniques described in Chapters 5 and 7. This chapter begins by allowing the user to conduct an initial screening and classification of several streams. The chapter then provides the user with step-by-step instructions on conducting the manual comparison techniques using Mylar or acetate overlays. 8.2 SCREENING TASK The first step in determining the potential for channel migra- tion and the migration history of a channel using the method- ology set forth in the Handbook is determining whether the channel is a meandering stream. An initial screening should be conducted based on the description of straight, braided, and anabranched streams in Section 2.2. Once a stream has been determined to be a single-thread meandering channel, the user proceeds to classify the stream based on the classification scheme shown in Figure 3.2. 8.3 CLASSIFICATION TASK Eleven channels are provided in Figure 8.1 to allow the user to conduct an initial screening and classification based on the descriptions of non-meandering stream planform in Chapter 2 and the modified stream classification scheme described in Chapter 3 and shown in Figure 3.2. The classification of each stream shown in Figure 8.1 is provided in Table 8.1. Stream Reaches 2, 6, and 11 in Figure 8.1 would be screened out from further analysis based primarily on plan- form characteristics. Stream Reach 2 (Cimarron River) would be screened out because it is a highly unstable, two-phase stream with a low-flow wandering channel. The Henrys Fork River, shown as Stream Reach 6, is classified as a braided river and would be removed from further analysis. Stream Reach 11 (Altamaha River) would be removed because it is an anabranched river; it is multichanneled, stable, and located in a low-energy environment. The remaining streams would be classified based on the stream classification scheme shown in Figure 3.2. Stream Reach 1 (Smokey Hill River) is classified as Class G2; it is a two-phase, bimodal sinuosity channel that is wider at the bends because of the presence of point bars. Stream Reach 3 (Pee Dee River) is classified as Class B1 because it is an equal-width, single-phase channel. Stream Reach 4 (Licking River) is classified as Class A because it is a deeply incised, single-phase, equal-width chan- nel. The narrow riparian fringe along both sides of the channel and the extensive agricultural fields surrounding the river indi- cate an incised or deep channel with a narrow terrace or berm and a dense growth of vegetation inset between both banks. The reach of the Sacramento River shown as Stream Reach 5 in Figure 8.1 is classified as Class D because it is a single- phase channel that is wider at the bends with extensive point bar development and well-developed chute channels across the point bars of both bends. The reach of the Sacramento River shown as Stream Reach 7 is near Stream Reach 5, but Stream Reach 7 is classified as a Class C because no chute channels are present on the point bars of the reach. The two-phase, equal-width, bimodal sinuosity channel shown as Stream Reach 8 (Saline River) is classified as Class G1. The channel shown as Stream Reach 9 (Little Pee Dee River) is classified as a single-phase, irregular-width Class E channel. Stream Reach 10 (Neches River) is classified as a Class B2 channel, which is a single-phase channel that is wider at the bends but has no apparent point bars. Once a channel reach has been classified, the prediction of the rate and extent of future migration of a given bend or channel reach for B2, C, D, E, or G2 channels can be con- ducted using the aerial photo and map comparison tech- niques described in Chapters 5 and 7 and in the example in the following section. 8.4 EXAMPLES USING AERIAL PHOTO COMPARISON AND PREDICTION TECHNIQUES The following sections provide steps for using aerial photo- graphs and maps to conduct a simple comparison of historic bankline positions and for using the manual overlay tech- niques described in Chapters 5 and 7 to make predictions on potential future migration. The reach of the Sacramento River shown as Stream Reach 7 in Figure 8.1 (a Class C channel) will be used as the example reach in the following discussion. 8.4.1 Manual Bankline Overlay and Circle Template Method The following steps include delineating the historic bank- line positions of the Sacramento River by hand, evaluating bankline shift in the reach using simple overlays, evaluating the historic migration rate and extent using circle templates, CHAPTER 8 ILLUSTRATED EXAMPLES

and predicting the potential position of the channel in the future. Step 1. The first step in evaluating the potential for channel migration is acquiring appropriate historic maps and aerial photography of the reach. These can be acquired from a num- ber of federal, state, and local government agencies and pri- vate vendors. The addresses for some of the more prominent distributors are provided in Table 4.1. Figure 8.2 shows two maps from a 1937 topographic survey of the Sacramento River 48 Valley conducted by the U.S. Army Corps of Engineers and acquired from the Sacramento District. Figure 8.3 shows blue- line sheets of 1972 aerial photographs of the Sacramento River acquired from the California Department of Water Resources. A 1998 NAPP black and white aerial photograph of the Sacra- mento River, shown in Figure 8.4, was acquired from the U.S. Geological Survey EROS Data Center. Step 2. After the appropriate historic maps and aerial photos have been acquired, the next step is to delineate the Figure 8.1. Various streams requiring classification by the user based on the initial screening described in Chapter 2 and the classification scheme shown in Figure 3.2. (continued on next page)

49 TABLE 8.1 Classification of streams shown in Figure 8.1 Stream Number Location Classification 1 Smokey Hill River near Chapman, KS G2 2 Cimarron River near Fairview, OK F 3 Pee Dee River Near Peedee, SC B1 4 Licking River near Romey, KY A 5 Sacramento River near Butte City, CA D 6 Henrys Fork River near Rigby, ID Braided 7 Sacramento River near Butte City, CA C 8 Saline River near Tescott, KS G1 9 Little Pee Dee River near Galivants Ferry, SC E 10 Neches River near Evadale, TX B2 11 Altamaha River near Darien, GA Anabranched Figure 8.1. (continued) Various streams requiring classification by the user based on the initial screening described in Chapter 2 and the classification scheme shown in Figure 3.2.

50 Figure 8.2. Two sheets showing U.S. Army Corps of Engineers topographic survey maps of the Sacramento River near Butte City, California, from 1937. Figure 8.3. Two sheets showing California Department of Water Resources aerial photographs of the Sacramento River near Butte City, California, in 1972.

51 Figure 8.4. Aerial photograph obtained from the USGS of the Sacramento River near Butte City, California, in 1998. banklines and identify the registration points that are common to all the maps and aerial photos. This can be done by hand or electronically. To do this by hand, a piece of Mylar or acetate is laid on top of the map or aerial photo, and the banklines are traced onto it. Registration points that are common to the pho- tos and maps being analyzed are delineated on the tracings as well. A registration point can be any geographic feature (such as a fence, road crossing, or a structure), and the point should bracket the reach at both ends to provide the best fit possible among the historic banklines. Figure 8.5 shows bankline tracings and registration points from the 1972 aerial photos hand drawn on Mylar. A particular site can often extend over two or more aerial photos or maps from each time period. In these cases, over- lapping aerial photographic coverage should be obtained so that the overlapping photos can also be registered together for delineating the channel banklines. When a site falls on two or more maps, the map edges are usually all that needs to be matched together. If adjacent photos and maps do not match, they are probably at a slightly different scale and will need to be enlarged or reduced to match the other photos or maps in the set. In many cases, the different historic aerial photos and maps used in the comparison may not have the same scale, and, therefore, they will need to be enlarged or reduced to a common scale. By enlarging or reducing the bankline tracings on a copier with a zoom feature, the user can closely match each subsequent bankline to the previous bankline with the registration points. The same steps can be accomplished more easily using graphics and photo-editing software such as Microsoft PowerPoint or Jasc Paint Shop Pro as discussed in Sections 5.3 and 7.3. The aerial photos and maps should be obtained in digital format for use with this software. This can be accom- plished by scanning the images on flatbed scanners or by obtaining the images in digital format directly from a vendor. The aerial photos and maps of the reach of the Sacramento River being used in this example were scanned and converted into digital format. The Sacramento River banklines and regis- tration points were delineated on the 1937 map (see Figure 8.6) and on the 1972 aerial photo (see Figure 8.7) using MS PowerPoint. Step 3. Once the banklines and registration points have been delineated on each map and photo, the registration points can be matched using a manual or digital overlay technique so that the banklines can be compiled onto a Mylar sheet (manual technique) or digital image (digital technique). Figure 8.8 shows the overlay of both the 1937 and 1972 banklines on the 1972 image of the Sacramento River. As seen in Figure 8.8, the left bank of the river has migrated laterally and down val- ley as much as 1,370 ft (417 m) in the 35 years between 1937 and 1972. This produces a maximum average migration rate of about 39 ft (12 m) per year. It is also apparent that the mean- der bend has increased in width over time as a result of the growth of the point bar. Step 4. Contemporary migration characteristics of this Sacramento River meander bend can be evaluated further by comparing the 1937 and 1972 banklines with the banklines of the river on the 1998 aerial photograph. Figure 8.9 shows the 1998 aerial photo of the river with the banklines and reg- istration points delineated. The banklines from 1937 and 1972 are then superimposed on the 1998 aerial photo (with

the 1998 banklines and registration points delineated) as shown in Figure 8.10. It is readily apparent that the contin- ued migration of this meander bend will soon threaten the flood control levee located less than 250 ft (76 m) from the left bank of the river. Step 5. The next step in evaluating the migration rate and extent of the bend of the Sacramento River involves delineat- ing the radius of curvature and centroid of a best-fit circle used to define the outer bank for each year in the analysis (see Figure 8.11). The measured radii of curvature of the outer bank of the meander bend in 1937 (RC1), 1972 (RC2), and 1998 (RC3) are 3,650 ft (1,113 m), 2,425 ft (739 m), and 2,650 ft (808 m), respectively. The rate of change in the radius of curvature of the outer bank can then be determined by figuring out how much the radius has contracted or expanded over each period and then dividing the amount of change by the number of years in the period. For the purpose of discussion, the period of time between 1937 and 1972 is defined as Period A, and the period of time between 1972 and 1998 is defined as Period B. For this bend of the Sacramento River, the outer bank radius of curva- ture contracted 1,225 ft (373 m) between 1937 and 1972. The resulting contraction rate (∆RCA), derived using Equation 7.1, is as follows: The radius of the outer bank expanded more than 225 ft (68 m) between 1972 and 1998. The resulting expansion rate (∆RCB), derived using Equation 7.2, is as follows: ∆R R R Y ft ft yr ft yr m yr CA C C A= −( ) = −( ) ( ) = −( ) 2 1 2425 3650 35 35 10 6. . 52 Step 6. After delineating the outer bankline radii of cur- vature with best-fit circles, the circle centroids are delin- eated, and their migration is plotted as shown in Figure 8.12. The centroid migration distance is defined as the length of the line between the centroid position defined by Period A (DA) and the centroid position defined by Period B (DB). The centroid migration distance (DA) from 1937 to 1972 is approximately 2,700 ft (823 m). The centroid migration distance (DB) from 1972 to 1998 is approximately 875 ft (267 m). The migration rate for Period A is 77 ft (23.5 m) per year, and the migration rate for Period B is 33.7 ft (10.3 m) per year. An arbitrary line is drawn vertically from the 1937 bend centroid and is used to define the change in direction of the 1972 and 1998 bend centroids (see Figure 8.12). Line DA defines the direction of migration of the bend during Period A. The angle θA, described by the intersection of line DA with the arbitrary line, represents the angle of migration of the bend centroid during Period A. The angle θB represents the bend centroid migration direction relative to the arbitrary line dur- ing Period B. The migration direction of the bend centroid between 1937 and 1972 as defined by angle θA is 13.5 degrees. The migration direction of the bend centroid between 1972 and 1998 as defined by angle θB is 46 degrees. Step 7. The final step in the analysis is the prediction of future meander migration. Using the rates and direction of ∆R R R Y ft ft yr ft yr m yr CB C C B= −( ) = −( ) ( ) = ( ) 3 2 2650 2425 26 8 7 2 6. . . Figure 8.5. Bankline tracings (black lines) and registration points (circled red crosses) delineated on a Mylar sheet laid on top of the 1972 aerial photograph of the Sacramento River near Butte City, California.

53 Figure 8.6. U.S. Army Corps of Engineers 1937 map of surveyed topography of the Sacramento River Valley near Butte City, California. The banklines of the river are defined by blue lines. Common registration points are defined by the red circled crosses. (RM = river mile.) past migration, a prediction can be made on the position of the outer bankline of a bend in the future. For example, a pre- diction of where the outer bankline will be in the year 2028 for the example reach of the Sacramento River is provided in Figure 8.13. Prediction of the future migration rate of the bend centroid should use the migration rate for Period B because the conditions associated with this rate are likely to be more closely related to those in the future. Thus, the pre- dicted migration distance (DC) during Period C (1998 to 2028), derived using Equation 7.5, is the following: D D Y Y ft yr yr ft mC B B C= ( ) = ( )( ) = ( )875 26 30 1010 308, .

54 Figure 8.7. California Department of Water Resources 1972 aerial photography of the Sacramento River Valley near Butte City, California. The banklines of the river are defined by green lines. Common registration points are defined by the red and blue circled crosses. (RM = river mile.) If only one period is available, then it can be assumed that the migration direction does not change (θC = θB). If two or more periods are available, a decision should be made as to whether the prediction of future meander migration should include change in migration direction, as described by Equation 7.4 in Chapter 7. It is assumed for the purposes of this example that meander migration will follow the same direction as that of the previous period. The incorporation of change in migration direction for this problem can be found in Appendix E. Consideration should be given to applying both methods in order to define an area over which future meander migration might occur.

55 Figure 8.8. Overlay and comparison of the 1937 and 1972 banklines and common registration points on the 1972 aerial photography. (RM = river mile) Finally, the radius of curvature for the outer bank of the meander bend in 2028 (RC4) is predicted based on the assump- tion that the bend will continue to expand (or contract) at the same rate as the previous period (Period B). This is calculated using Equation 7.3, which is the rate of change per year of the radius during Period B (the expansion rate) times 30 years plus the 1998 radius of curvature. Thus, the radius of curvature for the outer bank of the meander bend of the river in the year 2028 (RC4) is predicted to be The outer bankline position for the river bend in the year 2028 is defined by a circle with the predicted radius, RC4. R R R R Y Y ft ft ft yr yr R ft m C C C C B C C 4 3 3 2 4 2 650 2 650 2 425 26 30 2 910 887 = + −( )  ( )   = + −( ) ( )     ( )     = ( ) , , , , .

56 Figure 8.9. Aerial photograph of the Sacramento River near Butte City, California, showing the bankline positions of the river in 1998. (RM = river mile)

57 Figure 8.10. Overlay and comparison of the 1937 (blue), 1972 (green), and 1998 (black) banklines on the 1998 aerial photo of the Sacramento River near Butte City, California. The red arrows show the maximum bankline migration display. (RM = river mile.)

58 Figure 8.11. Circles of known radius of curvature fitted to the 1937, 1972, and 1998 outer banklines (dotted lines) of the Sacramento River. (RM = river mile.)

59 Figure 8.12. Delineation of the bankline radii of curvature (RC1, RC2, and RC3), bend centroid migration distances (DA and DB), and angles (θA and θB) for Period A (1937 to 1972) and Period B (1972 to 1998). (RM = river mile.)

60 Figure 8.13. Prediction of bend migration for the period from 1998 to 2028 and the predicted position of the outer bank of the bend in 2028. (RM = river mile.)

61 Based on past channel configurations, the predicted channel banklines for the year 2028 can be drawn and then superim- posed on the 1998 aerial photo. The predicted channel banklines for the year 2028 can be drawn based on past channel configurations and then superimposed on the 1998 aerial photo (see Figure 8.14) to determine the potential haz- ard to local structures and features posed by continued active migration over the next 30 years. On the basis of the predicted position of the channel in 2028, it appears that a local levee is in the direct path of the migrating bend. 8.4.2 Method Using GIS Applications All of the steps described in the previous paragraphs can be accomplished using CAD and GIS software. For example, two and three historic aerial photos and maps for more than 1,500 bends were used to evaluate meander migration under NCHRP Project 24-16. In order to accomplish this rapidly and accu- rately, Bentley’s MicroStation was used to import and register corrected and georeferenced aerial photos and topographic maps, and Bentley’s Descartes was used to adjust scanned (uncorrected) aerial photos for distortion during the registra- tion process. The banklines for each year were delineated and general measurements were taken using MicroStation. The historic banklines were then exported to files that could be used with the ArcView-based Data Logger and Channel Migration Predictor described in Sections 5.4 and 7.4. 8.5 EXAMPLE USING FREQUENCY ANALYSIS The results of the previous example of the Sacramento River bend near Butte City, California, are shown below compared with a frequency analysis. To use the frequency results, the rates of extension and translation are selected from Table 7.1 or from Figures 7.31 and 7.32 for a specific channel class (in this case a C site). These rates are then mul- tiplied by the channel width and time period of interest to arrive at amounts of bend translation and extension. The cal- culations are shown in Table 8.2 using the C site percentages, a 30-year period, and a channel width (average of the upstream and downstream crossing widths) of 1,000 ft for this site. The 30-year movement of the bend, excluding change in radius, can be plotted by shifting the existing bend by the amounts com- puted for extension and translation. These amounts are shown in Figure 8.15. Figure 8.15 shows that the predicted rate of movement for this bend matches the 75-percent frequency bankline very well. At the 75-percent level, three-quarters of C Class bends have moved at a rate less than the predicted rate and one- quarter of C Class bends have moved at a rate higher than the predicted rate. The results do not indicate that the predicted amount is unreasonable, just that it is greater than average. If the predicted rate seems high, then the user should check the accuracy of photo registration, the accuracy and consis- tency of the banklines and circle fit, and the calculations of translation and extension. Even if the rates appear high, they may not be wrong. The bend may actually be migrating at a high rate. Whether this rate will continue into the future depends on hydrologic variability, bank material character- istics, and land use change, which are not considered in the extrapolation method or the frequency analysis other than by the assumption that past conditions will persist into the future. Another potential use of the frequency results would be at a site where historic photos are unavailable. If one had a cur- rent aerial photo and no other information at a site, one could use the frequency rates to assess potential bend migration. The only conclusion that could be drawn from the frequency analysis is that for this class of channel the measured bends in the database have moved at the indicated rates and fre- quencies. Clearly, if the Sacramento River bend near Butte City is allowed to migrate unconfined, channel migration into the levee is likely within 10 to 20 years.

TABLE 8.2 Frequency calculations for Sacramento River example 25 50 75 90 95 Cumulative Percent Extension C Sites (channel widths/yr) 0.0011 0.008 0.018 0.032 0.045 x 1,000 ft width (ft/yr) 1.1 8.0 18 32 45 x 30 years (ft in 30 yrs) 33 240 540 960 1350 Translation C Sites (channel widths/yr) 0.005 0.015 0.031 0.055 0.074 x 1,000 ft width (ft/yr) 5.0 15 31 55 74 x 30 years (ft in 30 yrs) 150 450 930 1650 2220 Figure 8.14. Overlay of the predicted channel position in 2028 on the 1998 aerial photograph of the Sacramento River near Butte City, California. Note the threat to the local levee. (RM = river mile.)

63 Figure 8.15. Bend movement for C Class frequencies. (RM = river mile.)