A. Steele Becker, Department of Geography and Earth Science, University of Nebraska - Kearney, Kearney, Nebraska 68849, (308) 865-8355.
The Big Bend reach of the Platte River, extending from Gothenburg to Grand Island Nebraska, has been a focus of attention for more than a century for its water resources. Irrigators, municipal dwellers, and recreational users all came to recognize the uniqueness, and finality, of this valuable resource. The construction of Interstate 80 in the 1960s brought even more attention, access, and use.
The present interest and issues center on a perceived controversy between environmental issues, agricultural and urban water users, recreational interests, and the Federal relicensing of Kingsley Dam. Kingsley Dam impounds Lake McConaughy, the eastern-most of a chain of reservoirs on the North Platte River that has largely controlled the flow of the Platte River since its completion.
In the more than 100 years of human utilization of the Platte River, considerable changes have occurred in channels, islands, vegetation, and volume of water flow. Much has changed in the Big Bend reach of the Platte since it was first seen by Europeans more than 200 years ago. While I have studied many facets of the Central Platte region for more than 30 years, my specific focus in this study is whether the Big Bend reach, or portions thereof, have reached stabilization and, if so, determine probable causes.
During two mapping projects in the early 1980s, I noticed an apparent relationship between bridge occurrence and the appearance of the river channels. Specifically, the reach from two miles east of the State 44 bridge south of Kearney to the Grand Island I-80 interchange was significantly narrower and had fewer channels (braiding) than the reach from Gothenburg to two miles east of Kearney. In 1985 and 1986 I conducted a preliminary grid analysis study of channel and island patterns in the reach of the river from Gothenburg to Grand Island, covering the period 1976 to 1985, to determine if significant differences in channel and island relations could be identified. The results of that preliminary study caused me to focus attention on the occurrence and frequency of bridges and their possible influence on river behavior.
In late 1987 I began the collection of a detailed data base on the 112 mile reach of the Platte River between Gothenburg and the east edge on Hall County, just east of Grand Island, using optical grid analysis. The reach was divided into 25 segments, based on the sections of the Township and Range system, of the U.S. Land Office Survey System, to facilitate data collection (Figure 1). I covered the twelve-year span 1976 to 1987, specifically collecting data from air photos for 1976 and 1982 to 1987. This time span would also allow collection of data covering the unusually high flow years of 1983 and 1984.
Results of the analysis completed on the collected data identified a gradual west to east change in the river channels and islands plus identifiable differences between segments of the study reach. A "normal" pattern of change for the entire reach over the 12 year period was a 13 percent increase in channel area accompanied by a 25 percent loss in island area. Channel and island changes from east to west are shown in Figure 2. Statistical analysis of the data also clearly indicated differences in the changes within each of the 25 segments.
Using Figure 2, three reaches of the river were identified based upon channel and island changes. While both channel and island changes were investigated and documented, the most striking contrasts were in the channel changes from west to east within the 112 mile study reach. Segments one through seven (Gothenburg to 3.5 miles east of Lexington) showed a channel gain above 13 percent; segments eight to eighteen, (3.5 miles east of Lexington to Gibbon), with two exceptions, were near or below the 13 percent figure. The most interesting results were in segments 19 to 25 (Gibbon to Grand Island) where five segments showed actual loss of channel area, the only such pattern within the 112 mile reach.
The preceding suggested that there are variables within the system, other than the high flows of 1983 and 1984, that were responsible for the changes observed. These changes were system-wide, but the pattern of change was not sectionally consistent. Five variables were identified as possible causative agents to changes within the system. Diversions/withdrawals, channel frequency (multiple channels or braiding), channel width, bridge occurrence and location, and river gradient.
Within the 112 mile study reach there are 29 bridges, 14 within the eastern reach, segments 19 to 25. This reach also contains five segments where the gradient exceeds the 6.6 fpm average for the entire reach. When gain to loss was compared segment by segment it became evident that the channel/island relationship was virtually stable from segment 18 (Gibbon) to the east end of the study reach. I then focused my attention on bridge occurrence and frequency as the possible primary cause of this "stability."
Nineteen bridge sites were selected at random (Figure 3) and data collected on channel and island area in half mile wide strips (2,640 feet) above and below the sites for a distance of three miles. Channel width was measured, and number of channels determined, for the same distance. Where there was another bridge within that linear distance, the dividing point for "above" or "below" was half the river distance between the two bridges.
The study sites were then subjected to comparative analysis to determine if there was an identifiable difference between the behavior of the six mile reaches centered on the bridges as compared to the entire study area. Results of the analysis indicated a difference in behavior of the bridge-dominated areas as compared to the entire study reach. A percentage comparison of total water/island area and water and island area alone for each study year confirmed the differences in behavior.
If the true controlling factor is the bridges, then statistical analysis seemed to offer a route to begin determining the extent of influence. To this end the following hypothesis were developed concerning the bridge-centered six mile reaches:(1) Total channel width above a bridge site is greater than or equal to the total channel width below the site. (2) Average channel width above a bridge site is greater than or equal to the average channel width below. (3) Total channels above a bridge site are greater than or equal to the total number of channels below. (4) Total channel widths for all bridge locations (sites) are not equal. (5) Total widths for all readings of channel widths are equal. (6) Total water area above a bridge site is greater than or equal to the total water area below a site. (7) Total island area above a bridge site is greater than or equal to the total island area below a site.
Hypothesis 1, 2, and 3 were subjected to a standard "T" test for the years 1976, 1983, and 1985, the years for which data was available for all 19 bridge study sites. Hypothesis 4 and 5 were tested using an Analysis of Variance (ANOVA) test for the same years. The hypothesis were found to be valid. Hypothesis 6 and 7 were tested using a standard "T" test for all years included in the study (1976 and 1982 through 1987). These hypothesis were also found valid.
The results of these analysis convinced me that the numerous bridges on the Platte River produced a stabilizing effect on river behavior, and that when changes did occur, they were more gradual in nature. In short, frequent bridge occurrence resulted in fewer separate channels and islands, a stable, and possibly deeper, main channel producing a much more stable general behavior.
The previously described study, and earlier grid analysis and mapping, produced intriguing insights and questions concerning the character and behavior of the Big Bend reach of the Platte River. This was particularly true concerning the possible effect of bridges which, to my knowledge, had not been considered before. Substantiation of the initial results, however, would require a more expanded and thorough approach.
In order to determine if the original results could be confirmed over a longer span of time than the original twelve year period, I undertook a new study covering a period of 50 years beginning with 1938, the first year air photos were taken of the river, and extending through 1988. This would allow the inclusion of a period prior to the closing of Kingsley Dam in 1941, the abnormally high flows in 1983 and 1984, and a sufficient period of time following these flows to determine if they produced any significant and/or permanent changes. In addition to the 1938 and 1988 data I included 1963 to have a representative year from the 1950s or 1960s. The same grid analysis collection techniques of the earlier studies were followed, but actual collection was done using an electronic planimeter, rather than optically, to improve accuracy.
One problem with the earlier study was that I possessed incomplete air photo coverage for all years for the reach between Gothenburg and Lexington (segments 1 to 6). I did, however, have sufficient data on the reach to show that the principal bridge affected segments were located considerably to the east of this reach. I thus decided to limit the new investigation to the approximately 88 miles of river between the Lexington bridge and the east edge of Hall county which is five miles east of the U.S. 281/34 Bridge at the Grand Island I-80 Interchange (original segments 7 through 25). This would include 16 of the original 19 bridge study sites. I do not believe this compromised the study and provided a more complete and uninterrupted data base over a longer period.
The earlier investigation utilized data collected for strips one-half mile wide for a distance of three miles above and below each bridge study site. Analysis of that data showed that the effect of the bridge extended for two miles above and below each site, not three. This was the distance included in the present study. For the present study data was collected in strips 1,320 feet wide to increase comparison accuracy. This data included channel and island area, channel width measurements, and channel numbers at the same intervals. For sections outside the four mile bridge-centered strips, channel width measurements and channel numbers were recorded at the upstream edge, middle, and downstream edge of each section. Channel and island area was measured per section. As stated earlier, the original grid analysis method was retained for this study, with a electronic planimeter replacing optical collection. This new procedure was followed for all years included in the study, not just 1938, 1963, and 1988. This ensured consistency of data, increased accuracy, and a greatly expanded data base.
In order to test the results of the earlier study in the greatest detail while making maximum use of the new and more detailed information, the collected data was entered into a new data base structured so that the original 19 segments between Lexington and the east end of Hall County were divided into 101 north-south tiers, based upon sections of the U.S. Township and Range System. This allowed for much greater detail of analysis and comparison. Each tier corresponded to a section one mile wide, or four data collecting strips 1,320 feet wide for the bridge-centered reaches and one mile wide strips for the nonbridge reaches, rather than the much wider segments of the earlier study.
The result of this new approach is that the entire study reach is subjected to much more detailed analysis, similar to the method used for the bridge-effected areas only in the earlier study. In that study each segment contained combined data for an average of four linear river miles, collected section-by-section, and averaged for the segment as a whole. Here the data was collected in strips one-fourth mile wide for the bridge-centered reaches and one mile wide for the nonbridge portions, structured in the data base so that each tier represented a cross-section of the river at that point.
I also included specific investigation of the impact of the abnormally high river flows of 1983 and 1984. The original study suggested the magnitude of change in channel and island area because of the high flows decreased from west to east, a pattern corresponding to the increased development of river stability. It also suggested that the system (channels plus islands) began to return to pre-1983 conditions in 1984, despite the repeated high flows of that year. In other words, significant channel/island area change was minimal and/or short lived. If this was true, and correlated with bridge affected areas, it would give further support to the importance of bridge control in river behavior. I believed that the much more detailed tiered data base would provide insight into this event.
The procedure was to average the changes in the entire system, tier by tier, for the 50 year time span of the study and then factor in the changes, tier by tier, for 1983 and 1984 to determine if significant departures from the long-term average did occur. The results were then displayed in bar graph form (Figures 4 and 5) for each tier.
The results showed that the greatest departures from the average pattern for the high flow years were in the areas roughly west of Kearney where channels are most numerous, the "system" (channels and islands) the widest, and bridges few, long, and widely separated. Stabilization begins to occur within tier 36, just east of the Overton Bridge, nine miles east of Lexington. This stabilization gradually increases to the east. At no point do the wide fluctuations from the long-term average, found to the west, occur. Where some departure does occur within a tier, it is due to specific channel conditions at that point or "pooling" of water above bridge sites, an effect that rapidly disappears below the bridge site unless bridges are so closely spaced that there is an overlapping affect. Table 1 gives an explanation of the behavior on a tier by tier basis.
The final step was to use the greatly expanded data base to construct a series of three- dimensional graphs to allow a comparison of four years, 1938, 1963, 1976, and 1988, tier by tier for the entire study reach from Lexington to the east edge of Hall county (Figures 6-11, at end of paper). The graphs compare land and water area for each tier, for each of the selected years. The results reinforce my earlier findings that channel stabilization is the result of bridge occurrence and frequency. The much more detailed tier analysis method actually shows that this process begins farther west than originally thought.
Careful examination of the data shows that the entire "system" was indeed significantly reduced over the 50 years following the closing of Kingsley Dam. The same data also shows that the principal loss between the Lexington and Minden I-80 interchange bridges (tiers 26-66) has been in channel area, accompanied by some actual island gain between the Odessa I-80 interchange bridge (tier 51) and the Minden I-80 bridge (tier 66). From tier 66 eastward to tier 76 (just west of the Shelton I-80 interchange bridge), channel area either is equal to island area, or exceeds it. This is the portion of the study reach where multiple bridges and "overlapping" of bridge effect first occurs.
Finally, the reach from the Shelton I-80 interchange bridge (tier 77) to the east end of Hall county (tier 101) generally shows a dominance of channel area to island area. This is particularly true from the Wood River I-80 interchange bridge (tier 85) eastward. Six of the 13 bridge study sites occur within this reach, which actually contains 12 bridges. I concluded, therefore, that based upon the data I have gathered and performed multiple analysis, that this reach of the river is in fact stable and that the primary controlling variable responsible is the occurrence and frequency of bridges.
Return to 1997 Platte River Basin Ecosystem Symposium
Last updated by Darren A. Jack on 4/28/97