Influence OF reservoirS ON SEDIMENT TRANSPORT AND EROSION PROCESSES BELOW THE DAMS

Zygmunt BabiŃski

Institute of Geography, Kazimierz Wielki University of Bydgoszcz

85-428 Bydgoszcz, Mińska St. 15, Poland, e-mail: zbabinski@o2.pl, phone: +48 603 58 94 38

Introduction

For almost 5.000 years man has influenced river systems by building single dams or multiple dams in cascade systems. This practice has been the most visible sign of human interference in the fluvial environment. Each barrier applied to a river bed causes direct and usually irreversible changes within and outside the ecosystem.

The river channel partition by the dam effects on one side by the stoppage of sediment transport, represented by suspended- and bed load ‑ accumulated in reservoir, and on the other, renewal it’s on the way of the channel-bed erosion below the dam. With intense erosion process downstream from the dam are usually associated channel transformation caused by sediment transportation, divided on two parts. The first area takes the form of a newly created flood-plain associated with the area of erosion. The second one, below the front of an erosion wave, attributes an aggradation character to that section of the river bed. The both opposite processes create new ecological conditions, not always favourable for natural environment and management.

The explanation of reasons and effects stoppage of sediment transport by dams is the main aim of the paper. Moreover, the qualitative and quantitative channel changes below the dams are considered. The influence of channel transformation on economy and ecology were presented.

The study of channel processes is based on the interpretation of air photographs, large-scale setting-altitude maps, bathymetric maps, cross-sections and longitudinal profiles of channel bed and water level. The statistic method and method of successive comparison have been used.

Moreover, the analyze of fluvial processes below the dams based on literature, mainly: A.B. Veksler and V.M. Donenberg (1983), G.P. Williams and M.G. Wolman (1984), S. Raynov, D. Pechinov and Z. Kopaliani (1986), A.B. Avakian, V.P. Soltankin and V.A. Sharapov (1987), on the background of 35. years investigations in Wloclawek Reservoir (Babiński, 1982, 1992, 2002).

Hydrological regime

The main task of dams is water retention and regulation of its current intensity. As a result, the annual amplitude of fluctuation of water level in a reservoir is lowered. Research carried out by American scientists has shown (Williams and Wolman, 1984) that the average annual high water level of 21 rivers studied dropped from 3% to 91% ‑ an average of 39% . However the so-called 5% water culmination was not reduced. As regards the Naser Reservoir (Raynov et al., 1986), it can accumulate up to 47% of the Nile’s annual effluent which has a significant effect on the flooding wave. In the case of the Danube below Iron Gate I, the 15% coefficient affects the decrease of annual amplitude of fluctuations only slightly. Such a low coefficient is characteristic of a number of lowland reservoirs of the Northern hemisphere (including the Włocławek Reservoir – 14%). A common feature is small effective capacity.

The operation of barrages usually influences the daily fluctuations of water level within a range of 0.5-4 m up to a maximum of 4.6 m below the Saratov Reservoir (Berkovich, 1992). In extreme cases fluctuations can be traced as far as 200 km downstream from the dam (Babiński, 1992). In comparison, the river Danube below Iron Gate I is similarly influenced at a distance of 250 km (Raynov et al., 1986). The Włocławek Reservoir on the Vistula has similar parameters (about 200 km) and only slightly smaller daily fluctuations reaching up to 3 m, 3.5 m at the extreme (Babiński, 1982). It should be noted that according to the demand for electric power (time of a day, kind of a river) there are 1-2 culminations in a 24 hour flow, seldom less than two culminations or no culminations at all.

Reservoir sedimentation

The building of barrages is inseparably associated with the process of accumulating bed-load and suspended-load in a reservoir. The amount of load subsumed by a reservoir and particularly the rate of flow intensity to tank capacity is essential factors when estimating its endurance. Research carried out on sedimentation showed that the amount of deposited load ranges from 80% in the Garrison Reservoir (the Missouri) to 99.5% in Canton ‑ North Canada (Williams and Wolman, 1984). In the Włocławek Reservoir, total load amounts to 90%, of which 100% is the bed-load and over 41% the suspended load. The youngest and greatest dam, creating of Three Canyons Reservoir on the Yangtze (China), is characterised by the last-mentioned value (The Three..., 1994). The downstream consequences of interrupting the flux of sand and gravel transport would argue for designing systems to pass sediment through reservoir (and thereby reestablish the continuity of sediment transport).

The sedimentation process leads to the loss of retention properties of reservoirs. The location and the kind of accumulated load affect the ecology and management of dams. The higher part of a reservoir constantly becomes shallower owing to the alluvial cone in the back flow area. Subsequently, abrupt changes develop in the hydrodynamic features of the river, making its outflow difficult. In favourable conditions the homogenous mass of water can split into several arms winding among emerging or submerging alluvial sandbanks. Some perceive the new hydromorphological situation in the back flow area as a positive phenomenon, while others claim that it is harmful to economy. Sedimentation processes help to form new settlements friendly to the local river fauna, hydrophilic animals and fish. This is regarded as a positive feature and an argument for justifying the building of reservoirs. As regards the negative influence, accumulation of the bed-load has the most harmful effect within the higher part of a reservoir. The shallow and very dynamic bed can make proper floating difficult or even impossible. Moreover, the constant “rising” of a bed reduces the area of a river or a reservoir cross-section. Hydrotechnical buildings built before the water was dammed up can be destroyed during a freshet. A result can be flooding, which would not occur with a “real” river. This fact is better illustrated in climatic zones susceptible to ice-jam. In favourable conditions, brash ice jams can dam up water above the safety level. The hydrotechnical buildings are then damaged and the previously safe regions become flooded. Such a disaster occurred in the proximity of the Włocławek Reservoir in the winter of 1982.

Channel incision below the dams

When flowing through a reservoir, water loses its clastic load. This is a direct cause of deep-seated erosion below dams (Babiński, 2002). This erosion varies in time and space and is estimated by the speed with which the river bed lowers and by the extension of the erosion area. In general we observe that these two processes slow down. In other words, erosion expands most dynamically at an early stage of the operation of dams and in their proximity. This observation is valid for the majority of dams discussed in this paper. However, a more detailed analysis suggests that the process is more complex. It has been found that there is a period of abrupt changes in a river bed attributable to the building of a dam. When water begins to fill a reservoir, the process of deep-seated erosion slows down but resumes when the reservoir is full. As time passes, the erosion process is increasingly modified by geological conditions of the river bed, by the local hydrographical and morphological net of the area and by man’s economic activity.

In all measured cases of channel incision below dams (tab. 1), the process was faster in the first years of damming up water and in the proximity of the dams. This was the period when the river beds lowered. Research into the American rivers showed they were lowered by 0.6 m to 5.8 m, with a maximum of 7.5 m (Williams and Wolman, 1984), in the case of other rivers in the Northern hemisphere, their beds lowered by 0.6 m to 5.6 m (fig. 1). When a flooding wave occurs, local deep waters up to 31 m can result at the entrance of electricity plants and weirs (Raynov et al., 1986). Such pools in a river bed endanger dams. They must be immediately protected against further degradation (usually they are filled with blocks of concrete).

An analysis of cross-section profiles of the American rivers helps to estimate the most frequent extreme depth of deep erosion in cross-sections of river beds, ranged from 2 to 4 m. In comparison, the stretch of the Vistula below the Włocławek Reservoir shows a much more active process. The river bed has lowered by an average 3.5 m in the course of the dam’s 27 years of operation (tab. 1, fig. 1).

Figure 1. The rate of deep erosion below the dams, after data presented in Table 1 (average in cross-profile in metres) and their straight regression

Complete presented in Z. Babiński (2002)

For a more comprehensive definition of deep erosion, it is necessary to examine the extent of the erosion area (the front of erosion zone) and the speed (shifting) of its advance. Although the extension of the erosion area is the most dynamic in the initial period of the operation of dams, it cannot be described as a vanishing straight line function (Galay, 1983, Raynov et al., 1986, Andrews, 1986, Belyj et al., 2000), complete presented in Z. Babiński (2002). The process varies and can be illustrated by the values achieved below the Włocławek Reservoir (0.7- 2.7 km per year) and on the Colorado River below Hoover’s Dam (Williams, Wolman, 1984). The latter’s erosion area below Hoover’s Dam advanced 21 km within six months of its construction. This erosion area has the greatest extension speed in the world (42 km per year). After a year it reached a distance of 28 km, after two years – 50 km, after three years – 85 km. Altogether after 14 years it extended to 111 km at an average speed of 10 km per year. The data concerning the examined rivers show a variety of erosion area lengths, within a range of 2.5 km (after observation of the Gerstheim Dam on the Rhine over 1.5 years) up to 551 km below the Aswan Reservoir on the Nile (after 18 years of observation). The rate of erosional zone shifting after authors referred above, introduced in fig. 2.

Table 1. Channel incision below the dams

Lp.	River, dam (reservoir), region (state)	Term of observation (years)	Degradation in cross-profile (m)
1	Colorado, Glen Canyon, Arizona	9	7.3
2	Colorado, Hoover, Arizona	13	7.5
3	Colorado, Davis, Arizona	26	5.8
4	Colorado, Parker, Arizona	27	4.6
5	Colorado, Imperial, Arizona	18	3.1
6	Jemez, Jemez Canyon, New Mexico	12	2.8
7	Arkansas, John Martin, Colorado	30	0.9
8	Missouri, Fort Peck, Montana	36	1.8
9	Missouri, Garrison, North Dakota	23	1.7
10	Missouri, Fort Randall, South Dakota	23	2.6
11	Missouri, Gavin’s Point, South Dakota	19	2.5
12	Medicine Creek, Medicine Creek, Nebraska	3	0.6
13	Middle Loup, Milburn, Nebraska	16	2.4
14	Des Moines, Red Rock, Iowa	9	1.9
15	Smoky Hill, Kanopilis, Kansas	23	1.5
16	Republican, Milford, Kansas	7	0.9
17	Wolf Creek, Fort Supply, Oklahoma	27	3.4
18	North Canadian, Canton, Oklahoma	28	3.0
19	Canadian, Eufaula, Oklahoma	6	5.1
20	Red, Denison, Oklahoma-Texas	16	3.0
21	Neches, Town Bluff, Texas	14	0.9
22	Chattahoochee, Buford, Georgia	15	2.6
23	South Canadian, Conchas, Nev Mexico	7	3.0
24	Salt Fork, Arkansas, Great Salt Plains, Oklahoma	9	0.6
25	Rio Grande, Elephant Butte, Texas	15	1.8
26	An Sabee, Foote Sariyar (Turkey)	15	1.5
27	Saskatchewan, Squaw Rapids	13	1.2
28	Cheyenne, Angostura	16	1.5
29	South Saskatchewan, Dietenbaker	12	2.4
30	Huang He, Sanmenxia, China	4	4.0
31	Syr-Daria, Farchacka	7	1.3
32	Murgab, Hindukuska	60	4.0
33	Murgab, Tadhenska	7	5.6
34	Isar, Dingolfing	14	2.8
35	Lech, Forgensee	10	0.6
36	Saalach, Reichenhall	47	4.6
37	Wertach, Schwabmunchen	5	1.8
38	Danube, Faimingen	12	1.0
39	Danube, Ingolstadt	14	1.8
40	Rhein, Gerstheim	1.5	2.5
41	Vistula, Wloclawek, Poland	27	3.5

After: G.P.Williams and M.G.Wolman (1984), V.J. Galay (1983), S. Raynov et al. (1986), E.D.Andrews (1986) – selected: detail in: Z. Babiński (2002)

The average annual speed of erosion is used as the criterion to compare the rivers in the study. For a period of 1.5 to 54 years the speed varied from 0.4 km (some rivers of the Northern hemisphere) to 36 km (below the Farcha Dam on the Syr-Daria River). The average annual speed was within a range of 1-3 km; the most quantity events (over 23) of front-erosion shifting ranged in zone 0-2 km per year (fig. 2A). The front edge of the erosion area below Włocławek Reservoir has advanced at an average annual speed of 1.2 km over 27 years.

Figure 2. The rate of erosion zone shifting (km per year); A – quantity events in respective erosion zone shifting group

After: V.J. Galay (1983), S. Raynov et al. (1986) and E.D. Andrews (1986), B.V. Belyj et al. (2000); complete presented in Z. Babiński (2002)

Morphological and s