This year the world population will reach 7B people, increasing to 9B by 2045. As the world population grows the demand for reliable water supplies increases. The sustainable and reliable supply of fresh water ensuring a net improvement in the global quality of life is accomplished through prudent resource use and through minimising the environmental footprint associated with developing those resources.

Prudent use

Of all the water on earth, river water is the only sustainable fresh water resource. Some estimates indicate that if one would be able to suddenly remove all water from the earth’s rivers, it will, on average, be replaced by natural processes in little more than two weeks. The replenishment rate for groundwater, on the other hand, is on the order of decades to millennia. River water use is sustainable, while groundwater use is unsustainable.

The total amount of fresh river water or earth is about 2120km3 (Gleick 1996). This water can fill a sphere with a diameter of only 16km, a distance that is less than most people’s daily commute. The fact that all of the world’s fresh river water can fill such a small sphere indicates that it truly is a scarce resource requiring careful management.

Another important characteristic is that the amount of water flowing in rivers varies from season to season, and from year to year. River flow never remains constant. During low flow seasons and multi-year droughts enough water might not be available to supply in demand. And during high flow periods, when floods occur, more water may be available than what can be used.

This varying availability of water therefore requires construction of dams in rivers to create enough storage space to capture flood waters for use during droughts. Historically, it is not uncommon to experience droughts of five to seven years in some parts of the world, including Southern Africa, East Africa, the American West and Australia. Large dams storing enough water to supply in demand during multiple-year droughts increase the reliability of water supply. Maintaining that storage in the long term should be a prime infrastructure management objective.

Historic storage trends

For the last 30 years the decreased rate of dam construction and the effects of reservoir sedimentation resulted in a net decrease in global reservoir storage volume, reduced reliability of water supply, and a deterioration of the quality of life for many.

When water flows into reservoirs it slows down and deposits the sediment it carries. As the amount of sediment depositing in reservoirs increases it occupies storage space and reduces the amount of water that can be stored. It is estimated that reservoir sedimentation annually leads to about 0.8% to 1% of global reservoir storage loss. The historic progression of the amount of storage that is internationally available is therefore determined by the rate of dam construction (the rate at which new storage is added) and the rate of storage loss due to reservoir sedimentation.

Figure 1 shows as a function of time the global population growth (left vertical axis) and the changes in global reservoir storage volume (right vertical axis). It shows that the rate of dam construction has decreased since about 1980. The combined effect of the reduced rate of dam construction and reservoir sedimentation resulted in the net storage volume loss.

The development level of communities and nations are tied to the amount of available water storage. This can be demonstrated by relating the amount of water storage per capita in various countries to their Human Development Index (HDI). The HDI, which ranges between 0 and 1, is used by the United Nations to quantify the relative development level of nations. High HDI values (i.e. index values approaching 1) indicate high levels of development and high standards of living. The bars in Figure 2 relate the per capita water storage to HDI, indicating high per capita storage for highly developed nations (HDI > 0.85). The figure also shows the available global per capita volume of water stored in reservoirs as a function of time (top, horizontal axis). It shows that the per capita storage added through dam construction (initial storage) decreased since about 1990. After allowing for storage loss due to reservoir sedimentation, it reduced since about 1980. It is noted that the current net per capita storage is about equal to what it was in 1965; a certain retrogression.

This leads to two conclusions. Firstly, for ensuring reliable water supply it is necessary to add more storage globally, ie to construct more dams. Secondly, implementation of reservoir sedimentation management technology is critical to maximise the longevity of water storage space behind dams. The amount of storage that should be added is determined by considering regional and national water supply needs, and is not dealt with here. Preserving storage space by countering the effects of reservoir sedimentation is presented in what follows.

Sustainable management of infrastructure

Conventional civil engineering design philosophy embraces the concept of a ‘design life’. More often than not, the design life approach works well for conventional civil infrastructure, such as roads. The reason for this is that although a road may have a design life of, say, 30 years it is easily refurbished by repaving it at the end of 30 years. It is therefore possible to use the road for periods longer than its design life, possibly in perpetuity. The ease of refurbishment of such infrastructure justifies the use of this design concept.

The problem of implementing the design life approach to dams, particularly when they are large, is that dams are difficult to refurbish once they are filled with sediment. In locations with high sediment loads, reservoir sedimentation leads to non-sustainable use of the infrastructure. Unlike roads, once reservoirs behind dams are filled with sediment they cannot be used anymore. The design life approach, although feasible for most conventional civil infrastructure, is not suitable for dams and their reservoirs. A paradigm shift is necessary: changing from a linear styled of thinking (design life approach) to a circular style of thinking (life cycle management approach).

A case in point is Tarbela Dam, Pakistan. This facility was constructed in the 1970s and has since filled with sediment at an alarming rate. The dam and reservoir supplies about 30% of the country’s irrigation water and about 30% of its electric energy. As it fills with sediment, the amount of water available for irrigation will reduce, as will its ability to generate electric energy. Eliminating these benefits due to the effects of reservoir sedimentation is undesirable and can have a significant adverse impact on the country’s national economy. Were sediment managed in this facility on a regular basis it would have been possible to use it in perpetuity, not losing the benefits it provides.

Figure 3 conceptually illustrates the differences between the design life and the life cycle approaches. Figure 3a shows the relationship between water supply reliability and time for a dam and reservoir that have been designed and operated using the design life approach and fills with sediment at the end of its design life. At that point in time the infrastructure has served its useful life and can no longer provide its original benefits. The infrastructure has been designed and is operated in a non-sustainable manner.

Figure 3b illustrates a dam and reservoir that has been designed to allow reservoir sedimentation management on a regular basis, ie following the life cycle management approach. By regularly removing deposited sediment from the reservoir it is, ideally, possible to retain the water supply benefits from the facility in perpetuity. The continued stream of benefits is therefore established, as is the case for other civil infrastructure, like roads.

Reservoir sedimentation management

A number of strategies exist and can be applied to minimise the effects of reservoir sedimentation. They can be applied either individually or in combination. The overall concept is very simple: reduce the amount of sediment flowing into a reservoir; create conditions that will prevent (or at least minimise) the deposition of sediment in a reservoir; and remove whatever sediment has deposited in the reservoir.

In order to minimise damage downstream of the dam to river morphology, nutrient supply and aquatic habitat, one may enhance the sediment supply to downstream river reaches; a practice known as sediment augmentation. The principal methods often proposed to reduce sediment yield from catchments are re-vegetation, contour farming, check dams and warping.

Re-vegetation – The general opinion of experts in reservoir sedimentation management is that re-vegetation is not an effective reservoir sedimentation management technique. This does not mean that it should not be implemented for other purposes, such as prevention of soil loss that is critical for food security. It merely means that other reservoir sedimentation management techniques are usually more economical, viewed purely from a reservoir sedimentation management point of view. For example, the catchment upstream of Mangla dam, Pakistan was subjected to re-vegetation efforts over the period 1960 to 1983. No noticeable reduction in sediment yield from this catchment, with an area of 33,360km2, was reported. It is generally believed that re-vegetation of small catchments (<150km2) can possibly reduce sediment yield by 30% to 70% depending on location and catchment conditions.

Warping – This is a technique often used in China where river water with high sediment loads is diverted onto agricultural land. The sediment depositing on the land enhances its agricultural value. However, in large rivers the amount of sediment diverted is only a small portion of the total sediment load. Therefore, although this practice adds value to agriculture it does not necessarily significantly reduce the amount of sediment carried by a river.

Contour farming – This is actively practised in some regions of the world, like in Nepal, and it benefits agriculture. However, the contribution it makes to reduce the sediment yield of major rivers is likely to be very small. This practice should obviously be encouraged from agricultural, social and environmental points of view; but it contributes little to reduce the sediment yield of major rivers.

Check dams – These have been implemented as a sediment management measure upstream of dams (Figure 4). The reduction in the river’s sediment transport capacity in the reaches immediately upstream of check-dams results in sediment deposition. In order to maintain their effectiveness they require regular maintenance, ie removal of deposited sediment. Check dams are generally applied in series to increase the amount of sediment they can capture. It is noted that even if they are filled with sediment, placing a large number of check dams in series in a river may reduce its sediment yield. To accomplish such a goal it is necessary to arrange the dams in a manner that will significantly reduce the energy slope in a river, and thus its sediment transport capacity. Check dams are most effective in catching large sediment particles, and are less effective in catching small particles.

The best known techniques to minimize the amount of sediment deposited in reservoirs are bypassing, sluicing and density current venting.

Bypassing – The objective with bypassing is to divert sediment carrying waters around reservoirs and prevent it from entering and depositing sediment in the reservoirs. Various schemes may accomplish this goal, but the most common entails the use of bypass tunnels. However, bypassing may also be accomplished by modification of river channels and using off-channel storage. At least five bypass tunnel schemes exist in Switzerland and four in Japan, with others planned.

A conceptual depiction of the bypass tunnel scheme for Miwa dam, Japan is shown in Figure 5. A diversion weir was constructed on the upstream end of Miwa dam’s reservoir. When floods containing high sediment loads are present the diversion structure diverts the flows into the tunnel for discharge downstream of the dam. During average flow conditions, the water is allowed to flow over the diversion weir into the reservoir, instead of into the tunnel. Average river flows contain low sediment loads, resulting in low volumes of deposited sediment within the reservoir.

An interesting bypass scheme that has been designed for Nagle dam in South Africa is shown in Figure 6. The dam and its reservoir are located in a river with high sediment loads and have been successfully operated since the 1940s. The dam is located at the downstream end of a large river meander and its reservoir is located in the meander bend, upstream of the dam. At the upstream end of the meander a flood weir was installed, which is open during average flow conditions but closed during floods. During average flow conditions the water flows through the open flood weir into the reservoir. These waters contain low sediment loads and the amount of sediment deposited in the reservoir during such conditions is small. When floods occur, the flood weir is raised and diverts the waters containing high sediment loads into the diversion channel that has been specially constructed to divert these waters to the river channel downstream of the dam. In this manner the high sediment loads are not conveyed into the reservoir during flood conditions and do not deposit within it.

Off-channel storage accomplishes sustainability by diverting clean water into an off-channel storage space, while allowing waters with high sediment loads to flow past the dam and reservoir. An example of such a project is shown in Figure 7 – the Fajardo dam in Puerto Rico. The photo shows the off-channel storage space.

Fajardo Dam receives its water via a pipeline conveying it from a higher elevation in the mountains where the water is free of sediment and essentially clean. The water is transferred by means of gravity and discharged into the storage space behind the off-channel dam lower down in the river. As the river water flows down towards the ocean the character of the river changes from carrying low sediment loads upstream to carrying high sediment loads downstream where the dam is located. Should the dam have been built in the river, these sediments would have deposited in the reservoir upstream of the dam. However, the design, as implemented, allows the river flows with high sediment loads to flow past the dam and reservoir without deposition, while only clean water transferred from the mountain is stored in the reservoir. This prevents the reservoir from filling with sediment.

Sluicing – This is an operational technique in which sediment laden flows are released through a dam before the sediment particles can settle. In essence, the sluicing concept consists of maintaining high sediment transport carrying capacities in the water flowing through a reservoir that will prevent or minimise the amount of sediment depositing in the reservoir. Ideally sluicing should be executed in a manner that will result in the amount of sediment entering the reservoir equalling the amount exiting. An example of a project where sluicing is practised is First Falls dam, South Africa.

Density Current Venting – When water with very high sediment concentrations flows into a reservoir it is possible that the density of the sediment laden water is higher than the water contained in the reservoir. Depending on local conditions very little mixing might occur between the density current and the reservoir water. This means that a dense, sediment laden current flows along the bottom of the reservoir towards the dam.

Deposition of this sediment can be prevented by releasing the density current downstream of the dam. This is accomplished by installing low level gates at the dam. When these gates are opened as the density current approaches the dam, the high sediment concentration water may be released downstream of the dam. This means that the heavy sediment loads contained in the density current is discharged downstream of the reservoir, without depositing in the reservoir. The process is known as density current venting.

An interesting approach to releasing density currents has been developed in Japan, and was implemented at Katagiri Dam. In that case a curtain wall that can direct the density current upwards and over the dam was provided. The concept is illustrated in Figure 8.

Once sediment has deposited in a reservoir it can be removed by means of dredging, hydro-suction, dry excavation, drawdown flushing and/or pressure flushing.

Dredging – As applied in reservoir sedimentation management, dredging generally consists of using hydraulic pumps on barges with intakes down to the reservoir bed, which can create enough suction to remove deposited sediment from the reservoir bottom. If the deposited sediment is very cohesive it is difficult to remove by suction only. In such cases it may be required to use cutter heads at the suction end of the pipe. Cutter heads loosen the deposited, cohesive sediment and allow it to be entrained by the suction created at the end of the dredge line.

Dredging is generally expensive and is rarely used to remove large amounts of sediment. More often than not it is used to perform ‘tactical dredging’; which means that it is only used to remove sediment from very specific areas. In the case of dams sediment removal will for example be focused to the immediate vicinity of intakes.

Dry excavation – Removal of deposited sediment by dry excavation consists of draining the reservoir and using conventional excavation equipment to load deposited sediment into trucks for removal from the reservoir. This approach to sediment management is usually suited for implementation at flood control dams and their reservoirs. In such cases it is often required that the reservoir remain empty or almost empty for most of the year. Such requirements provide favourable conditions for dry excavation. An example of a dam where dry excavation has regularly been implemented to manage reservoir sedimentation is shown in Figure 9. The normal procedure at Cogswell Dam in California, US, is to excavate the sediment using scrapers and loading the earth material in trucks by means of backhoes. The trucks then remove the excavated sediment to a landfill outside of the reservoir, where it is stabilised.

Hydro-suction – This is a technique that employs dredging equipment with sufficient hydrostatic head over a dam to create suction at the upstream end of the discharge pipe (Figure 10). This suction is then used to remove the deposited sediment. The upstream end of the pipe is typically moved around with a barge to remove sediment throughout various parts of the reservoir. A limitation to this kind of sediment removal is that it is principally used in short reservoirs (not exceeding 3km). The concept is also limited to low elevations due to the decrease in atmospheric pressure with increasing elevation, which adversely affects the performance of siphons. Sangroula (2005), who extensively studied alternative sediment management techniques for Kulekhani reservoir in Nepal, concluded that hydro-suction may be a viable technique for removing sediment from that storage reservoir.

Drawdown flushing – This is a technique requiring complete drawdown of a reservoir to re-suspend deposited sediment and flush it downstream. Therefore, an important requisite for drawdown flushing is that low-level gates of adequate size should be present in the dam. The gates should be large enough to freely pass the flushing discharge through the dam without upstream damming.

Successful flushing of sediment is best accomplished if river-like flow conditions can be re-established in the reservoir upstream of the dam. In order to accomplish this it is necessary for: the reservoir to have a narrow valley with steep sides; the riverbed to have a fairly steep longitudinal slope; to maintain river discharge beyond a minimum quantity so that it can mobilise and transport sediment; low-level gates to be installed in the dam.

Such gates should be sized in a manner that allows drawdown of the water level in the reservoir so that the water can flow freely through the gates, in an unimpeded manner. This means that the water level must be drawn down low enough so that the free surface of the water is at or below the gate soffit. Any damming is undesirable.

An example where drawdown flushing has been successfully implemented for many years is Gebidim dam in Switzerland. This dam is 120m high and contains two 2m x 3m low level gates at the base of the dam. Flushing occurs on an annual basis.

Pressure flushing – This is a technique that is used to remove sediment directly upstream of an outlet. It is implemented by opening the outlet without drawing down the water surface elevation. This action results in water accelerating for a limited distance upstream of the opening, causing removal of sediment for a limited distance upstream of the outlet only (Figure 11).

Sediment augmentation

It is known when sediment deposits in reservoirs upstream of dams that the water released downstream is ‘sediment hungry’. What this means is that the water that is released downstream of a dam contains little or no sediment. This water has the capacity to entrain larger net amounts of sediment when it flows over the river bed and along its banks than what would have been the case if the dam was not there. It causes riverbed and bank degradation, and adverse effects to the river morphology, consequently damaging aquatic habitat. The sediment that deposits in the reservoir upstream of the dam can also contain nutrients. This means that the nutrient loads in the river downstream of the dams can be reduced, adversely affecting aquatic life.

One way of dealing with this problem is to augment the sediment supply downstream of the dam. The sediment used for augmentation can either be obtained from remote sources or from the reservoir upstream of the dam itself. The intent with augmenting sediment to the river is to improve river morphology, and thus aquatic habitat. At this point in time it is unknown how sediment augmentation might improve nutrient supply and its potential success is site dependent. The locations of sediment augmentation projects under implementation in Japan are shown in Figure 12. The Orange County Water District in California is currently promoting this concept for implementation in the Santa Ana River, downstream of Prado dam.

Paradigm shift

Continued reservoir sedimentation and the reduced rate of dam construction since the 1980s has resulted in a net decrease in global reservoir storage volumes. Such a reduction in storage volume led to decreased reliability of global water supply and decreased quality of life. Rectifying this situation requires an increased rate of construction of new dams and a paradigm shift in dam design and operating philosophy. One of the important elements of this new perspective is to design dams that will allow operators enough flexibility to regularly remove sediment from reservoirs. Adopting this approach means that dams and their reservoirs can be used for periods extending beyond their normal operating period; matching the characteristics of other civil infrastructure such as roads.

Dr. George Annandale is a Principal at Golder Associates. He has worked in the field of reservoir sedimentation management for three decades and can be contacted at george_annandale@golder.com