Strategic planning of hydropower portfolios minimises dam sediment trapping and maximises economic benefits in large river basins
Rafael Schmitt is a winner of the 2017 IHA Young Researcher of the Year award. In this article from his award submission, he writes about his work focusing on reducing conflicts between hydropower and environmental objectives through basin-scale planning of dam portfolios in the world’s large river basins.
Sediment trapping: a key environmental impact of hydropower
Hydropower provides renewable energy at competitive production costs, making it a key component to match growing global energy demands, especially in emerging economies.
Producing hydropower requires the construction of dams and the impoundment of reservoirs that interrupt the continuity of sediment transport in a river. This dam-induced discontinuity of sediment transport has two key consequences.
First, sediment accumulates within the reservoir, endangering dam statics, damaging hydroelectrical and mechanical equipment, and reducing reservoir storage. Removing accumulated sediment from dams is often financially or technically unfeasible, leading to an annual storage loss of 1 per cent of the global reservoir storage capacity – storage that would be urgently needed in the face of a changing climate and increasing water demand.
Second, sediment trapping in dams deprives rivers, floodplains and deltas of their sediment load, resulting in changes in river morphology, which endangers infrastructures, habitats and many related eco-system services in the downstream river system.
The objective of my research is to reduce conflicts between hydropower and sediment trapping by making strategic infrastructure decisions. Specifically, my research provides evidence for how new numerical tools allow the identification of dam portfolios that minimise sediment trapping while maximising hydropower production.
How can basin-scale planning reduce sediment trapping?
What is fascinating about rivers is their spatial heterogeneity in terms of form and functioning. Rivers in a similar geographic setting, and even tributaries within the same catchment, can have widely different sediment loads in function of small differences in geology, climate or land use.
Thus, dams with similar technical layout and production capacity built at various locations in the river network can trap sediment at very variable rates.
In contrast to the spatial heterogeneity of river catchments, sediment trapping is commonly considered an engineering problem on the scale of single dam sites. Hence, when dam development begins there is limited understanding for how the final hydropower cascade will impact sediment transport in a river system, which dams in the final dam cascade will have the highest risk of storage loss, and which dam portfolio should be developed to maximise hydropower production and minimise sediment trapping.
Selecting such optimal dam portfolios is limited by two factors. First, data on sediment transport are notoriously scarce for most rivers. Second, even a small number of dam sites can be combined into many different hydropower portfolios, each with a specific performance in terms of hydropower production and sediment trapping.
To overcome these limitations, we develop new numerical models of network sediment transport and sediment trapping in dams, and integrate them with approaches from operations research that automatically identify optimal trade-offs between two (or multiple) objectives.
Finding optimal portfolios in large river networks
Recently, we introduced the CASCADE (CAtchment Sediment Connectivity And DElivery) framework. CASCADE is a computationally effective numerical model for network sediment transport and reservoir sediment trapping. CASCADE relies mostly on remotely-sensed globally available data-sets as input data. The numerical effectiveness of CASCADE is tailored to approach large-scale decision-making, and to screen a huge number of dam portfolios, even in data-scarce river basins.
Selecting dam portfolios with minimal sediment trapping is hindered by the limited availability of sediment transport data. In many basins, only few point observations of sediment transport are available. However, selecting optimal dam portfolios would require information on the spatial distribution of sediment transport in the river network. Recently, we showed for a major tributary of the Mekong River (Se Kong, Se San, Sre Pok Rivers or, in abbreviation, 3S), how CASCADE can be used to upscale available point sediment observations to network-scale estimates of sediment transport via an inverse Monte Carlo Approach (see Figure 1).
Figure 1: Inverse stochastic modeling of sediment transfers in the Se Kong, Se San, and Sre Pok basin in the Lover Mekong (c) using the CASCADE model1 reveals a clear spatial pattern of sediment transport in the three rivers based on single observation of sediment flux and grain size distribution at the basin outlet3 (a). The stochastic modeling approach, which is enabled by the computational efficiency of the CASCADE model, reduces uncertainty in modelled sediment fluxes to very low levels (c).
Sediment trapping in the 3S hydropower cascade is of concern, because of the key role of the 3S basin as sediment source for the lower Mekong and the Mekong Delta. Results of our model-based upscaling indicate that sediment is delivered at very variable rates from the three tributaries (Figure 1a), information that was not apparent from available point sediment observations.
In the 3S rivers, 42 dams with a total production of 30,000 GWh/yr are planned, built or under construction. We created 17,000 different hydropower portfolios (i.e. combinations of dam sites) and quantified the performance of each portfolio using the CASCADE model. Only 60 dam portfolios result in an optimal trade-off between sediment trapping and hydropower production, i.e. minimise sediment trapping for a given level of hydropower production (Figure 2a, dots).
At the current point of development, dams in the 3S will trap 90 per cent of the transported sediment, while exploiting only 30 per cent of the total production potential (Figure 2a, squares). Most of the trapping is in few dam sites in the Sre Pok River (Figure 2b, left panel) that add comparatively little production capacity compared to their sediment trapping. As opposed to the current development, there would have been portfolios of upstream dams (Figure 2b, right panel) that exploit 70 per cent of the full hydropower potential while trapping around 20 per cent of the total sediment flux.
Figure 2: Developing hydropower portfolios with an optimal trade-off between sediment trapping and hydropower production for the 3S basin (see Figure 1 for location). (a). The Eutopia Point (green star in panel a) presents the (unachievable) situation for which both objectives (sediment trapping and hydropower production) are optimised. Optimal portfolios (grey points in a) present achievable optimal trade-off between sediment trapping and hydropower production, i.e. the minimal level of sediment trapping for a given level of hydropower production. The current, site-by-site planning which focused on developing dams in the Sre Pok and upper Se San sub-basins (b, left panel) leads to a sub-optimal level of hydropower production regarding its high level of sediment trapping when compared to an optimal dam portfolio (see b, left panel, for the optimal configuration of sites, and a for performance).
The 3S exemplifies that a strategic analysis of basin scales has a great potential to identify dam portfolios that reduce economic costs and ecologic consequences of dam sediment trapping. Site-by-site planning without network-scale coordination will, in contrast, hardly result in an optimal trade-off between environmental and economic objectives.
New screening models, such as CASCADE, and approaches from operations research allow to identify optimal portfolios of dams that can be studied and designed in more detail to take final development decisions.
To conclude, making strategic decisions for planning optimal hydropower portfolios requires comprehensive assessments of a river basin’s natural functioning and a portfolio perspective on all potential hydropower projects. Challenging future research questions are posed, for example, in transnational river basins where the potential benefits of hydropower and its environmental impacts might be unequally distributed between nations.
The author acknowledges the contribution of Prof. Dr. A. Castelletti, Dr. S. Bizzi, Prof. Dr. G.M. Kondolf Dr. Z. Rubin, and many others at Politecnico di Milan and UC Berkeley. The author was supported by doctoral fellowship from the German National Academic Foundation.