Reservoir Emissions: Excerpt from Silenced Rivers

Patrick McCully
Wednesday, October 17, 2001

Excerpted from the new introduction of the updated 2001 edition of:
Silenced Rivers: The Ecology and Politics of Large Dams

It’s baloney and it’s much overblown . . . Methane is produced quite substantially
in the rain forest and no one suggests cutting down the rain forest.

Karolyn Wolf, spokeswoman for the US National Hydropower Association
responding to International Rivers press release on
greenhouse gas emissions from reservoirs, 1995

It seems difficult for many people to accept that the seemingly serene surface of a reservoir could be belching forth as much heat–trapping gas as a smokestack. Even the UN’s climate science panels have ignored the phenomenon. Measurements of methane (CH4) and carbon dioxide (CO2) emissions due to rotting organic matter in reservoirs date back only to 1993 and only around 30 reservoirs, mostly in Brazil and Canada, have been studied for their emissions.

The small cadre of scientists researching reservoir greenhouse gas emissions is deeply split. One group, which is largely funded by Hydro–Quebec and Brazilian hydropower interests asserts that reservoir emissions are far lower than emissions from equivalent fossil fuel plants. Other scientists who are affiliated with various universities and research institutes, mostly in Canada, Brazil and France, warn that reservoir emissions are much more significant than commonly assumed and that in the tropics they can exceed emissions from fossil fuel–fired power plants. In order to try and reach some consensus, the World Commission on Dams (WCD) brought together 17 of the leading researchers on reservoir emissions for a workshop hosted by Hydro–Quebec in Montreal. The participants agreed on a statement that provides an important summary of the current state of knowledge on the subject. The following paragraphs in italics are taken from the Montreal statement:

"Greenhouse gases are emitted for decades from all dam reservoirs in the boreal and tropic regions for which measurements have been made. Emissions result not only from vegetation and soils flooded by the reservoir, but also from the decomposition of aquatic plants and algae, and from organic matter washed into the reservoir from upstream. Reservoir emissions should be considered in assessments of individual dams and in global inventories of the sources and sinks for greenhouse gases. (1)

Until the last few years researchers had presumed that emissions from reservoirs would spike immediately after reservoir filling and then quite quickly decline to insignificant levels as the flooded biomass decomposed. However recent research has shown that while there may be an initial pulse of gases (especially for tropical reservoirs) emissions tend to decline over time only very slowly, if at all. The reason is mainly that methane and carbon dioxide are continuously produced from rotting plants and algae that have grown in the reservoir or been washed in from its catchment.

The realisation that reservoirs may be a significant source of greenhouse gases has major implications for national and global inventories of greenhouse gas emissions, which themselves have major implications in decisions on the most effective measures for reducing global warming. A paper published in 2000 by a team of Canadian researchers estimates that reservoir emissions contribute seven per cent of the total global warming impact of other known human–related releases of carbon dioxide and methane. (2) (This paper uses an estimate for the global surface area of reservoirs of all sizes is 1.5 million km2, far higher than the estimate of 400,000 km2 in the 1996 edition of Silenced Rivers).

– Methane and carbon dioxide are emitted from water passing through the turbines, over the spillway and downstream of the dam. These emissions may be significant.

Until recently researchers had only accounted for emissions from the reservoir itself. These are emitted by diffusion into the atmosphere from the reservoir surface and from bubbles rising from shallow reservoir zones. It is now realised that huge amounts of gases can also be emitted when water is discharged from the reservoir. Methane emissions from the turbines and spillway at Tucuruí Dam in the Brazilian Amazon are estimated to be up to eight times higher than those released by bubbling and diffusion from the reservoir. (3)

"Hydropower emissions should be evaluated on a net basis over the catchment in question – it is not the gross emissions from a reservoir which is most relevant, but the difference in emissions from the catchment before and after a dam is built.

Ecosystems are a complex and improperly understood mosaic of both sources and sinks of carbon dioxide and methane. Most forests act as sinks of both gases while natural lakes are sources of both. Northern peatlands are carbon dioxide sinks but important methane sources. (4) Evaluating net rather than gross emissions can therefore either increase or decrease the estimated global warming contribution of the reservoir depending on the characteristics of the area flooded.

the multiplier commonly used to convert methane emissions to "equivalent CO2" can significantly underestimate the climate change impact of reservoirs over the first several decades. Other time–dependent conversion methods such as that developed by Stuart Gaffin should be considered.

Methane is known to be a much more powerful greenhouse gas than carbon dioxide. But calculating exactly how much more a molecule of methane contributes to climate change than a molecule of carbon dioxide is fraught with difficulties. While methane is much shorter–lived in the atmosphere than carbon dioxide, each methane molecule is much more efficient at trapping heat. The methane multiplier commonly used is known as the 100–year Global Warming Potential (GWP) and represents the impact after 100 years of a one–time pulse into the atmosphere of a ton of methane compared to one of CO2. The UN’s Intergovernmental Panel on Climate Change currently estimates methane’s 100–year GWP as 21, meaning that a ton of methane in the atmosphere causes 21 times more warming than a ton of carbon dioxide.

If methane reservoir emissions were indeed a one–time event resulting from rotting biomass submerged when the reservoir was impounded, this "pulse" approach might be appropriate. However, because the emissions are continuous a different methodology is required. Atmospheric chemist Stuart Gaffin of the US group Environmental Defense has developed a model for assessing the climate change impact of continuous emissions of methane compared to CO2. According to Gaffin’s model, after 100 years the cumulative global warming effect of a constant methane emitter is some 39.4 times greater than that of a constant emitter of an equivalent quantity of CO2. (5) Methane, especially in the tropics, is a significant part of the emissions from a reservoir (up to three quarters of the total greenhouse gas impact in the case of Tucuruí). Using a larger methane multiplier can therefore greatly increase estimates of the total global warming impact of a reservoir.

the range of factors influencing greenhouse gas emissions include the reservoir’s depth, shape and size, the climate of its surrounding region, its operating regime and water residence time, the size and nature of the watershed, and the nature of human activities around the reservoir and upstream.

The single most important determinate of reservoir emissions is climate: emissions from tropical reservoirs are far higher than those from reservoirs in boreal zones. Shallow reservoirs also likely to have much higher emissions than deep ones. The contribution of a reservoir to climate change compared with other electricity sources will also depend on the amount of power it generates. A dam in the Amazon basin with a low installed generating capacity and a large shallow reservoir may have emissions several hundred times higher per kilowatt–hour generated than a dam in Canada with a small deep reservoir and high generating capacity.

According to Éric Duchemin of the University of Quebec at Montreal, mean net emissions from boreal reservoirs are equivalent to 20 to 60 grams of CO2 per kilowatt–hour generated. Net emissions from tropical reservoirs, according to Duchemin, range from 200 to 3,000 g/kWh. (6) By comparison, natural gas–fired combined cycle plants – currently the technology of choice for power generators in much of the world – emit around 430–635 g CO2–equivalent/kWh (including the warming contribution of methane leaks from gas extraction and transmission). (7)

These figures above for reservoir emissions are calculated using the methane GWP of 21. The results of calculating gross emissions from reservoirs and a gas combined cycle plant using Gaffin’s methane multiplier are given in Table 2 (methane is a very small part of the total emissions for most generating technologies so changing the methane multiplier would have little impact on their results). These estimates suggest that the warming impact of a modern gas plant is between five and eight times more than that of a high–emitting boreal reservoir, but that a tropical reservoir can have a warming impact up to 66 times greater than that of a gas plant. (8)

Table 2: Global warming impact of various electricity generation options. Hydro and combined–cycle natural gas emissions calculated using methane multiplier of 39.4. Hydro installation

Hydro installation Electricity generation Flooded area Emissions Ref.
  (TWh/yr)* (km2) (gCO2eq./kWh)  
Churchill Falls
35 6705 ≤90 1
Complexe La Grande 82 13000 ≤75 1
Balbina 1 3150 30250 2
Curua-Una 0.1 72 5700 1
Tucuruí 15.7 2250 3280 3
Non-hydro installation
Lignite (brown coal)     1150–1270 4
Coal (modern plant)     790–1200 4
Heavy oil     690–730 4
Diesel     555–880 4
Combined-cycle natural gas (550 MW)     460–760 5
Natural gas cogeneration
    300 6
Large fuel cell (natural gas–powered)     290–520 4
Photovoltaics     30–210 4
Biomass energy     17–120 4
Windpower     7–40 4
Nuclear     2–60 4

* estimated based on a capacity factor of 60% except Balbina and Tucuruí which are based on actual generation. At least for tropical dams capacity factor is likely to be closer to 50% than 60%.

1. P. Raphals (2001) Restructured Rivers: Hydropower in the Era of Competitive Markets, Helios Centre/International Rivers, Montreal/Berkeley, 2001.
2. Balbina emissions recalculated from data in P.M. Fearnside, "Hydroelectric Dams in the Brazilian Amazon as Sources of "Greenhouse Gases", Environmental Conservation 22(1) 1995.
3. Tucuruí emissions recalculated from P.M. Fearnside, "Greenhouse gas emissions from a hydroelectric reservoir (Brazil’s Tucuruí Dam) and the energy policy implications", Water, Air and Soil Pollution (in press). Tucuruí energy production (average 1984–1998) from WCD Case Study on Tucuruí.
4. IEA Implementing Agreement For Hydropower Technologies, Hydropower And The Environment: Present Context And Guidelines For Future Action. Main Report, May 2000, p.126. The figures here for wind and photovoltaics do not include high emission "outlier" estimates inconsistent with the other available estimates.
5. Recalculated from P. L. Spath and M. K. Mann, "Life Cycle Assessment of a Natural Gas Combined–Cycle Power Generation System", NREL, Colorado, 2000.
6. M. Rizau et al., "Clean Electricity Supply With Low Climate Impact and No Nuclear Power," Greenpeace, Hamburg, 1998.

The section of the WCD's final report, Dams and Development, on reservoir emissions notes that all reservoirs that have been studied emit greenhouse gases and that "in some circumstances the gross emissions can be considerable and possibly greater than the thermal alternatives". But the report does not say, as it should, that the evidence points to the climate impact of tropical hydropower often being far worse than the thermal alternative. The "Guidelines for Good Practice" in the WCD report recommend that estimates of net reservoir emissions be included in dam feasibility studies and notes the need for more research into reservoir emissions, especially in temperate and semi–arid regions.

A major report on hydropower and the environment was released in May 2000 by a hydro advocacy group in which Hydro–Québec plays a leading role known as the International Energy Agency Hydropower Agreement. (9) The report claims that hydro projects should receive subsidized loans from aid agencies "as a pay–back of the global community for the protection of nature and the world climate." It also states that there "is no doubt that Clean Development Mechanisms [sic] (CDM) will provide a stimulus for hydro." The CDM is the North–South emissions trading mechanism proposed under the Kyoto Protocol. This report claims that hydropower emissions are 2–48 g CO2/kWh – a gross underestimate which is several orders of magnitude less than emissions from tropical reservoirs.

While Hydro–Quebec and their colleagues are playing down the warming impacts of boreal dams, their impacts do appear much lower than those from fossil fuel plants. But the Clean Development Mechanism does not apply to projects in boreal countries. It does apply to countries in the tropics where the hydro industry sees the greatest opportunity to expand, and where reservoir emissions can be massive. Regardless of the emissions from large reservoirs, their multiple social and environmental impacts should in any case rule them out from subsidies that would better go to energy efficiency, sustainable energy sources like solar and wind, and other climate–friendly measures such as forest conservation and regeneration.

A Changing Climate for Dams

Just as dams are impacting the workings of the global climate, so changes in the global climate are impacting the workings of dams. Climate change is rendering obsolete one of the key assumptions used in dam planning and design – that the hydrological past is a reliable guide to the hydrological future. The 2001 assessment of the Intergovernmental Panel on Climate Change predicts that the planet will warm by 1.4–5.8 degrees Celsius by the end of this century. For every degree Celsius warming of the planet, global precipitation is likely to increase by 2–4%. The resulting changes in regional weather patterns will vary widely but there is widespread agreement among researchers that in many parts of the world the frequency and severity of both floods and droughts will increase.

Most dam spillways are designed to pass the estimated maximum flood which could occur in a catchment. But these maximum flood estimates do not allow for a changing climate. If spillway capacity is exceeded, water may flow over the top of the dam – "overtopping" is the single most important reason for dam failures. Dams and Development expresses concern over the adequacy of existing spillways given the likelihood of increased flood intensities, and the ability of flood control dams to perform as designed.

"Reservoir reliability" – the ability of a dam to meet its design objectives – will be affected both by changing patterns of river inflow and because hotter temperatures will increase reservoir evaporation. Hydropower generation, for example, could be seriously reduced by increased droughts and evaporation, although it would benefit from increased rainfall. Dams and Development recommends that planning and monitoring of dams should take account of the impact of potential climate changes on both dam safety and performance.

The impacts of climate change on water resources will vary widely between geographical regions and over time and are extremely difficult to predict – and will remain so for the foreseeable future. However, this uncertainty is no reason for ignoring climate change, which has largely been the response of dam operators until now. The WCD’s thematic review on climate change and dams states that the best way for water planners to deal with uncertainty will be to reduce vulnerability through reducing water demand, rather than attempting to increase supply. (10)

1 ’Dam Reservoirs and Greenhouse Gases: Report on the Workshop Held on February 24–25, 2000, Hydro–Québec, Montreal (Final Minutes),’ WCD, 2000. Many thanks to Philip Raphals of the Helios Centre in Montreal for helping me find my way through the complexities of this issue.

2 V.L. St. Louis et al., "Reservoir Surfaces as Sources of Greenhouse Gases to the Atmosphere: A Global Estimate", Bioscience 50(9), September 2000..

3 P.M. Fearnside, "Greenhouse gas emissions from a hydroelectric reservoir (Brazil’s Tucuruí Dam) and the energy policy implications", Water, Air and Soil Pollution (in press).

4 See St. Louis et al., "Reservoir Surfaces", p.769.

5 See P. Raphals, Restructured Rivers: Hydropower in the Era of Competitive Markets, Helios Centre/International Rivers, Montreal/Berkeley, 2001, pp.19–21.

6 E. Duchemin et al., "Hydroelectricity and greenhouse gas emissions: Emission evaluation and identification of the biogeochemical processes responsible for their production," Ph.D. dissertation, UQAM, Montreal, 1999, cited in Raphals "Restructured Rivers", p.20.

7 P. L. Spath and M. K. Mann (2000), "Life Cycle Assessment of a Natural Gas Combined–Cycle Power Generation System", NREL, Colorado.

8 Apart from the Tucuruí calculation, none of the hydro estimates in Table 2 allow for emissions from turbines and spillways.

9 IEA Implementing Agreement, ’Hydropower And The Environment: Present Context And Guidelines For Future Action: Main Report,’ May 2000, pp.41–2.

10 N. Arnell and M. Hulme, "Dams and global change: implications of climate change for large dams and their management", WCD Thematic Review, 1999