Greenhouse Gas Emissions from Reservoirs: Studying the Issue in Brazil

Although Brazilian utilities and research institutions have studied greenhouse gas emissions from hydroelectric reservoirs for more than ten years, they agree more work is needed to standardize methods for measuring and monitoring emissions and to more accurately model the presence of carbon throughout a watershed.

Increasingly, claims are made that reservoirs – especially in tropical and subtropical regions of the world – emit greenhouse gases, mainly methane and carbon dioxide.

In this article, representatives of two owners of hydroelectric projects and two research institutions in Brazil describe their studies of greenhouse gas emissions from hydro reservoirs and offer perspectives about additional work needed.

Representatives are:

  • Paulo Roberto Hall Brum de Barros, biologist, and André C. Prates Cimbleris, manager of the division of the environment, FURNAS Centrais Elétricas S.A., a major Brazilian utility.
  • Antonio Fonseca dos Santos, environment director of Brascan Energética S.A., the Brazilian branch of Brookfield Renewable Power.
  • Marco Aurélio dos Santos, professor, and Luiz Pinguelli Rosa, director of the Programa de Planejamento Enerjetico, Instituto Alberto Luiz Coimbra de Pos-Graducao e Pesquisa de Engenharia (COPPE), Federal University of Rio de Janeiro.
  • Alexandre Kemenes, PhD in Aquatic Sciences, researcher in the LBA program at the Instituto Nacional de Pesquisas da Amazà´nia (INPA/ CLIAMB).

HRW: Why did your organization initially become interested in studying this topic?

Paulo Brum and André Cimbleris: We became interested because of the increasing frequency, in the first few years of the twenty-first century, of far-reaching statements about the carbon footprint of hydroelectric reservoirs, mostly unsupported by research. FURNAS saw a gap that should be filled by the best available researchers, aiming at settling the controversy.

Antonio Fonesca: Brascan became interested in this topic in 2004 when we participated in an International Hydropower Association workshop on greenhouse gas emissions from reservoirs. The issue of emissions from small reservoirs was raised for the first time. Because Brascan is focused on the development of small hydroelectric power plants, we decided to study greenhouse gas emissions in a small reservoir.

Mario Aurélio and Luiz Pinguelli: The question of GHG emissions from energy production was discussed at a side event during Rio 92, the United Nations Conference on Environment and Development organized in June 1992 in Rio de Janeiro. At this event, Brazilian specialists on energy and environment discussed the possibility of GHG production by hydro reservoirs and decided more accurate information was needed.

Alexandre Kemenes: No one had conducted a complete study of GHG emissions in an Amazon hydroelectric system, including both upstream and downstream fluxes.

HRW: In your opinion, why is it important for the hydropower industry in Latin America to study this issue?

Measuring greenhouse gas emissions from reservoirs is a complicated process. More research and development is needed to develop standardized measurement and monitoring techniques.
Click here to enlarge image

Brum and Cimbleris: Through studies, we can prove the bias of older data that places excessive emphasis on atypical reservoirs, such as 250-mw Balbina and 216-mw Samuel in the Amazon. In the FURNAS carbon budget project, we studied a variety of reservoirs – big and small, isolated and in cascade, in preserved areas as well as close to urban centers. The fact that the reservoirs in our study are more typical is what allows us to conclude that the issue of greenhouse gas emissions from hydroelectric reservoirs is not a major issue in most cases.

Of course, there will always be very bad reservoirs from the viewpoint of GHG emissions, and planners everywhere should avoid building these kinds of reservoirs. We know by now what characterizes bad reservoirs (e.g., residence time, flooded biomass, low power density), and we should be able to prevent new “worst-case-scenarios” from happening.

Fonesca: We have to prove that hydroelectricity is a clean energy. The main characteristic that puts Latin America in the developing, instead of the developed, world is our poor infrastructure. This infrastructure must be upgraded and developed to be capable of bringing a better standard of living to our population.

Because more than 50 percent of Brazil’s hydro potential is yet untapped, hydro project construction must be one, if not the main, way to develop this infrastructure.

It is very important for the Latin American hydropower industry to have strong evidence that development and construction of power plants is not harmful to the planet and is a vital tool to diminish poverty.

Kemenes: In Brazil, the government is considering the possibility of developing more than 75 large hydroelectric projects in the Amazon. Natural wetlands currently occupy about 17 percent of the Amazon. With these new hydroelectric reservoirs, the total flooded area could grow to about 25 percent. It is important to consider the current functioning of the ecosystems that would be transformed into reservoirs. Studies at INPA/INPE show that the upland forests in the Amazon sequester carbon from the atmosphere on a net basis. Transforming these forests into reservoirs could alter the regional carbon cycle.

HRW: What are some of the major conclusions you’ve reached through your studies?

Brum and Cimbleris: We offer five major conclusions.

1) The budget approach is essential for a proper grasp of the processes going on in reservoirs. This approach involves taking into account the ways in which the system exchanged GHGs with the atmosphere before the reservoir was flooded. Older studies measured only the emissions of GHG from the reservoir surface or, more recently, from downstream de-gassing. But without the measurement of the inputs of carbon to the system, no conclusions can be drawn from surface measurements alone.

2) When you consider the total budgets, most reservoirs acted as sinks of carbon in the short run (our measurements covered one year in each reservoir). In other words, they received more carbon than they exported to the atmosphere and to downstream.

3) Smaller reservoirs are more efficient as carbon traps than the larger ones.

4) As for the GHG impact, in order to determine it, we should add the methane (CH4) emissions to the fraction of carbon dioxide (CO2) emissions which comes from the flooded biomass and organic carbon in the flooded (terrestrial) soil. The other CO2 emissions, arising from the respiration of aquatic organisms or from the decomposition of terrestrial detritus that flows into the reservoir (including domestic sewage), are not impacts of the reservoir. From this sum, we should deduct the amount of carbon that is stored in the sediment and which will be kept there for at least the life of the reservoir (usually more than 80 years). This “stored carbon” ranges from as little as 2 percent of the total carbon output to more than 25 percent, depending on the reservoirs.

5) When we assess the GHG impacts following the guidelines just described, all of FURNAS’s reservoirs have lower emissions than the cleanest European oil plant. The worst case – Manso, which was sampled only three years after the impoundment, and therefore in a time in which the contribution from the flooded biomass was still very significant – emitted about half as much carbon dioxide equivalents (CO2 eq) as the average oil plant from the United States (CO2 eq is a metric measure used to compare the emissions from various greenhouse gases based upon their global warming potential, GWP. CO2 eq for a gas is derived by multiplying the tons of the gas by the associated GWP.) We also observed a very good correlation between GHG emissions and the age of the reservoirs. The reservoirs older than 30 years had negligible emissions, and some of them had a net absorption of CO2eq.  

Fonesca: We learned that:

  • The pilot project showed strong evidence that the Salto Natal reservoir is not a methane source. Although the reservoir turned out to be a carbon dioxide source, we proved the reservoir is heterotrophic (i.e., the organic matter that causes the reservoir to be a carbon dioxide source is not from the reservoir but from the upstream hydrographic watershed);
  • Measuring GHG emissions from reservoirs presents many difficulties; and
  • More research and development is needed, including development of mathematical and physical carbon mass balance models; definition of model carbon variables, parameters and equations; and measurement and monitoring techniques.

Aurélio and Pinguelli: In general, the risk of greenhouse gas emissions can be reduced by:

  • Choosing sites for reservoirs that do not have low power densities (measured in watts per square meter). Power density is calculated by dividing the installed power at a hydroelectric project by the square meters of the reservoir. The executive board of the Clean Development Mechanism (CDM) of the Kyoto Protocol approved a consolidated baseline and monitoring methodology for the electricity sector. This document, “ACM0002 – Consolidated baseline methodology for grid-connected electricity generation from renewable sources,” defines 4 watts per square meter as the density that makes a hydropower project eligible to be considered a CDM candidate. In our opinion, this index is a good guideline to follow.
  • Clearing the forests from the reservoir area before flooding will not, alone, avoid emissions. Equally important is the management of land use in the upstream area of the watershed because there is a considerable amount of carbon that comes from the soils and flows into the reservoir water from the tributary rivers.

Kemenes: From my studies of the Balbina Reservoir in the Amazon, I offer the following conclusions:

  • Information on potential gas flux components in a hydroelectric system is especially important in tropical regions, like Amazà´nia.
  • Seasonal changes in wind and rainfall can have a strong effect on the vertical distribution and emission of GHGs in tropical reservoirs like Balbina. During the rainy season (December to June), wind mixing in the Balbina Reservoir was more frequent and deep GHG concentrations were reduced. When wind diminished in the dry season (July — November), deep mixing was reduced and methane and CO2 concentrations and emissions were higher.
  • Gas concentrations and diffusive emissions fell gradually until approximately 30 kilometers below the Balbina Dam, after which they remained relatively stable for the next 40 kilometers of the Uatum࣠River. All CO2 that was not released by ebullition at the turbine outflow and all CO2 produced by the oxidation of methane in the river channel below the dam were assumed to be released to the atmosphere by diffusion.

HRW: What are the biggest unknowns regarding greenhouse gas emissions from hydroelectric reservoirs?

Brum and Cimbleris: We still do not know as much as we would like about the three-dimensional structure and hydrodynamics of the reservoirs. We are dealing with huge water bodies, hundreds of square kilometers in area, with depths that can reach 100 meters. The distribution of our sampling stations can be quite inefficient if unknown currents take the gases produced in the sediment to specific spots (presumably uncovered by our stations). In the FURNAS carbon budget project, we did as well as we could with the resources available, but we need much more knowledge about this matter.

Also, measuring the emissions of a future impoundment, still in its terrestrial phase, is crucial. Without such a standard of comparison, we will never know whether the GHG emissions, even in the case of the young reservoirs, are a problem.

For example, it is possible that the moderately high emissions of the Manso Reservoir are smaller than the emissions prior to the impoundment, since there were many wetlands within the area before the reservoir was impounded.

Finally, we are still haunted by the lack of standardized methodologies. Even though we now have more than ten years of GHG measurements in reservoirs, further studies to cover this weakness are necessary.

Fonesca: The biggest needs include development of:

  • A unique mathematical or physical carbon mass balance model;
  • Standards for measurement and monitoring (today, depending on the measurements and monitoring techniques used, anyone can prove anything about GHG emissions);
  • Indicators to verify if a reservoir is a carbon source or a carbon sink; and
  • A technique to measure the natural GHG emission of a site before construction of a reservoir so we can compare the before and after situations.

Aurélio and Pinguelli: To accurately estimate GHG emissions from hydroelectric reservoirs, we need a better knowledge of the carbon cycle in the reservoirs, before and after flooding, and at various levels (reservoir level, watershed level, and after dam impoundment). We also need a comparison of various GHG measurements and analysis techniques, in terms of accuracy. Additionally, future studies should include emissions caused through downstream release of water from the turbines. Finally, we need to find a way to better account for CO2 emissions from the biomass existing prior flooding. One suggested approach is to separate emissions from the degradation of flooded organic matter and emissions from the decay of organic matter that comes from the watershed.

Kemenes: We are only beginning to understand all of the spatial and temporal variability in these emissions and develop the sampling and statistical methods necessary to represent them. We have also detected some major errors in current sampling protocols that must be dealt with if we are going to make further progress. Most studies until now have focused on measuring the emissions from hydroelectric systems, which, in itself, is difficult. However a more fundamental challenge is to understand all of the biological, physical, and chemical factors that produce those emissions. This information is critical if we are to develop predictive models of reservoir emissions that can be used to simulate the effects of impoundment.

HRW: Describe your organization’s current efforts and studies of GHG emissions from hydroelectric reservoirs, and what you hope to learn.

For five years, researchers from Brazilian utility FURNAS and the Universidade Federal de Juiz de Fora studied greenhouse gas emissions from a variety of reservoirs – big and small, isolated and in casacade, and in preserved areas as well as close to urban centers.
Click here to enlarge image

Brum and Cimbleris: Beginning in 2009, we will measure the emissions of two future impoundments: Simplà­cio and Batalha. This will be quite a challenge, since the mosaic of landscapes is more complex before the impoundment. We know that the temporal variability of the gas flows is quite high, and we intend to correlate it with environmental variables (e.g., winds and currents). We will upgrade our methodology, aggregating satellite data analysis. The satellite data will help to assess the land occupation and use around the reservoirs and how this has been changing with time. Hydrodynamic/water quality numerical modeling will be developed in order to get a better understanding of the stratification, currents, and water quality variability in the reservoirs.

Fonesca: A research and development effort is being initiated by the Brazilian Association of Independent Power Producers (APINE), in cooperation with the Federal University of Paranà¡ and Federal University of Rio de Janeiro — COPPE, to develop a carbon mass balance model. The model will be available for use in the reservoirs of the more than 40 APINE member companies. Together, these companies own 42,000 mw of installed capacity in Brazil, representing more than 40 percent of the country’s total installed capacity. Through this research effort, standards for GHG emission measurement and monitoring will be established, along with definitions of indicators to verify if a reservoir is a carbon source or a carbon sink.

Based on good results from its pilot project, Brascan is extending the project to all 28 of its reservoirs. We are expecting to find that our power plants are a clean source of energy.

Finally, Brascan is analyzing a proposal from the Federal University of Paranà¡ to transform the Salto Natal Reservoir into a “reservoir school,” where the university would perform needed research and studies.

Kemenes: I am studying the GHG emissions from four Amazon hydroelectric systems: 250-mw Balbina; 8,400-mw Tucuruà­, 60-mw Curuà¡-Una, and 216-mw Samuel. In these studies, I am measuring reservoir emissions, de-gassing at the turbine outflow, and emissions from the downstream river channel.

I am also testing a new sampling device, the Kemenes Sampler, designed to minimize de-gassing losses while collecting deep water samples. Both a Kemenes and a traditional sampler (Ruttner bottle) were used to collect water samples at multiple depths in Balbina, Tucuruà­, Curuà¡-Una, and Samuel reservoirs. The Kemenes sampler collected significantly more methane from all depths compared to the Ruttner bottle.

Finally, I am collecting data from several Amazonian hydroelectric projects to evaluate the potential of using their downstream biogas emissions to generate electric energy. This concept would increase the power generating capacity of reservoirs while reducing their emissions of methane. Stored gas could be used to meet daily and seasonal energy demands, resulting in a more efficient use of hydroelectric resources. s

Messrs. Brum and Cimbleris may be reached at FURNAS Centrais Electricas S.A., Rua Real Grandeza, 219, Sala 801 C, Rio de Janiero 22281-900 Brazil; (55) 21-25283259 or (55) 21-25285436; E-mail: or Mr. Fonseca may be reached at Brascan Energética S.A., Rua Padre Anchieta, n. 1856, Curitiba, Parana 80730-000 Brazil; (55) 41-33315594; E-mail: Prof. Aurelio dos Santos and Rosa may be reached at Federal University of Rio de Janeiro, Centro de Tecnologia, Bloco C, Sala 211, Rio de Janeiro 21941-972 Brazil; (55) 21-25628763 or (55) 21-25627024; E-mail: or Dr. Kemenes may be reached at the Instituto Nacional de Pesquisas da Amazà´nia, Av. André Araújo 2936, Aleixo, Manaus, AM 69083-000, Brazil; (55) 92-36433636; E-mail: or


Previous articleNiger seeks civil work, hydro-mechanical equipment for 125-MW Kandadji
Next articleU.S. solicits comments on marine, hydrokinetic report

No posts to display