Although the uncertainties are high, the potential emissions should be considered against emissions from other alternatives, the researchers say.
In Brazil, where the team focused the analysis, hydropower is the main source of electricity. It is often viewed as a low-carbon form of energy and, particularly where water is abundant, as cheap and reliable. It does not require the importation of fossil fuels, and is backed by a well-established industry.
"The government develops the electricity planning and organizes the auctions, and the current president is really pushing to build large dams," said Felipe Faria of Carnegie Mellon University, US.
But hydroelectric reservoirs are not as clean as they were once thought to be. Their construction requires extensive deforestation, with the cleared biomass being burned or dumped – either way, it generates carbon dioxide. What’s more, any biomass left in the reservoir area is likely to decompose, generating methane.
Previously, there have been no predictive emission models for hydroelectric reservoirs so no-one has known quite how bad these emissions are. That’s why Faria, together with colleagues from Carnegie Mellon and other institutions in the US and Brazil, devised two different approaches to estimate the future carbon balance of 18 new or forthcoming reservoirs in the Amazon.
In the first approach, the researchers compiled data on greenhouse gases measured just above the water surface. Based on the distribution of values and the physical features of the reservoir, the researchers could simulate the future emissions of a new reservoir.
This is the "top down" approach. In their second, "bottom up" approach, Faria and colleagues simulated how much carbon was available in each reservoir for producing greenhouse gases.
The results implied that the future emissions of some reservoirs were worse than previously thought, and possibly comparable to fossil fuel plants. The team estimated the average total emissions of carbon dioxide over 100 years for the 18 Amazonian reservoirs to be between 81 and 310 teragrams (Tg), compared with 35 Tg a year for the natural gas energy system for the entire US.
In general, Faria and colleagues believe that emissions from reservoirs could be reduced if the reservoirs have high energy densities – in other words, generate a lot of energy over a small area – and have their land cleared of vegetation beforehand. Although the latter generates carbon dioxide, it avoids the generation via wet decomposition of methane, a more potent greenhouse gas.
The researchers allowed for large uncertainties, which meant that their figures in some cases spanned two orders of magnitude. "The simulation results show that, even though the uncertainties are significant, for most of the cases, they are consistent enough to allow comparisons to other sources of electricity," said Faria.
The team published the analysis in Environmental Research Letters (ERL), where it is accompanied by a perspective article by Philip Fearnside.
Related links
- Estimating greenhouse gas emissions from future Amazonian hydroelectric reservoirs Felipe A M de Faria et al 2015 Environ. Res. Lett. 10 124019
- Greenhouse gas emissions from Brazil's Amazonian hydroelectric dams Philip M Fearnside 2016 Environ. Res. Lett. 11 011002
- Carnegie Mellon Electricity Industry Center