Summary of the Issues

All forms of waterbodies, including freshwater reservoirs, lakes, rivers, marshes, and wetlands, can emit carbon dioxide (CO2) and methane gases (CH4) due to the microbial processes in water and sediment. Freshwater networks—such as rivers, lakes, and wetlands—generate and transport organic matter that contains carbon from terrestrial ecosystems to the ocean; however, not all of that terrestrially derived carbon reaches the oceans. Organic carbon is also stored as sediments in lakes, wetlands, or behind manmade reservoirs.  At multiple points within these freshwater networks, organic carbon can also be converted and emitted to the atmosphere as CO2 and CH4

While it is important to improve our understanding of emissions from all waterbodies, there are initial indications that some manmade reservoirs may be more likely than natural water bodies to store and release gases (especially methane, which is of greatest interest and concern). In the United States, reservoirs are typically associated with dams and impoundments that provide a range of important services like flood control, navigation, irrigation, water supply, and recreation. A small percentage of U.S. dams and impoundments are equipped with hydroelectric generation (3%, or 2,200 facilities), and a subset of those rely on reservoirs to store water for greater flexibility in water usage and dispatchable power.

To date, emissions have been measured in several hundred water bodies globally, and while many of the studies have shown extremely low rates of emissions, others have found high levels of emissions at specific times or locations. It is important to better understand to what extent, and in what cases, isolated measurements are representative of typical conditions. However, even with the significant monitoring efforts and studies that have taken place to-date, the science is still extremely uncertain in its ability to predict the extent to which these gases are emitted by different reservoirs and other waterbodies—especially across varying climate conditions and reservoir types. For instance, some studies have shown that gross methane emissions per unit of surface area from marshes can also be high; in some cases up to an order of magnitude greater than those measured from some reservoirs.

Even for a single reservoir, estimates of CH4 emissions are challenged by highly variable temporal and spatial conditions. Many existing studies have also only been able to monitor for short periods of time and/or with limited spatial distribution, both within and among reservoirs.  Another reason for the difficulty in accurately measuring CH4 emissions—potentially contributing to varied results—is due to the multiple pathways by which CH4 is emitted. Capturing CH4 emissions via ebullition (or bubbling), which is a highly temporally and spatially variable process, is especially challenging. While statistical guidelines for conducting regional or global assessments from limited data now exist, they have not been widely adopted. Given this, previous measurements of CH4 emissions from reservoirs around the world and associated projections and extrapolations have shown wide ranges of results.

Even with significant uncertainty, researchers are continually improving their understanding of the patterns that may influence this variability. Generally, CH4 has been shown to be a smaller fraction of gases emitted from waterbodies in northern latitudes. In temperate regions, including most of the United States, CH4 emissions have generally been shown to be low, but higher outlying measurements have varied by two orders of magnitude across reservoirs with different characteristics such as size and the duration of hypoxic (low dissolved oxygen) conditions. For example, reservoirs that experienced warmer temperatures and received larger inputs of organic carbon tended to have higher CH4 emissions.   

Scientists are still working to improve and develop new methods for the efficient monitoring of emissions over broad temporal and spatial scales. Researchers are also exploring newer remote sensing technologies to see whether they can provide better and more efficient insight into the spatial and temporal variation of CH4 emissions.

Additional work is required to:

  • Better understand the physical processes driving emissions
  • Determine the extent to which emissions measured in manmade reservoirs represent ‘new’ sources that would not otherwise have been emitted elsewhere
  • Improve capabilities to predict and model the potential emissions from reservoirs
  • Reduce uncertainty in reservoir-scaled estimates of CH4 emissions.
Illustration showing carbon dioxide and methane pathways in a freshwater reservoir.

The light tan represents soils and organic matter present prior to reservoir construction, while the darker brown represents organic matter and sediments being added to the reservoir over time. 

How Is DOE Helping to Address Uncertainty?

In 2009 and 2010, DOE initiated research to test and improve measurement methodologies as well as provide more data relevant to understanding emissions from U.S. reservoirs. Oak Ridge National Laboratory (ORNL) and Pacific Northwest National Laboratory conducted field work to measure emissions and published their findings in 2016 and 2017 respectively. Researchers initially expected higher emissions in southern reservoirs than in northern reservoirs. However, emissions from a handful of reservoirs sampled across the Southeast were extremely low, whereas higher emissions were measured from two reservoirs in the Pacific Northwest. Relatively higher emissions in the two Pacific Northwest reservoirs were still moderate compared to many other measurements from around the world.

DOE's Hydropower Vision Report, which was published in 2016 with significant input and review from dozens of different organizations and stakeholders, also discussed the uncertainty in the science surrounding this issue. The report highlighted the ranges of results in measurements available at the time and noted additional areas of research needed; it concluded that:

“New deployments of low-impact hydropower on undeveloped streams that do not lead to large inundation areas are also likely to have low biogenic GHG emission impacts. Powering of existing NPDs (non-powered dams) is also unlikely to lead to changes in biogenic GHG emissions, since the dam has already been built.”

Currently, DOE and ORNL are consulting with the Environmental Defense Fund to review CH4 emission data from powered and non-powered reservoirs in the United States to better characterize the ranges of uncertainty and identify key processes contributing to this uncertainty. This research will also attempt to compare ranges in possible emissions to other major sources of CH4 emissions from across the U.S. economy.

DOE and ORNL are also partnering with the International Hydropower Association (IHA) to investigate the suitability of IHA’s new Greenhouse Gas (GHG) Reservoir (G-res) tool for applications at existing U.S. reservoirs. IHA, in partnership with the UNESCO Chair in Global Environmental Change, launched G-res in March 2017 to estimate potential GHG emissions from newly constructed reservoirs. G-res could prove to be a cost-effective and practical way for estimating nationwide emissions from existing reservoirs by drawing on readily available and remotely sensed data. To date, published studies of G-res applications have largely focused on new reservoirs in tropical climates. Additional research is needed to evaluate whether estimates from the tool can be refined by using more recent and higher-resolution information and to evaluate how the tool performs when applied to reservoirs across U.S. climates.

Additionally, DOE has engaged in international efforts to establish methodological guidelines and best practices for assessing potential reservoir emissions through the International Energy Agency’s Annex XII project: Managing the Carbon Balance of Freshwater Reservoirs.

In collaboration with relevant international and U.S. agencies and non-governmental organizations, DOE is seeking to reduce uncertainty in estimating and sampling CH4 emissions from waterbodies and contribute scientific advances in measurement, modeling, and assessment methods needed to understand the issue.