Hydropower serves an important role in the electrical grid. Not only is hydropower a widely available, reliable, and renewable resource, it is also one of few types of renewable power that can provide the energy storage and rapid-response generation capability needed to stabilize the grid during fluctuations in energy demand. However, growing concerns about greenhouse gas (GHG) emissions associated with hydropower reservoirs have motivated several studies into hydropower’s carbon footprint.  

As the United States continues to transition to a clean energy future, it will become increasingly important to accurately estimate GHGs, including carbon dioxide and methane, emitted from different energy sources, including hydropower. However, recent data has shown that current methods used to quantify the carbon footprint of hydropower facilities may be insufficient 

To further understand the state of hydropower GHG research, and to ensure accurate assessment of hydropower as part of a strong and clean electric grid, the U.S. Department of Energy’s Water Power Technologies Office (WPTO) commissioned Oak Ridge National Laboratory (ORNL) to review existing methods of measuring GHG emissions from hydropower facilities

ORNL staff identified several limitations with current accounting, as well as suggestions for improving future assessments. Addressing and correcting these limitations is crucial to ensuring a stronger and clean electric grid. 

Carbon Sequestration and Emissions from Reservoirs

All inland waters naturally produce some GHG emissions. However, when human-made reservoirs are constructed for hydropower facilities, they change the way carbon is emitted and stored in the river systems, sequestering some carbon, but also releasing some embedded carbon in the form of methane (CH4) emissions. When a reservoir is created and filled with water, submerged organic matter such as vegetation decomposes and breaks down, releasing carbon dioxide (CO2) and methane. These gases can then reach the atmosphere in multiple ways, including diffusion, ebullition (bubbles), transmission via vegetation, and degassing when water passes through a pump house or turbine. Both CO2 and methane contribute to atmospheric warming and climate change; however, methane has a much stronger immediate effect and lasts in the atmosphere about a decade, whereas CO2 can last for thousands of years and build up over time.  

Because GHG emission can occur naturally, it is important to distinguish what fraction of GHG releases can be attributed to hydropower reservoirs.  

Reservoirs are extremely effective at storing carbon. On average, human-made reservoirs are six times more effective at carbon burial than lakes , and in temperate climates, reservoirs can bury more carbon than they emit. However, conditions that promote carbon burial are similar to those that produce methane emissions.  

This balance between carbon storage and emissions is greatly influenced by several factors, including water temperature, sedimentation rate, organic matter inputs, and the shape, size, and depth of a reservoir. Nevertheless, only the fraction of methane generated by carbon decomposing in a reservoir represents an actual increase in net GHG emissions relative to potential carbon emissions that would have occurred without the reservoir. 

Challenges in Current Accounting

  1. Nearly all research into GHG emissions from reservoirs has focused on individual reservoirs without considering their place in the broader watershed. This approach assigns an average emission rate per square kilometer of water surface area. However, this approach is inaccurate when applied to open system freshwater networks. Reservoirs represent an aggregation of everything occurring in the watershed, including transportation of carbon from upstream sources. As such, emissions do not scale linearly with the surface area of lakes and reservoirs. Additionally, the depth-to-surface-area ratio of a reservoir has a large impact on overall emissions, making surface area measurements an incomplete picture.
  2. Existing global inventories of hydropower GHG emissions are estimated using a small dataset that was collected through inconsistent methodology and highly variable measurements. Measurements of hydropower GHG emissions were often taken from relatively few locations within a reservoir, and only a few studies were based on samples spanning an entire year. Current statistical analysis relies on extrapolated measurements from this non-representative sample of hydropower reservoirs, making conclusions unclear. Additionally, no national carbon emission inventory of hydropower reservoirs has yet to be attempted in the continental United States. 
  3. Current carbon accounting practices do not properly account for time scale or the effects of climate change. ​​It is estimated that low-lying floodplains and braided river channels located between two reservoirs can store carbon for well over 1,000 years. In the Missouri River, for example, carbon dating in the streams and floodplains showed oak logs could hold carbon for a median of 3,515 years (with few samples younger than 150 years). This also suggests that adjoining reservoirs can store carbon far below the surface even longer due to less oxygen exposure. Carbon accounting practices, however, do not account for these long burial rates. Current approaches also do not account for changes in climate conditions over time. For example, warmer climate conditions may increase mineralization rates of carbon into carbon dioxide for both land-based and aquatic carbon, which could decrease carbon storage. However, aquatic systems may also suffer from eutrophication under warmer conditions, which decreases mineralization rates and would increase potential carbon storage in reservoirs. Increased frequencies of wildfires and floods would further reduce land-based carbon storage and increase storage in aquatic systems.
  4. Current estimates of hydropower developments are calculating gross emissions instead of net emissions. When assessing GHG emissions from land-based sources, it is common to describe the net change in emissions. For example, the amount of carbon sequestered following reforestation is often expressed as the change in emissions before and after reforestation. In the case of hydropower, however, most published estimates report gross emissions without following the International Energy Agency’s recommendations to subtract the expected emissions from the land area before development.
  5. Hydropower’s carbon footprint is not placed in the context of the full electrical portfolio. Currently, both natural gas and hydropower have the flexibility to provide ancillary (grid-stabilizing) services when variable renewables, like wind and solar, are less available or not available. This makes natural gas the most likely energy source to be displaced by hydropower, but those displaced fossil-fuel emissions are not currently factored into hydropower’s carbon footprint. Assessments also do not properly compare cradle-to-grave assessments of hydropower versus other energy sources, as fossil fuel assessments frequently exclude historical emissions and the carbon footprint of exploration. As a result of these discrepancies, foot-printing methods sometimes erroneously suggest that fossil fuels have a smaller land and carbon footprint than renewable sources.

Suggested Improvements

  1. To accurately measure carbon emissions, reservoirs need to be considered within the context of broader watersheds. That means comparing the amount of organic matter versus surface area, accounting for size, shape, and/or slope of the watershed, and understanding the effects of upstream bodies of water on reservoirs. One improvement could be to report actual emissions relative to the expected influx of carbon, instead of reporting on the reservoir area. Previous modeling has shown dissolved organic carbon and the potential for erosion to be useful predictors of both CO2 and methane emissions. Studies consistently show reservoirs can be hotspots of methane generation and emission when upstream sources contain high carbon-laden sediment deposits.
  2. GHG emissions estimates must expand to represent a larger, more representative portion of the United States. Conducting probabilistic surveys with consistent measurements will minimize the risk of extrapolation and ensure proper statistical analysis. Additionally, more data is needed focused specifically on quantifying hydropower’s carbon footprint. For example, few studies have measured degassing emissions associated with turbines. Ideally, data should be collected consistently from probabilistic surveys of reservoirs and reference pre-dam ecosystems across time and seasons. Where model-based methods are used to extrapolate to a larger population of reservoirs, prediction uncertainties associated with models should be reported.
  3. Time scale needs to be considered. If the objective is to compare emissions across energy sources, it is also necessary to consider the length of time carbon will be sequestered and to weight the release of older carbon sources more heavily when calculating emissions. The amount of time reservoirs can store carbon needs to be considered when comparing hydropower with other energy sources.
  4. To improve attribution of emissions to reservoirs, a better understanding of methane emissions in pre-dam landscapes is needed. Watershed-scale ecosystem models could be used to quantify where high-emitting ecosystems with moderate carbon storage, such as wetlands and floodplains, and ecosystems with high carbon storage and low methane emissions, such as forests, were displaced. In addition, differences in the expected efficiency and duration of carbon burial should be estimated before and after constructing and flooding a reservoir. More research is needed to monitor emissions directly attributable to hydropower, such as degassing of methane from turbines, seasonal changes in degassing through the turbines as a function of water intake height, and changes in methane generation caused by daily and seasonal timing of flow releases to meet electricity demand.
  5. Several improvements are recommended to properly assign credit for hydropower’s benefits to the electricity portfolio. First, methods should be updated to account for raw materials, production, logistics, and byproducts, as well as full lifecycles of both renewable and non-renewable sources of electricity. Second, there is a need to better measure and model GHG emissions that are directly tied to hydropower operations. Third, total emissions of the electricity portfolio need to be understood in the context of their different ancillary advantages (e.g., storage, dispatchability, low cost, reduced emissions). Fourth, it is not enough to assess the GHG emissions from a single hydropower facility because facility siting and operations are designed to maximize generation across an entire system of hydropower facilities. Addressing this last item requires looking at GHG emissions across the entire power generation system, instead of individual hydropower facilities. There are two common approaches to assessing GHG emissions across the whole power generation system:  
    • Calculating the marginal increase in GHG emissions across the system for given changes in electricity demand (marginal emissions approach).
    • Accounting for hydropower’s ancillary services by computing variations in GHG emissions using different mixes of fuels at different times (e.g., hourly, daily, seasonally).

Better understanding GHG emissions from hydropower is essential to realizing the role it plays in a clean energy economy. However, the methods for scientifically analyzing GHG emissions need to be refined to ensure estimations are accurate and consistent. WPTO will continue to support research in this space and update this page as new research becomes available.  

Stay in the know with WPTO! Receive the latest information on funding opportunities, events, and other news by subscribing to the bimonthly Hydro Headlines newsletter, as well as the comprehensive, monthly Water Wire newsletter.