Greenhouse Gas Emissions from Reservoir Water Surfaces: A New Global Synthesis

When we think of greenhouse gases, the first thing that comes to mind is the pollution caused by our activities; emissions from cars, factories and landfills. But lower in our consciousness is the impact of a structure typically seen as green and eco-friendly, the reservoir. Dams are helpful, whether it’s for hydropower, flood control, water supplies, transportation or recreation.  Reservoirs create convenient deposit areas for large amounts of water, but they also create significant levels of greenhouse gases which haven’t always been explored in depth.


Learning from the Past

Because dams are artificial, they vary from natural water systems.  They have different nutrient levels, fluxes and ecosystems, which are caused by many factors. The rapid flooding involved in creating a reservoir causes high levels of microbial decomposition, the fluctuations in water levels (which are larger than in natural water bodies) cause greater ebullition (bubbling) of CH4 gases and their strategic/non-strategic placement near organic matter or human activity causes even more decomposition. These and other factors lead to greenhouse gas emissions that are quite different from natural bodies of water.

Common methods of measurement are concerned with quantifying the diffusion of gases across the air-water interface.  Which is fine for CO2 and N2O, which are soluble, but for CH4, which is comparably insoluble and bubbles up from sediments, this isn’t the best way of grasping what gases are actually being released. Methods that do capture ebullition, such as inverted funnel traps, aren’t left in for long enough or spread over a large enough area to attain accurate measurements, as these fluxes are highly variable over time and space. 

Do Past Methods Measure Up?

Sampling has been improved in a few ways, but even some of these improvements have their own problems. One possible solution is a modified funnel trap design which would allow for a higher degree of spatio-temporal coverage. A second option is echo-sounding; however, this method only has a certain operational depth range, doesn’t provide info on the concentration of bubbles and is an expensive piece of equipment which is difficult to operate. A third option is eddy covariance, but this method is also expensive, has worse performance in wet conditions and there is a difficulty in estimating the footprints of gases using this method. Overall these solutions are not perfect, but improve upon the current methods which overlook this important source of greenhouse gases.

In the collection of studies in this global overview, ebullition is only measured in 52% of the reports that measured CH4; the majority of these used funnel traps or simply included CH4 with diffusive flux. Only 4 studies used eddy covariance or acoustic methods.  But even with these few measurements, the mean of ebullition + diffusion fluxes were more than double that of those which only included diffusion – and CH4 fluxes were radically variable on the basis of whether or not ebullition was included.


This study gathered the areal CH4, CO2 and N2O flux estimates from 75, 229 and 58 systems respectively. For this synthesis, both hydroelectric and non-hydroelectric systems were included. Some important inclusions in this synthesis come from greenhouse gas estimates from temperate and subtropical reservoirs; these have not always been measured in the past. However, as many new reservoirs are being planned in these types of climates in the future, it is imperative that emissions in these areas be measured. These estimates indicate that midlatitude reservoirs can emit as much CH4 as tropical reservoirs, dispelling assumptions that lower latitude reservoirs have greater CH4 emissions than temperate reservoirs, especially in light of the fact that some Amazonian reservoirs were statistically indistinguishable from other regions.

This report also found that mean areal CH4 fluxes were approximately 25% larger than previous estimates, that CO2 estimates were approximately 30% smaller than previous, and also collected the first ever global mean estimate of reservoir N2O fluxes. When compared to fluxes in natural bodies of water, the areal CH4 emissions from reservoirs are higher than average.

The Role of Productivity
This study was specifically interested in the following hypothesis: “nutrient loading and the resulting increase in primary production stimulates GHG emissions from reservoir water surfaces, primarily via enhanced CH4 production”. To see if this hypothesis could be proven, the researchers used characteristics that were likely to co-vary with/control greenhouse gases; this included morphometric, geographic, and historical properties of reservoirs (ex: depth, residence time, volume, surface area, age, latitude, etc.) and measurements of primary productivity (trophic status, mean/modelled surface chlorophyll a concentrations, etc.)

In summary, what they found was:

  • CH4 emissions were best predicted by chlorophyll a concentrations
  • CO2 emissions were best predicted by report mean annual precipitation
  • N2O emissions were best predicted by NO3– concentrations


Very different CH4 emissions were found in systems with different trophic statuses.  Specifically, eutrophic systems (high in nutrients) emitted an order of magnitude (approximately ten times) more CH4 than oligotrophic ones (low in nutrients). This is concurrent with recent findings, which showed that CHemissions were best predicted by primary production – suggesting that low oxygen and high dissolved organic carbon, as they are found in a eutrophic system, create elevated CH4 levels compared to lower nutrient systems.

The Future is Now 

In terms of area, the study used 305,723 km2 of reservoir surface areas; this area represents 25% of global reservoirs. As reservoir areas are likely to increase substantially in approaching decades (847 large hydropower projects and 2,853 smaller projects already being planned/constructed), it is important more than ever to understand the greenhouse gases being produced by these systems.  Seeing as CH4 fluxes in this synthesis are higher than the per area fluxes of any other aquatic ecosystem, now is the time to develop more in-depth methods and promote interest in the studies of these artificial bodies of water.