GHG emissions from tropical reservoirs: state of the science and G-res tool

For a long time, hydroelectricity was presented as a naturally clean energy. The reality is more nuanced for reservoirs located in the tropics, where the decomposition of submerged biomass can generate far from negligible methane emissions. This article reviews the state of the science on reservoir emissions, presents the G-res tool developed by the IHA and the UNESCO Chair, distinguishes gross emissions from net emissions, and details how a practitioner estimates these fluxes and presents their uncertainty to lenders.

Why a tropical reservoir emits methane

When a valley is flooded, the submerged vegetation and soils cease to respire and begin to decompose under water. This decomposition consumes the oxygen present in the water column. At depth, the sediments become anoxic. It is under these oxygen-free conditions that bacteria produce methane.

Methane is a greenhouse gas far more potent than carbon dioxide over a given time horizon. A reservoir can therefore weigh on the climate balance even if the volumes of gas appear modest. Carbon dioxide is also present, resulting from the mineralisation of organic matter, but it is methane that defines the specificity of the subject.

The tropical climate accentuates the phenomenon. Warm waters accelerate bacterial activity. High biological productivity continuously supplies the reservoir with fresh organic matter. Finally, shallow, large-surface reservoirs, common on lowland rivers, offer a vast interface for exchange with the atmosphere. Three main pathways release methane: diffusion through the surface, ebullition from the sediments, and abrupt degassing when passing through the turbines and at the foot of the dam.

This last point deserves attention. Water drawn from depth, laden with dissolved methane, undergoes a pressure drop whilst passing through the structure. The gas is then released suddenly, like opening a bottle. This downstream degassing has long remained off the radar. Yet it counts in a serious estimate.

Gross emissions and net emissions: the right question

The first mistake is to measure what leaves the reservoir surface and stop there. This figure, the gross emission, says little about the project's actual effect on the climate.

The right question is that of the net footprint. Before flooding, the site was already emitting. A watercourse, a wetland, a forest naturally exchange gases with the atmosphere. The project must therefore be credited or debited with the difference between the situation with reservoir and the prior situation. This is the principle of net emission, which reasons in relation to a baseline state before impoundment.

Two corrections are added to this difference. On the one hand, sources unrelated to the project must be subtracted, such as wastewater discharges from an inhabited catchment, which would artificially inflate the reservoir's balance. On the other hand, allocation is necessary when the reservoir serves several uses: electricity generation, but also irrigation, drinking water, flood regulation, navigation. Not all emissions fall solely under the energy component.

This reasoning directly aligns with the boundary logic described in our article on theGHG balance of an infrastructure project. Reservoir emissions fall under the site's direct emissions, to be treated with the same boundary rigour as the rest of the inventory. They also articulate with scope accounting detailed in the article on thecarbon footprint of a project in scope 1, 2 and 3.

The state of the science: what we know, what we do not know

Research on reservoir emissions has progressed considerably over some twenty years. A few findings now enjoy consensus.

The first: tropical reservoirs on average emit more than reservoirs at temperate or cold latitudes, because of temperature and organic load. The second: emissions are high in the first years following impoundment, whilst the initial biomass decomposes, then tend to decline. The third: variability is enormous from one site to another. Two neighbouring reservoirs can present very different profiles depending on their depth, their water residence time and the nature of the catchment.

What remains uncertain matters just as much. The exact share of ebullition and downstream degassing, difficult to measure, varies greatly. The long-term trajectory of emissions, over several decades, still rests on extrapolations. Finally, the attribution between what would have been emitted without the dam and what is actually attributable to it remains a delicate exercise.

For a practitioner, this uncertainty is not an obstacle, it is a datum to be integrated. One does not present a single value as truth. One presents an estimate, its assumptions and a range. It is this methodological honesty that E&S teams at lenders know how to appreciate.

The G-res tool: what it does and what it does not do

Faced with this complexity, a need emerged: to have a shared method to estimate a reservoir's emissions without a heavy measurement campaign. This is the purpose of the G-res tool.

The G-res Tool is the fruit of long-term work conducted by the International Hydropower Association and the UNESCO Chair on Global Environmental Change, with the support of research institutions and the World Bank. It is presented as a free online tool that aims, according to its developers, to "estimate and report the net greenhouse gas emissions of a reservoir". Its method is based on empirical measurements conducted on more than two hundred reservoirs across the world.

In practical terms, the tool estimates the net footprint from accessible input data: reservoir characteristics, catchment, local climate, land use before flooding. It produces a net carbon footprint, distinguishes emission pathways and proposes allocation between the reservoir's uses. Its developers describe it as "a simple-to-use, web-based tool that can be used to predict GHG emissions from reservoirs without going to the field to take measurements".

One must understand what the tool is not. It is not a field measurement. It is a statistical model that estimates, from correlations observed elsewhere, what a given reservoir is likely to emit. Its value depends on the quality of the data entered and the representativeness of the site in relation to the training base. An atypical reservoir, very different from those used to calibrate the model, will produce an estimate to be taken with caution. G-res guides and frames an estimate. It does not dispense with reflecting on its limits, nor, on sensitive projects, with a targeted measurement campaign.

Risk factors to identify upstream

Certain configurations herald a high emissions profile. Identifying them early, from the pre-feasibility study, avoids discovering late a carbon issue that weighs on the project's attractiveness.

The signals to monitor are well identified. A large flooded surface for modest power degrades the ratio between emissions and energy produced. A shallow, warm reservoir with long residence time favours anoxia and methane production. Abundant submerged biomass, dense forest or peatland, constitutes a carbon stock that will decompose for a long time. A catchment bringing much organic matter and nutrients sustains emission over time.

Two design reflexes reduce risk. Prior clearing of the basin limits the stock of submerged matter. The position of the water intakes, if it avoids drawing the deepest, most laden layers, mitigates downstream degassing. These choices are considered upstream, not after impoundment.

This subject intersects other components of environmental assessment. Reservoir management interacts with water quality, sediment and hydrological regime issues. A project that seriously addresses itsenvironmental flow downstreamhas moreover often already gathered part of the data useful for estimating emissions.

Communicating the estimate and uncertainty to lenders

An emissions estimate has value only if presented in a defensible manner. Lenders do not seek a flattering figure. They seek a traceable approach.

The first principle is transparency of assumptions. Any input datum, any baseline state, any allocation rule must be explicit and sourced. The second is presentation in ranges rather than single values, with sensitivity analysis on the most uncertain parameters. The third is the trajectory over time: distinguishing the peak of the first years from the long-term situation avoids fixing an unfavourable figure for the entire duration of the concession.

This requirement aligns with the logic of climate reporting expected by financiers. Presenting a reservoir carbon footprint, its assumptions and its risks fits naturally into a TCFD-type framework, which requires documenting climate risks and the assessment method.

GHG emissions from tropical reservoirs are no longer a blind spot. Reservoir methane is an identified issue, which lenders now integrate into their reading of hydroelectric projects. Three reflexes structure a credible dossier. Estimate a net footprint, with a baseline state, and not a gross surface emission. Use a recognised tool such as G-res to frame the estimate, whilst acknowledging its limits. Present the result in ranges and trajectory, with transparent assumptions.

The right posture is not to minimise the subject, nor to dramatise it. It is to treat it like any other E&S risk: identified early, quantified methodically, reduced through design and communicated honestly. A project that does this work transforms a sensitive question into proof of seriousness.