<b>Uncovering the underestimated greenhouse gas emissions of hydropower reservoirs: A global synthesis</b>

<p dir="ltr">Hydropower reservoirs, long considered a clean energy solution, are increasingly recognized as major greenhouse gas (GHG) emissions hotspots. However, over 97% of these hydropower reservoirs remain unmonitored, creating major uncertainties in global carbon budgets. Here, we develop the first machine learning framework to predict GHG fluxes from unmonitored hydropower reservoirs leveraging widely accessible environmental data. Our gas-specific models not only quantify emissions but also reveal key environmental drivers controlling GHG dynamics, highlighting the predominant role of temperature, particularly in Africa’s reservoirs. We estimate that unmonitored reservoirs contribute an additional 95 Tg CO2eq annually, raising total global hydropower-related emissions to 504 Tg CO2eq year-1, a ~ 77% increase over previous estimates. Methane dominates emissions (55%), followed by carbon dioxide (42%) and nitrous oxide (3.4%), the latter often overlooked but increasingly important as eutrophication intensifies. Our findings call for a re-evaluation of the climate costs of hydropower and suggest that future warming may further amplify reservoir-based GHG emissions. The predictive capacity of our models provides a scalable tool for refining global GHG inventories and guiding low-carbon hydropower development under a changing climate.</p>