WATER GLOBAL PRACTICE TECHNICAL NOTE Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes DECEMBER 2017 About the Water Global Practice Launched in 2014, the Word Bank Group's Water Global Practice brings together financing, knowledge, and implementation in one platform. By combining the Bank's global knowledge with country investments, this model generates more firepower for transformational solutions to help countries grow sustainably. Please visit us at www.worldbank.org/water or follow us on Twitter @WorldBankWater. Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes DECEMBER 2017 © 2017 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Please cite the work as follows: World Bank. 2017. Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes. World Bank, Washington, DC. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights​ @­worldbank.org. Cover design: Jean Franz, Franz & Company, Inc. Contents Acknowledgments v Executive Summary vii Abbreviations xi Chapter 1  Purpose of this Technical Note 1 Notes 2 Chapter 2  Basic Overview of Greenhouse Gases from Reservoirs 3 2.1 The CO2 Cycle in a River Basin 3 2.2 Data on GHG from Reservoirs 5 2.3 Most Important Factors Influencing GHG Emissions from Reservoirs 8 2.4 Extreme Temporal and Spatial Variation 10 Notes 11 Chapter 3  What Does The Atmosphere See? 13 3.1 Gross and Net Fluxes of Greenhouse Gasess 13 3.2 The G-res Tool 14 3.3 The IEA Hydro Framework 16 3.4 Global Estimates of GHG Emissions from Reservoirs 16 Notes 22 Chapter 4  Recommendations for Preparing Dam Projects 23 4.1 General Framework 23 4.2 Documentation and Initial Screening 23 4.3 Quantitative Assessment of Net Emissions Using the G-res Tool 27 4.4 Assessment of Reliability in Estimated Emissions 28 4.5 Detailed Assessment Following the IEA Hydro Framework 29 4.6 Management of GHGs and Post-Impoundment Monitoring 30 Notes 31 Chapter 5  Future Research 33 Note 33 Appendix A  Conversion of GHG Units and CO2 Equivalents 35 Conversion from Moles to G 35 CO2 Equivalents 35 Carbon Dioxide Equivalents Vs. Carbon Equivalents 36 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes iii Appendix B  Example of G-res Application 37 Appendix C  Suggested Template for GHG Reporting 47 Glossary 49 References 53 Figures 2.1. Carbon Fluxes in Terrestrial and Aquatic Landscapes 4 2.2. Main Fluxes of Carbon for a Reservoir 5 2.3. Analysis of Scientific Papers Published Related to Measured GHG Data from Reservoirs 6 2.4. Measured GHG Emission Fluxes at Reservoirs 10 2.5. Measurement of the GHG Fluxes in the Batang Ai Reservoir, Malaysia, Showing the Extreme Spatial Variability 11 2.6. Measurement of Diffuse Fluxes of CH4 in the Petit Saut Reservoir (Impounded in 1994) 11 3.1. The Global Carbon Cycle 14 3.2. Displacement of CO2 when a Dam is Constructed in a River 15 3.3. Illustration of Changing Distribution of CO2 Emissions for All Reservoirs over the World When Considering Net Rather than Gross Values 19 3.4. GHG Emissions Estimated through the G-res Tool for a Global Set of Single-Purpose Hydropower Projects 21 4.1. Recommended Approach for Assessing GHG Emissions for New Dams 24 4.2. Proposed Thinking Process for Documentation and Initial Screening 25 4.3. The Interface of the Online G-res Tool 27 Maps 2.1. Location of Reservoirs with Measured Diffusive Flux of CO2 6 2.2. Location of Reservoirs with Measured Diffusive Flux of CH4 7 2.3. Location of Reservoirs with Measured Ebullition (Bubbling) Flux of CH4 7 2.4. Location of Reservoirs with Measured Degassing of CH4 8 3.1. Distribution of Reservoirs in the GRanD Database 18 3.2. Single-Purpose Hydropower Projects with Verified Installed Capacity and Power Production 20 Tables 2.1. Measurements of N2O and Comparison of CO2 Equivalents with Diffusive CH4 and CO2 11 3.1. Allocation of GHG Emissions for Multipurpose Reservoirs Followed with G-res Tool 17 3.2. Global Estimates of Reservoir Area and GHG Emissions 18 4.1. Basic Information to Assess Future Potential GHG from a Reservoir 24 4.2. Guidance for Screening Out Projects with Negligible Reservoir GHG Emissions 26 iv Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes Acknowledgments This technical note is the result of a collaboration Economist); Harikumar Gadde (Sr. Carbon Finance between the Water and Energy Global Practices of Specialist); Peter Bergsten (Environmental and Social the World Bank. It was prepared by Rikard Liden, Sr. Development Specialist, MIGA); Francisco Avendano Hydropower Specialist, Hydropower and Dams (Operations Officer, IFC); Amarilis Netwall Global Solutions Group. The report is, however, an (Environmental and Social Development Specialist, effort by a larger team from the Global Solutions MIGA); and Sean Nelson (Consultant). The work has Group, including William Rex (Global Lead); been overseen and approved by Pilar Maisterra Kimberly Lyon (Water Resources Management (Practice Manager, Global Water Practice – Strategy Analyst) and Karin Grandin (Consultant); and a mul- and Operations). Thanks also to Prof. Yves Prairie tidisciplinary advisory team from the World Bank (UNESCO Chair in Global Environmental Change), Group, including Karan Capoor (Sr. Energy Sara Mercier-Blais (University of Quebec at Montreal), Specialist); Marcus Wishart (Sr. Water Resources Richard Taylor and Mathis Rogner (IHA), and Niels Management Specialist); Samuel Oguah (Energy Nielsen (Kator Research) for their review and com- Specialist); Pablo Cardinale (Principal Environmental ments on draft versions of this technical note. This Specialist, IFC); Anne Schopp (Environmental note was edited by Inge Pakulski. Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes v Executive Summary Greenhouse gas (GHG) emissions from reservoirs— These predictive tools are based on the principle created to produce hydropower, achieve water secu- agreed by the Intergovernmental Panel on Climate rity, or provide flood protection—may be significant Change (IPCC) for the estimation of net reservoir emis- and should be considered in the planning and design sions. Rivers are major conveyors of carbon from of new dam infrastructure. This note provides guid- terrestrial areas to lakes and the sea. Terrestrial areas ance to World Bank Group (WBG) staff on how to assess are generally net carbon sinks while aquatic systems GHG emissions from reservoirs at an early stage of the are net carbon emitters. Changes in GHG fluxes to the preparation process. atmosphere resulting from the introduction of reser- voirs in a river system must therefore be viewed from a The emissions of carbon dioxide (CO2), methane (CH4), catchment perspective. Net GHG emissions caused by a and nitrous oxide (N2O) from reservoirs have been a reservoir are the difference between total fluxes of CO2 source of extensive debate, owing to the divergent equivalent emissions for the river basin before and results generated by the research community, which after the creation of that particular reservoir. has been examining this area for only the last three decades or so. The biogeochemical processes leading The G-res tool builds on this principle of calculating the to GHG emissions are very complex and emission net anthropogenic GHG emissions, that is, what the measurements are cumbersome. Consequently, it is atmosphere will see when a new, man-made reservoir difficult to estimate emissions from existing reservoirs is introduced into the landscape. A recent application of and even more difficult to predict them for future the G-res tool to a global database of reservoirs1 indicates reservoirs. that man-made reservoirs account for about 0.5 percent of global anthropogenic GHG emissions, which is less However, much research has been conducted and pub- than previously estimated. The main reasons for this lished during the last decade, and scientific papers lower estimate are that only emissions directly attribut- from 2017 indicate that the science and methodologies able to the reservoirs are considered and that site-specific to estimate GHG emissions from reservoirs are con- factors have governed the emissions rather than linearly verging. As a result, predictive tools are now available extrapolating average measured emissions from a few for practitioners involved in dam development. One reservoirs to a total global estimated area of reservoirs such tool is the GHG Reservoir Tool (G-res tool), devel- (as was done in the past). oped and launched in May 2017 by a research team led by the United Nations Educational, Scientific and The purpose of this note is to provide guidance to Cultural Organization (UNESCO) and the International WBG staff on how to assess GHG emissions from Hydropower Association (IHA). A complementary tool reservoirs in preparation of dam infrastructure proj- is the framework developed by the International ects, in accordance with the latest research and the Energy Agency Technology Collaboration Programme data and tools available today. This note provides a for Hydropower (IEA Hydro), which (i) recommends layman’s description of the major biogeochemical pro- procedures for primary data collection and pro- cesses responsible for GHG emissions from reservoirs cess-based modeling approaches to simulate reservoir and makes concrete recommendations for estimat- GHG emissions, and (ii) provides guidance on how to ing  the volume of GHG emissions caused by those manage and mitigate those emissions. biogeochemical processes for planned reservoirs. Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes vii Moreover, it includes a bibliography, listing key scien- Steps 2 and 3 should only be conducted if warranted tific studies for readers who seek more detailed infor- by step 1. The methods and results of the reservoir mation, a glossary, and directions for converting the emissions assessment should be reported as a sub- emission volumes of different GHGs into carbon chapter in the ESIA or as a separate dedicated dioxide–equivalent amounts. report, and they should be summarized in the project appraisal document (PAD). A template for In the case of dam infrastructure projects with inun- the presentation of this assessment in the PAD is dation for which the WBG may provide financing, provided (appendix C). GHG emissions from reservoirs should be analyzed as part of the Environmental and Social Impact For all projects, secondary data on key Assessment (ESIA). This analysis and estimation of variables affecting reservoir emissions should be GHG emissions from the reservoir should be based on compiled and documented to enable an initial data available from the early phases of project prepa- screening and, if required, to provide input data for ration (Prefeasibility study and Environmental further analysis. If the initial screening indicates that Screening). Doing so will allow managing and miti- reservoir emissions are not negligible, the G-res tool gating potentially significant reservoir emissions in should be applied to predict future net reservoir the project planning and design, as well as the inclu- emissions for the planned investment. However, the sion of specific actions in the Environmental and extent of uncertainty of the G-res tool results should Social Management Plan (ESMP) to address these be acknowledged and a thorough reliability check of emissions. the results should be conducted. If the assessment shows that the results of the G-res tool are highly Because the emissions of different reservoirs have uncertain, or if the reservoir emissions estimate has been shown to vary by several orders of magnitude, it to be highly reliable, more detailed assessments are is advisable to use a stepwise process in which the advised, including primary data collection and pro- complexity of the assessment of reservoir emissions cess-based modeling in accordance with the IEA is adjusted to reflect the severity of GHG emissions Hydro framework. and their importance to the specific investment proj- ect. It is suggested that the following steps be taken in For dam infrastructure projects with significant esti- the assessment and estimation of reservoir GHG mated reservoir GHG emissions, possible mitigation emissions: measures should be considered and specified in the 1. Secondary data compilation, documentation, and ESMP. It is suggested that a detailed GHG Management initial screening, which, in the case of hydropower Plan (a subplan of the ESMP) be prepared, which projects, would focus on the power density should include specific targets, actions, and moni- toring. GHG emissions management may include 2. Estimation of net emissions using secondary data infrastructure design—such as ensuring aerobic con- and the G-res tool, including reliability assessment ditions upstream of the intake—and management of of the result organic material and nutrient effluents from the 3. Refined estimation, based on primary data collection upstream catchment, aimed at reducing gross GHG and process-based modeling guided by the IEA emissions and improving the water quality of the Hydro framework. reservoirs. viii Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes Note Reference 1. The Global Reservoir and Dam (GRanD) Database (Lehner et al. 2011). Lehner, B., C. Reidy Liermann, C. Revenga, C. Vorosmarty, B. Fekete, This database includes more than 6,500 dams with a storage capacity P. Crouzet, P. Doll, M. Endejan, K. Frenken, J. Magome, C. Nilsson, larger than 1 km3 and was corrected to exclude regulated natural J.C. Robertson, R. Rodel, N. Sindorf, and D. Wisser. 2011. Global Reservoir lakes. The global estimate of GHG emissions was also corrected to and Dam Database, Version 1 (GRanDv1): Reservoirs, Revision 01. Palisades, include emissions from small impoundments. NY: NASA Socioeconomic Data and Applications Center (SEDAC). Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes ix Abbreviations CH4 methane CO2 carbon dioxide CO2e(q) carbon dioxide equivalent ESIA Environmental and Social Impact Assessment ESMP Environmental and Social Management Plan FAO Food and Agriculture Organization GHG greenhouse gas G-res GHG reservoir tool GRanD Global Reservoir and Dam (database) GWP global warming potential IEA Hydro International Energy Agency Technology Collaboration Programme on Hydropower IFI International Financial Institution IHA International Hydropower Association IPCC Intergovernmental Panel on Climate Change kWh kilowatt-hour MW megawatt N2O nitrous oxide UAS unrelated anthropogenic source UNESCO United Nations Educational, Scientific and Cultural Organization WBG World Bank Group Important definitions of gross and net emission for WBG the emissions that are measured from the reservoir staff to note: surface and the immediate river stretch downstream of the reservoir after impoundment. In the science of reservoir GHG emissions, “net emis- sions” refer to the difference between the volume of In World Bank accounting of GHG emissions for emissions measured after impoundment and the vol- investment projects (and similarly by other IFIs1), “net ­ ume of emissions (or uptake) that occurred prior to emissions” are the difference in emissions of the invest- impoundment, that is, the volume of additional emis- ment project and the counterfactual. In this context, the sions that is the result of introducing a reservoir into counterfactual may be either a “without project” sce- the landscape. By contrast, “gross emissions” refer to nario or an “alternative scenario” that reflects the most Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes xi likely alternative means of achieving the same project Notes outcomes or level of service. The “gross emissions” are 1. International Financial Institution Framework for a Harmonised the absolute emissions of the investment project.2 Approach to Greenhouse Gas Accounting, November 2015. In this technical note, “net emissions” allude to the 2. In case of a hydropower project for which the most likely alternative to produce the same amount of power is a coal power plant, the net definition used by the science of reservoir GHG emis- reservoir emissions become part of the gross hydropower project sions (that is, the first definition given above). emissions. Thus, the net project emissions are defined as the net res- ervoir emissions plus construction emissions, minus the emissions The unit “tons” refers to “metric tons” (“tonnes”) of a coal power plant producing the same amount of power as the throughout this report. hydropower project. xii Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes Chapter 1 Purpose of this Technical Note Reduction of greenhouse gas*1 (GHG) emissions is the complexity of the dynamic physical, chemical, and fundamental to the mitigation of climate change. biological processes, not all scientific processes are It has become increasingly important to estimate and described in detail and some are simplified. Further report on GHG emissions to enable the implementa- details and an in-depth description of these processes tion of mitigation measures to limit or reduce total may be found in the key references provided in the emissions. In most cases, such estimation is fairly bibliography. simple, using known emission factors per surface area Like the 2013 Interim Note, this updated version dis- or per produced energy unit. However, GHG emissions cusses: (i) the major biogeochemical processes causing from reservoirs created for the purpose of electricity GHG emissions from reservoirs; (ii) the state of current generation, water security, or flood protection are knowledge, and (iii) recommendations for assessing very difficult to estimate, and no single emission fac- GHG emissions caused by biogeochemical processes tor or formula can be applied. for planned reservoirs. Besides a general update on the The purpose of this note is therefore to provide state of the art, the main change with respect to the pre- guidance to World Bank Group (WBG) staff on how to vious version is the introduction of the G-res tool, devel- assess GHGs from reservoirs in preparation of dam oped by UNESCO/IHA, and the IEA Hydro framework as infrastructure projects. It is an update of the World the recommended tools for the quantification of reser- Bank (2013) Interim Technical Note with the same title. voir emissions. The note briefly describes these tools The note no longer has an interim status, which it was and explains how they can be applied to WBG dam given in 2013 on account of anticipated new research infrastructure investment projects. Moreover, it pro- published in recent years. vides a bibliography, listing key scientific studies for The technical note is limited to the GHG emissions readers who seek more detailed information, and a resulting from the biogeochemical processes that are glossary, as well as detailed directions for converting initiated when a river is dammed and the area upstream (emission) volumes of different GHGs into carbon is flooded. As GHG emissions are a vital part of GHG dioxide–equivalent amounts (appendix A). accounting for projects involving reservoirs (such as GHG emissions from reservoirs are still a relatively storage dams for flood management, irrigation, water new area of research. Therefore, it should be no sur- supply, or hydropower), the note provides input to the prise that research conducted over the last 20–30 years WBG’s methodology for estimating the carbon foot- has shown disparity in GHG emission magnitudes print* of a project.2 Yet it does not include guidance on from reservoirs, which has led to a debate on method- how to define the counterfactual scenario for alterna- ologies and the reliability of results. However, during tive development projects. the last decade, research has significantly improved The aim has been to create a short and concise note, our knowledge and understanding of the subject and written in easily understandable, not overly technical a  recently published scientific paper by a large language, covering the most important and relevant number  of recognized researchers in the field of facts relating to GHG emissions from reservoirs. Given GHG  emissions (Prairie et al. 2017) indicates that Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 1 research is converging. Unlike in 2013, when the Notes Interim Technical Note was published, various tools 1. Terms marked by an asterisk (*) at their first occurrence in the main and models are now available for use in the prepara- text are defined in the Glossary. tion of large dam infrastructure. Yet more research 2. For example, Guidance Manual: Greenhouse Gas Accounting for will be required to refine these tools,3 and WBG staff Energy Investment Operations, Transmission and Distribution Projects, Power Generation Projects, and Energy-Efficiency Projects, must take care when applying them. Staff must also Version 2.0, January 2015. ensure that they always use the latest (software) ver- 3. Few studies have been conducted on essential GHG pathways (such sions and be clear and frank in discussing the uncer- as methane bubbling and downstream degassing), where research is tainties still underpinning the science. still geographically uneven. 2 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes Chapter 2 Basic Overview of Greenhouse Gases from Reservoirs 2.1 The CO2 Cycle in a River Basin respiration* to the atmosphere, either directly from vegetation or through decomposition* of Changes in land use and/or changes to the natural dead organic matter. The balance of these CO2 cycles of water and energy affect the interactions fluxes creates the growth of biomass (live or among the terrestrial, aquatic, and atmospheric dead) in the terrestrial ecosystem, contributing environments, and therefore have an effect on GHG to the biomass* carbon pool. Live biomass, and emissions. When a river is dammed, the flow dynamics its carbon, may be removed, for example, through are  changed, riverine sediment and organic material harvesting or fires. In these cases, the carbon is are trapped, and terrestrial ecosystems* are flooded. eventually fed back to the atmosphere, primarily These changes alter the previous cycle and fluxes of in the form of CO2. carbon dioxide* and other GHGs within the project footprint. • Dead organic matter that has not been directly decomposed or respired is eventually absorbed into The main GHGs that may be emitted from a reservoir the soil or transported to the river through rainfall are carbon dioxide (CO2), methane* (CH4), and nitrous and overland flow. Carbon is thus either stored in the oxide* (N2O). CH4 and N2O have stronger warming soil or transported out of the terrestrial ecosystem effects than CO2 and may be important even if emitted to the riverine ecosystem as part of the erosion pro- in relatively small amounts. cess. Carbon can also be leached from dead organic To account for differences in the Global Warming material or soil and enter river systems directly in Potential* (GWP) of GHGs, the combined emissions of dissolved form. CO2, CH4, and N2O are expressed as CO2 equivalents* • In rivers and lakes (with or without reservoirs), car- (CO2eq).1 Since these three GHGs have different life­ bon can be leached from the bed sediment to the spans in the atmosphere, a specific period needs to be water phase, and the dissolved CO2 can be lost to set to compare their respective GWPs; this period is the atmosphere at the surface. As part of the aquatic normally 100 years. According to the 2013 IPCC Fifth ecosystem, CO2 from the atmosphere or dissolved in Assessment Report (AR5), to obtain CO2eq emissions for the water can also be consumed by aquatic plants a 100-year period, the quantities of CH4 produced and phytoplankton, feeding higher trophic level should be multiplied by 34 and those of N2O by 298. organisms (such as zooplankton and fish) that will To understand the impact of reservoirs on GHG emis- later decay and create new dead organic material, sions, it is essential to understand the main processes thereby adding to the bed sediments. in the cycle of CO2 and other GHGs in a river basin • CH4 is mainly created under anoxic* conditions (no (figure 2.1). oxygen available) in the soil or in bed sediments of • Atmospheric CO2 is taken up by plants through a water body. Such conditions also occur at the bot- photosynthesis* but is lost in parallel through tom of flooded areas. If the water column is strongly Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 3 FIGURE 2.1. Carbon Fluxes in Terrestrial and Aquatic situation before the reservoir was created. More specif- Landscapes ically, the following changes may occur: Upland soils Lowland soils River • The reservoir area changes from the previously terrestrial system into an aquatic system, thereby changing the conditions for interactions of GHGs with the atmosphere for this area. • Following inundation, conditions are created for decomposing vegetation and the carbon contained in the soil of the flooded area, thereby changing the SOM amount of GHGs released into the atmosphere from Sediment CO2 the reservoir area or in downstream rivers when the CH4 water is discharged. Organic and inorganic carbon Corbon cycle in the aquatic ecosystem • The reservoir may provide for seasonal growth and (organic matter mineralization, primary production, decomposition of vegetation in the drawdown zone, CH4 oxidation, etc.) resulting in the absorption and subsequent release SOM Soil organic matter of GHGs into the atmosphere. Source: UNESCO/IHA 2010. • The reservoir may provide anoxic conditions for Note: CO2 = carbon dioxide; CH4 = methane. creating CH4 rather than CO2, especially if the water column is seasonally stratified. stratified on a seasonal basis, CH4 can be produced • The reservoir partially traps riverine organic and accumulate in the anoxic zone. If the CH4 is material and nutrients transported in the river sys- released to the water column as dissolved CH4, it is tem, and may thereby change the circumstances either oxidized (and transformed to dissolved CO2) under which they are transformed into GHG emis- or, if there is a lack of oxygen in most of the water sions compared to where they otherwise would column, lost directly as CH4 to the atmosphere. CH4 have been carried—further down the river system may also be transported up through the water col- (until reaching a natural lake, wetland, or ocean). umn and into the atmosphere in gas form—either by In a reservoir, the flooded and inflowing carbon will be diffusion at the air-water interface or through ebul- exported to the atmosphere, stored in the bed sedi- lition* (bubbling*). ments, or transported further down the river system. • In some circumstances, N O is created as a by-product 2 These three processes occur in parallel to varying of nitrification*, under aerobic* conditions (relating to, degrees, depending on the topographical, geological, involving, or requiring free oxygen), or denitrification*, and climatological conditions, as well as the biological under anaerobic* conditions. As a result, creation of configuration of the water body. N2O mainly occurs in the riparian zones of water bod- Once a reservoir has been created, GHGs can reach the ies, where saturation varies with water levels. atmosphere through several pathways (figure 2.2). The The construction of a dam and impoundment of a res- main pathway is through diffusive flux* of both CO2 ervoir alters the GHG cycle. This will result in a change and CH4 from the surface of the reservoir. However, in flux of GHGs to the atmosphere compared with the significant GHGs can also be flushed through the intake 4 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes FIGURE 2.2. Main Fluxes of Carbon for a Reservoir 1. Bubbling 2. Di usive ux 3. Flux through macrophytes CH4 CO2 CH4 CO2 CH4 FLUVIAL OM (POC, DOC) O2 O2 Phyto Oxycline 4. Degassing 5. Di usive ux CO2 CH4 CO2 CH4 CO2 CH4 FLOODED OM (soils, plant material, wood) Methanogenesis Aerobic CH4 oxidation OM CH4 + CO2 CH4 + 2O2 CO2 + 2H2O Source: UNESCO/IHA 2010. Note: CO2 = carbon dioxide; CH4 = methane; DOC = dissolved organic matter; OM = organic matter; POC = particulate organic carbon. of a reservoir and be released into the atmosphere came from tropical climate zones in Brazil and through degassing* (due to change of pressure) at the French Guiana (the latter mainly related to just one outlet or as diffusive flux at the downstream river sur- reservoir—the Petit Saut). During the last decade, face. In shallow areas of the reservoir, methane can measurement campaigns have become increasingly also reach the atmosphere without being dissolved, distributed, spreading over boreal and temperate through bubbling. climate zones. However, the studied reservoirs are ­ still concentrated in specific regions, with very few 2.2 Data on GHG from Reservoirs observations of GHG emissions from reservoirs in Research on GHG emissions from reservoirs is a rela- Asia and Africa (maps 2.1–2.4). tively new scientific activity and most studies have Data availability also differs much depending on type been conducted during the last 25 years. A sample of and pathway. Of the 223 reservoirs analyzed in recent key references is given in the bibliography, including a research by Prairie (2017), the distribution of data is as short description of the key findings. follows: diffusive CO2 – 198 reservoirs; diffusive CH4 – Given the lack of data on GHG emissions, many stud- 137 reservoirs; bubbling CH4 – 39 reservoirs; and ies have focused on measuring the different forms of degassing CH4 – 35 reservoirs. Deemer et al. (2016) GHG emissions from reservoirs. An analysis of pub- used GHG emission data from a total of 267 reservoirs, lished studies related to observations of GHG emis- of which 229 yielded data on CO2, 142 on diffusive sions from reservoirs shows how the data compilation CH4, 50 on bubbling CH4, and 58 on N2O (maps 2.2-2.4) has developed (figure 2.3). Up to about 10 years ago, (figure 2.4). This shows that data on N2O and on CH4 the main data on GHG emissions from reservoirs bubbling and degassing are still relatively rare. Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 5 FIGURE 2.3. Analysis of Scientific Papers Published Related to Measured GHG Data from Reservoirs 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 No. of studies Brazil 20 French Guyana 7 Panama 2 USA 4 Canada 11 Australia 6 Finland 1 Sweden 2 Poland 2 Switzerland 1 Germany 1 China 7 India 2 Laos 4 Zambia/Zimbabwe 1 Kenya 1 10 Trend in no. 8 of studies 6 published 4 per year 2 0 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 Sources: World Bank 2013 and UNESCO/IHA 2017. MAP 2.1. Location of Reservoirs with Measured Diffusive Flux of CO2 Legend CO2 Di usive emissions 6 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes MAP 2.2. Location of Reservoirs with Measured Diffusive Flux of CH4 Legend CH4 Di usive emissions Source: Based on data from UNESCO/IHA. MAP 2.3. Location of Reservoirs with Measured Ebullition (Bubbling) Flux of CH4 Legend CH4 Bubbling emissions Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 7 MAP 2.4. Location of Reservoirs with Measured Degassing of CH4 Legend CH4 Degassing emissions Source: Based on data from UNESCO/IHA. Another challenge with the observations of GHG Further work on measurement programs and data emissions is that measurement methods had not analysis was published by IEA Hydro (2012). been standardized until recently. A standardized methodology for measuring GHG emissions from res- Estimates of total reservoir emissions are costly as ervoirs was only published in 2010 by UNESCO/IHA; it they require field campaigns that cover long periods of was based on consultations and is widely accepted in time and large areas. Thus, only a few reservoirs in the the scientific community. Diffusive fluxes are mainly world have records going back longer than a few years. measured with the help of floating chambers placed Some dam reservoirs for which more extensive mea- on the surface of a reservoir. Fluxes are quantified by surements have been done are Petit-Saut (French observing changes in the concentration of gases in Guiana), Nam Theun 2 (Lao People’s Democratic the chamber over time. Bubbling of CH4 is normally Republic), Eastmain 1 and La Grande 2 (both in measured through submerged funnels. Degassing Canada), and Tucuruí and Samuel (both in Brazil). occurring at low-level outlets of the dam is estimated by measuring differences in the respective CO2 and 2.3 Most Important Factors Influencing GHG Emissions from Reservoirs CH4 concentrations upstream and downstream of the outlet. Besides giving recommendations on the Thanks to research, a better understanding of the pro- equipment to use, the standardized methodology cesses governing GHG emissions from reservoirs has provides guidance on the temporal and spatial gradually emerged (see Bibliography). What follows is frequency of the measurements, which are essential ­ a summary of the main findings that are generally to getting reliable estimates of GHG emissions. agreed within the scientific community. Only the main 8 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes factors affecting GHG emissions from reservoirs are with higher temperatures governing higher rates. described in this section. Secondly, the lower the water temperature, the more oxygen can be  dissolved in water and, conversely, One of the principal factors affecting emissions from the higher the water temperature, the less oxygen reservoirs is the availability of carbon, the so-called can be dissolved. High air temperature at the water carbon stock.* The more carbon present in the soil surface also gives a  large temperature difference and in flooded biomass or transported into the reser- between surface and deeper water, which favors voir from upstream rivers, the more likely GHGs will stratification. Thus, temperature affects both the be emitted. Because the presence of carbon in soil production and emission of CO2 and CH4 from and biomass decreases as it is transformed into GHGs reservoirs. and released into the atmosphere, the rate of emis- Water quality and nutrient content (eutrophication sions normally exponentially decreases with age of status) also have a large effect on the concentration the reservoir. of dissolved oxygen in reservoirs. The poorer the quality of inflowing water (e.g., high content of nutri- Another major factor, which is especially important ents and organic matter), the higher the oxygen for determining the type of GHG produced and thus demand created in the reservoir, favoring anoxic the warming potential of CO2 equivalents, is the dis- conditions and methane production. For this reason, solved oxygen* concentration in the water of the reser- the land cover and land use in the upstream catch- voir. In reservoirs, seasonal stratification* due to ment areas affect the GHG emissions from reservoirs. temperature differences between surface and deeper Anthropogenic* sources of pollution such as efflu- water and poor vertical mixing may produce anoxic ents of untreated domestic and industrial sewage can conditions in the deeper, colder water. The water have a particularly large impact on GHG emissions depth and the stratification of the water column into from reservoirs. an anoxic zone (hypolimnion*) below the aerobic zone (epilimnion*) have a large impact on the emission of Similarly, inflows and the shape of the reservoir affect GHGs. Dissolved CH4 is produced in anoxic conditions the level and distribution of dissolved oxygen in the and can be oxidized to CO2 in aerobic conditions. If reservoir. The inflow and bathymetry influence the anoxic conditions exist in most of the water column, it water retention time* in the different parts of the res- allows fluxes of CH4 to the atmosphere. The greater ervoir. Water retention time in turn affects how much the thickness of the overlying epilimnion layer in the time is available for biological processes to occur. water column, the less likely it is that diffuse CH4 The volume and variation of inflows also affect how emissions will be produced. This is because the pro- much oxygen is transported into the reservoir and how duction area and volume become smaller while the well the inflowing fresh water is mixed with the water oxidizing area and volume (where CH4 can transform already present in the reservoir. to CO2) are enlarged. Water depth and extension of littoral zone are important The water and air temperatures have been found to for the amount of methane that can be transferred generally show the highest correlation with mea- directly from the bed sediment to the atmosphere sured GHG emissions. This is because higher tem- through bubbling. Bubbling is more likely to happen in perature affects many of the processes that contribute shallow waters, since solubility increases with pres- to higher emissions. First, temperature directly sure. At greater depths, the pressure is high and the influences the decomposition rate of organic matter, CH4 is more likely to be dissolved following its Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 9 creation. Because CO2 has much higher solubility than FIGURE 2.4. Measured GHG Emission Fluxes at CH4, bubbling of CO2 is low even in shallow waters. Reservoirs Mixing of surface and deeper water and stratifica- Observed GHG uxes (mg C per m2 and day) tion are also affected by wind speed. Furthermore, 10,000 wind speed at the surface affects the diffuse fluxes 1,000 of CO2 and CH4 from the water phase to the atmo- 100 sphere. Higher wind speed increases fluxes by 10 increasing the turbulence of the water at the air-­ 1 water interface. 0.1 0.01 Because of the sudden decrease in pressure when Di ussive Bubbling Degassing Di usive water is released from a low-level (deep) outlet of a CH4 CH4 CH4 CO2 dam, the solubility of gases will drastically decrease Source: From data by UNESCO/IHA (2017), based on values stated in and dissolved CH4 in particular may be degassed just published articles, recalculated to mg C per m2 total reservoir area and downstream of the reservoir. Even if oxygen is avail- day for comparison. Note: mg C = milligrams of carbon; CO2 = carbon dioxide; CH4 = methane. able in this environment, the depth and time avail- The boxes denote the first (25%), median, and third (75%) quartiles, able for oxidation is short, enabling CH4 to be while the whiskers denote the minimum and maximum values. transferred directly to the atmosphere. Therefore, the configuration of dam intake and outlets, especially seen in figure 2.4. If carbon burying is included as their position in relation to the thermocline depth in part of sediment deposition, the amount of CO2 emit- the reservoir, affects total GHG emissions. Other ted from reservoirs has in some cases even been infrastructure features, such as artificial aeration found to be negative over the period of the measure- weirs in the downstream river stretch, may also affect ment campaign.2 the ratio of CO2 or CH4 and thus the total GHG emis- sions expressed as CO2 equivalents. GHG emissions from new aquatic systems created by reservoirs will also change over the long term as the flooded organic material is decomposed and biochem- 2.4 Extreme Temporal and Spatial Variation ical conditions change. Upon inundation, easily decomposed organic matter starts decaying, causing All the above factors affect how the GHG stock is cre- high gross emissions during the initial phase. As this ated and released into the atmosphere. They interact matter is depleted, gross emission rates will increas- in a complicated manner to govern the biological pro- ingly depend on the amount of newly decaying mate- cesses such as organic matter production, respira- rial being transported into the reservoir by inflowing tion, methanogenesis*, CH4 oxidation, and gas rivers. exchange between the atmosphere and the reservoir. As a result, GHG emissions vary widely in time and Measured gross emissions have high spatial variabil- space. CH4 emissions, in particular, may vary ity within a reservoir (figure 2.5) and show large dif- extremely; this has a major impact on methane’s ferences between different reservoirs—in terms of warming potential, as reflected by the fact that the total emissions, type of GHG, and pathways by which factor 34 has to be applied to arrive at its CO2 equiva- the GHG is emitted. Measured emissions also show lent. Measurements of GHG emissions from reser- high seasonal variation and generally a decreasing voirs differ by several orders of magnitude, as can be trend with age of the reservoirs, the highest values 10 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes FIGURE 2.5. Measurement of the GHG Fluxes in the Batang Ai Reservoir, Malaysia, Showing the Extreme Spatial Variability a. Surface pCH4 (ppm) b. Surface pCO2 (ppm) 2,500 1,200 1,000 2,000 800 1,500 600 1,000 400 500 200 0 0 Source: University of Quebec at Montreal. Note: p = partial pressure; ppm = parts per million. FIGURE 2.6. Measurement of Diffuse Fluxes of CH4 in TABLE 2.1. Measurements of N2O and Comparison of the Petit Saut Reservoir (Impounded in 1994) CO2 Equivalents with Diffusive CH4 and CO2 In CO2 equivalents (mg CO2eq 700 mg N2O-N per m2 and day) 600 576 per m2 Di use CH4 ux (mg C / Diffusive and daya N 2 Oa CO2b 500 CH4a m2 and day) 400 Min –0.089 –27 1.6 85 300 25% 0.031 9.5 119 913 200 Median 0.057 17 275 1,585 100 35 34 6 75% 0.133 40 774 3,336 0 1 9 10 19 Max 5.768 1,718 26,134 20,014 Age of reservoir (years) Sources: Deemer et al. 2016 and UNESCO/IHA. Note: mg = milligrams; CH4 = methane; CO2 = carbon dioxide; Source: Based on data from UNESCO/IHA. Original data from CO2eq = carbon equivalents; N2O = nitrous oxide. Galy-Lacaux et al. 1997; Abril et al. 2005; Guérin et al. 2006; and a. Data from Deemer et al. 2016. Negative values mean the reservoir Cailleaud 2015. works as a sink. Note: mg C = milligrams of carbon; CH4 = methane. b. Data from UNESCO/IHA. being registered in the first 5–15 years after inunda- Sturm et al. 2014, and Guérin et al. 2008). Deemer et al. tion (figure 2.6). (2016) estimated that it only accounted for 4 percent of total global reservoir emissions. N2O emissions have been studied in a limited number of reservoirs so far. The results indicate that, similar to Notes methane, N2O creation and emission vary greatly 1. See appendix A for conversion from GHG units into CO2 equivalents. (table 2.1). In  general, N2O contributes less to GHG emissions from reservoirs than CO2 and CH4, even 2. See Sikar 2009 and Chanudet et al. 2011 for examples of reservoirs that have been reported as carbon sinks. However, whether carbon when expressed in CO2eq based on application of burial can be subtracted from atmospheric emissions is still being the  high warming potential (Descloux et al. 2017, debated (see Prairie et al. 2017). Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 11 Chapter 3 What Does The Atmosphere See? 3.1 Gross and Net Fluxes of A recent scientific paper by Prairie et al. (2017) high- Greenhouse Gasess lighted the importance of the concept of “what the atmosphere sees,” as the result of building new A fundamental concept for accurately describing dams. The paper was co-authored by 14 experts rep- GHG emissions from reservoirs created by biogeo- resenting 12 prominent global research centers on chemical processes is the difference in gross1 and net GHG emissions from reservoirs. Referring to recent fluxes. Rivers are major conveyors of carbon from publications showing that global emissions from terrestrial areas to lakes and the sea (figure 3.1). natural freshwater systems are larger than previ- Terrestrial areas are generally net carbon sinks* and ously estimated (e.g., Raymond et al. 2013), it argues aquatic systems are net carbon emitters. Changes in that only 25 percent of gross CO2 emissions measured GHG fluxes to the atmosphere because of the intro- at reservoirs are visible to the atmosphere; the duction of reservoirs in a river system must there- remaining 75 percent of emissions are simply dis- fore be viewed from a catchment perspective. Net placed and would have been emitted anyway, in the GHG emissions created by the reservoir are the dif- absence of the reservoirs. Prairie et al. (2017) also ference between total fluxes of CO2 equivalents for emphasize the need to estimate pre-impoundment the river basin before and after the reservoir has emissions to enable the calculation of the net GHG been constructed. emissions that are “visible” to the atmosphere. Reservoirs are one of many anthropogenic influences Accounting for pre-impoundment emissions can on the biogeochemical GHG cycle in a river basin. either increase the net value (e.g., in the case of Sources of carbon for a reservoir are normally both inundating a forest area previously working as a car- natural (e.g., soil and vegetation) and anthropogenic bon sink) or decrease the net value (e.g., if a dam (e.g., inflow of organic matter from untreated sewage floods vast wetlands or raises a natural lake). or flooded waste deposits). Changes in GHG fluxes due What matters to the atmosphere from the introduc- to the introduction of a reservoir must therefore also tion of a reservoir is, therefore, limited to the pro- be considered in the context of artificial influences cesses where the changes create a net increase in CO2 already in place in the river catchment. equivalent fluxes. It is important to note that the net In 2011, the Intergovernmental Panel on Climate emissions described above refer only to the biochemi- Change (IPCC) defined biogeochemically generated cal processes in the river system that affect the GHG net emissions from reservoirs as gross emissions fluxes to the atmosphere. In a complete life-cycle minus pre-impoundment emissions and minus unre- emissions assessment for a project involving a reser- lated anthropogenic sources (UAS). This definition voir, the baseline must be set according to alternative was also adopted by the International Energy Agency future scenarios (such as when hydropower replaces in their program on Managing the Carbon Balance in thermal power) and include emissions related to the Freshwater Reservoirs in 2012 (IEA Hydro 2012). implementation of the entire project (among other This is also the definition used in this technical note things, emissions deriving from the construction and the one recommended be adopted by the WBG. works themselves). Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 13 FIGURE 3.1. The Global Carbon Cycle Source: IPCC 2013. Note: 1 PgC = 1 Petagram of Carbon (= 1015 grams of carbon). 3.2 The G-res Tool emissions as defined by IPCC (2011). The objective of the tool is to “quantify the portion of GHG emissions that can Following the framework proposed by Prairie et al. be legitimately attributed to the creation of the reservoir (2017), the UNESCO/IHA research project of 2015–17 over its lifetime.” developed the GHG Reservoir Tool (hereafter called the G-res tool) (UNESCO/IHA 2017). The G-res tool builds on Average pre-impoundment emissions for the inundated the principles of the global carbon cycle* (figure 3.1) as area are calculated from land cover, soil, and climate. defined by IPCC (2013) and the definition of net reservoir If the area to be inundated works as a carbon sink, the 14 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes pre-impoundment emissions are negative. Unrelated emissions and the sum of pre-impoundment emis- anthropogenic sources are estimated based on land sions and UAS. As defined by IPCC, the annual emis- use, population, and known point sources in the sions are integrated over a 100-year period to estimate catchment area. the life-cycle emissions attributed to the creation of the reservoir. Annual post-impoundment emissions are estimated through carefully developed statistical models relating The G-res tool also enables the user to estimate the GHG emissions to key governing variables such as tem- construction- and material-related emissions for the perature, age of reservoir, littoral area, solar radiance, dam infrastructure by applying standard emission phosphorus concentration in the reservoir, and soil factors. These project-specific emissions are added to carbon content. Models are developed for different the reservoir emissions to arrive at an emissions esti- gases and pathways—diffusive CO2 flux, diffusive CH4 mate for the entire life cycle of the dam and reservoir. flux, bubbling of CH4, and degassing of CH4. The statis- The G-res tool goes further to suggest how the life- tical models are derived based on the measured gross cycle emissions should be allocated by sector in the emissions from 223 reservoirs studied over the last case of multipurpose dams and reservoirs. The alloca- 25 years (see section 2.2). Data have been standardized, tion is based on the operating regime of the reservoir, accounting for the different periods of the year during that is, on what uses are prioritized. which the measurements were made. All CH4 emis- The G-res tool is used through a web-based interface sions are attributed to the new reservoir, while attrib- and is available online.2 It is linked to global geo- utable CO2 emissions are reduced by the emissions that graphic databases to enable default estimations of are simply displaced by the new reservoir (figure 3.2). variables such as climate zone, land cover, and soil The factors used to determine the volume of displaced types. The user can introduce more detailed data, if emissions are the availability of the carbon stock in the available from primary data collection or studies, to reservoir bed soil and the shape of the exponential improve the GHG emission estimation. Use of decline in gross CO2 emissions. the  G-res tool is not very time-consuming—it can The annual net emissions of both CO2 and CH4 are cal- even require less than one day if all input data culated as the difference between post-impoundment are available. FIGURE 3.2. Displacement of CO2 when a Dam is Constructed in a River Pre-impoundment Post-impoundment Natural emissoins Reservoir emissoins Downstream emissoins Source: IHA/UNESCO 2017. Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 15 3.3 The IEA Hydro Framework of mechanistic models to describe the biogeochemical processes to create and emit CO2 and CH4 from a Based on the cumulative research, and acknowledging reservoir. the great complexity of estimating GHG emissions from reservoirs, the IEA Hydropower Implementing The recommendations given in the IEA Hydro frame- Agreement has developed guidelines for the quantita- 3 work are sourced from the experience of engineers, tive analysis of net GHG emissions from reservoirs. scientists, and academics, as well as experts from Three documents provide a framework for conducting the hydropower industry. The recommendations also site-specific primary data collection and modeling to build on process-based modeling applications for res- estimate and manage reservoir GHG emissions (see ervoirs, such as the one used in Nam Theun 2, in Lao Bibliography): PDR (Chanudet et al. 2016). Process-based modeling of water quality and GHG emissions from reservoirs is • Volume 1 – Measurement Programs and Data still a relatively young science, and therefore few appli- Analysis (October 2012) cations exist. Besides Nam Theun 2, a process-based model was recently applied to Eastman 1, in Canada, to • Volume 2 – Modeling (November 2015) simulate CO2 emissions (Kim et al. 2016), and modeling • Volume 3 – Management, Mitigation and Allocation is ongoing in Petit Saut, in French Guiana.4 (Draft, June 2017) The time frame for conducting primary data analysis Like the G-res tool, the IEA Hydro framework builds on and developing site-specific models is considerable— the principles of net emissions as defined by IPCC normally at least 1–2 years. The primary data collection (2011). While it does not provide a ready-to-use tool for usually needs to cover several seasons. The modeling estimating emissions, it describes in detail the steps involves various components, such as hydrodynamic involved in collecting field data, conducting data anal- modeling, water quality modeling, and GHG modeling, ysis, and developing process-based modeling tools for which are all very complex and time-consuming to set estimation of GHG emissions from a reservoir. up, calibrate, and validate. The framework is based on data collection and the five 3.4 Global Estimates of GHG Emissions components of a new reservoir project (based on EPRI from Reservoirs 2010): (i) the inundated area; (ii) the reservoir; (iii) the upstream catchment area; (iv) the reservoir outflow Several studies have been published in the last decades facilities; and (v) the downstream river. It provides a aimed at estimating the total contribution of GHGs list of environmental and technical descriptors that from reservoirs to global emissions (table 3.1). The first should be reported in an analysis of GHG emissions estimate of global emissions from reservoirs in the from a reservoir. year 2000 indicated a very high volume, in the order of 7 percent of global GHG emissions from all sources In the case of new reservoirs, it recommends proce- (St. Louis et al. 2000).5 The typical methodology used dures for primary data collection and suggestions for by these studies is to take the average specific emis- integrating these into pre-impoundment emissions. sions from observations (see sections 2.2. and 2.4) and For post-impoundment emissions, the framework pro- extrapolate these, in common units such as mg per m2 vides suggestions and requirements for modeling and year, to a global reservoir surface area. approaches to simulate GHG emissions over the life span of the reservoir. Moreover, it provides recom- One major source of uncertainty in these estimates is mendations on the setup, calibration, and validation the total area of (man-made) reservoirs. Because of 16 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes TABLE 3.1. Allocation of GHG Emissions for Multipurpose Reservoirs Followed with G-res Tool Importance Explicit prioritization Operating rule curve Primary Ranked 1 to 3 in operational hierarchy. Operating rules are designed to maximize these service ben- efits for part or all of the year. Secondary Ranked lower than 3 in operational hierarchy, or places The service places operational constraints on the operating constraints on operation. level of the reservoir for part or all of the year. Tertiary Provides benefits, but does not alter the operation of The service provides benefits but has little impact on the the reservoir. operation of the reservoir. Importance Apportionment (%) Notes Primary 80 If there is more than one service in the level, split equally. Secondary 15 Where there are no secondary services, the apportionment (15%) is split between the primary services. Tertiary 5 Where there are no tertiary services, the apportionment (5%) is split between the secondary services. Source: UNESCO/IHA 2017. Note: G-res = GHG Reservoir. incomplete records and the fact that the area varies excludes the relatively large number of small dams with the seasons, it is difficult to estimate the total area (those with a smaller than 0.1 km3 storage capacity). By of freshwater lakes and reservoirs. It is estimated that statistically extrapolating the distribution of dams, lakes cover a global area of 3.7–4.2 million km 2 Lehner et al. (2011) estimated the missing net addi- (Downing et al. 2006). In this context, a breakthrough tional water surface created by these dams6 to be was achieved for reservoirs with the Global Reservoir 306,723 km2. and Dam (GRanD) Database developed by the Global Starting from the framework proposed by Prairie Water System Project (Lehner et al. 2011). The GRanD is et al. (2017) for estimating the emissions that are a geographic database that contains many variables on attributable to reservoirs, Prairie (2017) applied the 6,862 reservoirs with more than 0.1 km3 of storage G-res tool individually to all the reservoirs in the capacity worldwide. It includes nearly all dammed res- GRanD database, excluding the largest natural lakes,7 ervoirs in the world (map 3.1). to estimate the global net reservoir emissions. To get an estimate of the net additional surface area The initial value obtained was further corrected to created by dams, the GRanD database had to be account for the high number of small reservoirs. adjusted somewhat. The database includes large natu- The results, presented in table 3.2, indicate that the ral lakes such as Lakes Victoria, Baikal, Winnipeg, global emissions from reservoirs are lower than pre- Onega, Vanern, Ontario, and Saima. These seven large viously estimated—in the order of 0.5 percent of lakes alone add up to 163,000 km2, or 36 percent of the global anthropogenic emissions. total area of 452,000 km2 covered by the GRanD data- base. As they are all partially regulated, they are con- Two factors in particular explain why the estimates sidered reservoirs. However, the regulation of these by Prairie (2017) are lower than previous global esti- lakes causes an increase in surface area that is insig- mates. Firstly, applying the principle to only include nificant compared to their previous natural state. emissions attributable to the reservoir means that Thus, these changes do not contribute to increases in displaced emissions8 are not included. This resulted the net surface area. On the other hand, the GRanD in a lower net carbon dioxide emissions from Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 17 MAP 3.1. Distribution of Reservoirs in the GRanD Database Legend Reservoirs Sources: Lehner et al. 2011; IHA/UNESCO. TABLE 3.2. Global Estimates of Reservoir Area and GHG Emissions Global reservoir area CO2 CH4a Global contribu- Reference (km2) (Tg CO2eq/year) (Tg CO2eq/year) tionb St. Louis et al. (2000) 1,500,000 1,000 2380 9.2% Lehner et al. (2011) 306,000 — — — Barros et al. (2011) 500,000 176 680 2.4% Bastviken et al. (2011) 340,000 — 136 — Hertwich (2013) 330,000 279 331 1.7% Deemer et al. (2016) 311,000 135 606 2.0% Prairie (2017) 350,000 45 105–164 c 0.4–0.6% Note: — = not available; n.a. = not applicable; CH4 = methane; CO2 = carbon dioxide; CO2eq = carbon dioxide equivalent; Tg = teragram (= 1 million metric tons). a. Using a warming potential of 34, as defined in the 2013 IPCC Fifth Assessment Report. b. Using a global estimate of anthropogenic GHG of 36.2 billion tons/year as of year 2015. c. High value if assuming CH4 degassing in all dams, low value if degassing is not included. Degassing only occurs if the dam outlet is located at low levels of the reservoir. Since the GRanD database contains no data on the location of the dam outlet, it is not possible to determine at which dams degassing occurs. reservoirs than previously estimated (figure 3.3). gives lower global emissions. This is because the Secondly, using site-specific factors to govern CH4 reservoirs where methane emissions have been mea- emissions for each dam in the GRanD, rather than sured are not fully representative of the global extrapolating average measured emissions from a distribution of reservoirs. Furthermore, because relatively few reservoirs to a total global surface area, distribution of CH4 emissions is highly skewed, 18 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes FIGURE 3.3. Illustration of Changing Distribution of CO2 Emissions for All Reservoirs over the World When Considering Net Rather than Gross Values 1,600 1,400 1,200 CO2 emissions (gCO2 m–2 yr–1) 1,000 800 600 400 200 0 Gross Integrated over 100 yrs Attributable Net Source: IHA/UNESCO. Note: CO2 = carbon dioxide. Previous global estimates CO2 emission considered the gross values. The reservoirs used for this figure have been taken from the GRanD database. individual high values in a small sample may have a lakes. Assuming that the Lehner et al. value is more very large effect on the average. reliable would give about 10 percent lower global GHG estimates. Furthermore, Prairie et al. do not consider The estimate by Prairie (2017) still has uncertainty UAS, as they argue that although emissions from UAS and is affected by the assumptions made. For example, are not directly attributable to the reservoir, they are it only includes emissions of CO2 and CH4, but ignores anthropogenic and thus “seen by the atmosphere.” N2O, for which data are still scarce. The limited According to verbal communication with UNESCO/ research on N2O has, however, indicated that nitrous IHA, applying the strict IPCC (2011) definition of net oxide emissions are generally small compared to the emissions would decrease the global estimate by emissions of CO2 and CH4. Deemer et al. (2016) esti- 10–15 percent. mated that N2O accounts for 4 percent of total reser- voir GHG emissions. On the other hand, their estimate The G-res tool used for the estimation of reservoir GHG of global reservoir area is higher than the estimate by emissions by Prairie (2017) is also associated with uncer- Lehner et al. (2011), which indicates that some of the tainties. The relationships between emissions and reservoirs in the GRanD database to which the G-res ­ governing variables are still based on a limited number tool has been applied are actually regulated natural of  observations, especially for CH4 bubbling and Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 19 degassing. The statistical models, however, are most Another reason is that it is difficult to correctly assign res- reliable for the center of the data distribution, which ervoir area to power produced because most reservoirs gives some robustness to the global average estimate. are multipurpose and a single reservoir often serves many hydropower plants in a cascade or complex trans- As figure 3.1 illustrates, the natural emissions from fer scheme. For example, assigning all reservoir emis- rivers and lakes form a large part of the carbon cycle. sions to a relatively small hydropower unit installed at the Raymond et al. (2013) estimate the total annual emis- spillway of an irrigation dam would misrepresent the role sions from streams, rivers, lakes, and reservoirs to be of power production as a driver for reservoir creation. 2,100 TgC (teragrams of carbon). The results from Prairie (2017) indicate that reservoir emissions are only a frac- Based on the work by UNESCO/IHA, the G-res tool was tion (less than 1 percent) of total emissions from fresh- recently applied (IHA 2017) to a global database of sin- waters bodies. Previous estimates (Barros et al. 2011) gle-purpose hydropower projects where the installed indicated a higher value of 4 percent. capacity and energy production had been verified and proven to be linked to the reservoir specifically created Estimating the emissions from hydropower reservoirs for the project (map 3.2). The database includes 180 and their relation to the power produced at the global hydropower projects with installed capacity ranging level is more difficult. One reason for this is that from 1.2 to 2,735 MW and reservoir areas ranging from detailed information on installed capacity (MW)—power 1.4 to 5,400 km2. The resulting GHG emissions from the generated (GWh/year) in particular—is not available for G-res tool are plotted against the power density (W/m2) reservoirs in global databases (such as GRanD). the ­ in figure 3.4. MAP 3.2. Single-Purpose Hydropower Projects with Verified Installed Capacity and Power Production Legend Hydropower projects Source: IHA. 20 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes FIGURE 3.4. GHG Emissions Estimated through the G-res Tool for a Global Set of Single-Purpose Hydropower Projects 10,000.0 1,000.0 Hydropower emmissions intensity 100.0 (gCO2/kWh) 10.0 1.0 0.1 0.01 0.10 1.00 10.00 100.00 1,000.00 Power density (W/m2) 20 18 16 Power density (W/m2) 14 12 10 8 6 4 2 0 0 50 100 150 200 250 300 350 400 Allocated GHG emissions intensity (gCO2/kWh) Boreal Temperate Subtropical Tropical 5 W/m2 100 gCO2/kWh Source: Data courtesy of IHA. Note: W/m2 = watt per square meter; gCO2/kWh = grams of carbon dioxide per kilowatt-hour. Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 21 The results indicate a very strong logarithmic Notes relationship between emissions and power density, 1. Gross emissions from reservoirs are defined as the emissions mea- although the envelope curves show that GHG emis- sured from the reservoir surface and the immediate river stretch sions may vary with more than one order of magnitude downstream of the reservoir; gross emissions are normally the ones measured (see sections 2.2 and 2.4). for a specific power density. Figure 3.4 indicates that 2. It was launched on May 10, 2017, and available at www.hydropower​ extremely high GHG emissions per kWh (same order .org/gres-tool. of magnitude as fossil fuel plants) can be produced by 3. The IEA Hydropower Implementing Agreement is a working group of a hydropower storage project, be it only at low power the International Energy Agency member countries and others that densities—that is, where installed turbine capacity is have a common interest in advancing hydropower worldwide. Under relatively small compared to the surface area created Annex XII (Hydropower and the Environment), this program has conducted a project on Managing the Carbon Balance in Freshwater by the reservoir. However, figure 3.4 also shows that a Reservoirs. low power density does not necessarily entail high 4. Ongoing work by the Environmental Defense Fund (EDF), verbal emissions, as most hydropower projects below 5 W/m2 communication V. Chanudet, May 2017. still have emission profiles below 100 g CO2eq/kWh, 5. The figure of 7 percent was derived on the basis of a lower GWP for which is considerably lower than any fossil fuel methane than is used today (21 instead of 34), and a total amount of alternative. global emissions of 33.9 billion tons/year. If a GWP of 34 is applied to the 2015 year global emissions, the resulting figure would be Figure 3.4 also confirms that there is no obvious rela- 9 percent, as indicated in table 3.2. tionship between the climate zone and reservoir emis- 6. Estimated 2.8 million impoundments larger than 0.1 ha worldwide, sions. In theory, temperature affects GHG emissions, 16.7 million impoundments when including dams larger than 0.01 ha. but this result indicates that temperature is only one of 7. Only the obvious large natural lakes were removed, which probably many parameters affecting the resulting emissions. explains the larger surface area estimated by Prairie (2017) and Lehner et al. (2011). Thus, great care should be taken before using just one factor to predict reservoir emissions. 8. These would have occurred even without the creation of the reservoir. 22 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes Chapter 4 Recommendations for Preparing Dam Projects 4.1 General Framework to be generous in the initial screening phase and only characterize GHG as negligible in cases where it is very In the case of dam infrastructure projects inundating obvious (as in the case of true run-of-river projects terrestrial landscapes for which the WBG may provide and retrofitting where the inundated area will not financing, it is recommended that the biochemically change). The next step is to apply the G-res tool using generated GHG emissions from the reservoirs be stud- secondary data to simulate future reservoir emissions ied as part of the Environmental and Social Impact for the planned investment. Considering the uncer- Assessment (ESIA). Doing studies of potential GHG tainty still remaining when using the G-res tool (see emissions from reservoirs as part of the ESIA enables section 3.4), it is essential to assess the reliability of comparison of alternative design options within the the results obtained with the G-res tool. If the esti- framework of the investment project and provides mate is deemed reliable, it can be used as input for the inputs for the economic analysis. economic analysis and for suggesting suitable options The inclusion of a GHG assessment in the ESIA does for managing the reservoir’s GHG emissions; this not change the boundaries of a regular impact assess- information should be included in the Environmental ment. It can be anticipated that changes to GHG emis- and Social Management Plan (ESMP). On the other sions will occur at the reservoir area and the river hand, if the results are attributed low reliability, it is stretch downstream of the reservoir. Thus, inputs for recommended that more detailed assessments be the GHG assessment include data from the catchment conducted, including primary data collection and area and the project area, which are normally also modeling according to the IEA Hydro framework. needed for the ESIA. Each of these steps is described in more detail in the following sections. The recommended framework for the GHG assessment is described in figure 4.1. Because GHG emissions have 4.2 Documentation and Initial Screening been shown to vary by several orders of magnitude for different reservoirs, it is prudent to follow a stepwise The first step for all infrastructure projects with inun- process in which the complexity of the analysis is dations is to provide an overall description of the main adjusted to reflect how severe and important GHG factors affecting future, potential GHG emissions from emissions are for the specific investment case. the planned reservoir options. Table 4.1 gives a list of standard information that is useful to compile. The first important step to take, required for all new dam investment projects, is proper documentation of The next step is to make a qualitative assessment of GHG aspects of the project. This step enables initial the reservoir’s capacity to supply a carbon stock and screening and provides input data for further ­analysis. to create and release different types of GHGs, based on The recent development of the G-res tool, which is the compiled information. A structured process as open-source and fairly easy to use, makes it possible described in figure 4.2 can be used. Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 23 FIGURE 4.1. Recommended Approach for Assessing GHG Emissions for New Dams Documentation and initial Yes Assume negligible reservoir emissions screening - is GHG likely to be Document in appraisal documents negligible? No Application of the G-res tool Con dence (UNESCO/IHA), including in results quality control Document in appraisal documents. Include GHG Low con dence reservoir emissions in economic analysis. If signi cant emissions, in results consider management and monitoring of GHG to be included in ESMP Data collection and process- based modelling following IEA Hydro framework TABLE 4.1. Basic Information to Assess Future Potential GHG from a Reservoir Factor to retrieve Proposed methodology Size and shape of inundated This information is available from the technical prefeasibility and feasibility studies. If not available, use area and volume of reservoir existing topographical maps or Digital Elevation Models, for instance, SRTM (https://www2.jpl.nasa.gov/ srtm/), to delineate inundated area up to planned full supply level.a When natural lakes are used as res- ervoirs, estimate how much new inundated area will be created by the damming. Calculate surface areas, volumes, and maximum and average depth. Climate, temperature, and These data are normally available from the first phases of the ESIA. If not available, use global databases rainfall in reservoir area such as the WorldClim (www.worldclim.org). River inflow to the reservoir These data are available from the technical prefeasibility and feasibility studies. If not available, use and water retention time records of river flows upstream from the reservoir to estimate inflow (see, e.g., www.bafg.de/GRDC/ EN/01_GRDC/grdc_node.html). Divide average annual inflows by the reservoir volume to get the average retention time. Type and extent of flooded This information is normally available from the first phases of the ESIA. If not available, use global maps of vegetation ecoregions (e.g., European Space Agency www.esa-landcover-cci.org). Type of flooded soil and This information is normally available from the first phases of the ESIA. If not available, use global data- extent of soil carbon bases such as the Harmonized World Soil Database by FAO, IIASA, ISRIC, ISSCAS, and JRC (http://webar- chive.iiasa.ac.at/Research/LUC/External-World-soil-database/HTML). table continues next page 24 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes TABLE 4.1. Basic Information to Assess Future Potential GHG from a Reservoir (continued) Factor to retrieve Proposed methodology Land cover and use in catch- This information is normally available from the first phases of the ESIA. If not available, use population den- ment area and water quality sity (http://sedac.ciesin.columbia.edu/data/collection/gpw-v4) together with global land cover maps (see of inflowing rivers above) to conduct an overall assessment of land use and potential sources of organic matter and nutrients. Characteristics of dam and This information is normally available from the technical prefeasibility and feasibility studies. Use infor- construction methods mation on civil works to describe the type of dam and provide a rough estimate of dam, excavation, and material volumes, as well as transport distances during construction. In the case of hydropower, informa- tion is needed on installed capacity and plant factor, which are used to calculate power density (W/m2). Retrieve basic geometry of the dam and intakes to assess how far from the reservoir bed level the bottom outlet and intake structures are located. Note: ESIA = Environmental and Social Impact Assessment; IIASA = International Institute for Applied Systems Analysis; FAO = Food and Agriculture Organization; ISRIC = International Soil Reference and Information Centre (World Soil Information); ISSCAS = Institute of Soil Science, Chinese Academy of Sciences; JRC = Joint Research Centre (of the European Commission); SRTM = Shuttle Radar Topography Mission. a. The full supply level corresponds to the normal maximum operating water level of a water storage body when not affected by floods; it represents 100% capacity. FIGURE 4.2. Proposed Thinking Process for Documentation and Initial Screening Ability to supply carbon stock Yes Ability to create GHG E.g. • Large reservoir area Yes • High percentage of forests Ability to release • High density of GHG vegetation E.g. • High carbon content in • High temperature soil Yes • Long retention time • Strati cation of reservoir • Large in ow of nutrients Potentially signi cant from upstream areas GHG emissions E.g. • Shallow reservoir • Anoxic conditions in most of the water column • Intake for downstream releases in anoxic zone Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 25 The compiled data and qualitative assessment into account. A few recommendations are given in should be used for the initial screening. The main table 4.2. purpose of this activity is to screen out any dam infrastructure that is likely to cause negligible GHG Since the G-res tool is fairly simple to use, it is best to emissions and where no further studies are required. be conservative in the screening. If it is difficult to Due to the extreme variability of GHG emissions qualitatively assess the risk of significant GHG emis- from reservoirs, no simple, single threshold can be sions, it is prudent to go to the next step and apply the used. Instead, a number of variables need to be taken G-res tool. If the information has been collected and TABLE 4.2. Guidance for Screening Out Projects with Negligible Reservoir GHG Emissions Hydropower projects The counterfactuals to hydropower project investments are normally other forms of power generation or energy demand management programs. The initial screening can therefore focus on the relative difference between potential reservoir emissions and the emission fac- tors of the likely counterfactual. Using the strong relation to power density (figure 3.4), an early estimate can be made based on the pro- posed installed capacity and the estimated reservoir surface area. • Irrespective of other factors, if the power density is higher than 100 W/m2, which would be the case in most run-of-river projects, the global data indicate that reservoir emissions are normally below 1 g CO2/kWh, and even in extreme cases below 10 g CO2/kWh. Compared to most counterfactuals for power production, this is relatively low (e.g., fossil fuel emission are in the order of 300–1,000 g CO2/kwh), and reservoir emissions can be assumed negligible since they would be within the error margin of the emissions of the counterfactual. • If factors clearly disfavor high GHG emissions (such as cold climate, low carbon stock, deep reservoir), which would indicate that extreme emissions are unlikely, the upper envelope curve does not need to be considered and a lower power density threshold can be used to assume negligible reservoir emissions. For instance, a power density of 20 W/m2 indicates a median reservoir emission of about 5 g CO2/kWh. • Should the counterfactual have negligible emissions and the power density lie below 100 W/m2, it is suggested that the threshold be set by the size of the reservoir. Using a threshold of 100,000 tonsa CO2eq (or 1,000 tons/year), as is used for other dam infrastructure, seems reasonable (see below). An analysis using the relation in figure 3.4 for different installed capacities (from 0.1 to 100 MW) shows that the resulting reservoir area threshold is fairly stable and varies only from 2.5–3.5 km2 during median conditions, and from 0.5–0.7 km2 under extremely favorable conditions for high GHG emissions (upper envelope). Other dam infrastructure projects For other dam infrastructure projects (e.g., water supply, flood control, irrigation), there is no obvious counterfactual and the reservoir emissions related to the project are generally assigned as a “cost” to the project. In this situation, the screening has to be based on the size of the reservoir and how significant the emissions may be relative to emissions caused by other sources. The construction emissions for large dam infrastructure lie in the order of 100,000–1,000,000 tons CO2eq. It is thus suggested that 100,000 tons CO2eq be used as the threshold when reservoir emissions may be negligible. This threshold corresponds to 1,000 tons CO2eq per year, which is equivalent to 0.03 PPM (parts per million) of the global anthropogenic emissions. Based on the distribution of measured gross emissions (figure 2.4), this threshold can be used to give default threshold values for reservoir surface areas delineating when GHG emissions can be assumed negligible: • If factors seem to favor high specific emissions, use the 75th percentile for total CO2 and CH4 emissions (~3,500 tons/km2 and year). A threshold of 100,000 tons for the lifetime emissions would correspond to a surface area of 0.3 km2. • If factors seem to favor low specific emissions, use the 25th percentile for total CO2 and CH4 emissions (~500 tons/km2 and year). A threshold of 100,000 tons for the lifetime emissions would correspond to a surface area of 2 km2. • If factors affecting specific emissions are mixed, use the median for total CO2 and CH4 emissions (~1,000 tons/km2 and year). A threshold of 100,000 tons for the lifetime emissions would correspond to a surface area of 1 km2. a. All tons refer to metric tons (or “tonnes”). 26 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes analyzed qualitatively, little additional time will be and free to use. The site includes technical documenta- required to enter input data and run the G-res tool. tion on the scientific basis for the tool and a guide for its step-by-step use (figure 4.3). The results of the screening should be reported as a subchapter of the ESIA. For the project appraisal docu- The G-res tool needs the following inputs: ment (PAD), a shorter version can be included. • Upstream catchment Appendix C gives a template for the presentation of input data and screening results in the PAD. • Catchment area (required) • Catchment annual runoff (required) 4.3 Quantitative Assessment of Net Emissions Using the G-res Tool • Land cover (required) If potentially significant GHG emissions cannot be dis- • Information on intensity of land use counted through the above initial screening assessment, it is recommended that the G-res tool be applied, for • Population in catchment area (required) which input is derived from secondary data. The G-res • Potential point sources and general level of waste tool is available online (www.hydropower.org/gres-tool) water treatment FIGURE 4.3. The Interface of the Online G-res Tool UNESCO/IHA GHG RESEARCH PROJECT G-res Tool Warning: Please never refresh the page with the Reload Page button of the browser. This web page will disconnect automatically after 30 minutes of inactivity. The G-res Tool works only with the following supported browsers : Safari 9.x, Chrome 48 or later, Microsoft Edge 25 or later. fil Export to .txt file Restart Analysis with a New Save Input Parameters Reservoir Rese ir Reservoir Name Technical pport Support Paramet s Import Saved Parameters Printable Reports Introduction Catchment Reservoir Reservoir services Construction GHG UAS Reservoir GHG Total GHG footprint Emission Factors Earth Engine GHG Reservoir Screening Test—Introduction Technical We strongly recommend to download and read the complete guidance here: User guide Begin Data Input > Document The GHG Reservoir Screening Tool (G-res Tool) provides an estimation of the level of net GHG footprint (CH4 and CO2) from freshwater existing and future reservoir through the following equation: Net GHG Footprint = [Post-Flooding Emissions] – [Pre-Flooding Emissions] – [Emissions from Unrelated Anthropogenic Sources (UAS)] The tool also includes the emissions from the construction phase. It also allocates the total emissions to the different purpose of the reservoir. To use the tool, click the “Begin Data Input” button. You will be directed to answer questions about the catchment, the reservoir, the purpose of the reservoir and the construction phase through a series of input tabs. The tool presents a summary results page including all emission sources and detailed calculation sheets which provides more information on some module. If you are missing data about some aspect of your reservoir or its catchment, you may extract it by following the instructions in the “Earth Engine” tab. Earth Engine Using this spreadsheet Cell Key: Cells where the user MUST input data for the calculations. Cells where the user may input data for the calculations. Cells that are calculated automatically by the model. Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 27 • Area to be inundated by reservoir tool to ensure that the quality of the input data is ­reasonable. It is equally essential that the G-res user be • Reservoir area (required) duly trained in the tool and have a basic understand- • Climate zone (required) ing of GHG emissions from reservoirs. Because of the nonlinearity and the existence of thresholds in the • Mean monthly temperature (required) underlying functions used to estimate reservoir emis- • Average wind speed (required) sions, the user’s ability to conduct sensitivity analyses • Mean global horizontal radiance (required) for key variables and assess the reasonability of the subresults (such as the different gases and pathways) • Soil type and soil carbon content (required) is important to understand and gauge the robustness • Land cover (required) of the resulting emissions. • Information on intensity of land use 4.4 Assessment of Reliability in • Reservoir Estimated Emissions • Reservoir volume The uncertainty associated with the G-res tool to quantify GHG emissions from a reservoir should be • Max and mean depth (required) noted. Because of the use of different data, the reli- • Planned normal operation level ability of the underlying models differs for different pathways. The estimated diffusive CO2 and CH4 have • Planned water intake elevation the highest reliability, while CH4 bubbling and degas- These data are generally available in feasibility and sing have the lowest reliability. The uncertainty also environmental impact studies or can be found in global increases toward the tails of the distribution, such as databases. The G-res tool offers possibilities to esti- for reservoirs with very high emissions. By contrast, mate the geographic data for the catchment and reser- estimates are more reliable when close to the median voir areas through open source Geographic Information (e.g., between the 25th and 75th percentile).1 The user System (GIS) software, which is linked to the latest must therefore be trained to note when there are global databases. An example of the application of the uncertainties in subtotals and overall estimates of G-res tool is given in appendix B. GHG emissions. Entering the above data into G-res will give the user an The user interface of the G-res tool makes it relatively estimate of reservoir emissions. It is also possible to easy to conduct a sensitivity analysis to understand include information on the dam construction (excava- what input parameters have the largest effect on the tion, cement and steel volumes, transport distances), estimated net emissions. If those input parameters are which would give a preliminary estimate of construc- uncertain, it is also an indication that the estimated tion emissions. The G-res tool further allows defining emissions have low reliability. ­ econdary, multipurpose uses of the reservoir (primary, s It is also important to put the G-res results into and tertiary use) and will allocate the total GHG emis- perspective—in terms of how they would affect the ­ sions to these different uses. intended investment. This requires assessing whether The G-res tool interface is fairly easy to use. However, the estimated level of emissions from the reservoir it is essential that a practitioner with experience in the would significantly contribute to GHG emissions and field of dam development and construction use this whether the emissions’ order of magnitude is such that 28 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes they would have a significant impact on the economic used for comparison are those based on the availability return of the project. If the latter case applies, the of carbon stock, which give an idea of possible net emissions may need to be assessed further to get more emissions.2 One such method is the methodology pro- reliable results. As in the case of the initial screening, posed by the Interim Technical Note on Greenhouse the GHG emissions estimated using the G-res tool for Gases from Reservoirs Caused by Biogeochemical the proposed investment should be compared with the Processes by the World Bank (2013), which uses the emissions caused by the counterfactual. It is therefore 2006 IPCC Guidelines for National Greenhouse Gas essential to identify the most likely counterfactual and Accounting together with assumptions on ratios roughly estimate its emissions before comparing it between CO2 and CH4 production to estimate the reser- with the G-res estimate for the proposed investment voir emissions. In the case of hydropower, median project. In the case of hydropower, this process is fairly emissions using the World Bank (2013) methodology straightforward, as the average emission factors per are available in table format for different power densi- produced kWh are well-known for the power genera- ties and plant factors.3 tion alternatives. If the G-res tool gives similar or Based on the above methods, the user needs to make a higher emissions per kWh than the counterfactual— qualitative judgment about the reliability of the reser- thereby highlighting the risk that the project may not voir emissions estimated with the G-res tool. Any pre- be a mitigation project—a high-reliability estimate is sentation of the G-res tool estimate should be desired, which may require further studies. accompanied with transparently acknowledging the An alternative to the above approach is to conduct a uncertainty associated with the results. rough economic analysis for the investment with and The results of the G-res tool, and the reliability assess- without a shadow carbon price. Moreover, such an ment, should be reported as a subchapter in the ESIA, analysis should not only be conducted for the reservoir or as a concise dedicated report. Sources and major emissions estimated with the G-res tool, but also for an assumptions for all input data should be included as interval that illustrates the uncertainty in this value, well as subresults such as the contribution from differ- for example, ±25 percent and ±50 percent. If the analy- ent GHGs and pathways. The reliability of the results sis shows that the economic cost of carbon emissions should not only be presented but also commented on. may be significant compared to the total cost of the For the PADs, a shorter summary of the input data and project, this suggests a highly reliable estimate is results is acceptable. A template for the presentation of desired and further studies may be required. the results of the reservoir GHG emission assessment Conversely, if the economic analysis indicates that is given in appendix C. even in the scenario of 50 percent higher emissions the impact on the project’s economic return is minimal, 4.5 Detailed Assessment Following the G-res tool estimate may suffice. the IEA Hydro Framework If significant reservoir emissions are indicated, esti- If the reliability assessment indicates a need for fur- mates derived through other means—to see if the ther assessment of reservoir GHG emissions, and if ­ values converge—should be considered. This is partic- the time (6 months to 2 years) and resources (order of ularly relevant if the main sources of the high reservoir magnitude of $50,000–$500,000) are acceptable emissions as indicated by the G-res tool are CH4 bub- considering the cost of the total investment, plans bling and degassing, and if the specific emissions lie should be drawn up for primary data collection above the 75 percentile. Suitable methods that can be th and modeling. Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 29 If the main reason for the unreliable estimation of res- The methods, data, and results of such a detailed ervoir emissions is the absence of or uncertainty in key assessment should be described in a dedicated, input data for the GHG emissions estimate, the pri- detailed report. It is envisioned that such a report mary data collection should focus on closing those would be a subreport of the ESIA. data gaps. The following types of field measurements should be conducted: (i) topographical surveys to con- 4.6 Management of GHGs and Post- firm maximum and medium depths, and littoral area Impoundment Monitoring extent, of the proposed reservoir; (ii) surveys to con- In the case of dam infrastructure projects for which firm vegetation, land cover, and land use in the reser- potentially significant GHG emissions have been esti- voir area; (iii) plot tests to confirm soil carbon content mated, possible mitigation measures should be con- in flooded area; (iv) water quality sampling to confirm sidered and specified in the Environmental and Social estimated organic material and nutrient concentra- Management Plan (ESMP). IEA Hydro (2017) provides a tions in inflowing water; and (v) climatological mea- general framework for managing and mitigating GHG surements to confirm solar radiation and wind speed. emissions. More specifically, it proposes that a detailed Verified data can be used to update the G-res tool to get GHG Management Plan (a subplan of the ESMP) be a more reliable estimation of the reservoir emissions. ­ prepared—including proposed mitigation actions and On the other hand, if the uncertainty is associated specific targets—for projects where net reservoir GHG with the G-res tool itself (e.g., due to very high levels emissions are estimated to be significant. The GHG of CH4 bubbling or degassing) or if the requirement for Management Plan should also include monitoring and reliability is high, more comprehensive primary data regular reporting of GHGs emitted as well as the miti- compilation is advised—to be able to estimate pre-­ gation measures applied. impoundment emissions and provide detailed input IEA Hydro (2017) gives a framework for systematically process-based data for the application of physical, ­ assessing possible mitigation measures for the five modeling for estimation of post-inundation emis- stages of project development: sions. This comprehensive data compilation should include additional, detailed bathymetric surveys, • Project planning and design; high-resolution climate data collection, and extensive water quality and soil sampling, as well as direct mea- • Project implementation (construction and reservoir impoundment); surement of pre-impoundment emissions. Based on these data, hydrodynamic and biogeochemical mod- • Dam, power plant, and reservoir operation, includ- els should be set up and calibrated to estimate water ing contributions of UAS; quality variables and reservoir emissions after impoundment. • Catchment management, including contributions of UAS; and Primary data compilation and modeling should be based on the guidelines and requirements of the IEA • Downstream management. Hydro framework and be informed by the experience GHG emissions management can, for example, and lessons learned from previous data measurements include infrastructure design. Measures to increase and modeling exercises (see Bibliography). Detailed oxygen concentrations upstream of intakes in Nam assessments would probably require the procurement Theun 2 have been shown to decrease emissions of of a dedicated team of experts in reservoir GHG emis- CH4 degassing (Deshmukh et al. 2016) downstream of sions to conduct data collection and modeling. the outlet. Keeping intakes above the level of the 30 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes hypolimnion is another mitigation measure that may Such a monitoring program should also be included in have large potential to decrease methane emissions. the GHG Management Plan and could have the follow- During operation, one possible mitigation measure is ing characteristics: to avoid a rapid drawdown, as this favors CH4 bub- bling. Further, although not considered part of the • Focuses on gross emissions of CO 2 and CH4 as mea- sured by surface floating chambers;4 net emissions, UAS can contribute to a high volume of reservoir GHGs but this is to a large extent manage- • Covers the reservoir along its longitudinal axis as able. Organic material and nutrient effluents from the well as the immediate river stretch downstream of upstream catchment can also be managed and cov- the reservoir; ered in the catchment treatment plans included in • Includes sampling of CH 4 concentrations upstream the ESMPs. Management of UAS reduces gross GHG and downstream of the intake/outlet; emissions and improves the water quality of the res- ervoirs, an aspect that has both recreational and O&M • Covers a period of at least 3 years and includes sea- benefits. Should mitigation measures be difficult sonal measurements. to  implement, offsetting reservoir GHG emissions through certified emission reductions (CER) or other Notes carbon credits is a further possibility to consider. 1. The G-res tool interface has a feature to show where the estimated GHG emissions (per m2) for each pathway of the studied reservoir are Moreover, in the case of dam infrastructure projects situated in the distribution curve of all dams in the GRanD database with potentially significant GHG emissions, it is rec- (with more than 6,500 reservoirs). This helps the user determine ommended that the post-implementation monitoring whether the results are extreme. of GHG emissions from the reservoir and immediate 2. Assuming that inflowing carbon is only displaced by the introduction river stretch downstream be streamlined. It should be of the reservoir, the amount of carbon stored in soil and vegetation in the flooded area represents the upper limit of the net emissions cre- emphasized that the purpose of this monitoring is not ated by the reservoir (see also World Bank 2013). to provide an accurate estimate of net GHG emissions 3. See Guidance Manual: Greenhouse Gas Accounting for Energy Investment but rather to allow a rough comparison of gross emis- Operations, Transmission and Distribution Projects, Power Generation sions with the estimates calculated during the prepa- Projects, and Energy-Efficiency Projects, Version 2.0, January 2015. ration phase and to monitor any changes (e.g., changes 4. See, for instance, UNESCO/IHA (2010) for a description of monitoring as a result of mitigation measures). methods. Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 31 Chapter 5 Future Research The research on GHG emissions from freshwater lakes The development of predictive, online models (such and reservoirs has made significant progress over the as the G-res tool) represent a major milestone in the last decade. However, data are still scarce and unevenly estimation of reservoir GHG emissions. As future data distributed, and the understanding of the complex and research are channeled toward these tools, the processes involved is still incomplete. Though there is practitioners in dams and hydropower will get access general agreement on the main factors affecting GHG to more reliable tools. UNESCO/IHA is in the process emissions from reservoirs, the statistical models of the of operationalizing the G-res tool so that it is duly G-res tool only consider a few of these factors signifi- maintained and updated as new research becomes cant for the prediction of emissions. This focus proba- available.1 In addition, the increased focus on reduc- bly reflects the scarcity of data (which allow statistically ing GHG emissions is expected to result in stakehold- significant improvements of the results) rather than the ers agreeing on what actually comprises significant assumption that the variables not taken into account emissions. are unimportant. Continued research is, therefore, It is believed that the recommendations given in essential to enhance knowledge and groundtruth new chapter 4 of this technical note will remain relevant for predictive models, such as the G-res, through measure- the most part, even as today’s predictive tools are ment campaigns. The IEA Hydro guidelines should also updated and improved, and agreed thresholds be supported by examples of their use. are introduced. The state of the art will certainly It is anticipated that more and more data will become change and global estimates of reservoir emissions available from measurements. Remote sensing of will be revised accordingly. It is important that WBG GHG fluxes using satellites is one method that could staff working in the field of hydropower and dam infra- substantially improve data collection over large reser- structure development keep up-to-date on the devel- voir areas and considerably increase the amount of opments in the area of GHG emissions from reservoirs available data. New data and new insights will allow and ensure that the latest research findings and tools the statistical models of the G-res tool to be updated are applied to WBG investments. and will be incorporated into similar predictive tools to be developed for practitioners. Process-based mod- eling is likewise expected to advance as more research Note and more powerful computers become available. 1. Verbal communication, Yves Prairie and Richard Taylor, May 2017. Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 33 Appendix A Conversion of GHG Units and CO2 Equivalents Conversion from Moles to G CO2 Equivalents In chemistry, a mole is considered Avogadro’s number The international practice is to express GHG emissions (6.02 × 10 ) of molecules (or anything) of a substance— 23 in CO2 equivalents (CO2eq or CO2e). Emissions of gases in other words, depending on the density of the sub- other than CO2 are converted into CO2eq by multiplying stance, the mass of that amount of the substance could their respective volumes by their respective Global vary widely. To convert from moles to grams you must Warming Potentials (GWPs). From the 2013 IPCC first find the molar mass of the element or compound. Report: Use the periodic table to read off the atomic mass of an GWP relative to CO2 at different time horizons for the element. If it is a compound, you must know the most common GHGs in reservoirs: molecular formula, and then you find the total molar mass of the compound by adding up the atomic masses of each atom in the compound. The unit of the molar GWP for given time Chemical Gas name horizon mass will be in grams per moles (g/mole). Once you formula 20-yr 100-yr have the molar mass, you can easily convert from Carbon dioxide CO2 1 1 grams to moles, and also from moles to grams. Methane CH4 86 34 Number of moles = (# of grams) ÷ (molar mass) Nitrous oxide N2O 268 298 Number of grams = (# of moles) × (molar mass) Source: 2013 IPCC Fifth Assessment Report (AR5). Conversion table for the most common GHGs in reservoirs: Conversion from “G of GHG” to “G of Carbon” The conversion between “g of GHG” and “g of carbon” is Element Atomic mass (g/mole) directly related to the ratio of the atomic mass of a GHG N 14.0067 molecule to the atomic mass of a carbon atom. Essentially, C 12.0107 this practice accounts for the carbon in the GHG mole- O 15.9994 cule, as opposed to counting the entire molecule. H 1.00794 For carbon dioxide, the ratio of the atomic mass of a CO2 molecule to the atomic mass of a carbon atom is 44:12. GHG Molar mass (g/mole) • To convert from “g of C” to “g of CO2” multiply CO2 44.0095 by 44/12. CH4 16.0107 • To convert from “g of CO2” to “g of C” multiply N2O 44.0128 by 12/44. Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 35 • Sometimes you find this noted as gC-CO 2 or tC-CO2 Carbon Dioxide Equivalents Vs. Carbon (to make clear that these “g of C” refer to carbon in Equivalents a CO2 molecule). While the international standard is to express emis- For methane, the ratio of the atomic mass of a CH4 sions in CO2 equivalents (CO2eq), many U.S. sources molecule to the atomic mass of a carbon atom is have expressed emissions data in terms of carbon 16:12. equivalents (CE) in the past. In particular, the United States Environmental Protection Agency (US EPA) has • To convert from “g of C” to “g of CH ” multiply 4 used the carbon equivalent metric in the past for bud- by 16/12. get documents. • To convert from “g of CH ” to “g of C” multiply 4 For the purposes of national GHG inventories, emissions by 12/16. are expressed as teragrams of CO2 equivalent (Tg CO2eq). • It is important to make clear that these grams of C One teragram is equal to 1012 grams, or 1 million tons. refer to carbon in a CH4 molecule (i.e., NOT CO2eq— in other words, not taking into account GWP). It is • To convert from CE to CO eq, multiply by 44/12. 2 common to use gC-CH4 or tC-CH4. • To convert from CO eq to CE, multiply by 12/44. 2 36 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes Appendix B Example of G-res Application Project Name Trung Son • Soil: Harmonized World Soil Database, FAO/IIASA/ ISRIC/ISSCAS/JRC (via Earth Engine linked to G-res) Country Vietnam Installed Capacity 260 MW • Temperature: Hijmans et al. 2005, Global Climate Yearly Generation 1,019 GWh/year database (via Earth Engine linked to G-res) Climate Tropical Reservoir area 13.1 km2 • Solar radiation: NASA - SSE 2008 (via Earth Engine linked to G-res) Sources of information: • Wind speed: GLOBE task team, NOAA (via Earth • Project information and hydrology: Supplementary Engine linked to G-res) Environmental and Social Impact Assessment, • Population density: Center for International Earth Trung Son Hydropower Project, 2009 Science Information Network, Columbia University • Land cover: ESA-CCI (via Earth Engine linked to G-res) (via Earth Engine linked to G-res) UNESCO/IHA GHG RESEARCH PROJECT G-res Tool Warning: Please never refresh the page with the Reload Page button of the browser. This web page will disconnect automatically after 30 minutes of inactivity. The G-res Tool works only with the following supported browsers : Safari 9.x, Chrome 48 or later, Microsoft Edge 25 or later. Export to .txt fil file Restart Analysis with a New Save Input Parameters Rese ir Reservoir Reservoir Name Technical Support pport Import Saved Paramet s Parameters Printable Reports Introduction Catchment Reservoir Reservoir services Construction GHG UAS Reservoir GHG Total GHG footprint Emission Factors Earth Engine GHG Reservoir Screening Test—Introduction Technical We strongly recommend to download and read the complete guidance here: User guide Begin Data Input > Document The GHG Reservoir Screening Tool (G-res Tool) provides an estimation of the level of net GHG footprint (CH4 and CO2) from freshwater existing and future reservoir through the following equation: Net GHG Footprint = [Post-Flooding Emissions] – [Pre-Flooding Emissions] – [Emissions from Unrelated Anthropogenic Sources (UAS)] The tool also includes the emissions from the construction phase. It also allocates the total emissions to the different purpose of the reservoir. To use the tool, click the “Begin Data Input” button. You will be directed to answer questions about the catchment, the reservoir, the purpose of the reservoir and the construction phase through a series of input tabs. The tool presents a summary results page including all emission sources and detailed calculation sheets which provides more information on some module. If you are missing data about some aspect of your reservoir or its catchment, you may extract it by following the instructions in the “Earth Engine” tab. Earth Engine Using this spreadsheet Cell Key: Cells where the user MUST input data for the calculations. Cells where the user may input data for the calculations. Cells that are calculated automatically by the model. Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 37 UNESCO/IHA GHG RESEARCH PROJECT G-res Tool Warning: Please never refresh the page with the Reload Page button of the browser. This web page will disconnect automatically after 30 minutes of inactivity. The G-res Tool works only with the following supported browsers : Safari 9.x, Chrome 48 or later, Microsoft Edge 25 or later. Export to .txt file Restart Analysis with a New p Parameters Save Input Reservoi Reservoir Reservoir Name Trung Son Technical Support Import Saved Parameters p Printable Reports Introduction Catchment Reservoir Reservoir services Construction GHG UAS Reservoir GHG Total GHG footprint Emission Factors Earth Engine Online Technical Current Totals tCO2e/yr Input Page 1/4—Catchment Data Document for Catchment Post-Impoundment 4643 On this sheet, enter the data on the land cover types in the catchment area and the reservoir area. Pre-Impoundment –97 UAS 177 Catchment Area (km2) 14660 User Guidelines Population in the Catchment 483323 The user should select land cover data Catchment Annual Runoff (mm/yr) 551 based on the most appropriate and relevant Community Wastewater Treatment—Please Select Primary data for the reservoir and catchment area. Where land cover categories differ with the Release of phosphorus from industrial sewage in the categories presented in the G-res Tool, the catchment (kg P/yr) 0 user should rationalize the data being used Industrial Wastewater Treatment—Please Select Primary into the same categories and check that the emission factors used in the G-res Tool are Land Cover in the Catchment Area Land Use Intensity applicable to those land cover types. Low = Unmanaged Land “Intensity” is used to describe the level of Please choose units of inputs: % High = Managed Land human in uence on the land use as part % km 2 Past Current of the UAS module. Broadly this means whether for agriculture and forest it is heavily Croplands 7% 1026.2 Low Low managed land, and for urban area whether the population density is high. Sensitivity Bare Areas 0% 0 analysis is encouraged. Wetlands 0% 0 Forest 40 % 5864 Low Low Grassland/Shrubland 53 % 7769.8 Low Low User Notices Permanent Snow/Ice 0% 0 WARNING—Please be sure to add 0% to all land cover with no Settlements 0% 0 Low Low value. Water Bodies 0% 0 Drained Peatlands 0% 0 Reset Catchment Land Cover No Data 0% 0 Pre-Impoundment Land Cover in the Reservoir Area Reservoir Area (km2) 13.1 % of Organic Past Land Use Intensity Soil that is Low = Unmanaged Land % Mineral Soil % Organic Soil Drained High = Managed Land % km2 Croplands 3.6 % 0% ( ) Low 3.6 % 1 Bare Areas 0% 0% ( ) 0.0 % 0 Wetlands 0% 0% ( ) 0.0 % 0 Forest 5.3 % 0% ( ) Low 5.3 % 1 Grassland/Shrubland 91 % 0% ( ) Low 91.0 % 12 Permanent Snow/Ice 0% 0% ( ) 0.0 % 0 Settlements 0% 0% ( ) Low 0.0 % 0 River Area before Impoundment 0.1 % ( ) 0.1 % 0 Drained Peatlands 0% 0.0% 0 No Data 0% Edit Emission Factors Reset Reservoir Land Cover Net Input Page 38 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes UNESCO/IHA GHG RESEARCH PROJECT G-res Tool Warning: Please never refresh the page with the Reload Page button of the browser. This web page will disconnect automatically after 30 minutes of inactivity. The G-res Tool works only with the following supported browsers : Safari 9.x, Chrome 48 or later, Microsoft Edge 25 or later. Export to .txt file Restart Analysis with a New p Parameters Save Input Reservoir Reservoi r Reservoir Name Trung Son n Technical Support Import Saved Parameters Printable Reports Introduction Catchment Reservoir Reservoir services Construction GHG UAS Reservoir GHG Total GHG footprint Emission Factors Earth Engine Online Technical Current Totals tCO2e/yr Input Page 2/4—Reservoir Data Document for Reservoir Post-Impoundment 4643 On this sheet, enter the key parameters that describe the reservoir. Pre-Impoundment –97 UAS 177 Country Vietnam User Guidelines Longitude of Dam (DD) 104.76 Project speci c information should be used. This may be obtained from current Latitude of Dam (DD) 20.59 operations or from feasibility studies. Climate Zone (Reservoir Area) Tropical For reservoirs that are expected to exhibit uctuations in certain parameters depending Impoundment Year 2016 on season or operating regime, the user should determine the ‘typical’ values and Reservoir Area (km2) 13.1 then undertake a sensitivity analysis to determine whether those variations affect Reservoir Volume (km2) 0.345 **To reset to automatically calculate value, the overall result. press the reset button ( ) associated. () Mean/Normal Operating Level (m above sea level) 159 1) If Reservoir Area and Volume are () available. Mean Depth will be calculated. Maximum Depth (m) 50 () 2) If Mean and Maximum Depth are Mean Depth (m)1 26.336 available, % Littoral Area will be calculated. () Littoral Area (%)2 5.408 3) If Reservoir Area, Maximum Depth, Mean () Depth, Annual Wind Speed and Monthly Thermocline Depth (m)3 0.5 Temperature are available, Thermocline Depth will be calculated. Water Intake Depth (m)4 14 4) If Mean/Normal Operating Level and Water Intake Elevation (m above sea level) 145 Water Intake Elevation are available, Water Intake Depth will be calculated. Soil Carbon Content Under Impounded Area (kgC/m ) 2 3.4 () 5) If Reservoir Area, Mean Depth, Runoff and Annual Wind Speed at 10 m (m/s) 1.7 Catchment Area are available, WRT will be () calculated. Water Residence Time (WRT, yrs)5 0.0427 () 6) If Reservoir Runoff and Catchment are Annual Discharge from the Reservoir (m/s)6 256.1 available, Discharge will be calculated. Phosphorus Concentration (µg/L) 7 26.4 7) If Catchment Land Cover, WRT, Runoff, Catchment Area and Population are Trophic Level Mesotrophic available, Phosphorus Concentration will be calculated. Reservoir Mean Global Horizontal Radiance (kWh/m2/d) 3.8 Mean Temperature per Month (°C) January 16.4 February 18.1 March 20.4 April 23.9 May 26.9 June 28 July 27.6 August 27.5 September 26.5 October 24.2 Next Input Page November 21.5 December 18 Mean Annual Air Temperature (°C) 23.3 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 39 UNESCO/IHA GHG RESEARCH PROJECT G-res Tool Warning: Please never refresh the page with the Reload Page button of the browser. This web page will disconnect automatically after 30 minutes of inactivity. The G-res Tool works only with the following supported browsers : Safari 9.x, Chrome 48 or later, Microsoft Edge 25 or later. Export to .txt file Restart Analysis with a New Save Input Parameters Reservoir Reservoi Reservoir Name Trung Son n Technical Support Import Saved Parameters Printable Reports Introduction Catchment Reservoir Reservoir services Construction GHG UAS Reservoir GHG Total GHG footprint Emission Factors Earth Engine Online Technical Document Input Page 3/4—Reservoir Services Data for Reservoir Services Many reservoirs provide multiple services. In order to assess the GHG emissions associated with each of the services, it is necessary to allocate a proportion of the total emissions across the relevant services. Allocation of Reservoir Purposes Percentage Allocation Flood Control Tertiary 5 Fisheries 0 Irrigation 0 Navigation 0 Environmental Flow 0 Recreation 0 Water Supply 0 Hydroelectricity Primary 95 Please indicate which allocation method was used to determine the importance of the services: Operating Rule Curve Please explain if another method was used: The de nitions of primary, secondary and tertiary services for these options are provided in the table below. For more information on the allocation method that can be used to determine the importance of the services (Explicit Prioritisation or Operating Rule Curve), please see the user guideline. Importance Explicit Prioritisation Operating Rule Curve Primary Ranked 1 to 3 in operational hierarchy. Operating rules are designed to maximize the bene ts of this service for part or all of the year. Secondary Ranked lower than 3 in operational hierarchy, The service places operational constraints on the operating level of the reservoir for part or or places constraints on operation. the whole of the year. Tertiary Does not alter the operation of the reservoir. The service has little impact on the operation of the reservoir. 40 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes UNESCO/IHA GHG RESEARCH PROJECT G-res Tool Warning: Please never refresh the page with the Reload Page button of the browser. This web page will disconnect automatically after 30 minutes of inactivity. The G-res Tool works only with the following supported browsers : Safari 9.x, Chrome 48 or later, Microsoft Edge 25 or later. Export to .txt file Restart Analysis with a New Save Input Parameters Reservoir Reservoir Name T Trung Son S Technical Support Import Saved Parameters p Printable Reports Introduction Catchment Reservoir Reservoir services Construction GHG UAS Reservoir GHG Total GHG footprint Emission Factors Earth Engine Online Technical Document Input Page 4/4—Construction Data for Construction If available, please input information that describes the amount of materials used in the construction phase. You can include your own value, or use the simple or more detailed parameters below. Please note that numbers included in each section will be added together. If the scheme used to 10,000 m3 of concrete, you only need to include it in the basic assessment or the more detailed assessment. Including it in both could lead to double counting. For transport, it is assumed that the delivery is by road so please include the last part of the journey to site, i.e., after any shipping. Own Assessment Emission output If you have undertaken your own assessment of GHG emissions associated with your scheme, you can include that value here. kgCO2e Known Value for Total Construction emissions Construction Basic Assessment These are the basic materials likely to make up a signi cant part of the construction phase GHG emissions. Earth and Rockfill Material excavated and/or used for construction 470000 m3 2 km moved 10685967 Concrete All concrete brought to site for the dam, tunnels, 960000 m3 80 km delivery distance 357709440 foundations Steel All steel brought to site for reinforcement, 11100 tonne 80 km delivery distance 29936522 pipelines, mechanical and electrical equipments More Detailed Assessment This provides a more detailed list of typical materials used. Use these values if you have more detailed information about the types of material used on the scheme. Earthworks Soft Excavation m3 km moved 0 Rock Excavation m3 km moved 0 Clearance and Removals ha 0 Fill Granular Fill m3 km delivery distance 0 Rock Armour m3 km delivery distance 0 Zoned Rock ll m3 km delivery distance 0 Rock bolts number km delivery distance 0 Concrete Works Formwork m3 0 Facing Concrete m3 km delivery distance 0 Mass Concrete m3 km delivery distance 0 Reinforced Concrete m3 km delivery distance 0 Shotcrete m3 km delivery distance 0 Reinforcement tonne km delivery distance 0 Steelworks Steel Penstocks tonne km delivery distance 0 Steel Liner tonne km delivery distance 0 Miscellaneous Steelwork tonne km delivery distance 0 Roads and Bridges New Roads km 0 Refurbishment of Existing Roads km 0 PCC Vehicular Bridge Deck m3 0 Equipment Power Generation MW 0 Power Connection kV km length 0 Please record any assumptions, limitations and data sources here: To UAS Summary Page Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 41 UNESCO/IHA GHG RESEARCH PROJECT G-res Tool Warning: Please never refresh the page with the Reload Page button of the browser. This web page will disconnect automatically after 30 minutes of inactivity. The G-res Tool works only with the following supported browsers : Safari 9.x, Chrome 48 or later, Microsoft Edge 25 or later. Export to .txt file Restart Analysis with a New Save Input Parameters Reservoir Reservoi r Reservoir Name Trung Son S n Technical Support Support s Import Saved Parameters Printable Reports Introduction Catchment Reservoir Reservoir services Construction GHG UAS Reservoir GHG Total GHG footprint Emission Factors Earth Engine Online Technical Unrelated Anthropogenic Sources Emissions Document for UAS Calculations of Phosphorus Loads Total P in the reservoir (µg/L) 26.4 Reference Level of P (µg/L) 25.7 User Notices P from industrial sewage (µg/L) 0.0 Trophic state: P from human sewage (µg/L) 0.0 If the reservoir is oligotrophic, no UAS contribution is identi ed P from human land use (µg/L) 1.5 P over Reference Land (µg/L) 0.7 Share of UAS of the P in reservoir, evaluated as P (µg/L) 1.5 Estimated Contribution of UAS to the GHG Emissions from the Reservoir Calculated CH4 emissions from the reservoir (gCO2e/m3/yr) 108.5 Amount of CH4 of total estimates due to UAS (%) 6% Estimated CH4 release due to UAS (gCO2e/m3/yr) 13.5 Weighted sum model risk 66 Sensitivity to Nutrient Load Comment on Risk Factor for the GHG Emissions: Climate Tropical High Climatic Sensitivity Water residence time (yrs) 0.0 Low to Moderate Sensitivity Share of Anthropogenic Impact % of total UAS emissions UAS emissions from Land Use (gCO2e/m3/yr) 13.5 100% Cropland Low to Moderate Risk Forestry Low to Moderate Risk Grasslands/Pasture Low to Moderate Risk Settlements Low to Moderate Risk UAS Emissions from Sewage (gCO2e/m3/yr) 0.0 0% Community sewage 0% Low to Moderate Risk Industrial sewage 0% Low to Moderate Risk 42 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes UNESCO/IHA GHG RESEARCH PROJECT G-res Tool Warning: Please never refresh the page with the Reload Page button of the browser. This web page will disconnect automatically after 30 minutes of inactivity. The G-res Tool works only with the following supported browsers : Safari 9.x, Chrome 48 or later, Microsoft Edge 25 or later. Export to .txt file Restart Analysis with a New Save Input Parameters Reservoi Reservoir Reservoir Name Trung Son Technical Support Import Saved Parameters Printable Reports Introduction Catchment Reservoir Reservoir services Construction GHG UAS Reservoir GHG Total GHG footprint Emission Factors Earth Engine Online Technical Document Predicted Emissions for Reservoir GHG Net Predicted Annual CO2e Emission Unrelated Post-Impoundment – Pre-Impoundment – Anthropogenic = Net GHG Footprint Sources Emission Rate (tCO2e/yr) 4 643 – –97 – 177 = 4 563 of which CO2 1 537 – –97 n/a = 1 634 of which CH4 3 106 – 0 – 177 = 2 929 Emission Rate (gCO2e/m2yr) 354 – –7 – 14 = 348 of which CO2 117 – –7 n/a = 125 of which CH4 237 – 0 – 14 = 224 Percentile of Net GHG emissions within the database 56% 0 25 50 75 100 0 25 50 75 100 Relative contribution to CH4 Post-Impoundment Emissions Percentile of each CH4 contribution within the database Fraction of CH4 diffusive ux from Total Reservoir CH4 Emission (%) 46% 59% 0 25 50 75 100 0 25 50 75 100 Fraction of Degassing of CH4 from Total Reservoir CH4 Emission (%) 48% 10% 0 25 50 75 100 User Notice 0 25 50 75 100 WARNING Deep Water Intake. Likely Important Contribution of Degassing. Fraction of Bubbling of CH4 from Total Reservoir CH4 Emission (%) 7% 20% 0 25 50 75 100 0 25 50 75 100 Unrelated Anthropogenic Sources Potential amount of UAS as % of post-Impoundment emissions 6% Weighted sum total model risk result 66 To Total GHG Summary Page Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 43 UNESCO/IHA GHG RESEARCH PROJECT G-res Tool Warning: Please never refresh the page with the Reload Page button of the browser. This web page will disconnect automatically after 30 minutes of inactivity. The G-res Tool works only with the following supported browsers : Safari 9.x, Chrome 48 or later, Microsoft Edge 25 or later. Export to .txt file Restart Analysis with a New Save Input Parameters Reservoir Reservoi Reservoir Name Trung Son Technical Support p S Import aved Parameters Saved p Printable Reports Introduction Catchment Reservoir Reservoir services Construction GHG UAS Reservoir GHG Total GHG footprint Emission Factors Earth Engine Online Technical Document Predicted Emissions for Total GHG footprint Net Predicted Annual CO2e Emission Unrelated Post-Impoundment – Pre-Impoundment – Anthropogenic + Construction = Net GHG Footprint Sources (Reservoir) Areal Emission (gCO2e/m2/yr) 354 – –7 – 14 + n/a = 348 Reservoir Wide Emissions (tCO2e/yr) 4 643 – –97 – 177 + 3 983 = 8 546 Total Lifetime Emission (tCO2e) 464 339 – –9 683 – 17 706 + 398 332 = 854 648 Net GHG Emissions Contribution for Each Reservoir Services GHG Emissions GHG Emissions GHG from Reservoir from Construction Footprint Percentage Reservoir Service (tCO2e/yr) (tCO2e/yr) (tCO2e/yr) Allocation (%) Flood Control 228 199 427 5 Fisheries 0 0 0 0 Irrigation 0 0 0 0 Navigation 0 0 0 0 Environmental Flow 0 0 0 0 Recreation 0 0 0 0 Water Supply 0 0 0 0 Hydroelectricity 4335 3784 8119 95 Allocation Method Used: Operating Rule Curve Emission Factor Used: Default Emission Factors Used Construction Comments: 44 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes UNESCO/IHA GHG RESEARCH PROJECT G-res Tool GHG Results Report Name of reservoir: Trung Son Back to G-REs Tool Back to Reports Reservoir GHG Information Net Predicted Annual CO2e Emission Unrelated Post- Pre- Anthropogenic Net GHG Impoundment – Impoundment – Sources = Footprint Emission Rate (tCO2e/yr) 4 643 – –97 – 177 = 4 563 of which CO2 1 537 – –97 n/a = 1 634 of which CH4 3 106 – 0 – 177 = 2 929 Emission Rate (gCO2e/m2/yr) 354 – –7 – 14 = 348 of which CO2 117 – –7 n/a = 125 of which CH4 237 – 0 – 14 = 224 Percentile of Net GHG emissions within the database 56% Relative contribution to CH4 Post-Impoundment Emissions Fraction of CH4 diffusive ux from Total Reservoir CH4 Emission (%) 46% Fraction of Degassing of CH4 from Total Reservoir CH4 Emission (%) 48% Note: WARNING Deep Water Intake. Likely Important Contribution of Degassing. Fraction of Bubbling of CH4 from Total Reservoir CH4 Emission (%) 7% Unrelated Anthropogenic Sources Potential amount of UAS as % of post-impoundment emissions 6% Weighted sum model risk result 66 Total GHG footprint information Unrelated Post- Pre- Anthropogenic Construction Net GHG Impoundment – Impoundment – Sources + (Reservoir) = Footprint Areal Emissions (gCO2e/m2/yr) 354 – –7 – 14 + n/a = 348 Reservoir Wide Emissions (tCO2e/yr) 4 643 – –97 – 177 + 3 983 = 8 546 Total Lifetime Emission (tCO2e) 464 339 – –9 683 – 17 706 + 398 332 = 854 648 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 45 Appendix C Suggested Template for GHG Reporting The following page shows an example of a template that can be included in appraisal documents for hydro- power and dam infrastructure projects. By way of illustration, the template has been filled out based on  the application of the G-res tool to Trung Son (see appendix B). General Project Description Project name: Trung Son, Vietnam Purpose: Main purpose: Hydropower. Project also has flood control benefits. Reservoir Area (km ): 2 13.1 Climate: Tropical Vegetation: Mixed forest and shrubland If hydropower Installed capacity (MW) 260 Annual energy (GWh) 1019 Power Density (W/m ) 2 19.8 Initial screening Likely counterfactual: Because of the high growth in electricity demand in Vietnam, if Trung Son would not be built, it is likely that power would be produced through a mix of gas and coal powered plants. These would produce emissions in the order of 450−900 g CO2/kWh Gross emissions applying 25−75th percentile of global measurements (t CO2eq/year): 6,000−46,000 If hydropower Power density below 100 W/m ? 2 Yes Global envelope for Power density (g CO2/kWh): 1−40 Conclusion on screening: Significant reservoir emissions cannot be dismissed Comment: Although it is clear this project is a mitigation project by replacing fossil fuel power generation, the reservoir emissions could be significant in absolute terms Estimation of reservoir emission (leave empty if initial screening concluded negligible emissions) Catchment area (km2): 14,660 Description of land cover in CA: 53% shrubland, 40% forest, 7% cropland Reservoir volume (million m ):3 349 table continues next page Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 47 Average depth (m): 27 Description of land cover in reservoir area: 91% shrubland, 5% forest, 4% cropland Type of soil and soil carbon content: Mineral soil, 3.4 kg C/m2 Solar Radiation (kWh/m2/day): 3.8 Average temperature (oC): 23 Method used for estimation: G-res (UNESCO/IHA 2017) Net reservoir emissions (t CO2eq/year): 4,563 GHG from unrelated anthropogenic sources (t CO2-eq/year) 177 Other comparative methods: World Bank (2013): 15,300 t CO2eq/year Confidence in Reliable Yes results Explanation The G-res shows specific emissions per m2 (56th percentile) in the normal range where the tool is most reliable. The G-res and World Bank (2013) results indicate 5-15 g CO2/kWh, which are both in the same order of magnitude (<3%) compared to the counterfactual. Further measures to be taken Further detailed measurements and/or modeling required: No GHG management measures to Although not a major issue, the ESMP should consider reducing inflow of anthropogenic organic material consider: and nutrients, which will improve water quality and reduce GHG as a result of UAS Sources: Project information from Supplementary Environmental and Social Impact Assessment 2009. Carbon content from FAO Harmonized Soil Database, Solar Radiation from NASA - SSE 2008, Land cover from ESA-CCI, Temperature from Global Climate Database (2005). 48 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes Glossary Aerobic Describes conditions or processes in water or sediments in which oxygen is present. Anaerobic Describes conditions or processes in water and sediments in which oxygen is absent. Anoxic Synonymous with anaerobic: describes conditions in water and sediments in which oxygen is absent. Anthropogenic Resulting from or produced by human beings. Biomass The total mass of living organisms in a given area, volume, or ecosystem at a given time; recently dead plant material is included in dead biomass. The quantity of biomass can be expressed as a dry weight or as the energy or carbon content. Bubbling See ebullition. Carbon cycle The process of carbon flow through the atmosphere, ocean, terrestrial biosphere (including freshwater systems), and sediments, as well as its transformation pro- cesses (chemical alteration, photosynthesis, respiration, decomposition, air-sea exchange, etc.). Carbon dioxide (CO2) A naturally occurring GHG fixed by photosynthesis into organic matter and released during respiration. It is a by-product of fossil fuel combustion, biomass burning, land use changes, and other industrial processes. CO2 equivalent (CO2e(q)) The amount of CO2 emission that would have the same GWP, over a given time horizon, as an emitted amount of a GHG or a mixture of GHGs. The CO2equivalent emission is obtained by multiplying the emissions of a GHG by its GWP for the given time horizon. For a mix of GHGs it is obtained by adding up the CO2equivalent emissions of each gas. Carbon footprint A form of carbon calculation that considers the net emissions of GHG throughout the life cycle of a project or investment. Carbon mass flow Carbon in a water body can be in particulate or dissolved form and can be organic or inorganic. The forms to be measured are: Total Organic Carbon (TOC), Dissolved Organic Carbon (DOC), Dissolved Inorganic Carbon (DIC), and Particulate Organic Carbon (POC). Carbon inputs and outputs to be considered are: carbon brought in by macrophytes*, carbon exchanges with groundwater, carbon lost permanently to sedi- ment, carbon exchanged with atmosphere in form of CO2 and CH4, and humic substance income and output. Carbon sequestration Buildup of GHG concentration in vegetation, water or sediments. Carbon sink See sink. Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 49 Carbon stock In the context of this note, the quantity of carbon in a water body and its sediments. [In the context of forests, it refers to the amount of carbon stored in the world’s forest ecosystem, mainly in living biomass and soil, but to a lesser extent also in dead wood and litter.] Decomposition Chemical processes by which organic matter in a water body is transformed into gas- eous end products. Major processes are oxidative decomposition, methanogenesis, and denitrification and their end products: CO2, CH4, and N2O. Degassing GHG flux induced by dramatic pressure change immediately after water discharge from reservoir outlets. Denitrification Denitrification describes the conversion of nitrate into nitrite, then to N2O, and finally to nitrogen gas. This process happens in the slightly anoxic upper layer of sediment. Diffusive flux Discharge of GHG from the air-water interface of a water body. Dissolved Oxygen (DO) The oxygen in a water body in its dissolved form. Dissolved oxygen influences organic matter decomposition processes and serves fish and other aquatic organisms for respiration. Ebullition (bubbling) Discharge in form of bubbles of gaseous substances from a water body, which results from carbonation, evaporation, or fermentation. Ecosystem The interactive system formed from all living organisms and their abiotic (physical and chemical) environment within a given area. In the context of this note, a difference is made between aquatic and terrestrial ecosystems. Epilimnion The dense top-most layer of water in a thermally stratified water body. Global Warming An index, based on the radiative properties of GHGs, measuring the radiativeforcing of Potential (GWP) a unit mass of a given GHG in today’s atmosphere integrated over a chosen time hori- zon, relative to that of CO2. The GWP represents the combined effect of the differing lengths of time that these gases remain in the atmosphere and their relative effective- ness in absorbing outgoing infrared radiation. The IPCC considers the GWP of GHG within a 100-year time frame. Greenhouse gas (GHG) GHGs are those gaseous constituents of the atmosphere, both natural and anthropo- genic, that absorb and emit radiation at specific wavelengths within the spectrum of thermal infrared radiation emitted by the Earth’s surface, the atmosphere itself, and by clouds. In the context of this note, the evaluation of net emissions from water bodies includes the three GHG species CO2, CH4, and N2O. Hypolimnion The dense bottom layer of water in a thermally stratified water body. Macrophyte Rooted plant that grows in or near water. Methane (CH4) A naturally occurring GHG, a main component of natural gas, and an end product of animal husbandry and agriculture. 50 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes Methane oxidation Process by which CH4 is oxidized to CO2 and that occurs in aerobic conditions. Methanogenesis Production of CH4 by anaerobic bacteria and microbes present in the anoxic layers of a water body, which feed on the detritus of organic matter and respire CH4. Nitrification An aerobic process in which bacteria change the ammonia and organic nitrogen in water and decomposed matter into oxidized nitrogen (nitrate). Nitrous oxide (N2O) A naturally occurring GHG which is produced through bacterial nitrification and deni- trification processes. Oxic Synonymous with aerobic: describes conditions in water and its sediments in which oxygen is present. Photosynthesis Process driven by solar energy by which atmospheric CO2 is fixed by plants and algae for the primary production of organic matter and oxygen as a by-product. Residence time Average time a water molecule spends in a reservoir; used to describe the flow rate of the water through the reservoir. Its value can vary inside one reservoir. Respiration Heterotrophic respiration is the process whereby micro-organisms grow by convert- ing organic matter to sugars; autotrophic (or maintenance) respiration is the process through which plants and animals burn sugars to give energy. Both reactions pro- duce CO2. Sink A natural or artificial reservoir that accumulates and stores some carbon-containing chemical compound for an indefinite period. Stratification A water body can be stratified in layers of temperature, salinity, or chemical composi- tions, which can act as barriers to water mixing. Water retention time A calculated quantity expressing the mean time that water spends in a particular water body. Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 51 References Abril, G., F. Guérin, S. Richard, R. Delmas, C. Galy-Lacaux, P. Gosse, Guérin, F., G. Abril, S. Richard, B. Burban, C. Reynouard, P. Seyler, and A. Tremblay, L. Varfalvy, M. A. Santos, and B. Matvienko. 2005. “Carbon R. Delmas. 2006. “Methane and Carbon Dioxide Emissions from Tropical Dioxide and Methane Emissions and the Carbon Budget of a 10-year Old Reservoirs: Significance of Downstream Rivers.” Geophysical Research Tropical Reservoir (Petit-Saut, French Guiana).” Global Biogeochemical Letters 33: L21407. Cycles 19: GB 4007. Guérin, F., G. Abril, A. Tremblay, and R. Delmas. 2008. “Nitrous Oxide Barros, N., J. J. Cole, L. J. Tranvik, Y. T. Prairie, D. Bastviken, Emissions from Tropical Hydroelectric Reservoirs.” Geophysical Research V. L. M. Huszar, P. del Giorgio, and F. Roland. 2011. “Carbon Emission Letters 35 (6). doi:10.1029/2007GL033057. from Hydroelectric Reservoirs Linked to Reservoir Age and Latitude.” Hertwich, E. 2013. “Addressing Biogenic Greenhouse Gas Emissions from Nature Geoscience 4: 593–96. Hydropower in LCA.” Environmental Science and Technology 47 (17): Bastviken, D., L. J. Tranvik, J. A. Downing, P. M. Crill, and A. Enrich- 9604–11. Prast. 2011. “Freshwater Methane Emissions Offset the Continental Carbon Sink.” Science 331 (6013): 50. Hijmans, R. J., S. Cameron, J. Parra, P. Jones, A. Jarvis, and K. Richardson. 2005. WorldClim, version 1.4. University of California, Berkeley. Cailleaud, E. 2015. “Cycles du carbone et de l’azote et émissions de gaz à effet de serre (CH4, CO2 et N2O) du lac de barrage de Petit Saut et du fleuve IEA Hydro. 2012. International Energy Agency, Hydropower Annex XII: Sinnamary en aval du barrage (Guyane Française).” PhD dissertation. Guidelines for Quantitative Analysis of Net GHG Emissions from Toulouse: Université de Toulouse. Reservoirs – Volume 1: Measurement Programs and Data Analysis. Chanudet, V., S. Descloux, A. Harby, H. Sundt, B. H. Hansen, O. Brakstad, ———. 2017. International Energy Agency, Hydropower Annex XII, D. Serça, and F. Guérin. 2011. “Gross CO2 and CH4 Emissions from the Guidelines for Quantitative Analysis of Net GHG Emissions from Reservoirs, Nam Ngum and Nam Leuk Sub-Tropical Reservoirs in Lao PDR.” Science Volume 3, Management, Mitigation and Allocation, Draft Version. of the Total Environment 409 (23): 5382–91. IHA. 2017. “Hydropower Emissions.” Presented at World Hydropower Chanudet, V., J. Smits, J. Van Beek, P. Boderie, F. Guérin, D. Serça, Congress, Addis Ababa, May 2017. C. Deshmukh, and S. Descloux. 2016. “Hydrodynamic and Water Quality IPCC. 2006. IPCC Guidelines for National Greenhouse Gas Inventories. 3D Modelling of the Nam Theun 2 Reservoir (Lao PDR): Predictions Prepared by the National Greenhouse Gas Inventories Programme. and  Results of Scenarios Related to Reservoir Management, Eggleston, H. S., Buendia, L., Miwa, K., Ngara, T. and Tanabe, K. (eds). Hydrometeorology and Nutrient Input.” Hydroécologie Appliquée 19: Published by IGES, Japan. 87–118. ———. 2011. Special Report on Renewable Energy Sources and Climate Deemer, B. R., J. A. Harrison, S. Li, J. J. Beaulieu, T. DelSontro, N. Barros, Change Mitigation, edited by O. Edenhofer, R. Pichs-Madruga, Y. Sokona, J. F. Bezerra-Neto, S. M. Powers, M. A. dos Santos, and J. A. Vonk. 2016. K. Seyboth, P. Matschoss, S. Kadner, T. Zwickel, P. Eickemeier, G. Hansen, “Greenhouse Gas Emissions from Reservoir Water Surfaces: A New S. Schlömer, and C. von Stechow. Cambridge University Press, Cambridge, Global Synthesis.” BioScience 66 (11): 949–64. United Kingdom. Descloux, S., V. Chanudet, D. Serça, and F. Guérin. 2017. “Methane and ———. 2013. Climate Change 2013: The Physical Science Basis. Contribution Nitrous Oxide Annual Emissions from an Old Eutrophic Temperate of Working Group I to the Fifth Assessment Report of the Intergovernmental Reservoir.” Science of the Total Environment, 598 (2017): 959–72. Panel on Climate Change, edited by Stocker, T. F., D. Qin, G.-K. Plattner, Deshmukh, C., F. Guérin, D. Labat, S. Pighini, A. Vongkhamsao, M. Tignor, S. K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P. Guédant, W. Rode, A. Godon, V. Chanudet, S. Descloux, and D. Serça. P. M. Midgley. Cambridge, United Kingdom: Cambridge University Press. 2016. “Low Methane (CH4) Emissions Downstream of a Monomictic Kim Y., N. T. Roulet, C. Li, S. Frolking, I. B. Strachan, C. Peng, C. R. Subtropical Hydroelectric Reservoir.” Biogeosciences 13: 1919–32. Teodoru, Y. T. Prairie, and A. Tremblay. 2016. “Simulating Carbon Dioxide Downing, J. A., Y. T. Prairie, J. J. Cole, C. M. Duarte, L. J. Tranvik, R. G. Exchange in Boreal Ecosystems Flooded by Reservoirs.” Ecological Striegl, W. H. McDowell, P. Kortelainen, N. F. Caraco, and J. M. Melack. Modelling 327: 1–17. 2006. “The Global Abundance and Size Distribution of Lakes, Ponds, and Lehner, B., C. Reidy Liermann, C. Revenga, C. Vörösmarty, B. Fekete, Impoundments.” Limnology and Oceanography 51 (5): 2388–97. P. Crouzet, P. Döll, M. Endejan, K. Frenken, J. Magome, C. Nilsson, J. C. EPRI. 2010. The Role of Hydropower Reservoirs in Greenhouse Gas Robertson, R. Rödel, N. Sindorf, and D. Wisser. 2011. “High-Resolution Emissions. Electric Power Research Institute, Palo Alto, CA, USA. Mapping of the World’s Reservoirs and Dams for Sustainable River-Flow Management.” Frontiers in Ecology and the Environment 9(9): 494–502. Galy-Lacaux, C., R. Delmas, C. Jambert, J. F. Dumestre, L. Labroue, S.  Richard, and P. Gosse. 1997. “Gaseous Emissions and Oxygen Prairie, Y. T. 2017. “The Global GHG Footprint of Reservoirs: What Does Consumption in Hydroelectric Dams: A Case Study in French Guiana.” the Atmosphere See?” IHA World Hydropower Congress, Addis Ababa, Global Biogeochemical Cycles 11: 471–83. Ethiopia. Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes 53 Prairie, Y. T., J. Alm, J. Beaulieu, N. Barros, T. Battin, J. Cole, P. del Gases to the Atmosphere: A Global Estimate.” Bioscience 50 (9): Giorgio, T. DelSontro, F. Guérin, A. Harby, S. Mercier-Blais, D. Serça, 766–75. S.  Sobek, and D. Vachon. 2017. “GHG Emissions from Freshwater Reservoirs: What Does the Atmosphere See?” Ecosystems. https://doi​ Sturm, K., Z. Yuan, B. Gibbes, U. Werner, and A. Grinham. 2014. “Methane .org/10.1007/s10021-017-0198-9. and Nitrous Oxide Sources and Emissions in a Subtropical Freshwater Reservoir, South East Queensland, Australia.” Biogeosciences 11 (18): Raymond, P. A., J. Hartmann, R. Lauerwald, S. Sobek, C. McDonald, 5245–58. M. Hoover, D. Butman, R. Striegl, E. Mayorga, C. Humborg, P. Kortelainen, H. Dürr, M. Meybeck, P. Ciais and P. Guth. 2013. “Global Carbon Dioxide UNESCO/IHA. 2010. GHG Measurement Guidelines for Freshwater Reservoirs. Emissions from Inland Waters.” Nature 503: 355–59. Paris: UNESCO/IHA. Sikar, E., B. Matvienko, M. A. Santos, L. P. Rosa, M. B. Silva, E. O. Santos, ———. 2017. UNESCO/IHA Research Project on the GHG Status of Freshwater C. H. E. D. Rocha, and A. P. Bentes, Jr. 2009. “Tropical Reservoirs Are Reservoirs, The GHG Reservoir Tool (G-res). www.hydropower​.org​/­gres​ Bigger Carbon Sinks than Soils.” Verhandlungen des Internationalen -tool. Paris: UNESCO/IHA. Verein Limnologie 30: 838–40. World Bank. 2013. “Greenhouse Gases from Reservoirs Caused by St. Louis, V. L., C. A. Kelly, E. Duchemin, J. W. M. Rudd, and D. M. Biochemical Processes.” Interim Technical Note. World Bank, Rosenberg. 2000. “Reservoir Surfaces as Sources of Greenhouse Washington, DC. 54 Greenhouse Gases from Reservoirs Caused by Biogeochemical Processes W17099