VALUING THE ECOSYSTEM SERVICES PROVIDED BY FORESTS IN PURSAT BASIN, CAMBODIA JU LY 20 20 © 2020 The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved This work is a product of the staff of The World Bank. The findings, interpretations and conclusions expressed in this work do not necessarily reflect the views of the Executive Directors of The World Bank 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. Attribution - Please cite the work as follows: “Rawlins, Maurice, Stefano Pagiola, Kashif Shaad, Mahbubul Alam, Rosimeiry Portela, Srabani Roy, Derek Vollmer and Werner Kornexl. 2020. ‘Valuing the Ecosystem Services provided by Forests in Pursat Basin, Cambodia.’ World Bank: Washington D.C. All 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. The team thanks Chloe Hill, Herbert Navasca, and Topi Carlos for editing services. Cover photo: Aerial shots of a remote waterfall in the Cardamom Mountains © Conservation International/photo by David Emmett CONTENTS LIST OF TABLES iv LIST OF FIGURES v ACRONYMS AND ABBREVIATIONS vi GLOSSARY OF KEY TERMS vii ACKNOWLEDGEMENTS viii EXECUTIVE SUMMARY 1 1 BACKGROUND & OVERVIEW 12 1.1 How are Forest Ecosystem Services Contributing to Cambodia’s Economy? 14 1.2 Valuing Cambodia’s Ecosystem Services is Relevant and Timely 16 1.3 About this Analytical Work and Report 18 1.4 Overall Methodological Approach to Valuing Ecosystem Services 20 2 HYDROLOGICAL ANALYSIS 24 2.1 Overview 24 2.2 Pursat River Basin 26 2.3 Key Hydrological Datasets 27 2.4 Dam Characteristics 29 2.5 Area Under Irrigation 30 2.6 Modeling Approach 31 2.7 Hydrological Modelling Results 34 2.8 Limitations of the Analysis 43 2.9 Summary of Key Results 44 3 MONETARY VALUATION OF HYDROLOGICAL SERVICES 45 3.1 Overview 45 3.2 Methodology 46 3.3 Results 50 3.4 Summary 51 4 CARBON VALUATION 53 4.1 Background 53 4.2 Estimating Carbon Stocks 54 5 TOURISM BENEFITS 57 6 RECOMMENDATIONS 62 7 CONCLUSION 70 REFERENCES 71 Appendix 1: Data Requirements 76 Appendix 2: SWAT Parameters 78 Appendix 3: Scatter Plot of Monthly WYLD from Sub-basins as Directly Contributing to 79 Monthly Discharge Appendix 4: Forest Conversion for Agriculture and Charcoal, and Forest Protection Estimates 80 Appendix 5: Tourism and Carbon NPV Analysis 82 TABLES Table 1.1: Large Hydropower Dams in Cambodia 15 Table 2.1: Water Availability at Bac Trakuon Station 28 Table 2.2: Dam Characteristics 30 Table 2.3: Target Irrigated Areas in the Pursat River Basin 30 Table 2.4: Summary of Main Projected Hydrological Impacts of Land Use Change in 44 the Pursat River Basin Table 3.1: Value of Water as an Input to Production in Different Farming Categories 48 Table 3.2: Key Assumptions Used in Valuation of HEP 49 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Table 3.3: Effect of Land Cover Change on Value of HEP 50 Table 4.1: Land Cover in the Pursat River Basin, 2016 54 Table 4.2: Carbon Stocks by Forest Type 55 Table 4.3: Estimated Carbon Emissions from Deforestation in the Pursat River Basin 56 BOXES Box 1.1: Drivers of Degradation in Protected Areas 17 Box 1.2: Cambodia Sustainable Livelihood and Ecotourism (CSLE) Project 19 iv FIGURES Figure 1.1: Forest Land Cover Change in Cambodia, 2006-2018 13 Figure 1.2: Natural, Produced and Human Capital in Cambodia, 1995 – 2014 14 Figure 1.3: Overlaps Between Freshwater Habitats and Forest Landscapes in 16 the East and West Mountainous Areas of Cambodia Figure 1.4: Schematic Diagram of the Sequence of Biophysical and Economic Valuation in 23 SEEA EEA Context, with an Illustrative Example of Water Supply for Irrigated Agriculture Figure 2.1: Land Cover 2016 Map and Distribution 26 Figure 2.2: Comparison of ERA5 Estimates with Values Measured at Local Gauge 27 Figure 2.3: Comparing Monthly Discharge Statistics for Two Time Periods 29 Figure 2.4: Estimated Maximum Area Under Cultivation in the Pursat River Basin 31 Figure 2.5: Data for Validation of Hydrological Model 35 Figure 2.6: Results from Daily and Monthly Calibration and Validation 36 Figure 2.7: Average Water Yield Distribution over Sub-basins and Quarters Over a Year 37 Figure 2.8: Water Yield Distribution by Sub-basins Linked to the Dams and Tributary St Prey Khlong 38 Figure 2.9: Average Annual Sediment Yield for SWAT Sub-basins 39 Figure 2.10: Comparison of Current Hydrological Flows to those of No-forest Counterfactual 40 Figure 2.11: Spatial Changes in Water Yield Under the No-forest Counterfactual Compared 40 to Current Conditions Figure 2.12: Comparison of Inflows and Outflows from Each Dam Operations 41 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Under Current Conditions and Under the No-forest Counterfactual Figure 2.13: Effect of Dam Operation on Projected Water Availability, Selected Years 42 Figure 2.14: Comparison of Estimated Water Demand and Estimates of Water Available at 42 Bac Trakuon Station, Selected Years Figure 2.15: Estimated Irrigation Water Demand Deficit Under Current Conditions 43 Figure 3.1: Paddy Rice Production Data, by Season 47 Figure 3.2: Key Characteristics of Farming Systems 47 Figure 3.3: Methodological Process for Valuation of Hydrological Services Used for HEP 48 Figure 4.1: Carbon Stocks for Major Forest Types in the Pursat Basin 55 Figure 5.1: Projected Visitors to PAs in the upper Pursat River Basin 58 Figure 5.2: Projected Increase in Tourism and Carbon Benefits from the CSLE Project 59 Figure 6.1: Areas of High water Yield within PAs in the Cardamom Mountains 64 Figure 6.2: Precipitation is Projected to Increase in the Wet Season Over Cardamom Mountains 65 Figure 6.3: Precipitation is Projected to Decrease in the Dry Season with Climate Change 65 v ACRONYMS AND ABBREVIATIONS CMTS Cardamom Mountains-Tonle Sap NSE Nash–Sutcliffe Efficiency ECMWF European Centre for Medium-Range NTFP Non-timber forest product Weather Forecasts PA Protected area ERPA Emission Reduction Payment Agreement PPU Percent prediction uncertainty FAO Food and Agriculture Organization REDD+ Reduced Emissions from Deforestation GHG Greenhouse gas and Forest Degradation ha Hectares RGC Royal Government of Cambodia HEP Hydroelectric power SEEA EEA System of Environmental-Economic Accounting Experimental Ecosystem HRU Hydrologic response units Accounts MME Ministry of Mines and Energy SCS Soil Conservation Service MoE Ministry of Environment SME Small and medium enterprise MoWRAM Ministry of Water Resources and SVC Social value of carbon Meteorology SWAT Soil and Water Assessment Tool MUSLE Modified universal soil loss equation tCO2e Tons of carbon dioxide equivalent VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A NCA Natural Capital Accounting VCS Verified Carbon Standard NDC Nationally Determined Contribution VSSPNP Veun Sai-Siem Pang National Park NPASMP National Protected Areas Strategy and Management Plan vi GLOSSARY OF KEY TERMS Biodiversity Biodiversity reflects the number, variety and variability of living organisms and how these change over time from one location to another. It forms the basis of the multiple benefits provided by ecosystems to humans. Ecosystem services The benefits to people from ecosystems, such as timber, pollination, water supply, water regulation, climate regulation, recreation, mental health and others. Foregone benefits Benefits (costs) of an action that are not received because an alternative action (costs) is undertaken. Landscape A heterogeneous land area comprising of a cluster of interacting ecosystems that are repeated in a similar form throughout. Natural Capital The stock of renewable and non-renewable natural resources (e.g. plants, animals, air water, soils, minerals) that combine to yield a flow of benefits to people. Natural Capital Natural capital accounting integrates natural resources and economic analysis, providing Accounting a broader picture of development progress than standard measures such as GDP. Natural capital accounts are a set of objective data showing how natural resources contribute to the economy and how the economy affects natural resources. Natural hazard Hazard regulation or disaster mitigation is the function of ecosystems in modulating the regulation effects of extreme events like droughts, floods and fires and in particular in protecting VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A human well-being from the impacts. Sediment regulation Sediment regulation refers to the capacity of ecosystems to regulate the quantity of eroded sediment reaching the stream network and thus delivering key benefits, like maintaining soil and water quality and reservoir functions. Value The worth, importance or usefulness of something – and is often categorized as anthropocentric and instrumental. Anthropocentric values reflect human needs and preferences. Instrumental values serve a specific goal, use or need. Water flow Water flow regulation is an ecosystem service that can be defined as the ability of regulation watersheds and catchments to capture and store water from rain storms, reducing the direct runoff and flood peaks as well as releasing the water more slowly so that flows are sustained into or through the dry season. vii ACKNOWLEDGEMENTS The Valuing the Ecosystem Services Provided by Forests in Pursat Basin, Cambodia is a World Bank advisory product developed for the Royal Government of Cambodia (RGC) to support the sustainable management of the country’s natural capital. The advisory work was undertaken as part of a broader effort of the World Bank in Cambodia to provide guidance to the RGC through technical assistance and analytical and advisory services on managing its natural capital through landscape approaches. Working in the Cardamom Mountains, the Mekong Delta and in the upstream Mekong, the Bank is helping the RGC to better manage and add value to its natural capital through strengthening its links to the economy. This report is the output of rich dialogue and exchange with a number of institutions and individuals in the RGC. This work is part of the technical assistance (TA) provided under the Bank’s Enhancing Environmental Sustainability and Resilience in Cambodia technical assistance Program. The work was funded by a World Bank multi-donor trust fund, Global Program on Sustainability (GPS) (previously known as Wealth Accounting and Valuation of Ecosystem Services or WAVES) and the former Program on Forest (PROFOR). A World Bank team carried out this work in cooperation with Cambodia’s Ministry of Environment (MoE), Ministry of Water Resources and Meteorology (MOWRAM), Ministry of Economy and Finance (MEF), Ministry of Mines and Energy (MME) and Conservation International. The World Bank team would like to thank in particular, Excellency Sao Sopheap, Secretary of State MoE, Excellency Chann Sinath, Under Secretary of State MOWRAM, and Excellency Tin Ponlok, Chair National Council for Sustainable Development and their dedicated teams for their support in data collection and technical guidance. World Bank management, including Inguna Dobraja, Country Manager for Cambodia; Christophe Crepin, previous Practice Manager, and Stephen Ling Acting Practice Manager for Environment and Natural Resources and Blue Economy for East Asia and the Pacific; and Mark Austin, Program Leader for Sustainable Development, South East VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Asia, provided strategic guidance. The World Bank task team for Forest Services Valuation in Cambodia is led by Maurice Andres Rawlins (Natural Resource Management Specialist) and Werner Kornexl (Senior Natural Resource Management Specialist) and at the time of delivery includes Stefano Pagiola (Senior Environmental Economist) and Elisabeth Steinmayr (Junior Professional Officer). The analysis and valuation of hydrological benefits was conducted by a team from Conservation International team led by Srabani Roy and comprising of Rosimeiry Portela, Kashif Shaad, Mahbubul Alam, and Derek Vollmer. The World Bank team appreciates the inputs provided by World Bank peer reviewers Raffaello Cervigni, Glenn- Marie Lange and Claire Honore Hollweg. viii EXECUTIVE SUMMARY VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Valuing Ecosystem Services is relevant for Cambodia now Ecosystem services, such as water flow regulation, erosion reduction and biodiversity conservation that are provided by Cambodia’s forests, underpin the country’s economy. Agriculture, which depends heavily on water flow regulation, erosion reduction and nutrient retention services all of which are provided by forests, contributed to GDP in 2018 by 22 percent. Tourism and ecotourism, which contributed about 18 percent towards Cambodia’s 2018 GDP, is dependent on the country’s considerable biodiversity in forest landscapes. Forests also help to sustain hydropower in Cambodia, an essential power source to the garment and other economic sectors. Waterfall in the Central Cardamoms. © Conservation International/photo by Jake Brunner 1 The Royal Government of Cambodia (RGC) has improved decision-making on protected areas, forest strategically placed the majority of Cambodia’s and natural resources more generally; but the lack of forestlands under the protected areas (PAs) system data, accepted methodologies and technical capacity to better protect forest resources. Cambodia’s forests have prevented it from using ecosystem valuation for cover about 8.1 million ha (45 percent of the country). decision making on meaningful scale. There is strong Approximately 67 percent of this forest area is under interest among the Ministry of Environment (MoE) the country’s PA system which covers 7.4 million ha and the Ministry of Water Resources and Meteorology (41 percent of Cambodia). Cambodia’s forest cover (MOWRAM) for integrating ecosystem service values has declined since 2006 because of pressure for land in decision-making as it pertains to forests, protected and unsustainable natural resource use. By putting areas and watershed management. MoE’s foremost forest resources under the PA system, the RGC, has policy documents on PAs – The Protected Areas Law taken key steps to ensure the reversal of this declining of 2008 and the National Protected Areas Strategy trend in forest cover. Under the PA system, forests will and Management Plan (NPASMP) 2017-2031 – strongly be subject to stricter management that includes no advocate for the use of ecosystem service values in timber harvesting, sustainable harvesting of non- conservation planning and prioritization of areas timber forest products (NTFPs), patrolling by rangers in PAs. By utilizing such values, not only are strong and forest communities to prevent illegal activities protection efforts cultivated but drivers of forest and forest restoration where needed. Furthermore, degradation and loss can also be more effectively the RGC has made a recent significant investment addressed. However, data and capacity are lacking in in forest and protected areas management through Cambodia that are essential for assessing and valuing the World Bank-financed Cambodia Sustainable the benefits provided by ecosystem services and Landscape and Ecotourism (CSLE) project. integrating these data and information into decision- making like PA zoning and management planning. Cambodia’s impending water crisis, impacts of Economic analysis has therefore played a limited role climate change and the onset of COVID-19 will put in determining how forests are managed and financed increased attention on forests. Seasonal water in Cambodia. scarcity exists in the river basins of the Tonle Sap River Basin Group (18 out of 25 provinces were Meeting the RGC’s need for values-driven VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A affected by droughts in 2016, for example, with decision-making on forest and PAs around 2.5 million people lacking water) and existing through World Bank analytical support reservoir storage capacity (less than 10 percent of total water generated) is insufficient to redistribute This analytical work is being undertaken as part of a water significantly between seasons. Drought and broader effort of the World Bank to provide guidance flooding are expected to worsen because of climate to the RGC on managing its natural capital through change, with an additional 1.5 months of drought and landscape approaches. By focusing on the Pursat more extensive flooding, particularly in the eastern River Basin in the Cardamom Mountains, the case areas of the country. Forests are a safety net during study presented in this report is intended to provide socioeconomic shocks and are important for linked a practical illustration of how the ecosystem services sectors like ecotourism, as they can help to stimulate that are provided by a forest can be valued and then jobs and rural economies as well as tourism value compared to the benefits that would otherwise be chains which have been negatively affected by obtained if the forest was converted to other uses. COVID-19. The study provides evidence of the importance of forests in providing ecosystem services that are The RGC is increasingly recognizing that adequate important for Cambodia’s economy as well as for valuation of ecosystem services is a key input to the country’s climate and disaster resilience. The 2 results therefore intend to help the RGC quantify 2. Quantification of benefits of forest ecosystem and communicate the value of its natural capital to services, which provide evidence of the returns on Cambodia’s economy. investments that the RGC will gain by strengthening forest ecosystems through use of public resources The main outputs of the study are: for forest and protected areas (PAs) management. 1. Methodology on how to undertake measurement 3. The causal link between forest conservation/ and valuation of hydrologic ecosystem services that degradation and regular water flow has been can be repeated in other locations in Cambodia. established through the analysis. The work will also provide the underpinnings for 4. Recommendations for scaling-up ecosystem tools (e.g. Payment for Ecosystem Services (PES) services assessment and valuation are provided which will be developed under the Cambodia as a road map within this report and actions Sustainable Livelihood and Ecotourism (CSLE) proposed for how this work would inform project) as well as investments in forest and PA investments for forest and PA management namely management also under the CSLE project and conservation, protection and restoration in the natural capital accounting. Cardamom Mountains. KEY MESSAGE 1 The economic benefits from intact forests (estimated at US$99 million) are almost five times higher than the gains from cutting them down for small-scale agriculture and charcoal production estimated at US$22 million. The standing forests in the Pursat River Basin (RB) provide benefits worth at least US$99 million through the provision of ecosystem services like water and sediment flow regulation and tourism. Converting these forests for charcoal production and agriculture would provide benefits worth about US$22 million to only a few individuals, whilst the benefits of maintaining the forests intact, which are worth about US$99 million and which would reach a broader segment of the Cambodian population, would be lost. It is clear, therefore, that Cambodia benefits more VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A from keeping forest ecosystems in the Pursat RB intact than from cutting them down. Considering just agriculture, the findings show that there are more benefits derived from maintaining forests to facilitate water for irrigation (US$28 million) than cutting down the forests for agriculture (US$22 million); see Table ES1. TABLE ES1: SUMMARY OF ESTIMATED BENEFITS PROVIDED BY FORESTS IN THE PURSAT RIVER BASIN Value of benefits Beneficiaries Net Present Value (US$million) Annual (US$million/yr) Irrigation 28 1.6 Hydropower (HEP) 18 1.1 Tourism 53 3 Carbon storage - 2.5 Source: Authors’ calculations Note: 1. Net present value estimates use a 6 percent discount rate. A value of US$5 t/ CO2e was used for the analysis of carbon. The data and calculations underlying these findings are in Chapters 2, 3, 4 and 5 of the report. 2. The counterfactual approach (if forests were removed) was used to estimate net present value of benefits of forest for irrigation, HEP and tourism. 3. The deforestation scenarios approach was used to estimate annual benefits of forests for irrigation, HEP, carbon storage and tourism. 3 Water flow regulation is one of the key benefits better defend its budget requests and justify its provided by forest ecosystems and understanding its expenditure on PA management for protecting forests high monetary value creates a strong incentive for in watershed areas like the Pursat Basin. investing resources into the sustainable management of the forest ecosystem asset. This study makes the Results also help to illustrate that investing in link between upstream forest management, water yield protecting forest resources is part of the economic and use of water by downstream users. recovery from COVID-19. This will help the RGC to ensure that agricultural growth is not limited by a Without forests dry season flows would be reduced lack of water and that the ecotourism potential of by 25 percent, which would exacerbate the water forest biodiversity is preserved. Importantly, keeping deficit already being experienced in drier periods, hydropower plants functioning, even during the water like February and June; see Figure ES1. With this crisis, will rely to some extent on the hydrological information, the Ministry of Environment (MoE) can flows facilitated by forests. FIGURE ES1: CHANGES IN WATER YIELD IN THE ABSENCE OF FORESTS THE MAP ILLUSTRATES THE IMPACT ON WATER YIELD IN THE PURSAT RIVER BASIN IF THE FORESTS WERE NOT THERE. THE RED AREAS INDICATE THAT WATER YIELD COULD DECREASE BY MORE THAN 30% IF FOREST WERE NOT THERE. THIS MAP WAS DERIVED FROM THE NO-FOREST COUNTERFACTUAL COMPARED TO CURRENT CONDITIONS. VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Source: Authors’ creation 4 KEY MESSAGE 2 Investing in the maintenance Forest loss or degradation would reduce the benefits of forest is good business. that they provide. Such loss and/or degradation Annual public expenses would tend to increase peak river flows and erosion, to maintain the forest in which would lower water availability in the dry season the Pursat Basin are about (when water is particularly valuable for irrigation), 20 times lower than the increase flood risk and increase siltation of reservoirs, benefits provided by them. thus reducing their useful life. The actual losses would depend on the rate and extent of forest loss. The forests in the Pursat RB provide annual benefits Upstream deforestation has a high impact on peak worth an estimated US$8.2 million from tourism, river flows and floods. Forests slow the flow of water carbon and water and sediment flow regulation. This is through a watershed (from rainfall event to the river). about 20 times the US$0.4 million that the RGC spends In the absence of forests, water would run off faster, annually on forest protection in Pursat. This indicates resulting in higher flood risk in the wet season and that the return on spending on forest protection is an reduced availability in the dry season, as shown in efficient use of the government’s budget. figure ES2 below (left panel). FIGURE ES2: CHANGES IN WATER AND SEDIMENT FLOWS IN THE ABSENCE OF FORESTS THE CHART ON THE LEFT SHOWS THE EFFECT OF REMOVING THE FOREST ON RIVER DISCHARGE AND THE CHART ON THE RIGHT SHOWS THE LARGE INCREASE IN SEDIMENT FLOWS THAT ARISE BECAUSE OF REMOVAL OF THE FOREST. Source: Authors’ calculations 2010 300 900 800 Dam1 sediment input (ton) x 10000 250 700 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Discharge (cumecs) 200 600 500 150 400 100 300 200 50 100 0 0 Oct Oct Jan Feb Mar Apr May Jun Jul Aug Sep Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Nov Dec Current No Forest Current No Forest Upstream deforestation has a high impact on cover conditions, erosion is generally low. The few sediment output. In the absence of forests, areas with high erosion rates are in parts of the sedimentation would increase 30-fold, as shown in Samkos Wildlife Sanctuary, where forest conversion Figure ES2 above (right panel). Under current forest has taken place, as shown in Figure ES3 below. 5 FIGURE ES3: CURRENT SEDIMENT YIELD IN THE PURSAT RIVER BASIN THE MAP SHOWS THE AVERAGE ANNUAL SEDIMENT YIELD IN THE PURSAT RIVER BASIN. THE ORANGE AND RED AREAS ARE IN THE SAMKOS WILDLIFE SANCTUARY, WHERE MANY AREAS OF FOREST HAVE BEEN CONVERTED TO AGRICULTURE. Source: Authors’ creation LEGEND SEDIMENT YIELD t/ha/year (<2) (2-5) (5-10) (10-25) Deforestation would cause reservoir sedimentation, but not entirely - even the current and relatively low reducing electricity production. In the absence of deforestation rates in the Pursat Basin of 0.25 percent forests, the reservoir of the hydroelectric power (HEP) per year (below the national average) would reduce station at Dam 1 in the Pursat River Basin would be the present value of electricity production by US$0.8 completely silted up in 65 years, reducing the present million. Should deforestation accelerate to 1 percent value of electricity production by US$18.2 million, as a year, as in the rest of the country, the loss would VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A shown in Table ES2. The extensive forest cover in the increase to US$2.8 million. These results demonstrate Pursat Basin protects the reservoir from this fate, the benefits from protecting forest resources for HEP. TABLE ES2: IMPACT OF FOREST LOSS ON HYDROELECTRIC POWER PRODUCTION Current High deforestation No-forest deforestation (1%/year) counterfactual (0.25%/year) Change in reservoir capacity (%) 23 61 100 (65 years) Reduction in value of HEP (US$million) 0.8 2.8 18.2 Note: Present value of reduction in value of HEP computed over 100 years with a discount rate of 6 percent. These figures provide insight into the values at increased flood risks. Electricity users would either risk from forest loss. Farmers would be worse off, face shortages of electricity or be forced to switch by only being able to irrigate smaller areas and to other, more expensive sources. And the entire would also face higher maintenance costs. The global community would suffer from increased population in the lower part of the basin would face emissions of GHGs. 6 These benefits of forests are not included in With clear beneficiaries of services in the Pursat standard national accounts. National policymakers Basin identified, a PES scheme could be devised to have often not been as concerned with forest loss provide payments from the beneficiaries (farmers as they might have been. Conversion of forests to and hydropower companies) that would help support other uses will often appear to be beneficial in upland forest conservation. Carbon emission reduction national accounts, as crop production will add to payments under Cambodia’s Reducing Emissions from them while losses either do not appear at all (carbon, Deforestation and Forest Degradation (REDD+) program biodiversity), or appear but in ways that seem could bring in financing from international private unrelated to forest loss (irrigation). sector for the maintenance of upland forest. KEY MESSAGE 3 Funding for the PES schemes involve payments to the managers maintenance of those forests in the long of land or other natural resources in exchange run can be captured for the provision of specified ecosystem from private and services (or actions anticipated to deliver these international sources. services) over-and-above what would otherwise be provided in the absence of payment. Payments are made by the beneficiaries of the Although public resources are needed in the short services in question, for example, individuals, term to finance maintenance of forests, a range of communities, businesses or governments acting national and international private financing options on behalf of various parties. can be brought to scale to support appropriate forest management. National private financing Source: Smith et al. 2013: 9 options include local non-timber forest product (NTFP) enterprises, the Ibis Rice program and companies investing in sustainable plantations like CamAgra Key Recommendations of the Study and Grandis Timber. International private financing options include conservation trust funds like the The results in this report lead to two sets of VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Conservation International trust fund , private 1 recommendations: (i) policy recommendations for the equity financing and private financing for carbon Pursat River Basin (Recommendations 1 and 2) and (ii) emission payments. This approach is an important recommendations aimed at scaling up the analysis to emerging trend in conservation and Cambodia has ultimately cover all of the country (Recommendation 3). the opportunity to bring this to scale to alleviate public funding sources. With the prospect of more Recommendation 1: Focus forest stable and longer-term funding other than reliance protection and restoration efforts on on traditional grants, this approach can provide more upstream watersheds in the Cardamom effective conservation as well as more positive impacts Mountains Protected Area Landscape on livelihoods. PES for water and payments for carbon to enhance the resilience of water under REDD+ (Reducing Emissions from Deforestation resources. and Degradation) are already being undertaken in Cambodia, but these now need a comprehensive Cardamom forested lands act to slow down high approach to explore those opportunities fully. discharges during wet season and supplement low 1 See more on Conservation International at https://www.conservation.org/projects/cambodias-central-cardamom-protected-forest 7 flows during dry season. Upstream deforestation and conversion is one of the key objectives of MoE’s in the Sam Kos Wildlife sanctuary, which is linked to management of PAs. As agriculture is a key driver improving accessibility by access roads above Dam 1 of forest change in the Cardamoms, it is important (see Figure 2.1), possesses a quantifiable risk to the that this be addressed in a manner that is pro-poor, operation of irrigation and hydropower infrastructure recognizing that forest communities need livelihood downstream. These risks include changes to the pattern support. Developing agroforestry is a way of creating of seasonal water yield, higher consumption of water additional value on lands that have been converted for irrigation in upstream areas during water -stressed to agriculture and restoring trees that provide some months, and increased sediment accumulation in important forest ecosystem services like sediment downstream infrastructure. Measures to arrest the regulation. The interventions that are decided from rate of deforestation and engage in options for the assessment should also be included in the PA afforestation would be recommended to protect this management plan. resource. The results showed that important areas for water yield are overlapping with PAs in the Cardamom Action 3: Develop interventions for reducing Mountains but are also overlapping with an area of the pressure on forest resources from charcoal, the Cardamom Mountains that has experienced high including more sustainable charcoal production rates of degradation due to agricultural encroachment. and environmentally friendly alternatives to wood Deforestation and forest degradation on the other charcoal. The charcoal industry can be a significant hand, will reinforce the impact of the climatic trends, opportunity for Cambodia’s rural PA economies if magnifying the risk of extreme events like floods in done right. GERES (2015) assessed the industry to be the basin. Protection of upstream forestlands takes on worth about US$177 million per year and the RGC will renewed importance in this light. need to play a key role in leading the organization of this industry in order to reduce its potentially negative The RGC has some important decisions to make with impacts on forests. Sourcing wood for charcoal from regards to prioritizing areas in the PA landscape existing plantations can help to address the wood- that require attention, additional resources supply needs for the charcoal industry and improving for management and the implementation of the wood to charcoal conversion efficiency could also restoration activities. In this context and under help to reduce pressure on forests for wood. Four VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A this recommendation there are also some strategic key measures proposed for further action on moving actions proposed for the government as well. towards more sustainable charcoal production are: (i) formalization of existing small-scale charcoal Action 1: Prioritize zoning and development of PA producers and linking these with private sector management plan in Samkos Wildlife Sanctuary and plantations who can provide a consistent source Biodiversity Conservation Corridors. MoE is already of wood through which a certification system for taking a step in this direction through the inclusion of charcoal could be developed; (ii) refining existing Samkos Wildlife Sanctuary as a priority PA for zoning, charcoal producing techniques and technologies to PA management planning and boundary demarcation. improve energy efficiency of the firewood conversion The PA management planning should assess the into charcoal in a cost-effective way; (iii) developing best options for reducing forest degradation and small-scale woodlots within community use zones conversion and should include a plan for restoration and sustainable use zones to meet wood needs of degraded landscapes. This analysis provides for charcoal; (iv) exploring opportunities for more information that MoE can use such as areas of high environmentally-friendly options for charcoal like sediment yield and high-water yield in prioritizing coconut husks. Again, the link to the private sector areas within Samkos for interventions. would be strategic for investments and management of the wood lots. Action 2: Assess opportunities for agroforestry on existing agricultural lands. Halting forest degradation 8 Recommendation 2: Explore the potential PES program that stopped all deforestation completely, for private financing to support PA including both the costs of the payments to participants and the costs of implementation of the program. The management. corresponding figure on damages that irrigated farmers would suffer from deforestation would add to this Ecosystem service values can help inform amount, and, if an ERPA can be negotiated, carbon government decision-makers, but it will not, by itself, payments would increase it even further. change the incentives facing actors on the ground. In the Pursat RB, these actors receive only a small subset Designing and implementing a PES program of forest benefits but stand to receive the bulk of the to reduce deforestation in the upper benefits from alternatives such as agriculture. Pursat River Basin would require: 1. Using the hydrological model developed for PES and REDD+ are realistic opportunities in this analysis to identify the critical areas in the Cambodia for directing private financing to support upper basin: the areas which, if they were to be PA management. There is strong interest of the MoE deforested, would result in the greatest impact on in establishing PES with ongoing pilots helping to hydrological flows and sediment loads; inform development of PES. The success of PES in 2. Undertaking a threat assessment of these areas, to capturing private payments for conservation has see how likely they are to be actually deforested, been shown through several international examples based on factors such as their suitability for including Mexico, Vietnam and Costa Rica. Payments agriculture, proximity to roads, etc., and quantifying from international private sector under the REDD+ the potential benefits to local people of converting mechanism are already being received in Cambodia these areas to other uses; and could scale-up. There are already some excellent 3. Measuring any benefits that retaining forests studies that exist on evaluating carbon resources could generate for local communities, for example (stocks), which can be useful when it comes to through the sustainable collection of NTFPs and developing the right benefit-sharing mechanisms and through activities such as ecotourism; ensuring that these link to an overall revenue system 4. Estimating the cost of a PES program to protect that support protected areas. these critical areas, based on their size (number of VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A The present study contributes to the development ha to be protected), the size of payments needed of PES in several ways. First, it clearly documents to induce forest conservation, based on the net the benefits provided by forests. Second, it identifies costs to local communities of conserving them several important groups of beneficiaries of these (potential benefits from conversion to agriculture services (irrigated farmers and electricity users) and minus local benefits from retaining forests), and quantifies the damages that they would face if these the likely implementation costs of the program, services were lost or reduced. Third, the hydrological based primarily on the cost of monitoring, which is model developed for the analysis provides tools that affected by the size of plots to be monitored, their would permit PES conservation efforts to the areas dispersion and their accessibility; where they would be most effective. 5. Determining whether the program, based on these estimates, is feasible (i.e. the total costs are less To be sure, much more needs to be done. The than total willingness to pay); putting in place quantification of benefits provided by this analysis, for arrangements to collect funding from service users example, only provides an upper bound of willingness (such as irrigated farmers, HEP producers and/or to pay to avoid damages. For example, the analysis carbon buyers) and make appropriate payments shows that at a deforestation rate of 1 percent a year, to service providers (upstream communities who the losses suffered by HEP producers would be about refrain from deforesting). The analysis conducted US$0.17 million a year. This figure is the maximum provides a substantial start on this road map, but amount such producers would be willing to pay for a clearly much more needs to be done. 9 Lastly, it would be important for any PES or REDD+ of the RGC, for example in determining national budget scheme to be a part of ongoing plans within the allocation for MoE for PA management; and (ii) that data government for strengthening the institutional and information will be more reliable and less costly if framework for emissions reduction payments. The data collection, analysis and access are standardized national REDD+ strategy and REDD+ nesting framework under an NCA approach. The analysis of the Pursat Basin that is being established currently provides the demonstrates the potential benefits of undertaking opportunity to provide payments from reducing carbon NCA both to identify the need for interventions and emissions through undertaking forest conservation, to help design them and elucidates some key lessons conservation compatible livelihood activities etc. The for replication and scaling up. Lessons include the: kind of analysis undertaken in this study provides a (i) need for a thorough analysis of the interactions of basis for the levels of investment needed to holistically beneficiaries with ecosystem services; (ii) importance of ensure and incentivize more emissions reduction a robust data collection plan and early commitment on payments. In addition, we recommend the RGC to: data sharing from relevant ministries. 1. Provide oversight and management of REDD+ activities as is being proposed Moving from a single case study to a comprehensive in the REDD+ Regulatory Framework approach requires a road map that includes: (Prakas), which is being developed. 1. Conducting a ‘scoping’ exercise that identifies (i) 2. Set up the national system for emissions reduction policies, decision-making and planning processes payments that includes a benefit sharing for which the implementation of NCA could provide mechanism that will make clear investments critically important information; and (ii) data for forest conservation and protected area availability/needs, institutional framework, financial, management, including co-management, etc. technical resources and capacity required for NCA; 3. Enhance and promote the attractiveness of 2. Identifying and informing key institutional Cambodia for REDD + payments with clear rules partners that should be engaged (i) at the ministry and regulation for the system. level, such as MoE, Ministry of Water Resources 4. Ensure that PES and REDD+ payments are well and Meteorology (MOWRAM), Ministry of Rural integrated into the overall financing mechanism for Development (MRD), Ministry of Agriculture, Forestry and Fisheries (MAFF), Ministry of Interior VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A PAs. Revenues from tourism in PAs should also be considered for supporting PA management. Recent (MoI), Ministry of Economy and Finance (MEF); work for the Greening Prey Lang project identified a (ii) at the provincial level, including Provincial number of fund sources including the Environment Department(s) of Environment (PDoE); and (iii) and Social Fund, the Forestry Administration NGOs and Development Partners like Conservation National Forest Development Fund, ecotourism International, Flora and Fauna International (FFI) revenues and private conservation funds that and World Wildlife Fund (WWF) who currently do needed to be managed and used in an integrative related work on ecosystem valuation; way to be efficient and effective. 3. Considering a phased approach, starting with basin-specific accounting-compatible assessments Recommendation 3: Develop a road map with a small set of key ecosystem services - such for scaling up assessments of economic as those in the current report - and evolving towards a more encompassing exercise that would benefits provided by forest ecosystems in time be extended to the country’s national across Cambodia using a Natural Capital boundaries. A phased approach could initially Accounting (NCA) approach. focus on representative watersheds where there are clear beneficiaries, as in the case of the Pursat The advantages for Cambodia of a using a NCA Basin. Criteria for prioritizing areas for undertaking approach verses one-off economic valuation studies ecosystem service accounting-compatible are: (i) standardizing how ecosystem service values are assessments may include: areas that are most at determined and integrated in regular decision-making risk from degradation and forest loss; watersheds 10 important for hydropower and irrigation, water assessment to include watersheds that feed into production and sediment regulation. Additional the Tonle Sap and Kulen Mt. and Kbal Chay where ecosystem services that should be considered the government is pursuing pilot PES projects; include water flow regulation with a particular focus 5. The scoping and road map would most certainly on drought. Estimation of economic benefits of highlight the need for enhancement on monitoring hydrological, carbon and tourism ecosystem services and generation of data for similar assessments and should be prioritized for the following reasons: ultimately for accounting efforts: Cambodia is experiencing a serious water a. a. As changes in soil erosion and sediment shortage, which is expected to be exacerbated accumulation are significantly affected by by climate change. It is therefore important that forest change and can result in large costs, it is the RGC, through MoE and MOWRAM, strengthen recommended that monitoring of suspended management of important watersheds like the sediment and bed load, at least at the site of Cardamom Mountains and Kulen Mountains the future dams or at the main gauging station with protection of existing forest resources and be undertaken. Additionally, experimental plots restoration of degraded important watershed to monitor soil erosion rates could be helpful in areas. Analysis of hydrological ecosystem services, verifying soil loss projections. as undertaken for this study, will be instructive for b. Rainfall variability in the mountainous region prioritizing areas for watershed management. is high. Weather monitoring needs to be b. Cambodia has invested significantly in strengthened to derive accurate estimates to hydropower plants on the Mekong River, as water resources available. This will become well in the Cardamom Mountains. Ensuring especially important as precipitation patterns a close to maximum operation capacity of continue to shift with a changing climate. A more these hydropower plants will be important extensive network of rainfall gauges is needed. for energy security in Cambodia especially in c. Groundwater often plays an important role, but the dry season, which means protecting forest data on groundwater are even less available watersheds that are upstream of this dam. than for surface water. Improved groundwater Valuation of hydrological services can support mapping and monitoring is needed to better VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A schemes for hydropower companies to provide understand the role that it plays. The first finance that can assist with the management essential step of this should be to map of forest resources that provide critical water the major aquifers, followed by identifying flow regulation and sediment regulation for the important recharge areas and travel time. operation of hydropower plants. Without this, we risk incomplete protection c. Carbon storage, as an ecosystem service, for water source areas, as we are limiting our is strategic to analyze, as there are well- source region of the rivers based on visible established methodologies for doing terrain slope contributing to the river water. this. The carbon market establishes a 6. The benefits of forest for disaster reduction – flood price which often accounts for regulating mitigation and forest fire prevention – would also ecosystem services (which have no be important to capture in subsequent analysis. market and/or are difficult to value) and Data on the flood, drought and fire damages would biodiversity that facilitate carbon storage. be important for determining the benefits provided d. Ecotourism development in PAs is a priority of by forest in terms of disaster risk reduction; the RGC to boost the overall tourism sector in 7. Additional benefits that would be important terms of jobs and value added, provide income to capture include: water used for domestic for rural and forest communities and generate purposes; recreational ecosystem services resources that can help with PA management. from ecotourism should be considered for 4. We would recommend initial consideration areas where this is significant; non-timber for geographic priorities for undertaking such forest products (NTFPs); and charcoal. 11 1 BACKGROUND & OVERVIEW VA L U I N G T H E E CO SYST E M S E RV I C E S P R OV I D E D BY F O R E STS I N P U R SAT BA S I N , C A M B O D I A Terrestrial ecosystems such as forests provide important ecosystem services, including freshwater-related services such as sediment regulation, preventing too much soil erosion from filling reservoirs and other irrigation works and reducing their capacity; and water-regulating services affecting the timing of supply - satisfying irrigation water demand, recharging storage tanks and reservoirs, maintaining minimum flows in the river and reducing the likelihood of flooding. Unsustainable land use can drastically alter the ability of these ecosystems to continuously provide these ecosystem services and climate change can exacerbate these effects. Assessing the magnitude of these benefits, both in biophysical and monetary terms, is therefore critically important for decision-making, as it enables a better understanding of changes and trade-offs, leading to more informed and sustainable economic development and planning. The Chambok water fall at Kirirom National Park © Conservation International/photo by Sokhorn Kheng 12 Forests occupy a significant area of Cambodia and for land and unsustainable use of natural resources the majority of the country’s forests and important (Figure 1.1). Forest degradation is pervasive and watersheds are in Protected Areas (PAs). Forests cover attributed to unsustainable logging practices, salvage about 8.1 million ha (45 percent of the country) with logging and fuelwood extraction. Land expansion has intact forest patches interspersed with secondary been a major factor, with cultivated land for crops forest and areas that have been cleared for agriculture increasing by 50 percent between 2002 and 2012, or other activities (MoE 2018). The PA system covers partly at the expense of forests (Forest Trends 2015). 7.4 million ha (41 percent of Cambodia) and includes In addition, approximately 5.5 million tons of fuelwood about 67 percent of Cambodia’s forest area. Many are used each year by households and small and protected areas (PAs) face threats from encroachment medium enterprises (SMEs), with 88 percent of the for cultivation, charcoal production, illegal timber population still relying on traditional biomass for harvesting and wildlife poaching. cooking (GERES 2015). Under the PA system, forests will be subject to stricter management that includes By putting forest resources under the Protected Areas no timber harvesting, sustainable harvesting of non- system, the Royal Government of Cambodia (RGC), timber forest products (NTFPs), patrolling by rangers has made key steps to ensure the reversal of the and forest communities to prevent illegal activities declining trend on forest cover. Cambodia’s forest and forest restoration where needed. cover has declined since 2006 because of pressure FIGURE 1.1: FOREST LAND COVER CHANGE IN CAMBODIA, 2006-2018 65% 60% VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A FOREST COVER ( PERCENT) 55% 50% 45% 40% 2006 2010 2014 2016 Source: MoE 2018 13 1.1 How are Forest Ecosystem Services This demonstrates an increase in gross national Contributing to Cambodia’s Economy? per capita from US$300 in 1994 to US$1,623 in 2019 (current prices) and a graduation to lower-middle- Natural capital such as forest resources account income status in 2015. Cambodia’s forest ecosystems for about 40 percent of Cambodia’s wealth and provide important services like water and sediment contributes significantly to its economy; see Figure flow regulation, biodiversity disaster mitigation 1.2. This capital, which includes agriculture, forests that underpin Cambodia’s key sectors – agriculture, and fisheries have contributed to Cambodia’s strong garments and tourism. economic growth at 7.6 percent from 1994–2019. FIGURE 1.2: NATURAL, PRODUCED, AND HUMAN CAPITAL IN CAMBODIA, 1995 – 2014 100% 90% 80% 70% % OF TOTAL WEALTH 60% 50% 40% 30% 20% 10% VA L U I N G T H E E CO SYST E M S E RV I C E S P R OV I D E D BY F O R E STS I N P U R SAT BA S I N , C A M B O D I A 0% 1995 2000 2005 2010 2014 Natural Capital Produced Capital Produced Capital Terrestrial forests in Cambodia are important contributor to the agricultural economy. Cambodia habitats for freshwater. All of the headwaters of the relies on irrigation for about 16 percent of its total country’s five major river basins are under forest cover cultivated area, and about 10 percent of its rice crop. (ADB 2014). Analysis by Bottrill et al. (2015) presented Water for irrigation is drawn from surface water in Figure 1.3 showed that intact forests in the sources and, increasingly, groundwater is being used Cardamom Mountains, and in the eastern mountains for irrigation, especially in the south of the country. of Cambodia, are important for regulating water flows Studies indicate that wet season irrigation has very and facilitating important water purification processes. little impact on rice yields and is more important for dry-season rice. The Ministry of Water Resources and Agriculture, which depends heavily on water flow Meteorology (MOWRAM) plans to increase its irrigated regulation, erosion regulation and nutrient retention area to 872,000 ha (by 2025) from 672,000 in 2015 with services provided by forests, contributed 22 percent surface water structure initiatives like large irrigation of GDP in 2018.2 Rice production is the most important canals which underscore the importance of sustaining 2 https://data.worldbank.org/indicator/NV.AGR.TOTL.ZS?locations=KH 14 surface water flows (Raju and Taron 2018).3 As much Cardamoms hosts the longest wild elephant track as 85-90 percent of household income depends (Koh Kong, southern Cardamom) in the world and has directly on ecosystem services (fisheries, timber, wild successfully preserved elephants from poaching over food, crops and firewood).4 Forest products that are the last ten years. The RGC is exploring ecotourism as collected frequently include timber, bamboo, rattan, a driver to strengthen management of its rich natural other edible plants and wild food such as snails, frogs, capital and boost economic prosperity. With a captive eels and crabs for household consumption and sale. tourism market focused on the Angkor temples, the MoE and Cambodia’s Ministry of Tourism (MoT) are Fisheries are also dependent on forest services. now developing management policies, regulatory Freshwater mangroves and flooded forests provide frameworks, strategies and guidelines and making habitat support for more than 300 species of fish and investments to expand ecotourism. crustaceans and thereby help support the fisheries industry. Freshwater fisheries in the Mekong River and Forest also help to sustain hydropower in Cambodia Tonle Sap are a vital part of the country’s economy, which provide power to the garment and other food security and culture. The fisheries sector provides industries. The RGC has indicated that investing in employment to 2 million people, accounting for 75 hydropower, and sustainable energy is a national percent of households’ animal protein intake and priority and have already made significant investments contributes to about 12 percent of GDP. in dams in large hydropower plants in Koh Kong and Kampot provinces. The country’s growing energy Tourism and ecotourism, which contributed to needs are met in part by hydroelectric power (HEP), approximately 18 percent of Cambodia’s 2018 where forest systems are important for water flow GDP, is dependent on the country’s considerable and sediment regulation in rivers. The RGC estimates biodiversity in forest landscapes. Cambodia, which that HEP will become an increasingly important part sits within the Indo-Burma biodiversity hotspot, is one of its energy supply mix in the medium term (from 26 of the most biodiverse countries in Southeast Asia. percent in 2013 to about 50 percent by 2020) and have In total, an estimated 53 percent of this biodiversity made significant investments already in this (see Table is contained in the country’s protected areas (Bottrill 1.1). However, Cambodia is experiencing increasingly and others 2015). The Cardamom Mountains forest severe droughts which affect the performance of the VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A landscape, in particular, has a remarkable diversity hydropower plants. A reduction of 75 percent of rated of animal species, including elephants, bears, gaur capacity of hydropower plant occurs during the dry (the world’s largest bovine) and freshwater fish and season and this will be exacerbated under climate new species are regularly being discovered. The change (RGC 2014). TABLE1.1: LARGE HYDROPOWER DAMS IN CAMBODIA Hydropower Project Installed Capacity (MW) Cost (US$million) Year of Operation Lower Se San II 400 MW 800 Dec. 2018 Stung Tatai 246 MW 540 Aug. 2014 Russei Chrum Krom 338MW 500 Jan. 2015 Sources: Dreher and others 2017; Khmer Times 2015. 3 Erban and Gorelick (2016). Closing the irrigation deficit in Cambodia: Implications for transboundary impacts on groundwater and Mekong River flow 4 Persson and others (2010) examined the connection between ecosystem services and livelihoods in Kratie, Kampong Thom and Battambang provinces. 15 FIGURE 1.3: OVERLAPS BETWEEN FRESHWATER HABITATS AND FOREST LANDSCAPES IN THE EAST AND WEST MOUNTAINOUS AREAS OF CAMBODIA VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Notes: Darker blue areas show important freshwater habitats which overlap Source: Bottrill et al. 2015 with forest landscapes in the east and west mountainous areas of Cambodia 1.2 Valuing Cambodia’s Ecosystem Valuations should include both provisioning services Services is Relevant and Timely (e.g. timber, firewood, NTFPs and agriculture) and regulating services (e.g. carbon storage and water). “A priority for resource assessments is the valuation This type of economic analysis is key to sustaining of the natural assets present in key protected areas to financial support for protected areas as part of a document their importance and to better evaluate the national Green Growth strategy.” Excerpt from NPASMP, relative benefits (and costs) associated with investment MoE (2017:7) for development purposes versus protection. 16 Influencing the RGC’s decision-making on illegal timber harvesting and wildlife poaching.5 MoE’s investments in Protected Areas conservation and foremost policy documents on PAs – The Protected protection is among the core reasons for valuing Areas Law of 2008 and the National Protected Areas ecosystem services in Cambodia. A large proportion Strategy and Management Plan (NPASMP) 2017-2031 – of Cambodia’s natural landscape and rich biodiversity strongly advocate the use of ecosystem service values is contained within PAs, which makes them important in conservation planning and prioritization of areas for conservation and protection but still they remain in PAs for strongest protection efforts and helping to susceptible to illegal activities. For example, even address drivers of deforestation. The NPASMP also though 62 percent of the Cardamom Mountains contains performance targets for valuation studies are under the PA system, it still faces threats from influencing PAs development which emphasize the encroachment from cultivation, charcoal production, importance among the RGC for valuation.6 BOX 1.1: DRIVERS OF DEGRADATION IN PROTECTED AREAS Forest degradation in PAs, like the Cardamom Mountains Protected Landscape, is mostly driven by small-scale clearings in the forest for the cultivation of crops such as rice, land grabbing, and charcoal and fuelwood collection. The principal cause of this type of forest clearing is the combination of in- migration and poorly defined property rights, compounded to an extent by local population growth. Infrastructure development is increasing access to the area for in-migration from outside. Land grabbing and speculation by powerful local actors has also been identified as a proximate cause of deforestation, with land often being cleared and sold to businessmen from nearby towns. Reliance of local communities on forest resources for their subsistence, such as timber for housing, firewood, non-timber forest products, fish and bushmeat, can also have detrimental effects on forest biomass, although less severe. Charcoal production is a significant cause of deforestation at eastern edge of the Cardamom Mountain VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Protected Landscape due to demand for charcoal in Phnom Penh, Pursat and Koh Kong. Despite the official logging moratorium, there are indications of small-scale (illegal) logging in the Cardamoms. Illegal logging is carried out by villagers who log and sell timber to make a living, and by migrants who are often employed by powerful local actors. It is mostly undertaken in the form of small-scale selective logging, targeting mainly luxury woods. Sources: Conservation International (2012); Initial Forest Reference Level for Cambodia under the UNFCCC Framework (2016) Cambodia’s impending water crisis and the impacts Cambodia’s economic activities. Forests, in helping to of climate change will put increased attention on maintain water flows especially during the dry season, forests and the water flow regulation and disaster help to reduce drought conditions. Seasonal water mitigation services that they provide. Forests also scarcity exists in the river basins of the Tonle Sap provide disaster mitigation benefits that help protect River Basin Group (e.g. in 2016, 18 out of 25 provinces 5 A National REDD+ Strategy undertook an analysis (for the entire country) on drivers of degradation, which helped identify the threats being faced in PAs (RGC 2017). This (government-led) analysis was used in the formulation of the government’s forest investment plan in 2018. 6 A target of ‘5 PAs that have completed valuation studies’ by 2021 has been set by MoE as part of its NPASM (2017-2031). It is expected that 20 PAs would have completed valuation studies by 2031 (MoE 2017). 17 are affected by droughts with around 2.5 million stakeholders and user groups, and critically also people lacking water) and existing reservoir storage on the willingness to pay.7 This analysis will add capacity (less than 10 percent of total water generated) to the growing body of work needed to design a is insufficient to redistribute water significantly comprehensive PES mechanism in Cambodia. between seasons. In the wet season, forests also play important roles by acting as stores of water during In order to meet the demand of the RGC for values- rainfall events and regulating (slowing) the speed at driven decision-making on forest and PAs, there which water enters rivers and streams. Drought and needs to be development of government-endorsed flooding are expected to be exacerbated in Cambodia valuation methodology, development of case because of climate change, with expected changes to studies of ecosystem services value assessments be around an extra 1.5 months of drought, and with and capacity development, among other things. flooding expected to be more extensive particularly in Economic analysis has played a limited role in the eastern areas of the country. Impacts of climate determining how forests are managed and financed change on forests have not been studied in Cambodia, in Cambodia. Data, information and analysis by NGOs, but evidence from similar tropical environments including some economic analysis, have, over the indicates that with drier conditions forests will be years, influenced various national policies, including more susceptible to fires, and excessive rain can on forest management, and also in the development increase landslide occurrence. Conservation of forests of the draft Environment Code. However, there has as freshwater habitats is therefore critical. been little systematic economic valuation work at the national level that has determined decisions around Strong interest of the RGC in Payment for Ecosystem how resources are allocated for forest protection Services (PES) to support natural resources at the national or subnational levels. There are two management underscores the need for ecosystem important opportunities now for the RGC integrating service valuation studies. PES initiatives are ecosystem service values into forest and PA decision- being piloted in Cambodia, with the support of making: (i) the Cardamom Sustainable Landscape and Conservation International (CI) and UNDP. Following a Ecotourism (CSLE) project which focuses on improving visit of MoE officials to Costa Rica for a PES study tour, the management and value addition of the natural there was strong interest in developing a PES scheme capital in the Cardamom Mountains and Tonle Sap VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A in Cambodia. Two pilot sites have been selected by landscape, and (ii) a new landscape project being MoE to look at how to implement PES – Phnom Kulen prepared which will focus on landscape management watershed which provides water for Siem Reap, and with a focus on integrated watershed management Kabal Chay watershed which feeds Sihanoukville. in PAs of the eastern river basins – Sen, Chinit, Upper UNDP commissioned a study to look at what PES Mekong and 3S (Sekong, Sesan, Srepok). mechanisms may be put in place for these two sites and some recommendations were made. These were 1.3 About this Analytical Work and Report more general recommendations, which provided ideas about the application of possible tourism The analytical work is being undertaken as part of fees to tourist areas (e.g. for Khulen mountain, or a broader effort of the World Bank in Cambodia a surcharge for hotels in Sihanoukville), although to provide guidance to the RGC through technical practicalities on how to design and implement a assistance and analytical and advisory services on PES system were not covered. Indeed there was a managing its natural capital through landscape lack of funds to provide a detailed analysis on the approaches. The World Bank in Cambodia has agreed 7 A national dialogue on PES was organized by UNDP and CI in September 2019 to keep the dialogue on PES moving forward. It brought together people across different ministries as well as stakeholders from both Khulen and Kabal Chay (e.g., water utility in Kabal Chay, Angkor beer, etc.). It was a way to explain PES to various stakeholder and ministries as well as to gauge their thoughts on willingness to pay. Since then, UNDP did provide MoE with recommendations for steps going forward (e.g. a fund for Kabal Chay), but there’s been no further action on this to date. 18 within its Country Partnership Framework (CPF) to impacts of deforestation on forest ecosystem services. support the Cambodian government on strengthening Methodologies for measuring and valuing forest the sustainable use of natural resources. Working ecosystem services have also been developed as well through a landscape approach in the Cardamom as discussed with the RGC, and documented for the Mountains, the Mekong Delta and in the upstream government as a tool to assist in decision-making. Mekong, the Bank is helping the RGC to better These methodologies are expected to contribute to manage and add value to its natural capital through the growing body of literature on science-based policy strengthening its links to the economy. and decision-making. Recommendations for scaling- up ecosystem services assessment and valuation Accordingly, this study provides evidence of the are provided as a road map within this report and importance of forests in the Cardamom Mountains actions are proposed for how this work can inform in providing ecosystem services, inter alia, that are investments for forest and PA management, namely important for Cambodia’s economy and climate conservation, protection and restoration in the and disaster resilience and helps the RGC to Cardamom Mountains. The work will also provide the quantify and communicate the value of its natural underpinnings for tools like PES as well as investments capital to Cambodia’s economy. This is done by in forest and PA management, both of which will be quantifying ecosystem services provided by forests developed under the CSLE project (Box 1.2 below), as in biophysical and economic terms, and evaluating well as tools such as natural capital accounting. BOX 1.2: CAMBODIA SUSTAINABLE LIVELIHOOD AND ECOTOURISM (CSLE) PROJECT (US$55 MILLION) This project’s objective is to improve PA management and promote ecotourism opportunities and non- timber forest product (NTFP) value chains in the Cardamom Mountains Tonle Sap (CMTS) landscape in Cambodia. The CSLE project will also support the RGC in strengthening the legal and regulatory framework for the management of ecotourism investment projects (EIPs) in PAs. VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A To achieve this, the project will strategically invest in areas that are strongly aligned with RGC’s development plans. There are three components: • Strengthen Capacity for PA Landscape Planning and Management • Strengthen Opportunities for Ecotourism and NTFP Value Chains • Improve Access and Connectivity. The analysis focused on selected ecosystem services it was selected to strategically demonstrate how provided by forests in the Pursat River Basin of the the valuation of ecosystem services is undertaken, Cardamom Mountains: (a) seasonal water-regulation particularly where there is a stream of benefits service, which affects the availability of water for from the ecosystem flowing to well-defined users, irrigation; and (b) sediment regulation service, which and also to demonstrate how valuation can inform affects the operability of hydropower. Another analysis PA management. The biophysical and monetary was also done to assess the value of tourism and assessments follow approaches consistent with forest carbon in the same area. The selection of the guidelines proposed by the System of Environmental- Pursat watershed does not indicate its importance Economic Accounting Experimental Ecosystem relative to other watersheds in Cambodia. Instead Accounts (SEEA EEA). 19 Focus on the Pursat River Basin The Pursat River Basin is at an interesting juncture of its development trajectory. For millennia, the floodplains of this river have been used for cultivation. Dominated mainly by rice, the basin’s agricultural rhythm had been dictated by seasonal changes. Agricultural irrigation has relied on rainfall, the flood pulse of the Tonle Sap and water from the Pursat River – which is fed by rainfall high in the Cardamom Mountains. Today, the river offers an opportunity to propel this region’s economic and human development forward, by constructing irrigation infrastructure to stabilize and extend the growing season and also by harnessing the power generating potential of the river. As built infrastructure is developed in the region, it is important to also account for the basin’s natural characteristics (in the form of terrestrial ecosystems in the Cardamom Mountains) and understand how ecosystem degradation could change the ability of the basin to reach its potential. The Pursat River Basin provides an excellent case study for ecosystem service valuation as it is an economically important area and there are clear beneficiaries of services being provided by the forest ecosystem there. As the basin supports a range of different uses from irrigation for rice production to hydropower and ecotourism as well as PAs, it can be considered a representative for other basins in Cambodia and so methodologies applied here can be used and replicated in other river basins in Cambodia. The report puts forward three key messages (KMs) 1.4 Overall Methodological Approach to and three recommendations: Valuing Ecosystem Services KM 1: The economic benefits from intact forests The major analytical activities and methodological (estimated at US$99 million) are almost five approach (described below) for this study focus on the times higher than the gains from cutting them measurement and valuation of hydrological services down for small-scale agriculture and charcoal provided by forests. Work was also done to estimate production, which is estimated at US$22 million. the value of carbon and tourism benefits from forests, KM 2: Investing in the maintenance of forest and the methods that were used are described in VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A is good business. Annual public expenses to more detail in both Chapters 4 and 5 as well as in maintain the forest in the Pursat Basin are Appendices 4 and 5. about 20 times lower than the benefits provided by them. Using a Natural Capital Accounting Framework KM 3: Funding for the maintenance of those forests in the long run can be captured from Environmental and natural resources make important private and international sources. contributions to economic activity and human well- Recommendation 1: Focus forest protection and being, but these roles are often not recognized nor restoration efforts on upstream watersheds in the included in national accounts but rather attributed to Cardamom Mountains protected area landscape other activities. As a result, these important resources to enhance resilience of water resources. are often thought to be without value and damage Recommendation 2: Explore the potential for to them is not taken into consideration when policy private financing to support PA management. decisions are made. Recommendation 3: Develop a road map for scaling up assessment of economic benefits A system of environmental accounts has been provided by forest ecosystems across Cambodia developed to complement standard national using a Natural Capital Accounting (NCA). accounts to remedy this problem.8 These accounts 8 See https://seea.un.org/ for more information and technical manuals on the system of environmental accounts. 20 begin by identifying the stocks of environmental forests in the Cardamom Mountains (which compares resources and the flow of services that they provide, current benefits to those that would be received if the then they track how these stocks and flows change forests did not exist) with an analysis of the economic over time. Taking the additional step of estimating the costs resulting from current deforestation rates (which value of these stocks and flows allows their benefits compares current benefits to those that would be to be quantified in a form that enables comparison received if forests were partially lost). to other economic activities. This is what this report seeks to do in the case of the Cardamom Mountains. Biophysical Assessment and Economic Valuation of Hydrological Ecosystem Services The economic value of an ecosystem — such as a forest — is estimated by summing the values of the various The core component of the hydrological analysis was flows of benefits it is providing. For comparability with the setup, calibration and validation of a hydrological the value of other assets in the national accounts, model for the river basin that could be used to this value is usually expressed as the asset value: the analyze spatial and seasonal variation to water and present value of the flows of benefits expected to be sediment yield. provided over a given time horizon. • After a brief survey of potential hydrological For comparability of the forest asset with other models that could be suitable for this study, a assets in the national accounts, the flow of benefits Soil & Water Assessment Tool (SWAT) hydrological is measured against the counterfactual in which the model was selected and setup using data derived asset does not exist. This is relatively straightforward from local and global sources. Daily discharge from when it comes to the ecosystem’s flows of benefits. Bac Trakuon monitoring station was used to ensure Without a forest, there would be no timber for the hydrological model for Pursat Basin provided harvesting, so the flow of that service would disappear. river flow estimates that are representative of Other services, however, would not necessarily known records. disappear. Even if there were no forests, rain would • The impact of current and planned infrastructure continue to fall and so water would continue to flow on the river discharge (mainly 3 dams currently — but it’s flow would be different and the water would designated as Dam 1, 3 and 5) was achieved VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A be dirtier, which would affect the level of benefits it through a simple dam model that ran on the time provides. Estimating how these benefits would change series results extracted from the hydrological in the absence of forests is one of the main challenges model. This was done specifically because (1) at the we face in a work such as this. time of model-setup, the operation rules for the 3 dams were not known; and (2) an external dam It is important to note that this asset value, model allowed us the option to modify operation while useful for comparison to the value of other rules in real-time during stakeholder gatherings assets, does not speak directly to the decisions and examine together the impact downstream. facing policy-makers. The forests in the Cardamom • Once the hydrological model performance was Mountains, while under pressure, are in no immediate deemed satisfactory, the resulting time series danger of vanishing entirely. For policy purposes, it of water and sediment yield (both overall and is more relevant to examine how the value of the associated with different land covers) for two basin benefits they provide would change under likely conditions was extracted to facilitate the monetary changes, such as current deforestation rates. This evaluation. The first condition represents the is akin to asking how the value of the house would current land cover in the river basin and was based change if there was a leak in the roof that lets rain on official land cover for 2016 received from the come in; the value of that damage would be less Cambodia’s Ministry of Environment. The second than the entire value of the house. In this report, we condition represents a hypothetical change - a ’no complement the analysis of the asset value of the forest’ counterfactual. Under this, all mature forest 21 in the basin was replaced by barren or bare earth • For valuation of soil erosion control services, we land cover. As a note of explanation, the rationale used soil erosion rates for different land cover of the counterfactual is to provide an indication types derived from hydrological analysis. These of the total economic value of having intact rates were used to simulate total sediment forests and should not be confused with scenarios that is deposited in the dam in a progressive developed from policy and planning perspectives deforestation at 0.25 percent, 1 percent and that will apply more realistic deforestation/ no-forest counterfactual. This simulation was afforestation rates. run for 100 years – which is assumed to be the • A complementary aspect of deriving seasonal lifecycle of the dam. The corresponding loss (monthly) water availability was estimating in hydro energy values was then estimated seasonal variation in water required for irrigation. based on reduction in dam capacity due to For this, the net projected command area for the sedimentation. We estimated total value of wet and dry season paddy crop was adjusted forest for provision of this service and at the with known cropping patterns to an approximate same time estimated how much of the value is total area requiring irrigation at any month in a threatened by different rates of deforestation. year. Then, using rainfall and evapotranspiration Because of a long-time horizon of dam operation, estimated (from the hydrological model), crop we reported economic value in net present water requirement from irrigation (for each month) value (NPV) term as well as in annual values. was calculated using FAO’s irrigation water demand assessment method. Estimating tourism and carbon benefits. Data on visitors to tourism sites in the Pursat Basin Economic valuation in this analysis takes inputs (arrivals, visitor spending and length of visit) was from hydrological modeling. Two specific inputs are used to estimate benefits from tourism where forest water supply to meet rice irrigation demand and soil ecosystems are the key tourism asset. Data was erosion control services to avoid dam sedimentation. obtained from the MoE ecotourism records and assumptions about changes in visitor spending, • For water supply services a production function visitor arrivals and length of day were agreed with method (called Residual Imputation Method) MoE. Estimating carbon benefits followed standard VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A was used. In this case, water is considered an Verified Carbon Standard (VCS) methodology VM0015 input to rice production process alongside other for assessing carbon stocks. Data on carbon stocks inputs such as fertilizer, seed, chemicals labor by forest type was obtained from forest carbon stock etc. Although the prices of most inputs are known assessment done for Cambodia’s REDD+ program from market information, the price of water is not and the price of carbon determined from emission known because it is not traded in the market like reduction programs in Cambodia. other commodities. The purpose of the production function method is to tease apart the value of The estimated economic benefits of forests are the water based on information such as rice yield, summation of benefits derived from the three studies market price and cost of all known production on hydrological, tourism and carbon benefits. The inputs. This value of water was then used to counterfactual approach (if forests were removed) was estimate the economic value of forests which affect used to estimate net present value of benefits of forest water supply. This was executed by comparing for irrigation, HEP and tourism. The deforestation water supply and demand in the presence and scenarios approach was used to estimate annual absence of forest as described above. We then benefits of forests for irrigation, HEP, carbon storage estimated how much of that value is threatened and tourism. by ongoing deforestation (at 0.25 percent rate) and also in a more aggressive deforestation scenario Estimates of benefits of forest conversion were (at 1 percent rate). determined by assessing the economic returns that 22 could be gained if forests in the Pursat River Basin are Pursat Basin like flood regulation and biodiversity converted at an annual rate of 0.25 percent (about 980 regulation. If assessed and valued, these values will hectares) for charcoal and agriculture. The stream of lead to a more comprehensive assessment of the benefits and costs of forest conversion were compared benefits provided by forest ecosystems. This is a using net present value (NPV) analysis for a period of limitation of the overall study and it is recommended 50 years at 6 percent discount rate. to be looked at in further work. The study acknowledges that there are more A schematic of the key tasks for biophysical modeling ecosystem services provided by forests in the and ecosystem valuation is provided in Figure 1.4 FIGURE 1.4: SCHEMATIC DIAGRAM OF THE SEQUENCE OF BIOPHYSICAL AND ECONOMIC VALUATION IN SEEA EEA CONTEXT, WITH AN ILLUSTRATIVE EXAMPLE OF WATER SUPPLY FOR IRRIGATED AGRICULTURE Biophysical Economic assessment valuation Ecosystem Ecosystem services use services Ecosystem Ecosystem and benefits supply and Ecosystem condition (e.g., volume use values services supply extent (e.g., land Ecosystem (e.g., volume of of water (e.g., monetary (e.g., forested degradation) Supply-Use water supply) supply for value of water sub-basins) (km2 of intact Account (m3 of water irrigation) for agriculture) (km2) forest per (m3/ha of ($/ha of supply) sub-basin) irrigated irrigated agriculture) agriculture) VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Data for Analysis for the Pursat Basin; results of the valuation of hydrological services in the Pursat Basin are presented The data used in the analysis is listed in Appendix 1. Chapter 3; results of carbon and tourism valuation are presented in Chapters 4 and 5; and in Chapter Structure 6 a discussion on policy implications of this work including recommendations for next steps is provided. The rest of the report is structure as follows: Chapter An extensive methodology for the analysis is included 2 presents the results of the hydrological analysis as a separate document. 23 2 HYDROLOGICAL ANALYSIS VA L U I N G T H E E CO SYST E M S E RV I C E S P R OV I D E D BY F O R E STS I N P U R SAT BA S I N , C A M B O D I A 2.1 Overview This chapter outlines the development of the mathematical modeling of seasonal flow variation and soil erosion in the Pursat River Basin, including data preparation, set up and results. The analysis derives seasonal estimates of changes in water availability, irrigation demand and sediment flows. The results from this analysis - specifically, estimates of water deficit relative to irrigation demand and sediment yield for different land cover scenarios— provides the basis for the economic evaluation in the next chapter. The overall objectives for the hydrological analysis can be summarized as: • Develop a hydrological model of the basin to estimate current water and sediment yield; • Understand spatial contributions and seasonal changes to these quantities; • Understand the relation of water and sediment yield to land cover Sebastian Troeng records drone photos and video footage while traveling across Tonle Sap Lake during the 2018 Asia Pacific Executive Leadership (APEX) meetings in Cambodia. © Conservation International/photo by Sebastian Troeng 24 25 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A STUDY AREA MAP 2.2 Pursat River Basin Pursat River, as the Svay Daunkeo and Moung Ruessei streams provide relatively little water. The Pursat River originates in the Eastern slopes of the Cardamom Mountains and flows into the Tonle Official land cover for the year 2016 was obtained Sap. The basin has an area of roughly 5960 km2; from the MoE. A significant proportion of the basin the river is approximately 150 km long. Its two main (~ 68 percent) is covered by natural forested land - tributaries, the Stung Peam and the Stung Santre (Prey predominantly evergreen forest. However, the area Khlong), join the mainstream just above the gauging upstream of the proposed Dam 1, within the Sam station of Bac Trakuon (Figure 2.1). Kos Wildlife Sanctuary, appears to have already experienced significant changes in land cover, Water storage infrastructure in the Pursat River Basin driven by conversion of forests to cropland and supports hydropower and irrigation. Two dams (Dam rubber plantations. It should be noted that the MoE 3 and Dam 5) have been constructed in the basin and classification system defines “cropland” as land cover are operational, and a third, larger hydropower dam that “includes arable and tillage land and agroforestry (Dam 1) has been approved for construction. About 20 systems where vegetation falls below the thresholds irrigation infrastructure works are at varying degrees of used for the forest land category.” Unlike the “paddy implementation, including irrigation schemes (Damnak field” land cover category, this land is not associated Chheur Kram & Damnak Ampil) in which water from with any specific crop. In at least some cases, these the Pursat River is to be diverted to the smaller Svay cropland parcels are likely to be in transition from Daunkeo and Moung Ruessei streams, west of the natural land cover to agriculture. The area cleared for Pursat Basin. The total command area for the schemes agriculture in Sam Kos Wildlife Sanctuary is largely of is listed in Section 2.4. In this study, we assume that this land class and is expected to include cash crops, most of the water for these schemes depends on the cassava and orchards. FIGURE 2.1: LAND COVER 2016 MAP AND DISTRIBUTION VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A 26 2.3 Key Hydrological Datasets estimates from the ERA5 global dataset at 13 points covering the basin have been extracted using the Precipitation Data Google Earth Engine. ERA59 is a climate reanalysis dataset developed through the Copernicus Climate Daily precipitation data covering the period 2004- Change Service (C3S). After comparison of the two 2018 is available from Cambodia’s Ministry of Water datasets (Figure 2.2), the ERA5 dataset was selected Resources and Meteorology (MoWRAM) for a single as the primary precipitation dataset used in the study precipitation gauge in the basin. Given the variability owing to its superior coverage. of altitude within the basin, daily precipitation FIGURE 2.2: COMPARISON OF ERA5 ESTIMATES WITH VALUES MEASURED AT LOCAL GAUGE (A) MONTHLY AVERAGE (B) MAXIMUM RECORDED 12 120 ERAS data Pursat Gauge ERAS data Pursat Gauge 10 100 8 80 P (MM) P (MM) 6 60 4 40 2 20 0 0 Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Notes: (a) ERA5 monthly averages are in general higher than the local gauge while (b) ERA5 generally underestimates peak precipitation rate. Given that ERA5 provides areal averages rather than values at a specific point, both these trends are consistent with expected behavior. VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Source: Authors’ creation Discharge Data The overall behavior is consistent between these two datasets (Figure 2.3). However, extremely low MoWRAM reported (pers. comm.) monthly water flows are recorded in the first quarter for the years availability and discharge for the Bac Trakuon 2014-2016. While drought-like conditions have monitoring station based on monitoring from 1997 to been reported over this same period, the daily 2011 (Table 2.1). Actual daily gauge data for the same gauge reading of no flows (0 m3/s) from 7 January station was made available by MoWRAM covering the 2015 to 13 June 2015 appear uncharacteristic. period from 1 January 2007 to 31 December 2016. 9 Data processing for ERA5 is carried out by the European Centre for Medium-Range Weather Forecasts (ECMWF), using ECMWFS’ Earth System model IFS, cycle 41r2. The name ERA refers to ‘ECMWF ReAnalysis’, with ERA5 being the fifth major global reanalysis produced by ECMWF. 27 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A 28 TABLE 2.1: WATER AVAILABILITY AT BAC TRAKUON STATION Water availability 1997-2001 Discharge 1997-2001 (million m3) (m3/s) Average discharge (m3/s) Max Average Min Max Average Min 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Jan 76.1 29.2 4.4 28.4 10.9 1.7 5.2 2.7 5.2 6.3 7.6 6.5 3.4 6.8 0.1 0.5 Feb 45.9 13.9 1.2 19 5.8 0.5 2.2 0.5 2.1 2.7 1.9 3.6 0.1 0.5 0 1 Mar 59.7 20.7 1.3 22.3 7.7 0.5 2.9 13 2.5 3.4 6 5.6 0.4 0 0 0.1 Apr 405 78.9 7.1 156.3 30.4 2.3 6.1 156.1 51.1 4.2 11.6 8.9 2.7 3.8 0 0.1 May 853 180 13.4 318.5 67.2 5 144.7 318.4 101.2 10.1 41.5 16 13.2 24.2 0 3.5 Jun 461 170 32.4 177.9 65.6 12.5 165.6 177.9 83.7 26.3 62.5 32.1 79.5 34.6 25.4 5.2 Jul 627 262 54.9 234.1 97.8 20.5 234.1 100.5 94 63.4 44.5 41 129.6 47.7 12.4 40.9 Aug 607 317 160 226.7 118.4 59.8 118 226.7 70.9 125.1 131 54.9 134 76.6 23 34.4 Sep 873 435 216 336.8 167.8 83.3 299.5 336.7 153.5 249.6 172.3 226.8 176 121.8 122.4 108.5 Oct 1146 617 270 427.9 230.4 100.8 156.7 120.2 193.1 358.7 281.6 242 286.8 178 151.1 250.3 Nov 864 246 55.3 333.3 94.9 21.3 84 133.9 102.5 66.3 49.5 95.7 243.1 57 70.1 110.1 Dec 212 89.7 23.5 79.2 33.5 8.8 11.1 64.7 32.9 23.3 13.9 24.2 29.7 14.6 3.5 124.2 Source: MOWRAM FIGURE 2.3: COMPARING MONTHLY DISCHARGE STATISTICS FOR TWO TIME PERIODS 2007-2016 Max 2007-2016 Max 2007-2016 Max 1997-2011 Max 1997-2011 Max 1997-2011 Max 12 10 8 12 Discharge (m3/sec) 10 8 6 4 2 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Source: MOWRAM Other Input Data operational. These two dams are focused on strengthening irrigation water supply. It is understood Additional datasets used in this analysis include that permission has been granted for the construction a 30m resolution digital terrain model (SRTM), of the third (Dam 1) that will include the generation 30 arc-second resolution soil layer (Harmonized of hydro-energy as one of its objectives. All three World Soil Database v 1.2) and humidity, wind of these dams are upstream of the irrigation speed and solar radiation for the period 1979- infrastructure outlined in the next section. Information 2014 (National Centre for Environmental Prediction on the characteristics of these dams (Table 2.2) were Climate Forecast System Reanalysis). obtained from MoWRAM (2013) and Ministry of Mines and Energy (MME) (pers. comm.). Dam operation 2.4 Dam characteristics rules are currently unavailable and so simplifying assumptions were used in order to simulate the Construction of Dams 3 and 5 has been completed functioning and outflow from these dams. and in this analysis they are assumed to be 29 TABLE 2.2: DAM CHARACTERISTICS Dam 1 Dam 3 Dam 5 Storage volume (million m3) 1385.6 25.5 24.5 Capacity (MW) 80 Rated gross head (m) 113 Dead storage (million m ) 3 228 Normal storage level (masl) 180 Low supply level (masl) 160 Reservoir area at normal storage levels (NSL) (km2) 85.91 2.5 Area Under Irrigation Target irrigated areas for wet and dry season River were obtained from a MoWRAM report paddy cultivation using water from Pursat (pers. comm.). These are shown in Table 2.3. TABLE 2.3: TARGET IRRIGATED AREAS IN THE PURSAT RIVER BASIN Project Name Wet season (ha) Dry season (ha) Damnak Chheur Kram 16,100 16,100 Damnak Ampil 27,000 3,500 Charek 11,000 3,500 Chheur Touk (Santre) 1,142 50 Kampeng reservoir 380 100 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Baktra 102 Our Roka Reservoir 4,700 1,000 Lor Lork Sor 1,167 Ang Andoung Wat Luong 2,410 Khnorng Porpol reservoir 1,315 Prey Nhi Reservoir 1,519 10 Preah Chambok 145 50 Koh Svay 350 160 Our Tatong 956 300 Roneam Chhlech reservoir 713 Kbal Hong 7,270 Kandieng station 120 Phum Stueng 222 Ou Sanlung 120 Thlork 1,000 410 Tuol Kour reservoir 722 171 Total area 78,453 25,351 30 June to September are the core months for wet a paddy crop for at least the core months of each season paddy production in the Pursat Basin season. In the wet season, 50 percent of the area is and December to March are the core months assumed to still be under cultivation in the other for dry season production (MoWRAM, 2013). Wet months (given the wider range of cropping days in season production can use either direct sowing or the wet season), while in other months of the dry transplanting techniques, and the growing season season, only 20 percent of the areas is assumed to ranges from 105 to 145 days. Dry season cropping be cultivated. As April does not fall in either cropping is reported to cover 90 days. The exact irrigation season, only 10 percent of the wet season cropping area for each cropping pattern is not available, so area is assumed to be cultivated. Figure 2.4 tabulates the maximum demand over a year is estimated by the area derived for each month. assuming that the total area in Table 2.3 is under FIGURE 2.4: ESTIMATED MAXIMUM AREA UNDER CULTIVATION IN THE PURSAT RIVER BASIN 90, 000 Wet season Dry season 80, 000 70, 000 60, 000 Net area to irrigate (ha) 50, 000 40, 000 30, 000 20, 000 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A 10, 000 0 Apr Jul May Jun Aug Sep Oct Nov Dec Jan Feb Mar 2.6 Modeling Approach different sub-basins of the Pursat River Basin was developed. The objective of the modeling in the hydrological 2. Monthly dam operations: Outputs of water yield analysis is to derive seasonal estimates of changes from the hydrological model were used to drive a in water availability, irrigation demand and sediment monthly dam operations model for the three dams, flows from different parts of the river basin. to provide estimates of water availability at the Bac These variables will form the basis of the economic Trakuon Station. analysis in the next chapter. To achieve these goals, 3. Irrigation water demand: Outputs of potential the modeling can be broken down into four sub- evapotranspiration and precipitation over components as follows: agricultural lands (from the hydrological model) 1. Hydrological Response: A hydrological model were used with FAO’s irrigation water demand capable of estimating the spatial and temporal assessment method to estimate monthly water response of water and sediment yield from demand for paddy in the projected command area. 31 4. Soil erosion: Sediment yield estimates from the be extracted at daily or monthly time-steps over the hydrological model were used to derive average period January 2007 to July 2014. Spatially, the basin soil erosion rates from different land cover types to was divided into 125 sub-basins. The MoE 2016 land lead to an estimate of sediment accumulation rates cover was mapped on SWAT land cover codes (classed in the proposed Dam 1 reservoir under current under 10 different categories; Appendix 1) and with conditions and a “no forest” counterfactual. slope and soil type was the main basis of deriving The sections below describe each step in more detail HRUs. The SWAT model for Pursat had 1656 HRUs in and provide references where applicable. total. Hydrological Response Modeling Model calibration and validation SWAT model The model was calibrated and validated based on daily and monthly discharge at the Bac Trakuon The Soil and Water Assessment Tool (SWAT) is a Station. The calibration used discharge data from semi-distributed, continuous-time, process-based January 2007 to December 2011 and was carried out model (Neitsch et al. 2011). SWAT’s hydrological module using the sequential uncertainty fitting algorithm allows explicit calculation of different water balance (SUFI-2) in SWAT-CUP (SWAT Calibration & Uncertainty components and subsequently water resources at a Programs) (Abbaspour 2012). Fourteen parameters sub-basin level (Abbaspour et al. 2015). SWAT divides linked to discharge, which were shown to be sensitive watersheds into multiple sub-basins, which are then for the basin (Oeurng et al. 2019), were selected for further subdivided into hydrologic response units the calibration. The initial ranges of values for each (HRUs). These HRUs form the basic unit of assessment parameter were set from the likely maximum range in SWAT and consist of unique land use, topographical suggested in SWAT. Validation was carried out using and soil characteristics. discharge from the period January 2012 to July 2014. Watershed hydrology is simulated in two phases. The Abbaspour et al. (2012) describe the application of land phase controls the amount of water, sediment, the SUFI-2 algorithm to SWAT models as a method nutrient and pesticide loadings to the main channel to map all uncertainties (parameter, conceptual VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A in each sub-basin; and the routing phase controls model, input, etc.) on the parameters (expressed as the movement of water, sediments, etc., through the uniform distributions or ranges) and capture most of streams of the sub-basins to the outlets. the measured data within the 95 percent prediction uncertainty (95PPU) of the model through an iterative SWAT simulation for water yield from sub-basins and process. The 95PPU is calculated at the 2.5 percent and HRUs involves using the Soil Conservation Service 97.5 percent levels of the cumulative distribution of (SCS) curve number method and the Green-Ampt an output variable obtained through Latin hypercube infiltration method to simulate runoff. Peak runoff sampling. For the goodness of fit, two indices referred rate is estimated using a modification of the Rational to as “P-factor” and “R-factor” (Abbaspour et al. Method, while groundwater flow contribution to total 2012) are used. The P-factor varies from 0 to 1; and river flow is simulated by creating shallow aquifer is defined as the fraction of measured data (and its storage (Arnold et al. 1993). Erosion and sediment error) bracketed by the 95PPU band. A P-factor value yield are simulated for each HRU using the modified of 1 indicates 100 percent of the measured data within universal soil loss equation (MUSLE). model prediction uncertainty i.e., a perfect model simulation considering the uncertainty. For discharge, For the Pursat Basin, the SWAT model was set up using the recommended value for P-factor is >0.7. The the hydrological datasets described above and with R-factor on the other hand is the ratio of the average the QSWAT interface. The data from the first 3 years width of the 95PPU band and the standard deviation of (January 2004 – December 2006) was used to spin the measured variable. Abbaspour et al. (2012) reports up the hydrological model. Simulation outputs can 32 a value of <1.5 as a desirable value for this index. Monthly Dam Operations Nash-Sutcliffe model efficiency coefficient (NSE) was an additional objective function used to assess the A simple model of dam operation is used, which predictive power of the hydrological model. 10 assumes targeting a stable monthly outflow, if reservoir conditions and inflow allow it.11 Monthly Under the calibration process, the following general inflows into the 3 dams’ sites was extracted from the approach was used: SWAT model and processed using the equations below in MS Excel/VBA script to estimate outflow from the 3 Set up a SWAT-CUP run for the selected parameters 1. dams. The monthly dam operations model is adapted and initial range. 200-300 iterations are performed from the equations proposed by Yassin et al. (2019). in each batch, with parameter values extracted i Dam outflow for any month i(S r ,actual) is estimated from the provided range through Latin hypercube from the capacity of the dam (Vc), its dead storage sampling. As a post-processing step, 95PPU and i) capacity (VD) and the design discharge per month (S r objective function values are calculated. using the following equations: Using the previous batch of iterations, the 2. parameter ranges are updated based on new i > Si : If S r n ranges suggested by the program. Once the i — S i ,V i — V ) ∆out = min (S r n r D new ranges have been reviewed, a new batch of iterations can be executed. i +1 = V i — ∆ Vr r out The above step is repeated until satisfactory 3. i i results in terms of the P-factor and R-factor are Sr ,actual = Sn + ∆out reached or no significant improvements are seen i < Si : Else if S r in the NSE. This takes generally 3-4 batches of n iterations. The set of parameters, which gives i +1 = min (V i + S i - S i , V ) Vr r n r c the best objective function fit, are used as the i + Si V ) Spi ll = max (0,V r parameters for the calibrated model. n - c i +1 - V i ∆in = V r Hypothetical counterfactual r VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A i is the inflow (in m3/month); V i is the actual where, Sn Valuing forests requires comparing current conditions r to hypothetical counterfactual without forests. This volume at start of the month (initial condition = VC) i +1 is the volume at end of month. and V r is not intended as a realistic scenario, but solely as a counterfactual against which to measure the total This simple dam operations model was applied value of forest benefits. This “no forest” counterfactual instead of SWAT’s reservoir and dam modeling options was created in the SWAT model by replacing all forest because these options require more parameters to related land cover codes with a barren (bare earth) sufficiently describe the dam operating rules and land cover. No other changes were made to parameters reasonable estimates of these parameters were not selected from the calibration and validation exercise, available, especially for Dams 3 and 5. In the monthly as described in the previous sub-section. dam operations model, VC and VD were extracted from 10 DNash–Sutcliffe efficiency can range from −∞ to 1. NSE = 1 corresponds to a perfect match of modeled discharge to the observed data, while NSE = 0 indicates that the model predictions are as accurate as the mean of the observed data. Positive values of NSE indicate a better model simulation while negative values of NSE indicate that the observed mean is a better predictor than the model being used. Most hydrological models are considered to have a ‘good’ fit if NSE is between 0.5 to 0.65. 11 More sophisticated models might seek to maximize outflows during the dry season, or to weigh the possibly competing demands of HEP and irrigation, as well as flood prevention. Additionally, since the model used runs as a post-processing step to the hydrological model results, it can easily be replaced by alternative models. 33 Table 2.2 (VD was set to zero when not available). WL is water needed to establish a water layer Design discharge per month (S r i ) was set at a constant during transplanting or sowing and maintained discharge throughout the year. For Dam 1, this value throughout the growing season. Value used is based on MME specification of annual average here is 100 mm for months of June and discharge for the dam (42.3 m3/s). For Dams 3 and 5, December. this was set at the average annual discharge at the dam site as obtained from the SWAT model (1 and 21 SAT is water needed to saturate the root zone m3/s respectively). a month before sowing or transplanting. Direct rainfall is assumed here as the source of this Estimating Irrigation Water Demand and SAT is set as 0 mm. The methodology proposed by FAO (Brouwer and Once demand per hectare has been established Heibloem 1986) was used to estimate water demand using the method above, the estimates of area under for irrigation. This methodology calculates irrigation cultivation per month was used to estimate total water requirements as the difference between irrigation water demand. crop-specific water needs and the portion that can be meet by rainfall, taking into account losses to Estimating Soil Erosion Rates evapotranspiration, deep seepage, evaporation, percolation, etc. The steps for calculating water The SWAT model generates a sediment yield estimate demand for irrigation are: for each HRU based on its application of the MUSLE equation. As a post-processing step, HRUs in sub- Step 1. Estimate potential evapotranspiration (ETO): basins above Dam 1 were sorted based on land cover Average monthly estimates for agricultural and average sediment yield rates per month for lands were extracted from the SWAT model. each land cover type extracted. These values were further averaged over time to produce average annual Step 2. Estimate average crop evapotranspiration sediment yield rates per land cover type (in tons per coefficients (Kc) based on known cropping hectare per year). pattern and cropping stage: Coefficients for VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A paddy were derived from Chapter 6 of Brouwer 2.7 Hydrological Modeling Results and Heibloem (1986) and adjusted to cropping pattern. Performance of the Hydrological Simulations Step 3. Calculate for each month the effective rainfall The model was calibrated over January 2007 to using the formulae: December 2011 and validated over January 2012 to Pe = 0.8*P - 25 if P > 75 mm/month July 2014 period. The 14 parameters that the model Pe = 0.6*P - 10 if P < 75 mm/month used for the calibration-validation of the model are documented in Appendix 2. Batches of iteration Average monthly estimates of actual with daily discharge as the calibration variable wer precipitation (P) over the cropping area were carried out using SUFI-2 in SWAT-CUP until NSE value extracted from the SWAT model. stopped showing significant improvement (NSE=0.45). The R-factor obtained for this final batch of iterations Step 4. Calculate the irrigation water need: IN = ETO* Kc (=0.73) is satisfactory, however, the P-factor (=0.5) – Pe + PERC + WL + SAT is less than desirable when using discharge as the calibration parameter. A closer look at the discharge Where, PERC is the percolation and seepage plot in Figure 2.5 shows that while low and mean flows losses depending on the type of soil; here generally fall within the 95PPU band, the observed estimated as 6mm/day on average. peak discharges (especially for the period before 2009) fall outside. The explanation for this behavior appears 34 to be beyond model parameter uncertainty and to is suitable to carry out the water allocation studies be linked to one of the key underlying datasets. The for irrigation and hydropower supply. The model comparison between gauge precipitation data and might underestimate sediment yield as soil erosion ERA5 global dataset (outlined above) notes the bias rates are a function of rainfall intensity, but without in maximum precipitation rate in the global dataset monitored datasets to ground-truth the results, this when compared to local gauge data. This was a trade- will remain an area of high uncertainty. For using off to achieve improved spatial coverage by using the the model for flood magnitude or design related global dataset. The global dataset was found suitable studies, it would be advisable to explore further in capturing average flows as seen in the model improvement in the precipitation datasets. This performance for monthly flows (NSE > 0.5;Figure 2.5). could be achieved through obtaining additional gauge data (if available) or use of statistical methods Figure 2.6 summarizes the model fit at both daily to merge local gauge data with remotely sensed and monthly timescales. The satisfactory fit for precipitation products (Xie et al. 2011; Verdin et al. monthly flows suggests that the current SWAT model 2015), which was beyond the scope of this project. FIGURE 2.5: DATA FOR VALIDATION OF HYDROLOGICAL MODEL (A) DAILY DISCHARGE WITH 95PPU PLOT FOR CALIBRATION 95PPU Observed Best Simulation 1000 800 Discharge [cumess] 600 400 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A 200 0 1-Jan-07 1-Jul-07 1-Jan-08 1-Jul-08 1-Jan-09 1-Jul-09 1-Jan-10 1-Jul-10 1-Jan-11 1-Jul-11 (B) MONTHLY DISCHARGE DATA OVER CALIBRATION PERIOD Observed Simulated 400 NSE: 0.62 Discharge [cumess] 200 0 Jan-07 Jul-07 Feb-08 Aug-08 Mar-09 Sep-09 Apr-10 Nov-10 May-11 Dec-11 Source: Authors’ calculations 35 FIGURE 2.6: RESULTS FROM DAILY AND MONTHLY CALIBRATION AND VALIDATION CALIBRATION [DAILY DISCHARGE] CALIBRATION [MONTHLY DISCHARGE] 1000 400 NSE: 0.45 NSE: 0.62 800 300 600 Simulated Simulated 200 400 100 200 0 0 0 200 400 600 800 1000 0 50 100 150 200 250 300 350 400 Observed Observed VALIDATION [DAILY DISCHARGE] VALIDATION [MONTHLY DISCHARGE] 1000 400 NSE: 0.47 NSE: 0.67 800 300 600 Simulated Simulated 200 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A 400 100 200 0 0 0 200 400 600 800 1000 0 50 150 250 300 400 Observed Observed Source: Authors’ calculations Water Yield from Watershed 2-4 times the downstream regions, depending on the seasonal pattern. For 3 of the 4 quarters in a year, the Water yield was analyzed for the period 2007 to 2014 upstream source areas had a water yield exceeding using the SWAT model. The projected water yield at a 100 mm/month on average. These upstream source sub-basin level (Figure 2.7) indicates a clear correlation areas, therefore, are clearly important for securing of important source areas with elevation. Regions good quality year-round supply of water. Land cover of the watershed that fall within the Cardamom change that could disrupt the quality and timing of Mountain PAs can have a water yield ranging from flows will have a direct impact on use downstream. 36 37 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Source: Authors’ creation FIGURE 2.7: AVERAGE WATER YIELD DISTRIBUTION OVER SUB-BASINS AND QUARTERS OVER A YEAR Based on the modeling results, elevation appears although confirmation through ground observations to be the primary driver of water yield in this basin would build confidence in this outcome. (i.e., more precipitation at higher elevations), with land cover having a secondary but still discernible Another approach for analyzing water yield is influence. As an example, water yield for the driest comparing the cumulative yield for sites of interest month in the simulation (January 2014) showed that in the basin. Figure 2.8 breaks down the portion water yield in land parcels converted from forests of the Pursat River Basin upstream of Bac Trakuon to agricultural land in the source region (within Station into 5 sub-basins. These are: The sub-basin Sam Kos Wildlife Sanctuary) was about 4-8 percent for Dams 1, 3 and 5 (marked on figure as a, c, and lower than the surrounding forested regions, but d, respectively), the sub-basin for one of the main these yields were still higher than the downstream tributaries, Stung Prey Khlong (marked as sub-basin (lower elevation) sections of the basin. This drop e) and the remaining portion of the basin between in yield is likely linked to the model’s estimate the dams and the gauging station (marked as sub- of land cover’s role in groundwater recharge. In basin b). Stung Prey Khlong is the largest tributary dry months, groundwater stores provide the bulk of the Pursat with no known major infrastructure of the water yield in the model dynamics. While works on its main stem. From the water yield results land covers with a healthy soil layer are generally of the model, it is seen that the sub-basin for Dam known to facilitate higher infiltration rates, this 1 has the largest contribution to the flows at the is not necessarily the same as high groundwater gauging station, followed by the sub-basin for Dam recharge rates (Portela et al. 2019: Filoso et al. 5. The conclusion that can be drawn from this is 2017). In some cases, higher infiltration rates do that besides the regulating effect of the natural lead to higher storage in the groundwater layer, infrastructure on the water flow, the set-up and thus allowing greater water yield in drier months. operating rules of these two dams can potentially This appears to be the case indicated by the model, have a large impact on water availability downstream. FIGURE 2.8: WATER YIELD DISTRIBUTION BY SUB-BASINS LINKED TO THE DAMS AND TRIBUTARY ST PREY KHLONG VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Above Dam1 Above Dam3 Above Dam5 St Prey Khlong Remaining 60 50 x 10000000 40 30 cub.m of water 20 10 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Source: Authors’ calculations 38 Sediment Yield and Erosion Rate year for other vegetated land covers like agricultural land use above the three dams. The loss of protective The annual average sediment yield estimate from the cover over regions with high rainfall intensity, creating model, showed that the majority of the basin still has these high sediment yield situations, will likely have low rates of erosion (<5 t/ha/year). As the precipitation numerous negative impacts downstream. Increased dataset likely underestimates rainfall intensity, this sedimentation rates for the downstream reservoir is likely an underestimate or lower bound of actual is the most tangible of these. Others could include erosion. However, the influence of land cover change sedimentation of irrigation channels leading to drop in the upper-western region of the basin is clearly in conveying capacity of irrigation water and higher identifiable with these results (Figure 2.9). The model propensity to flood, drop in water quality impacting estimates that sediment yield may change from the household and industrial usage as well as fisheries, for range of 0.3 t/ha/year for forested lands to 20t/ha/ which these streams are important breeding grounds. FIGURE 2.9: AVERAGE ANNUAL SEDIMENT YIELD FOR SWAT SUB-BASINS VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Source: Authors’ calculations Role of Forests in Hydrological Flows no-forest simulations. The spatial pattern (Figure 2.11) for yield change in a dry month is similar. The change in seasonal pattern of flows between current conditions and no forest counterfactual can As expected, sediment yield rates would see a marked be seen in panels a and b of Figure 2.10. The increase change. The annual average sediment yield over the in discharge during the earlier months of the wet parcels where land cover type was changed would season indicates a higher propensity to generate go from 0.3 t/ha/year for current conditions to 175 t/ surface runoff from the cleared land. The rise of ha/year under the no-forest counterfactual, resulting the hydrograph is sharper and the fall is steeper, in high sediment inflows into the reservoir for Dam in line with increased surface flow. The drop in dry 1 (seen in section c of Figure 2.10). On average, the season flows is relatively moderate, extending up to model estimates that the sediment accumulation rate 25 percent lower flows when comparing current and in the dam’s reservoir could increase by 30-fold. 39 FIGURE 2.10: COMPARISON OF CURRENT HYDROLOGICAL FLOWS TO THOSE OF NO-FOREST COUNTERFACTUAL (A): PERCENTAGE CHANGE IN DISCHARGE AT BAC TRAKUON STATION. COMPARISON BETWEEN (A) MONTHLY DISCHARGE HYDROGRAPH FOR BAC TRAKUON STATION AND (C) MONTHLY SEDIMENT INPUT FOR DAM 1 2007 2008 2009 2010 2011 2012 2013 2014 Jul Jan Jul Jan Jul Jan Jul Jan Jul Jan Jul Jan Jul Jan Jul 250 200 % Change in discharge 150 100 50 0 -50 2010 300 900 Dam1 sediment ionput (ton) x 10000 800 250 700 Discharge (cumecs) 200 600 500 150 400 100 300 200 50 100 0 0 Oct Jan Feb Mar Apr May Jun Jul Aug Sep Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Current No Forrest Current No Forrest Source: Authors’ calculations FIGURE 2.11: SPATIAL CHANGES IN WATER YIELD UNDER THE NO-FOREST COUNTERFACTUAL COMPARED TO CURRENT CONDITIONS Source: Authors’ calculations 40 Water Availability and Demand months but is not large enough to have a major impact downstream. Dam 5, on the other hand, receives Monthly dam operations significant flows, but its storage capacity is too small to significantly alter the flow. Consequently, inflows With 62 percent of the water yield originating in and outflows from dam 5 are nearly identical (Figure sub-basins upstream of them, dam operations and 2.12). Dam 1 appears to have both the storage capacity storage can significantly alter the water availability and the inflows to have a significant impact on water downstream. As indicated in Section 3.2, current availability downstream. The water availability at Bac information about the dams is incomplete and hence Trakuon Station will be sensitive to the operations a simple monthly dam operations model is applied. rules set for this dam, so improving the representation of this dam in the model should be a priority for future Dam 3 only controls a small portion of total flow. iterations of this work. Under current operations, It appears to have just enough storage to satisfy the discharge only managed to match the specified irrigation demand in its local vicinity during dry monthly target of 42.3 m3/s about half of the time. FIGURE 2.12: COMPARISON OF INFLOWS AND OUTFLOWS FROM EACH DAM OPERATIONS UNDER CURRENT CONDITIONS AND UNDER THE NO-FOREST COUNTERFACTUAL DAM 1 DAM 2 DAM 3 inflow outflow inflow outflow inflow outflow 140 6 70 120 5 60 100 50 4 80 40 3 60 30 2 40 20 20 1 10 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A 0 0 0 inflow outflow inflow outflow inflow outflow 140 6 60 120 5 50 100 4 40 80 3 30 60 2 20 40 20 1 10 0 0 0 Source: Authors’ calculations 41 FIGURE 2.13: EFFECT OF DAM OPERATION ON PROJECTED WATER AVAILABILITY, SELECTED YEARS Source: Authors’ calculations 2007 2008 80 80 60 60 x 10000000 x 10000000 40 40 20 20 0 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 2009 2010 80 80 60 60 x 10000000 x 10000000 40 40 20 20 0 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Natural flow Water availability (Dam operational) Natural flow Water availability (Dam operational) Outflow from Dams 1, 3 and 5 is aggregated with the relatively short stream distance, the introduced error water yield from sub-basins below the dam and before is generally low. In test conditions, this error was 1.01 Bac Trakuon gauging station in order to estimate percent on average (see Appendix 3). Ideally, use of total water availability at that gauging station. This more complete dam data will allow dam operation requires the simplifying assumption that monthly simulation to be included within the hydrological water yield in these sub-basins is equal to monthly model and remove the need for this simplifying discharge and we therefore do not consider effects of assumption. Figure 2.13 shows how water availability at stream routing. But given the monthly timescale and the gauging station changes due to dam operations. VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A FIGURE 2.14: COMPARISON OF ESTIMATED WATER DEMAND AND ESTIMATES OF WATER AVAILABLE AT BAC TRAKUON STATION, SELECTED YEARS 2009 2010 80 80 60 60 x 10000000 x 10000000 40 40 20 20 0 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Estimated Demand Water availability (Dam operational) Estimated Demand Water availability (Dam operational) 2011 2012 80 80 60 60 x 10000000 x 10000000 40 40 20 20 0 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Estimated Demand Water availability (Dam operational) Estimated Demand Water availability (Dam operational) Source: Authors’ calculations 42 Estimating irrigation water demand Irrigation demand deficit Over the simulation period, a spatially-averaged time The results of estimated water availability and series of precipitation and evapotranspiration values irrigation demand form the inputs for the economic over cultivated lands is extracted from the hydrological analysis, which will continue in the next chapter. model as an input to estimate irrigation requirement. Over the 7-year simulation period, demand exceeded This season-specific water requirement estimate varies supply multiple times, most often in February significantly from month to month and year to year, and June (5 instances over the 7-year simulation) ranging from 5.8 mm/month/ha to 453 mm/month/ha, and sometimes in January (twice over the 7-year with an average irrigation requirement of 273.4 mm/ simulation), May, July and August (once each). The month/ha. A time series of total demand for irrigation average and maximum magnitude of deficit is plotted water per month was estimate by combining the in Figure 2.15. Matching the frequency, February and monthly per hectare requirement with the estimate June have the highest average deficit, while January of area under cultivation derived above. Figure 2.14 has the maximum-recorded deficit when normalized below compares the estimated water demand with the by area under irrigation. estimates of water available at Bac Trakuon Station. FIGURE 2.15: ESTIMATED IRRIGATION WATER DEMAND DEFICIT UNDER CURRENT CONDITIONS 100 75 MAXIMUM DEFICIT 50 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A AVERAGE DEFICIT 25 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Source: Authors’ calculations 2.8 Limitations of the Analysis (Jan 2007- July 2014) and the daily model fit, even though comparable to a published model for the The main limitation and caveats in the modeling same region, was below the range accepted as a approach used for this study, either from the data “good” fit. Improvement in precipitation data is perspective or from under-lying assumption in model likely key to improve these model results; setup, have been explored largely alongside the 2. Lack of sediment yield data: Absence of any results in Section 2.7. In summary, the key limitations monitored sediment data – either at the plot scale and caveats linked to the hydrological analysis are: or at downstream gauging stations or local soil maps, leads to sediment yield results that could 1. Model fit: The total period of data for which the not be compared to real measurements. Although model could be calibrated and validated was short the estimates are close to the values used by 43 Ministry of Mines and Energy (MME) in planning the necessarily the same as high groundwater recharge hydropower dam – other studies in the region have rates. Our ability to represent this dynamic in shown a large variation in this estimate; hydrological models generally remains poor 3. Caveat on characterization of dam operation: (disciplinary constraint). Additionally, in this region, Partially because the main dam (Dam 1) has not the extent and state of aquifers, recharge areas, been constructed yet and limited data available on etc., have not been surveyed (data constraint). its design, dam operations rules were not available to incorporate in the models. Consequently, dams 2.9 Summary of Key Results were represented by simplified operation rules and added as a post-processing step to facilitate The erosion rates estimated for different land use improvement through stakeholder interaction. as well as the water availability/demand results Moving these dams from a post-processing from both the current conditions and no forest step and into the hydrological model would be counterfactual will be used as inputs for the economic advisable in a future iteration of this work; analysis of the hydrological ecosystem services 4. Caveat on link between shallow infiltration and (to continue in the next chapter). Table 2.4 below groundwater recharge: As noted earlier, while land summarizes the main outcomes from the hydrological covers with a healthy soil layer are generally known analysis when considering the impact of changes to facilitate higher infiltration rates, this is not occurring in the Pursat River Basin. TABLE 2.4 SUMMARY OF MAIN PROJECTED HYDROLOGICAL IMPACTS OF LAND USE CHANGE IN THE PURSAT RIVER BASIN Impact of Level (uncertainty) Comments/Explanation Upstream deforestation on water Mid to low impact Water yield maps correlate more to elevation yield during low flows and rainfall patterns than land cover. However, (high uncertainty) the information on aquifer extents and groundwater drainage is poor in the model, VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A contributing to high uncertainty. Establishment of agriculture High impact Based on estimates of irrigation water demand, upstream on water yield during the region being converted to agricultural (low uncertainty) low flows land use above Dam 1 could divert low flows significantly. The “no-flows” in recorded discharge from 2014 onwards for first quarter of the year (if not an instrumentation error) could be likely linked to this. Upstream deforestation on High impact Monitored sediment data is unavailable, so this sediment output is largely modeled figures without real-world (high uncertainty) validation. The results do indicate that each land parcel deforested, could for a period increase the sediment output by at least one order of magnitude. Upstream deforestation on floods/ High impact The no forest counterfactual is not realistic, peak flows however, it does confirm a pattern of higher (low uncertainty) discharge in rainy season. The actual magnitude will not be like the 100 percent to 150 percent increase, but the trend is robust. 44 3 MONETARY VALUATION OF HYDROLOGICAL SERVICES 3.1 Overview VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A The annual water deficit estimated in the hydrological analysis was used to estimate the forgone agricultural revenue based on area under cultivation. This information, in combination with additional data on crop yield and production cost was used to estimate the economic value of water. Detailed methods for calculating unit monetary value of water for irrigation are described under section 2.1. Soil erosion and sediment deposition in the reservoirs were similarly quantified using different deforestation scenarios and the estimates were used to calculate reduction in water storage and resulting changes in hydro energy production. The monetary value of water for hydro energy production was estimated using the methods described in section 2.2. All the modeling and analysis was done using standard Python programming libraries and MS Excel. Overall objectives are the following: • Estimating impact of upstream forest cover on water supply in monetary terms; • Measuring long term impact of sediment erosion on hydro dam capacity; • Estimating monetary value of sediment retention by upstream forests and avoided hydropower loss due to sediment deposition). Central Phnom Penh City in Cambodia 45 3.2 Methodology production. Many of those variants were extensively discussed with case studies and examples in the Deforestation Scenario for Irrigation System of Environmental-Economic Accounting for Water SEEA-Water (UNSD 2012). From the 7-year hydrological simulation results from both current conditions and no-forest counterfactual, In RIM, the total value of a product is divided into spatially and temporally averaged monthly water yield the opportunity cost of all inputs, until the total response curves (cub.m per hectare per month) were amount is completely exhausted. The assumption is derived for 4 land cover types above Bac Trakuon that in a perfectly competitive market the value of a monitoring station, namely for: Forest, Agricultural, commodity (i.e. output X price) is exactly equal to the Barren and Others. Based on deforestation rate opportunity cost of all inputs. So, if opportunity cost considered in the scenario, the area under each land of all non-water inputs is known, the shadow price of cover type is calculated. It is assumed that as part water would be the difference between the value of of the process of forested land being converted to outputs and cost of non-water inputs. agricultural, any deforested plot remains without vegetation (barren) for that year and becomes part of We used data collected in a nation-wide farm survey agricultural usage the following year. With area under held between 2012 and 2013 (World Bank 2015) each land cover known, the monthly water yield time to value water for irrigation services. This survey series is recombined using area as weights to estimate collected data on a range of agricultural products the overall water availability in each month. using a structured survey of individual farmers (drawn from a random sample stratified by provinces, The economic analysis shows that out of the total districts, communes and villages) in combination with estimated value of irrigation benefits US$16 thousand focus group discussions and key informants. Data per year is threatened if current rate of deforestation were disaggregated according to farm size, technology continues at 0.25 percent. In an accelerated used and seasonality. The data collected include a deforestation of 1 percent a proportionate value of range of financial and farm characteristics such as US$64 thousand per year will be under threat. cultivated area, yield, input cost, labor cost, cost of services and irrigation. VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Valuation of Water in Irrigation Farm characteristics Several methodological frameworks and techniques for estimating the monetary value of water for In the Pursat River Basin, the average farm cultivates irrigation have been proposed (Speelman et al. 4 ha of rice in the dry season and about 2 ha in the 2008; Mesa-Jurado et al. 2010; Berbel et al. 2011). wet season (Figure 3.1). Most farms are small (average The most common approach is the “production size is less than 1ha) and they tend to grow rice only function” approach. There are a few variants of this during the wet season, using traditional techniques methodology known by different names, such as (Figure 3.2).12, 13 Large farms, on the other hand, are Residual Imputation Method (RIM), Net-back Analysis, more likely to grow under both wet and dry seasons Net Return to Water, Net Income, etc. Each variant and use modern techniques. The total area under rice takes a residual value approach, where the value production in the wet season (78,000 ha) is more than of a given input is inferred from other inputs to three times higher than in the dry season (25,000 ha). 12 <1 ha is for small farms only, and 2-4 ha is averaged across all small and big farms. Although the number of large farmers is comparatively smaller, this size differences made the overall average larger. 13 (a) Average farm size (cultivated areas) is about 3 ha, with large farms having >6 ha and smaller farms <1 ha. On average technologically improved farms are bigger than traditionally cultivated farms. (b) The yield of rice across farm types falls between 3-4 tons/ha. Yields are not significantly different across farm types. (c) Total variable cost was estimated to be around US$600 per ha rice production and do not vary much according to farm types. 46 FIGURE 3.1: PADDY RICE PRODUCTION DATA, BY SEASON 5 700 4 4 600 3 500 3 400 t/ha $/ha 2 ha 2 300 200 1 1 100 0 0 0 Dry Wet Dry Wet Dry Wet Average area of rice cultivated per farm Average paddy yields Average variable cost of production Source: Authors’ calculations FIGURE 3.2: KEY CHARACTERISTICS OF FARMING SYSTEMS 4.0 600 6 5 3.5 500 5 3.0 400 4 2.5 4 2.0 $/ha 300 t/ha ha 3 1.5 200 2 1.0 100 1 0.5 0 .00 0 Large Modern Overall Traditional Small Large Modern Overall Small Traditional Large Traditional Overall Modern Small Farming system Farming system Farming system Average area Average yield Cost of production VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Source: Authors’ calculations Dry season rice cultivation has higher yields (4.5 tons/ studies (Srean et al. 2018). Under current condition ha) than in the wet season (3 tons/ha), a difference total value added from rice cultivation is US$40.7 also confirmed by other studies (Lee and Kobayashi million (gross income US$89 million). 2017). High productivity in the dry season is generally attributed to the use of high-yielding seeds and Unit monetary values better irrigation management. The unit value of water estimated using the RIM Total variable cost was estimated to be around approach ranges from 0.05-0.09 US$/m3, the higher US$696 per ha in the dry season and US$510 per values being for dry season rice cultivation. These ha in the wet season. This difference is primarily values are within the range estimated in other studies due to higher inputs use. Because of this high cost (Hussain et al. 2007). Correspondingly, the economic of production farmers, dry season returns are low return to water ranged from 175 US$/ha for wet season despite the higher yields, as also noted in other rice to 311 US$/ha for dry season rice (Table 3.1). 47 TABLE 3.1: VALUE OF WATER AS AN INPUT TO PRODUCTION IN DIFFERENT FARMING CATEGORIES Farm categories Unit Small Large Traditional Modern Dry Wet US$/m3 0.089 0.075 0.069 0.063 0.089 0.05 US$/ha 185 261 242 220 311 175 Source: Authors’ calculations Valuation of Sediment Impacts on Hydroelectric which in turn affects the dam’s ability to generate Power Production electricity. The approach is depicted in Figure 3.3. Note that because sedimentation is a cumulative The economic return to water for hydroelectric power problem, its impact is small initially, but increases (HEP) production was estimated based on the effect over time. Therefore, results are presented in present of sedimentation on the capacity of the reservoir, value terms rather than in annual terms. FIGURE 3.3: METHODOLOGICAL PROCESS FOR VALUATION OF HYDROLOGICAL SERVICES USED FOR HEP Reservoir vol. at time t-1 Reservoir vol. Electricity at time t produced Sediment Net present Revenue trapped at t Value (NPV) VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Water Electricity productivity value Erosion rate Reservoir vol. based Piece of PV of revenue at Sum of PV (land cover -> erosion -> on storage at t electricity time t, interest i for T years sediment delivery to reservoir) Source: Authors’ creation The rate of sedimentation depends on erosion rates, (high deforestation). In each case, the mathematical which depend on land cover in the upper watershed. model was run twice, comparing the sedimentation As before, we compare current land cover to a impacts to those that would be experienced under hypothetical no-forest counterfactual, which enables no deforestation, the difference between each pair of us to value the role of forest as a whole in protecting values indicating the impact. HEP production. We also examine two policy- relevant scenarios, one with a deforestation rate of The process of valuation starts by estimating the 0.25 percent per annum (business as usual) and on sedimentation rate and how it affects reservoir with a deforestation rate of 1 percent per annum capacity. In any year t, the reservoir volume Vt was 48 estimated by reducing the previous year’s active in the SEEA-EEA methodological guidelines was storage by the amount of sediment deposited (Sedt) in used.15 A rent in this case is the residual value, that year.14 calculated as the surplus generated by producing one unit of HEP, above the cost of inputs. However, Vt = V t-1 - Sedt as data on energy production cost and other inputs are available for the study site, however, This reservoir volume at time t was used to the unit value was derived based on the consumer estimate total amount of electricity generated price of electricity (US$0.25 /kwh) and assuming for that year using an estimate of the water an average cost of production US$0.04/kwh. productivity of electricity (i.e. kwh produced per m3 of water; Miglietta et al. 2018) and of We made several assumptions for calculating the the unit value of electricity. This process was economic value of water for HEP production. Key repeated for a 100-year dam lifecycle (T) with data required were water supply, water demand corresponding changes in forest cover and soil and soil erosion (estimates described in biophysical erosion. Finally, a net present value (NPV) was chapter). Additional data on dam parameters (such estimated using a discount rate (i) of 6 percent. as electricity price, dam storage, electricity capacity, dam lifetime) were collected from MOWRAM (2013). To calculate a unit monetary value of water for HEP The following table provides key modeling parameters production, the Resource Rent approach proposed and assumptions. TABLE 3.2: KEY ASSUMPTIONS USED IN VALUATION OF HEP Assumptions Unit Baseline Electricity capacity Kwh/y 400 million Electricity price US$/Kwh 0.25 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Electricity cost of production US$/Kwh 0.04 Deforestation rate percent 0.25 Erosion rate t/ha/y 0.3 forested 20 other land use 175 barren land Discount rate percent 6 Water productivity of electricity Kwh/m3 0.47 Time frame years 100 Sediment trapping efficiency percent 90 percent Soil bulk density g/ml 1.2 14 Sediment loads estimated by the SWAT model (in tons) were converted to a volume equivalent (m3) using a constant bulk density of 1.2 g/l. 15 An alternative approach would be to compare the cost HEP to that of other energy sources such as diesel, renewable energy, thermal sources, or firewood, that would used if HEP was not available. 49 3.3 Results forests and their soil retention role, the reservoir of Dam 1 would lose all its capacity for electricity Economic Value of Ecosystem Services to Farmers generation in just 65 years (Table 3.3). Under the gradual deforestation scenarios, the erosion would The hydrological modeling calculated the quantity of increase more slowly and so the reservoirs would water supplied by the Cardamom forest ecosystems fill more slowly, but they would still be significantly by comparing water yield under existing land cover affected. With current annual deforestation rates of with that under a hypothetical no-forest condition. 0.25 percent, reservoir capacity would be reduced by As described above, water availability is better 23 percent by the end of its 100-year life cycle; at a with forests than it would be without them. Indeed, higher annual deforestation rate of 1 percent, storage without forests the total value of irrigated crop would be reduced by more than 60 percent. production would be US$0.6 million per year lower. In other words, the presence of forests in the Pursat This reduction in capacity would, in turn, reduce the River Basin makes farmers better off by US$0.6 ability to generate electricity. Without forests, the million a year. value of electricity production would decline from about US$76 million a year to almost nothing after Moreover, in the absence of forest, increased soil 65 years. In present value terms, the value of erosion would cause further losses to farmers, electricity production would be US$18.2 million lower estimated at US$1 million per year, by reducing the than it would be with forests. That is, the presence of ability of the dams to regulate the flow of water to forests increases returns to HEP in the Pursat River irrigation. Increased sedimentation in the absence Basin by US$18.2 million over the reservoir’s lifetime. of forests would also increase maintenance costs This is equivalent to an annual benefit of about in the irrigation systems, as sediment would clog US$1.1 million. distribution canals. However, these costs could not be estimated for lack of data. Gradual deforestation would have the same effect, but more slowly. At current rates of forest loss Thanks for the presence of forests in the Pursat River (0.25 percent a year), the annual value of electricity Basin, farmers in the irrigated areas in the lower basin would decline from US$76 million to US$58 million in VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A are thus better off by at least US$1.6 million per year 100 years. As most of the losses would come in later than they would be in the absence of forests. The years, however, the reduction in present value of present value of this benefit stream over a 100-year electricity production would be US$0.8 million than time horizon, discounted at 6 percent, is US$28 million. with no deforestation. At a high deforestation rate of 1 percent a year, the annual value of electricity would Economic Value of Ecosystem Services to Hydropower decline to US$30 million in 100 years - a reduction Operators in present value of electricity production of US$2.8 million (equivalent to an annual cost of about The presence of forests in the Pursat River Basin US$0.17 million). protects the soil from erosion. In the absence of TABLE 3.3: EFFECT OF LAND COVER CHANGE ON VALUE OF HEP Business as usual High deforestation No-forest deforestation (1%/year) counterfactual (0.25%/year) Change in reservoir capacity (%) 23 61 100 (65 years) Reduction in value of HEP (US$million) 0.8 2.8 18.2 Note: Present value of reduction in value of HEP computed over 100 years with a discount rate 6 percent. 50 3.4 Summary Limitations/Caveats The analysis shows that forests in the Pursat River This study is subject to a number of limitations, Basin have a value of at least US$46 million dollars, particularly due to data availability. Some of these based on the net present value of the services limitations result in entire problem areas being that they provide to irrigation and HEP. The main omitted from the analysis. Others (such as limited beneficiaries of these forests are the farmers in the rainfall and sediment load data) affect the reliability Pursat plain, who are able to irrigate a greater area of the hydrological modeling. Resolving these than they would if there were no forests as well as the problems would lead to both more precise and more HEP producers (and so, also their customers) who are reliable estimates. Some of the data used to estimate able to produce more electricity. the value of irrigation was collected some years ago and may not fully reflect current conditions. As This estimated value is an underestimate for two noted, the results are underestimating the true value reasons. First, the estimated value does not include of services, so with better data the estimated value the effects of climate change, which are likely to would almost certainly increase. exacerbate the damage that would be suffered in the absence of forests. Climate change is likely to result There are also some methodological limitations. in more intense rainfall and in higher temperatures, Water demand for paddy rice production is met by which would tend to increase erosion and reduce both surface and ground water, for example, but the dry season flow and probably even total water analysis focuses solely on surface water. Although availability. Under these conditions, the role of ecosystems play a role in ground water infiltration, forests would become even more important. Second, surface water is directly regulated by ecosystems, the estimated value does not include some aspects which is the focus of this study. A simple model of of the hydrological services that forests provide. dam operations is used; a more sophisticated model For example, the savings in maintenance costs that would better capture the dam’s role in ensuring the forests’ erosion reduction service generates are water supplies for irrigation and HEP. The analysis not included, and neither is the reduction in flood uses average value as a proxy for a marginal value, assuming a constant return to scale. Determining VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A risk that forests provide. Again, these benefits are likely to become even more important with climate return to scale is a difficult exercise, as this needs to change (the following section examines the carbon be determined empirically for each crop and each sequestration benefits that forests provide). site. Therefore, to simplify, a constant return to scale is assumed in this analysis. Of course, this value would not be lost all at once. With gradual deforestation, benefits to both farmers The deforestation scenario is clearly a simplified and HEP producers would gradually decline, with approach but is helpful to provide an initial rough the extent and rapidity of this impact dependent on estimate of likely continued loss of seasonal water the rate of deforestation. It should also be stressed regulation service over time. Two main caveats to that this estimate only applies to the forests within note, alongside the results are that the method: the Pursat River Basin, which is only one of the 42 (1) assumes water yield, at any point in the basin, watersheds in the Cardamom Mountains Tonle-Sap becomes available as discharge downstream within basin. The forests in the rest of the Cardamoms the same month; (2) ignores the spatial variability of are also generating substantial benefits. This study water yield. Even the agricultural lands, above the provides a methodological framework that could gauging station, are on comparatively lower altitude be replicated in other watersheds of the Cardamom than the forested lands and so any average water Mountains, or elsewhere in Cambodia. yield curve derived will also be biased by lower 51 precipitation rates of the agricultural lands, rather Economic analysis is tightly interlinked to biophysical than impact of land cover alone. analysis and physical quantification of ecosystem services. Therefore, the outcome of economic analysis One of the key limitations of the study is the ability is naturally sensitive to underlying assumptions of the hydrological model to reliably link large made and inputs used into biophysical modeling. changes in land cover and hydrologic variables. In this Specific to economic analysis, we made several study, SWAT was calibrated and validated for current assumptions and inputs choices that will have conditions, however, due to equifinality (where the different level of outcomes at different scales and potential output - in this case, basin discharge - can locations. An example is the unit value of water that be met by many possible combinations of input is derived from a residual value approach that is states), as well as unavailability of data to validate highly sensitive to opportunity cost of labor. Since the model in the new (no forest) conditions, the labor cost varies a lot, depending on location and outputs derived for the new state are reliant on employment opportunities, this needs to be carefully the rigor of mathematical representation of the chosen in scaling up efforts. Similarly, sediment relevant processes in the model as well as their maintenance cost of irrigation channels was not parameterization. New physically-based models available in our analysis but it is an important part that improve the representation of vegetation of measuring impacts of upstream land management are continuously being developed and refined on downstream ecosystem services assessment. In – however, data required to run these models, cases of HEP, the determination of the unit value is, at many times, hard to obtain from existing of hydropower has one of the largest impacts on monitoring programs. Going forward, improvement economic value. Furthermore, and as valuation takes in data (as per some of the recommendations in into consideration a long time horizon (100 years this study) will first likely lead to improvement in our case), discount rates and other associated of the characterization of the current model uncertainties (e.g. future demand, supply, policy, and consequently help refine the methods and incentives etc.) therefore need to be carefully models as this field of study further develops. considered in scaling up. VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A 52 4 CARBON VALUATION 4.1 Background Carbon storage in forest ecosystems is globally significant because it reduces greenhouse gases (GHG) in the atmosphere and reduces the warming effect. Cambodia is one of the many signatories to the Paris VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Agreement, which states that “Parties should take action to conserve and enhance, as appropriate, sinks and reservoirs of the GHGs … including forests” (UNFCCC 2015). Cambodia’s Nationally Determined Contribution (NDC) contains explicit commitments to reducing GHG emissions through forest conservation and restoration. Enhancing storage and sequestration of carbon in forest ecosystems in Cambodia is a priority of the Royal Government of Cambodia (RGC) (NCSD 2019). The analysis in this chapter focuses on the carbon storage ecosystem service. The value of carbon in forests is determined through the potential benefits for human well-being that come from minimizing the stock of carbon in the atmosphere where it can contribute to global warming.16 Aerial shot of the Cardamom Mountains, Cambodia. © Conservation International/photo by David Emmett 16 Notwithstanding the lack of consensus among practitioners about whether carbon storage should be considered an ecosystem service determining the value of carbon storage is useful for building consensus on the value of forest ecosystem services. See discussions on carbon storage as an ecosystem services in Keith et al. (2019). 53 4.2 Carbon Pool and Stocks Estimating Carbon Stocks The average carbon stock of an ecosystem is Data on carbon pools is drawn from analysis determined by the environmental conditions, land conducted by Flora and Fauna International (FFI) use and regime of natural and anthropogenic as part of a REDD+ feasibility study for the Central disturbances (Keith et al. 2019). Cardamom Protected Area (Kempinski and Ramos 2013). FFI determine carbon pools based on the Land Cover general requirements of the Verified Carbon Standard (VCS) methodology VM0015 (Pedroni 2012). The Land cover in the Pursat Basin in 2016 is shown in carbon pools included in this methodology are: Table 4.1, based on official data obtained from the above ground biomass (trees); below ground biomass MoE. Forest cover was determined from Landsat 8 (roots); dead wood; and wood products. A biomass satellite images from October 2015 to May 2016, as inventory was used to calculate carbon stock and well as from data from RapidEye, SPOT5, Sentinel-2 estimate 95 percent confidence intervals of the and images from Google Earth for verification of sample.17 These figures were found to be comparable land use/cover classification with 1651 verified points with previous carbon inventories for similar areas covering 25 capital-provinces nationwide. The 2016 and forest types; including those reported by Sasaki land cover was generated with 22 categories of cover, and Yoshimoto (2010), which suggests a stocking of in which forest classes that fell under 13 categories 172 t C/ha (Pedroni 2012) (see Table 4.2). An estimate and non-forest were in 9 categories with minimum of the carbon pool for mangrove forest of 1,094 t C/ mapping 5 ha (MoE 2018). ha was drawn from Kauffman and Bhomia (2017) who determined an average carbon pool for mangroves in East Asia based on several studies. TABLE 4.1: LAND COVER IN THE PURSAT RIVER BASIN, 2016 Land cover Area (ha) Percent of total VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Evergreen forest 242,989 40.8 Deciduous forest 100,903 16.9 Cropland 93,453 15.7 Paddy field 66,812 11.2 Semi-evergreen forest 53,223 8.9 Wood shrub 12,337 2.1 Flooded forest 7,033 1.2 Village 7,033 1.2 Grassland 4,053 0.7 Bamboo 2,324 0.4 Water 1,788 0.3 Forest regrowth 1,490 0.2 Rubber plantation 1,132 0.2 Sand 894 0.1 Built-up Area 536 0.1 Rock 60 0.0 Total 596,060 100.0 54 TABLE 4.2: CARBON STOCKS BY FOREST TYPE Evergreen Semi-evergreen Deciduous Lower Limit (t C/ha) 140.3 109.7 73.7 Upper Limit (t C/ha) 183.0 166.7 94.0 Average (t C/ha) 161.7 138.2 83.9 Average (t CO¬2e/ha)* 593 507 308 Note: C to CO2e is determined with a conversion factor of 3.67 Source: Pedroni 2012. Carbon stocks for the forest assets in the Pursat River and deciduous forest types. The total estimated forest Basin were estimated for evergreen, semi-evergreen carbon stock for each forest type is shown in Figure 4.1. FIGURE 4.1: CARBON STOCKS FOR MAJOR FOREST TYPES IN THE PURSAT BASIN 160 140 120 100 80 60 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A 40 20 0 Evergreen forest Semi-evergreen forest Deciduous forest Source: Authors’ calculations Annual Carbon Value To obtain the total value, these quantities are then multiplied by the unit value of carbon emissions. There are two options for estimating the value The unit values to be used depend on whether of carbon storage ecosystem services provided the value is being considered from the national or by the forest in the Cardamom Mountains. The the global perspective. Carbon emissions resulting first is to value the entire stock of carbon stored from deforestation would affect the entire world by in these forests; this option is appropriate when contributing to GHG concentrations in the atmosphere estimating the value of these forests as an asset. and hence to global climate change. Various efforts The second approach is to value the change in have been made to estimate the cost of the damage stock resulting from current deforestation, or the that would be caused by incremental emissions. The sequestration that would result from reforestation; World Bank, for example, has prepared a range of this approach is more policy relevant. estimates of the Social Value of Carbon (SVC). These 55 range from a low estimate of US$37/tCO2e in 2017 estimates are of potential benefits: How much Cambodia rising to US$78/tCO2e in 2050 and a high estimate could realistically expect to receive if it succeeded in of US$75/tCO2e in 2017 rising to US$156/tCO2e in avoiding forest loss and the resulting emissions. 2050 (World Bank 2017). However, the value of these emissions to Cambodia itself is much smaller: Even From an asset value perspective, the value of the though Cambodia is, of course, heavily threatened by entire stock of carbon stored in forests in the Pursat climate change, the contribution to climate change of River Basin is US$1 billion if valued at US$5/tCO2e and these specific emissions is small. From Cambodia’s of US$7.5 billion if valued at the SVC. perspective, these stocks have concrete value if the country receives compensation for maintaining them About US$2.5 million in carbon value is lost (that is, for avoiding emissions). Thus, from Cambodia’s every year in the Pursat Basin. At the current 0.25 perspective, the stocks should be valued at prices percent annual deforestation rate, the estimated such as it could realistically receive under an emission value of emissions is US$2.5 million/yr if valued at reduction payment agreement (ERPA). A value of US$5/ ERPA prices of US$5/ tCO2e and of US$19 million/ tCO2e is applied by the Green Climate Fund for REDD yr if valued at the SVC (average of high and low projects and can be used to provide a first estimate estimates). If there was an emissions reduction of the value of the forest carbon stocks in the Pursat program that would reduce deforestation in the Basin. Note that as there is no ERPA in place for Pursat watershed to zero, an annual payment of avoided deforestation in the Pursat Basin, the resulting about US$2.5 million could be negotiated.18 TABLE 4.3: ESTIMATED CARBON EMISSIONS FROM DEFORESTATION IN THE PURSAT RIVER BASIN 0.25% deforestation scenario Forest type 2016 forest cover (ha) Average annual forest Estimated annual CO2 loss (ha)* loss (t CO2e) Evergreen 242,989 601 360,000 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Semi-evergreen 53,223 132 67,000 Deciduous 100,903 250 77,000 Total 397,115 504,000 Notes: This analysis assumes the same deforestation rate for all forest types. In reality, there may be some differences in the deforestation rates of the different forest types. * Avg. over 10 years (2016-2025) 18 This payment would be contingent on reducing deforestation to zero. Should deforestation be only partially reduced, the payment to the country would be correspondingly smaller. 56 5 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A TOURISM BENEFITS In addition to the downstream benefits they provide to water users, the forests in the upper part of Pursat River Basin also provide benefits to those who come to visit them. Cambodia’s spectacular and pristine natural assets are exactly what ecotourists look for and the opportunities for supporting the expansion of this industry are great. There are already examples of successful mid- to high-end ecotourism operations in Cambodia which suggest that ecotourism products have great potential. In addition, iconic landscapes like the Cardamom Mountains offer the opportunity to develop new and exciting multi-day itineraries that take advantage of the biodiversity, lush forests and rugged terrain that is perfect for adventure tourism. Shinta Mani Wild resort in the Cardamom Mountains © Shinta Mani Wild 57 Estimates prepared for the Cambodia Sustainable at 200,000 a year to avoid placing excessive pressure Landscape and Ecotourism (CSLE) Project show that on the PAs), the present value of tourism revenue over ecotourism in the PAs of the upper Pursat River Basin 50 years, at a 6 discount rate, is estimated to be about is generating revenue of about US$3 million a year US$210 million.20 However, part of this revenue is spent (about 125,000 visitors a year, each spending US$22/ on providing these visitors with accommodation, food, day and staying a little over a day).19 Moreover, this facilities, guides, etc. Assuming conservatively that revenue is projected to gradually increase over time as about a quarter of revenue represents a net benefit the number of visitors, their spending and their length to Cambodia, the net value of ecotourism benefits of stay are all slowly increasing. Assuming current generated by the forests in the upper Pursat River trends continue (but capping the number of visitors Basin is estimated at about US$53 million. FIGURE 5.1: PROJECTED VISITORS TO PAS IN THE UPPER PURSAT RIVER BASIN 225, 000 200, 000 150, 000 150, 000 Projected visitors 125, 000 100, 000 75, 000 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A 50, 000 25, 000 0 2020 2030 2040 2050 2060 Source: Estimates prepared for the CSLE project.21 Benefits of Improved Tourism investments could help increase them. The recently approved Cambodia Sustainable Landscape and Continued deforestation, even at current low rates, Ecotourism (CSLE) Project, for example, aims to would result in a reduction of the benefits provided increased benefits from tourism and carbon storage by forests in the Pursat River Basin. Conversely, services by supporting RGC investments in ecotourism 19 Visitor numbers were estimated from data provided by the Ministry of Environment on tourist arrivals to the Chrok La Eng Waterfall, Anlong Svay and Phnom Chreav Waterfall in Phnom Aural Wildlife Sanctuary. The estimate of average visitor spend is based on data from community-based ecotourism operations (CBETs). 20 These estimates do not take into account the impact of the COVID emergency, which will reduce ecotourism significantly in the short term. 21 CSLE project: http://documents.worldbank.org/curated/en/934211559527272828/Cambodia-Sustainable-Landscape-and-Ecotourism-Project 58 development and protected areas management continue to decline, with continued forest loss and and law enforcement. The CSLE project, which is degradation, declining soil quality and agricultural being implemented over a six-year period from 2020 output and continued high emissions of GHGs. Poor to 2025, is expected to cost US$55 million. It will access to tourism sites and limited tourism options support PA management, planning and enforcement would limit both the number of visitors, as well as at seven sites, which will reduce forest cover loss by the length of their stay and daily spending in these more than 2 million ha in the Cardamom Mountains destinations. Communities and smallholder farmers Tonle Sap landscape, thereby increasing carbon would continue subsistence agricultural practices, storage in the Cardamom forests. The project will with low productivity and value added and limited also increase access to ecotourism sites from main diversification toward other economic opportunities. hubs by improving infrastructure, thereby increasing In these conditions, project interventions would be the number of visitors. New ecotourism activities considered positive even if they only slowed the will also be developed at the project sites (visitor continuing negative trends. The analysis compared the and interpretation centers, hiking trails, tourism balance of the “without” and “with” project scenarios facilities), thereby increasing both the visitors’ length for each benefit stream, discounted over a 20-year of stay and their daily spending. Increased income time horizon using a 6 percent discount rate to in the beneficiary communities will lead to greater determine the value addition of the project in relation tax revenue for the RGC. There will also be multiplier to its financial input. To place a value on the benefits effects of the tourism spending in the country and in of reducing carbon emissions a conservative value of the target landscapes. US$5 t/ CO2e was used. Explanations of the scenarios developed for each benefit stream and further details A model was developed to compare the stream of of the benefits analysis are provided in Appendix 5. benefits and costs under various scenarios to assess Based on these scenarios, the benefits of improved the incremental benefits generated by the project in tourism and carbon storage from the forest in Pursat the Pursat RB. In the “without project” scenario, in the River Basin are estimated to be US$95 million (see target areas, the environmental situation is likely to Figure 5.2).22 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A FIGURE 5.2: PROJECTED INCREASE IN TOURISM AND CARBON BENEFITS FROM THE CSLE PROJECT 1, 000, 000, 000 Tourism Carbon 900, 000, 000 800, 000, 000 700, 000, 000 600, 000, 000 500, 000, 000 400, 000, 000 300, 000, 000 200, 000, 000 100, 000, 000 0 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 Source: Authors’ calculations 22 Note that these estimates assume that tourism would begin to grow beginning in year 2 of the project. The COVID emergency will almost certainly delay the tourism benefits, perhaps significantly. 59 Limitations As a direct result, the analysis has aimed to significantly reduce the original assumptions Determining the real revenue value of project sites of the impact of the project. However, it was the greatest constraint to this analysis. Data for would be important to address these data the analysis was limited and incomplete and have challenges during the implementation not been subject to validation or ground-truthing. stage to validate the NPV assumptions. VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A 60 6 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A RECOMMENDATIONS As earlier stated, this analytical work is being undertaken as part of a broader effort of the World Bank in Cambodia to provide guidance to the RGC through technical assistance and analytical and advisory services on managing its natural capital through landscape approaches. The earlier chapters presented results of the benefits of forest ecosystem services, which provide evidence of the returns on investments that the RGC will gain through the strengthening of forest ecosystems through the Cambodia Sustainable Landscape and Ecotourism project. Methodologies on how to undertake measurement and valuation of ecosystem services were presented so that this work can be repeated in other places. In this chapter, recommendations to guide the RGC in some important next steps are provided. Flooded Forests, Tonle Sap © Conservation International/photo by Tangkor Dong 61 Key Messages of the Study 1. The public benefits from intact forests estimated at US$99 million are nearly five times higher than the private gains had from cutting them down for small-scale agriculture or charcoal production estimated at US$22 million. 2. Investing in the maintenance of forest is good business. Annual public expenses to maintain the forest in the Pursat Basin are about 20 times lower than the public benefits provided by them. 3. Funding for the maintenance of those forests in the long run can be captured from private and international sources. This work proposes three recommendations that a quantifiable risk to the operation of irrigation and are intended to: (i) help the RGC better integrate hydropower infrastructure downstream. These are, ecosystem service values into forest and PA in the form of changes to the pattern of seasonal management decision-making; (ii) indicate some water yield, higher consumption upstream in water priority geographic areas within the Cardamom stressed months for irrigation and increased sediment Mountain landscape for action on prioritization accumulating in downstream infrastructure. Measures of forest protection and management efforts; and to arrest rate of deforestation and engage in options (iii) expand efforts to address drivers of forest for afforestation would be recommended to protect degradation in Cambodia with financial instruments this resource. The results showed that important like payments for ecosystem services (PES). areas for water yield are overlapping with PAs in the Cardamom Mountains, but also overlapping The results in this report lead to two sets of with an area of the Cardamom Mountains that recommendations: (1) policy recommendations has experienced high rates of degradation due to VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A for the Pursat River Basin (Recommendations agricultural encroachment (Figure 5.1). With climate 1 and 2) and (2) recommendations aimed at change projections for the region suggesting that scaling up the analysis to ultimately cover there will be lower rainfall during dry season and all of the country (Recommendation 3). higher rainfall in wet season (Figures 5.2 and 5.3), these forests act as natural protection infrastructure Recommendation 1: Focus forest and therefore will contribute towards countering protection and restoration efforts on the trend enforced by the shifting rainfall pattern. upstream watersheds in the Cardamom Deforestation and forest degradation, on the other Mountains protected area landscape to hand, will reinforce the impact of the climatic trends, magnifying the risk of extreme events like floods enhance resilience of water resources in the basin. Protection of upstream forestlands Cardamom forested lands act to slow down high takes on renewed importance in this light. discharges during wet season and supplement low The RGC has some important decisions to make with flows during dry season. Upstream deforestation in regards to prioritizing areas in the PA landscape for the Sam Kos Wildlife sanctuary, above Dam 1 linked attention, additional resources for management to improved accessibility by access roads, possesses 62 and even restoration activities. In this context, Action 3: Develop interventions for reducing and under this recommendation, there are some the pressure on forest resources from charcoal, strategic actions proposed for the government. including more sustainable charcoal production and environmentally friendly alternatives to wood Action 1: Prioritize zoning and development of PA charcoal. The charcoal industry can be a significant management plan in Samkos Wildlife Sanctuary and opportunity for Cambodia’s rural PA economies if Biodiversity Conservation Corridors. MoE is already done right. GERES (2015) assessed the industry to taking a step in this direction through the inclusion of be worth about US$177 million per year and RGC will Samkos Wildlife Sanctuary as a priority PA for zoning, need to play a key role in leading the organization of PA management planning and boundary demarcation. this industry in order to reduce its potential negative The PA management planning should assess the best impacts on forests. Sourcing wood for charcoal from options for reducing forest degradation and conversion existing plantations can help to address the wood and should include a plan for restoration of degraded supply needs for the charcoal industry. At the same landscapes. This analysis provides information that time, improving the wood-to-charcoal conversion MoE can use such as areas of high sediment yield efficiency could also help to reduce pressure on and high-water yield (Figures 2.7, 2.9 and 2.11) in forests for wood. Four key measures proposed for prioritizing areas within Samkos for interventions. further action on moving towards more sustainable charcoal production are: (i) formalization of existing Action 2: Assess opportunities for agroforestry on small-scale charcoal producers and linking these existing agricultural lands within forested areas. with private sector plantations who can provide Halting forest degradation and conversion is one a consistent source of wood and through which a of the key objectives of MoE’s management of PAs. certification system for charcoal could be developed; As agriculture is a key driver of forest change in the (ii) refining existing charcoal producing techniques Cardamoms, it is important that this is addressed and technologies to improve energy efficiency of in a manner that is pro-poor, recognizing that forest the firewood conversion into charcoal in a cost- communities need livelihood support. Developing effective way; (iii) developing small-scale woodlots agroforestry is a way of creating additional value within community use zones and sustainable use on lands that have been converted to agriculture VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A zones to meet wood needs for charcoal; (iv) exploring and restoring trees to provide some important opportunities for more environmentally-friendly ecosystem services like sediment regulation. The options for charcoal, like coconut husks.23 Again, interventions that are decided from the assessment the link to the private sector would be strategic for should also be included in the PA management plan. investments and management of the wood lots. 23 Green Fuel is a company operating in Cambodia that is producing charcoal from coconut husks. More information available in https://www.khmertimeskh.com/57549/converting-coconut-husks-into-charcoal/ 63 FIGURE 6.1: AREAS OF HIGH WATER YIELD WITHIN PAS IN THE CARDAMOM MOUNTAINS IMPORTANT AREAS OF WATER YIELD OVERLAP THE SAMKOS WILDLIFE SANCTUARY, CENTRAL CARDAMOM NATIONAL PARK, AURAL WILDLIFE SANCTUARY AND BIODIVERSITY CORRIDOR. THIS TYPE OF MAP CAN BE USED AS AN INPUT IN THE ZONING ACTIVITIES FOR THESE PROTECTED AREAS UNDER THE CSLE PROJECT, TO INFORM THE DESIGNATION OF CORE ZONES. VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Source: Authors’ creation 64 FIGURE 6.2: PRECIPITATION IS PROJECTED TO INCREASE IN THE WET SEASON OVER CARDAMOM MOUNTAINS Source: International Centre for Environmental Management (ICEM) 24 FIGURE 6.3: PRECIPITATION IS PROJECTED TO DECREASE IN THE DRY SEASON WITH CLIMATE CHANGE VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A 22 See ICEM maps at https://icem.com.au/portfolio-items/mekong-arcc-climate-change-maps/ 65 Recommendation 2: Explore the services (irrigated farmers and electricity users) and potential for private financing quantifies the damages they would face if these services were lost or reduced. Third, the hydrological to support PA management model developed for the analysis provides tools that would permit PES conservation efforts to the areas NCA can help inform government decision-makers, where they would be most effective. but it will not by itself change the incentives facing actors on the ground. In the Pursat RB, these actors To be sure, much more needs to be done. The receive only a small subset of forest benefits but stand quantification of benefits provided by this analysis, for to receive the bulk of the benefits from alternatives example, only provides an upper bound of willingness such as agriculture. to pay to avoid damages. For example, the analysis shows that at a deforestation rate of 1 percent a year, Public sector resources are needed in the short the losses suffered by HEP producers would be about term to finance maintenance of forests, but the US$0.17 million a year. This figure is the maximum such RGC should consider developing over the medium producers would be willing to pay for a PES program term a financing approach that integrates financing that stopped all deforestation completely - including from private sources for the management of forests. both the costs of the payments to participants and This approach is an important emerging trend in the costs of implementation of the program. The conservation. If undertaken effectively, private corresponding figure on damages that irrigated capital investment can serve as a complement or farmers would suffer from deforestation would add to alternative to traditional conservation funding. With this amount and if an ERPA can be negotiated, carbon the prospect of more stable long-term funding than payments would increase it even further. traditional granting, this approach can provide deeper conservation and livelihood results. Payment for Designing and implementing a PES program ecosystem services (PES) for water and payments to reduce deforestation in the upper for carbon under REDD+ (Reducing Emissions from Pursat River Basin would require: Deforestation and Degradation) are already being undertaken in Cambodia and a critical step for the 1. Using the hydrological model developed for VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A RGC is organizing these as part of an overall system of this analysis to identify the critical areas in the financing for forest management. upper basin, the areas which, if they were to be deforested, would result in the greatest impact PES and REDD+ are realistic opportunities in on hydrological flows and sediment loads; Cambodia for directing private financing to support 2. Undertaking a threat assessment of these PA management. There is strong interest of the MoE in areas, to see how likely they are to be actually establishing PES with ongoing pilots helping to inform deforested, based on factors such as their development of PES. Payments from international suitability for agriculture, proximity to roads, private sector under the REDD+ mechanism are already etc. and quantifying the potential benefits to being received in Cambodia and could scale-up. There local people of converting these areas to other are very good studies already on evaluating carbon uses (taking into account likely crop yields, resources (stocks) and important is developing the costs of production, etc.); right benefit-sharing mechanisms and ensuring that 3. Measuring any benefits that retaining forests these link to an overall revenue system that support could generate for local communities, for protected areas. example through the sustainable collection of NTFPs and through activities such as The present study contributes to the development ecotourism; of PES in several ways. First, it clearly documents 4. Estimating the cost of a PES program to protect the benefits provided by forests. Second, it identifies these critical areas, based on their size (number several important groups of beneficiaries of these 66 of ha to be protected), the size of payments 3. Enhance and promote the attractiveness of needed to induce forest conservation, and the Cambodia for REDD + payments with clear rules net costs to local communities of conserving and regulation for the system; them (potential benefits from conversion to 4. Ensure that PES and REDD+ payments are well agriculture minus local benefits from retaining integrated into the overall financing mechanism for forests). Also the likely implementation costs PAs. Recent work for the Greening Prey Lang project of the program, based primarily on the cost identified a number of fund sources including of monitoring, which is affected by the size of the Environment and Social Fund, the Forestry plots to be monitored, their dispersion and Administration National Forest Development Fund their accessibility; and private conservation funds that needed to 5. Based on these estimates, determining whether be managed and used in an integrative way to be the program is feasible (i.e. the total costs are efficient and effective. less than total willingness to pay); putting in place arrangements to collect funding from Recommendation 3: Develop a road map service users (such as irrigated farmers, HEP for scaling up assessment of economic producers and/or carbon buyers) and; make benefits provided by forest ecosystems appropriate payments to service providers across Cambodia using a Natural Capital (upstream communities who refrain from Accounting (NCA) Approach deforesting). The analysis conducted provides a substantial start on this road map, but clearly The advantages for Cambodia of a natural capital much more needs to be done. approach (NCA) verses one-off economic valuation studies are: (i) standardizing how ecosystem Lastly, it would be important for any PES scheme service values are determined and integrated into to be part of ongoing plans within the government regular decision-making of the RGC, for example in for strengthening the institutional framework determining national budget allocation for MoE for for emissions reduction payments. The national PA management; and (ii) that data and information REDD+ strategy and REDD+ nesting framework, that will be more reliable and less costly if data collection, is currently being established, also provides the VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A analysis and access are standardized under an opportunity to provide payments from reducing carbon NCA approach. The analysis of the Pursat Basin emissions through undertaking forest conservation, demonstrates the potential benefits of undertaking conservation compatible livelihood activities etc. NCA both to identify the need for interventions and The kind of analysis undertaken in this study provides to help design them and elucidates some key lessons a basis for the levels of investment needed to for replication and scaling up. Lessons include: (i) holistically ensure and incentivize more emissions the need for thorough analysis of the interactions reduction payments. In addition, we recommend the of beneficiaries with ecosystem services; and (ii) the RGC to: importance of a robust data collection plan and early commitment on data sharing from relevant ministries. 1. Provide oversight and management of REDD+ activities as is being proposed in the Moving from a single case study to a comprehensive REDD+ Regulatory Framework (Prakas), which is approach requires a road map that includes: being developed; 1. Conducting a “scoping” exercise that identifies (i) 2. Set up the national system for emissions reduction policies, decision-making and planning processes payments which includes a benefit sharing for which the implementation of NCA could mechanism that will make clear investments provide critically important information; and (ii) for forest conservation and protected area data availability/needs, institutional framework, management, including co-management, etc.; 67 financial, technical resources and capacity required instructive for prioritizing areas for watershed for NCA; management. 2. Identifying and informing key institutional b. Cambodia has invested significantly in partners that should be engaged (i) at the ministry hydropower plants on the Mekong River, and level, such as MoE, Ministry of Water Resources as well in the Cardamom Mountains. Ensuring and Meteorology (MOWRAM), Ministry of Rural as close to maximum operation capacity of Development (MRD), Ministry of Agriculture, these hydropower plants will be important for Forestry and Fisheries (MAFF), Ministry of Interior energy security in Cambodia especially in the (MoI), Ministry of Economy and Finance (MEF); dry season, and this means protecting forest (ii) at the provincial level, including Provincial watersheds that are upstream of this dam. Department(s) of Environment (PDoE); and (iii) Valuation of hydrological services can support NGOs and Development Partners like Conservation schemes for hydropower companies to provide International, Flora and Fauna International (FFI) finance that can support the management of and World Wildlife Fund (WWF) who currently forest resources that provide critical water flow undertake related work on ecosystem valuation; regulation and sediment regulation for the 3. Considering a phased approach, starting with operation of hydropower plants. basin-specific accounting-compatible assessments c. Carbon storage, as an ecosystem service, with a small set of key ecosystem services – such is strategic to analyze as there are well- as those in the current report - and evolving established methodologies for doing this. The towards a more encompassing exercise that well-established carbon market establishes would in time be extended to the country’s a price which often accounts for regulating national boundaries. A phased approach could ecosystem services (which have no market and/ initially focus on representative watersheds where or are difficult to value) and biodiversity that there are clear beneficiaries, as in the case of facilitate carbon storage. the Pursat Basin. Criteria for prioritizing areas d. Ecotourism development in PAs is a priority of for undertaking ecosystem service accounting- the RGC to boost the overall tourism sector in compatible assessments may include: Areas that terms of jobs and value added, provide incomes VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A are most at risk from degradation and forest for rural and forest communities and generate loss; watersheds important for hydropower resources that can help with PA management. and irrigation, water production and sediment 4. We would recommend that initial regulation. Additional ecosystem services consideration for geographic priorities for that should be considered include water flow undertaking such an assessment to include regulation with a particular focus on drought. the watersheds that feed into the Tonle Estimation of economic benefits of hydrological, Sap and Kulen Mt. and Kbal Chay where the carbon and tourism ecosystem services should government is pursuing pilot PES projects; be prioritized for the following reasons: 5. The scoping and road map would most certainly a. Cambodia is experiencing a serious water highlight the need for enhancement on monitoring shortage which is expected to be exacerbated and generation of data for similar assessments and by climate change. It is therefore important that ultimately for accounting efforts: RGC, through MoE and MOWRAM, strengthen a. As changes in soil erosion and sediment management of important watersheds, like the accumulation are significantly affected by Cardamom Mountains and Kulen Mountains, forest change and can result in large costs, it is with protection of existing forest resources and recommended that monitoring of suspended restoration of degraded important watershed sediment and bed load, at least at the site of areas. Analysis of hydrological ecosystem the future dams or at the main gauging station services, as undertaken for this study, will be be undertaken. Additionally, experimental plots 68 to monitor soil erosion rates could be helpful in important recharge areas and travel time. verifying soil loss projections. Without this, we risk incomplete protection b. Rainfall variability in the mountainous for water source areas, as we are limiting our region is high. Weather monitoring needs source region of the rivers based on visible to be strengthened to derive accurate terrain slope contributing to the river water. estimates to water resources available. 6. The benefits of forest for disaster reduction – flood This will become especially important as mitigation and forest fire prevention – would also precipitation patterns continue to shift be important to capture in subsequent analysis. with a changing climate. A more extensive Data on the flood, drought and fire damages would network of rainfall gauges is needed. be important for determining the benefits provided c. Groundwater often plays an important role, but by forest in terms of disaster risk reduction; data on groundwater are even less available 7. Additional benefits that would be important than for surface water. Improved groundwater to capture include: Water used for domestic mapping and monitoring is needed to better purposes; recreational ecosystem services understand the role that it plays. The first from ecotourism should be considered for essential step of this should be to map areas where this is significant; non-timber the major aquifers, followed by identifying forest products (NTFPs); and charcoal. VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A 69 7 CONCLUSION VA L U I N G T H E E CO SYST E M S E RV I C E S P R OV I D E D BY F O R E STS I N P U R SAT BA S I N , C A M B O D I A The development of a methodology for b. Identifying and prioritizing areas in the undertaking an ecosystem services assessment landscape for forest conservation because and valuation, including the results of this they are important for water provisioning. work, creates several strategic opportunities 2. On the economic side, this work shows that for the RGC to enhance its decision- the costs of losing forests are significant, and making capability on investing in forests they negatively impact parts of the economy. and Protected Areas and to quantify and Beneficiaries of the services are identified communicate the value of its natural capital and these benefits quantified in monetary to Cambodia’s economy. The contribution of terms that allow for easier comparison with this work is summarized as the following: the costs of investment. 8. The hydrologic model developed in this 3. Upstream activities like land use conversion analysis is a key contribution for: (e.g. forest to agriculture) have significant a. Quantifying what would happen if forest impacts for people in the downstream ecosystem were lost and therefore impacted by flooding, sedimentation and understanding the contribution of forests in impaired water flows and there upstream- helping to provide water flow regulation and downstream linkages have been spatially sediment regulation services; identified and quantified in this work. Akal village on Tonle Sap lake 70 © Conservation International/photo by Tangkor Dong REFERENCES Abbaspour, K.C. 2012. “SWAT-CUP 2012: SWAT calibration and uncertainty programs–a user manual.” Dübendorf, Switzerland: Swiss Federal Institute of Aquatic Science and Technology. Abbaspour, K.C., E. Rouholahnejad, S. Vaghefi, R. Srinivasan, H. Yangand B. Kløve. 2015. “A continental-scale hydrology and water quality model for Europe: Calibration and uncertainty of a high-resolution large-scale SWAT model.” Journal of Hydrology, 524:733-752. Arifanti, V.B. 2017. “Influences of Land Use on Carbon Cycles in Mangrove Ecosystems of the Mahakam Delta, East Kalimantan, Indonesia.” PhD Dissertation. Corvallis: Oregon State University. Asian Development Bank (ADB). 2014. “Cambodian Water Resources Profile”. Water Resources Management Sector Development Program ADB Loan 2673-CAM and TA 7610-CAM. Arias, M.E., T.A. Cochrane, K.S. Lawrence, T.J. Killeen and T.A. Farrell. 2011. “Paying the forest for electricity: a modelling framework to market forest conservation as payment for ecosystem services benefiting hydropower generation.” Environmental Conservation, 38(4):473-484. Arnold, J.G., P.M. Allen and G. Bernhardt. 1993. “A comprehensive surface-groundwater flow model.” Journal of Hydrology, 142:47-69. Berbel, J., M.A. Mesa-Jurado and J.M. Pistón. 2011. “Value of irrigation water in Guadalquivir Basin (Spain) by residual value method.” Water Resources Management, 25(6):1565-1579 Bogdan, S.M., I. Pătru-Stupariu and L. Zaharia. 2016. “The Assessment of Regulatory Ecosystem Services: The Case of the Sediment Retention Service in a Mountain Landscape in the Southern Romanian Carpathians.” VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Procedia Environmental Sciences, 32:12-27. Bottrill, Madeleine, Rachel Neugarten, Leonardo Saenz, David Hole, Bambi Semroc, Trond Larsen, Steven Panfil, Curan Bonham, Marc Steininger, Wu Yang, Natalia Acero, Andres Carlos Cano, Annette Olsson, Matthew Gibb, Kellee Koenig, Timothy Max Wright, Will Turner, Bunra Seng, and Tracy Farrell. 2015. “Metrics for Green Growth in Cambodia: Demonstration of Metrics for Conservation and Human Well-being”. Conservation International, Washington D.C. Brouwer, C. and M. Heibloem. 1986. “Irrigation Water Management: Irrigation Water Needs.” Training Manual No.3. Rome: FAO. Available at: http://www.fao.org/3/s2022e/s2022e00.htm#Contents CDRI (Cambodia Development Resource Institute). 2010. “Empirical Evidence of Irrigation Management in the Tonle Sap Basin: Issues and Challenges.” CDRI Working Paper Series No.48. Phnom Penh: Cambodia Development Resource Institute. Conservation International (CI). 2012. “REDD+ in the Central Cardamoms: A Feasibility Study.” Phnom Penh. Donato, D.C., J.B. Kauffman, D. Murdiyarso, S. Kurnianto, M. Stidham, M. Kanninen. 2011. “Mangroves among the most carbon-rich forests in the tropics.” Nature Geoscience, 4:293-297. Dong, Z., Y. Yan, J. Duan, X. Fu, Q. Zhou, X. Huang, X. Zhu and J. Zhao. 2011. “Computing payment for ecosystem 71 services in watersheds: An analysis of the Middle Route Project of South-to-North Water Diversion in China.” Journal of Environmental Sciences, 23(12):2005-2012. Dreher, A., A. Fuchs, B.C. Parks, A.M. Strange and M.J. Tierney. 2017. “Aid, China and Growth: Evidence from a New Global Development Finance Dataset.” AidData Working Paper #46. Williamsburg, VA: AidData. Erban, Laura E. and Steven M. Gorelick. “Closing the irrigation deficit in Cambodia: Implications for transboundary impacts on groundwater and Mekong River flow.” Journal of Hydrology, 535: 85–92. Filoso, S., M.O. Bezerra, K.C. Weiss and M.A. Palmer. 2017. “Impacts of forest restoration on water yield: A systematic review.” PloS ONE, 12(8):e0183210. Forest Trends. 2015. “Conversion Timber, Forest Monitoring and Land-Use Governance in Cambodia.” Forest Trends Report Series July 2015. Available at: https://www.forest-trends.org/wp-content/uploads/2015/07/ Cambodia20Concessions20Report20small20size.pdf Group for the Environment, Renewable Energy and Solidarity (GERES), 2015. Biomass Energy Consumption Patterns in Cambodia: Challenges and Opportunities. Available at: http://www.eurocham-cambodia.org/ uploads/85746-geres.pdf Hussain, I., H. Turral and D. Molden. 2007. “Measuring and enhancing the value of agricultural water in irrigated river basins.” Irrigation Science, 25(3):263-282. Karn, P.K. 2008. “Making Payment for environmental services (PES) work: A case study of Shivapuri National Park, Nepal.” In: S.B. Bajracharya and N. Dahal (Eds.), Shifting Paradigms in Protected Area Management. Kathmandu: National Trust for Nature Conservation. pp.171-185. Kauffman, J.B., V.B. Arifanti, H. Hernandez-Trejo, J.G.M. del Carmen, J. Norfolk, M. Cifuentes, D. Hadriyanto and D. Murdiyarso. 2017. “The jumbo carbon footprint of a shrimp: carbon losses from mangrove deforestation.” VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Frontiers in Ecology and the Environment, 15(4):183-188. Kauffman, J.B. and R.K. Bhomia . 2017. “Ecosystem carbon stocks of mangroves across broad environmental gradients in West-Central Africa: Global and regional comparisons.” PLoS ONE, 12(11): e0187749. Kauffman, J.B., C. Heider, T.G. Cole, K.A Dwire and D.C. Donato. 2011. “Ecosystem carbon stocks of Micronesian mangrove forests.” Wetlands, 31:343-352 Kaura, M., M.E. Arias, J.A. Benjamin, C. Oeurng and T.A. Cochrane. 2019. “Benefits of forest conservation on riverine sediment and hydropower in the Tonle Sap Basin, Cambodia.” Ecosystem Services, 39:101003. Keith, H., M. Vardon, D. Lindenmayer and B. Mackey. 2019. “Accounting for carbon stocks and flows: Storage and sequestration are both ecosystem services.” Paper presented at the 25th Meeting of the London Group on Environmental Accounting, Melbourne, 7-10 October 2019. Kempinski, J. and J. Ramos. 2013. “Central Cardamom Protected Forest REDD+ Feasibility Study. Cambridge: Flora and Fauna International. Khmer Times. 2015. “Cambodia’s Biggest Dam Boosts Power Grid.” January 13, 2015. Available at: https://www. khmertimeskh.com/53789/cambodias-biggest-dam-boosts-power-grid/ 72 Lee, Y.H. and K. Kobayashi. 2017. “Rice yield increase by the system of rice intensification is dependent on supplementary water availability in rainfed lowland fields of southern Cambodia.” Tropical Agriculture and Development, 61(1):48-55. Le Maitre, D.W., I.M. Kotzee and P.J. O’Farrell. 2014. “Impacts of land-cover change on the water flow regulation ecosystem service: Invasive alien plants, fire and their policy implications.” Land Use Policy, 36:171-181, Mesa-Jurado, M.A., J. Berbel and F. Orgaz. 2010. “Estimating marginal value of water for irrigated olive grove with the production function method.” Spanish Journal of Agricultural Research, 8(S2):197-206. Miglietta, P.P., D. Morrone and F. De Leo. 2018. “The water footprint assessment of electricity production: an overview of the economic-water-energy nexus in Italy.” Sustainability, 10(1):228 MoE (Ministry of Environment). 2018. Cambodia Forest Cover 2016. Phnom Penh: Kingdom of Cambodia. MoE (Ministry of Environment). 2017. National Protected Area Strategic Management Plan 2017-2031. MOWRAM (Ministry of Water Resources and Meteorology). 2013. “Water demand analysis within the Pursat River catchment.” Phnom Penh: Royal Government of Cambodia. Available at: https://wle-mekong.cgiar.org/ download/mk16-evidence-based-iwrm-catchment/MK16_Water-Demand-Analysis-Pursat-river-catchment. pdf Murdiyarso, D.M., J. Purbopuspito, J.B. Kauffman, M.W. Warren, S. Sasmito, D. Donato, S. Manuri, H. Krisnawati, S. Taberima and S. Kurnianto. 2015. “The potentials of Indonesian mangrove forests for global climate change mitigation.” Nature Climate Change, 5:1089-1092. NCSD (National Council for Sustainable Development). 2019. “NDC Road map and Stakeholder Engagement Plan 2019-2030.” Phnom Penh: Royal Government of Cambodia. VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Neitsch, S.L., J.G. Arnold, J.R. Kiniry and J.R.,Williams. 2011. “Soil and Water Assessment Tool: theoretical documentation.” Version 2009. Texas Water Resources Institute. Neugarten, R.A., K. Moull, N. Acero Martinez. L. Andriamaro, C. Bernard, C. Bonham, C.A. Canoa, P. Ceotto, P. Cutter, T.A. Farrella, M. Gibb, J. Goedschalk. D. Hole, M. Honzák, T. Kasecker, K. Koenig, T.H. Larsen, J.C. Ledezma, M. McKinnon, M. Mulligan, R. Nijbroek, A. Olsson, Z.L. Rakotobe, A. Rasolohery, L. Saenz, M. Steininger. T.M. Wright and Turner. 2020. “Trends in protected area representation of biodiversity and ecosystem services in five tropical countries.” Ecosystem Services, 42:101078 Oeurng, C., T.A. Cochrane, S. Chung, M.G. Kondolf,T. Piman and M.E. Arias. 2019. “Assessing climate change impacts on river flows in the Tonle Sap Lake Basin, Cambodia.” Water, 11(3):618. Pedroni, L. 2012. “Approved VCS Methodology VM0015: Methodology for Avoided Unplanned Deforestation.” Version 1.1, 3 December 2012 Sectoral Scope 14. Carbon Decisions International. Available at: https://verra.org/wp- content/uploads/2018/03/VM0015-Methodology-for-Avoided-Unplanned-Deforestation-v1.1.pdf Persson, Linn, Nang Phirun, Chanrith Ngin, John Pilgrim, Chanthy Sam and Stacey Noel. 2010. “Ecosystem Services Supporting Livelihoods in Cambodia.” Stockholm Environment Institute, Project Report – 2010. Plummer, M.L. 2009. “Assessing benefit transfer for the valuation of ecosystem services.” Frontiers in Ecology and 73 the Environment, 7(1), pp.38-45 Richardson, L., J. Loomis, T. Kroeger and F. Casey. 2015. “The role of benefit transfer in ecosystem service valuation.” Ecological Economics, 115:51-58 Portela R., M.O. Bezerra, M. Alam, K. Shaad, O. Banerjee and M. Honzák. 2019. “Water Supply Services: Biophysical Modeling and Economic Valuation in Ecosystem Accounting.” Discussion Paper 8. Submitted to the Expert Meeting on Advancing the Measurement of Ecosystem Services for Ecosystem Accounting, New York, 22-24 January 2019 and subsequently revised. Version of 15 March 2019. Available at: https://seea.un.org/events/ expert-meeting-advancing-measurement-ecosystem-services-ecosystem-accounting Raju, K V and Avinandan Taron. 2018. “Enabling sustainable and inclusive irrigation development in Cambodia”. Munich Personal RePEc Archive, 90172. Available at: https://mpra.ub.uni-muenchen.de/90172/ Rohan, Brian. 2019. “Prey Land Extended Landscape Sustainable Finance Plan”. Tetra Tech, Vermont USA Royal Government of Cambodia (RGC). 2014. National Strategic Development Plan 2014-2018. Ministry of Planning, Phnom Penh. Royal Government of Cambodia (RGC). 2017. National REDD+ Strategy 2017-2026. Phnom Penh. Available at: https:// redd.unfccc.int/files/20180813_national_redd__strategy_cambodia.pdf Sasaki, Nophea and Yoshimoto, Atsushi. 2010. Benefits of tropical forest management under the new climate change agreement—a case study in Cambodia. Environmental Science & Policy. 13. 384-392. 10.1016/j. envsci.2010.04.007. Smith, S., Rowcroft, P., Everard, M., Couldrick, L., Reed, M., Rogers, H., Quick, T., Eves, C. and White, C. 2013. “Payments for Ecosystem Services: A Best Practice Guide”. Defra, London VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Sok, K., P. Méral, D. Pillot and S. Defossez. 2019. “Ecosystem Services from Tonle Sap Flood Pulse: Spatial and Economic Analysis in Aek Phnom and Sangkae Districts of Battambang Province, Cambodia.” In: M. Stewart and P. Coclanis (Eds), Water and Power. Advances in Global Change Research, Vol.64. Springer. Speelman, S., S. Farolfi, S. Perret, L. D’haese and M. D’haese. 2008. “Irrigation water value at small-scale schemes: evidence from the North West Province, South Africa.” International Journal of Water Resources Development, 24(4):621-633. Srean, P., B. Eang, R. Rien and R.J. Martin. 2018. “Paddy rice farming practices and profitability in northwest Cambodia.” Asian Journal of Agricultural and Environmental Safety, 1:1-5. United Nations. 2014. System of Environmental-Economic Accounting 2012: Experimental Ecosystem Accounting. New York: UN, EU, FAO, OECD, WBG. UNSD (United Nations Statistics Division). 2012. “System of Environmental-Economic Accounting for Water (SEEA- Water). Department of Economic and Social Affairs.” ST/ESA/STAT/SER.F/100. New York: UNSD. Available at: http://unstats.un.org/unsd/envaccounting/seeaw/seeawaterwebversion.pdf United Nations, European Commission, Food and Agricultural Organization of the United Nations, Organisation for Economic Co-operation and Development, The World Bank (2014b) System of Environmental–Economic 74 Accounting 2012—experimental ecosystem accounting. United Nations, New York UNFCCC (United Nations Framework Convention on Climate Change. 2015. “Adoption of the Paris Agreement.” UNFCCC Conference of the Parties, 21st session, Paris, November 30-December 11, 2015. UN-REDD (United Nations Programme on Reducing Emissions from Deforestation and Forest Degradation). 2016. “Initial Forest Reference Level for Cambodia under the UNFCCC Framework”. Available at: https://redd. unfccc.int/files/cambodia_frl_rcvd17112016.pdf Verdin, A., B. Rajagopalan, W. Kleiber and C. Funk. 2015. “A Bayesian kriging approach for blending satellite and ground precipitation observations.” Water Resources Research, 51(2):908-921. World Bank. 2015. “Cambodian Agriculture in Transition: Opportunities and Risks.”, Report No.96308-KH. Washington: World Bank. World Bank. 2017. “Cambodia: Strategic Framework for the Irrigation Sector.” Technical Assistance Note. Water and Agriculture Global Practices, East Asia and Pacific Region. Washington: World Bank. World Bank. 2017. “Guidance Note on Shadow Price of Carbon in Economic Analysis.” Washington: World Bank. Available at: http://pubdocs.worldbank.org/en/911381516303509498/2017-Shadow-Price-of-Carbon- Guidance-Note-FINAL-CLEARED.pdf World Bank. 2020. World Bank Open Data. Available at: https://data.worldbank.org/ Yassin, F., S. Razavi, M. Elshamy, B. Davison, G. Sapriza-Azuri and H. Wheater. 2019. “Representation and improved parameterization of reservoir operation in hydrological and land-surface models.” Hydrology and Earth System Sciences, 23(9):3735-3764. Xie, P.and A.Y. Xiong. 2011. “A conceptual model for constructing high-resolution gauge-satellite merged VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A precipitation analyses.” Journal of Geophysical Research: Atmospheres, 116(D21). 75 APPENDIX 1: DATA REQUIREMENTS Date requirements Description Data used in this study Precipitation & other Rainfall data from precipitation • Daily precipitation data made available by climate data gauges and observatories in basin. Cambodia’s Ministry of Water Resources Alternately, global remotely sensed and Meteorology (MoWRAM) for a single datasets. precipitation gauge in the basin with coverage 2004-2018. Ideal coverage: at least most • Daily precipitation estimates from the recent 10 years. ERA5 global dataset at 13 points covering the basin has been extracted using Google Earth Engine (this was used for the hydrological model). • Humidity, wind speed and solar radiation for the period 1979-2014 from National Centre for Environmental Prediction Climate Forecast System Reanalysis. Discharge Observed discharge data at least • Discharge data for single downstream recorded daily and based on gauge (Bac Trakuon monitoring station) segment downstream of the area made available by MoWRAM & covering under study. the period from 01-Jan 2007 to 31-Dec 2016. Ideal coverage: at least most recent 10 years. Sediment yield and soil Sediment yield studies, Turbidity • 30 arc-second resolution soil layer erosion data data, Basic data layers for (Harmonized World Soil Database v 1.2) VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A • No local monitoring data available Universal Soil Loss Equation (USLE) model Land cover/ land use Validated land cover map of the • Official land cover for the year 2016 maps and terrain data region under study obtained from the Cambodia Ministry of Environment (MoE) • 30m resolution digital terrain model (SRTM) Irrigation demand Agricultural production maps, • List of irrigation works in the progress and Irrigation requirements, Formally total target irrigated areas for wet and dry Irrigated land, Informally irrigated season paddy cultivation (from MoWRAM). land and Rainfed irrigation land, Rough information of local paddy crop Drought yields. pattern derived from MoWRAM reports Physical inputs to production and Inputs: market price of those inputs • Seeds • Manure • Fertilizers • Herbicide and insecticide 76 Date requirements Description Data used in this study Labor: • Land preparation • Plantation & transportation • Weed control • Crop management • Harvest & post-harvest Services: • Transportation • Irrigation Area of rice under cultivation in the study area Size of a typical rice farm (ha) Irrigation infrastructure Scope, plan and status; Maintenance cost, Damage and repair costs, operation status of infrastructure Hydroelectricity Location of hydropower dams • Location of 3 dams from global datasets (operational plus those under • Some operating parameters for the development) planned hydropower dam (Dam 1) from Ministry of Mines and Energy (MME) For each dam unit: • Electricity price VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A a. Reservoir capacity/operation • Cost of production rules if available • Total electricity generation b. Actual electricity output (not • Soil bulk density capacity) per year • Discounting information c. Active lifetime • Dam lifecycle d. Water use statistics e. Price of electricity f. Any data on sediment dredging dredging cost (if any), incl. dredging frequency g. Operating cost Extreme events Floods and droughts in the basin • Anecdotal, from newspaper articles 77 APPENDIX 2: SWAT PARAMETERS TABLE A2-1: SWAT LAND COVER CODES LC Name PAST Pasture RNGE Range-Grass FRST Mixed Forest FRSD Forest-Deciduous FRSE Forest-evergreen AGRL Agricultural Land-Generic WATR Water WETF Wetland-forested SWRN South Western Range BERM Urban Medium Density Source: https://oldgeni.isnew.info/landuse.html TABLE A2-2: SWAT CALIBRATION PARAMETERS SWAT Parameter Min Max Best Simulation VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A v__ALPHA_BF.gw 0 1 0.743932 v__GW_DELAY.gw 0 500 4.276659 v__GWQMN.gw 0 5000 2950.77832 v__GW_REVAP.gw 0.02 0.2 0.147688 v__REVAPMN.gw 0 500 423.24704 v__RCHRG_DP.gw 0 1 0.180437 v__LAT_TTIME.hru 0 180 105.478706 v__SLSOIL.hru 0 150 95.762405 v__CANMX.hru 0 100 61.229294 v__ESCO.hru 0 1 0.793162 v__CH_N2.rte 0 0.3 0.174933 v__CH_K2.rte 0 50 39.94957 v__CH_N1.sub 0.01 0.3 0.091888 v__CH_K1.sub 0 50 35.175396 78 APPENDIX 3: SCATTER PLOT OF MONTHLY WYLD FROM SUB-BASINS AS DIRECTLY CONTRIBUTING TO MONTHLY DISCHARGE 300.00 250.00 Estimated discharge at downstream gauge by aggregating outflow from locationm of Dam 1, 3 & 5 - and - water yield from remaining intermediate sub baisins 200.00 150.00 100.00 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A 50.00 0.00 0 50 100 150 200 250 300 Discharge at downstream gauge directly from SWAT model (with stream routing) 79 APPENDIX 4: FOREST CONVERSION FOR AGRICULTURE AND CHARCOAL AND FOREST PROTECTION ESTIMATES A spreadsheet model was developed to compare the stream of benefits and costs of forest conversion using net present value (NPV) analysis. NPV was determined to be US$22.1 million for 50 years at 6 percent discount rate. The assumptions and data tables that follow pertain to this. The analysis pertains to a 0.25 percent deforestation scenario which estimates that about 980 hectares of forest in the Pursat RB are converted each year. 0.25% deforestation scenario 2016 forest cover Average annual forest Estimated annual CO2 Forest type (ha) loss (ha)* loss (t CO2e) Evergreen 242,989 601 360,000 Semi-evergreen 53,223 132 67,000 Deciduous 100,903 250 77,000 Total 397,115 983 504,000 * Avg. over 10 years (2016-2025) Assumptions: i. Forestland conversion is done mainly for agriculture: 90 percent of conversion is to agriculture and 10 percent to charcoal. ii. Cost of production is 50 percent lower than lowland paddies. Upland rice farming still follows traditional practices such as slash and burn for ethnic people while some have already adopted advanced technologies VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A including land preparation, fertilization and pest control. Wet season yield - 1.2 t/ ha Source: Var et al. (2016) from MAFF Statistic iii. Rice is grown only in the wet season as it is rainfed. Price is based on average price of Ibis Rice (Jan-Mar 2020) from Cambodia Rice Federation http://www.crf.org.kh/?page=api_location_detail&menu1=592&menu2=1110&menu3=&menu4=&menu5=&id= 1071&lg=en iv. 6.41 kg of wood is needed to produce 1 kg of charcoal (GERES 2015). v. Share of total above ground biomass used for charcoal of 80 percent (GERES 2015). vi. This analysis assumes the same deforestation rate for all forest types. In reality, there may be some differences in the deforestation rates of the different forest types. vii. Average price of charcoal at US$350/t (GERES 2015). Inputs to Charcoal Production Unit Unit Cost (US$) Number of Units Total (US$) Labour Man-days 7 28 196 Fees Kowyun 0.50 50 25 Equipment Kiln 50 2 100 Total T Charcoal 25 321 Source: UNDP 2017 80 Biomass Estimation ABV Biomass Avg annual forest loss ABV Biomass loss Forest type (t/ha) (ha) (t) Evergreen 163 601 97,963 Semi-evergreen 243 132 32,076 Deciduous 85 250 21,250 151,289 Source: Cambodia Forest Reference Level 2016 Charcoal Production Estimates Est. ABV Biomass lost (t/ yr) 151,289 Est. Charcoal produced (t)/ yr 3,776 Est. benefits from Charcoal US$/ yr 1,321,714 Rice Production Costs Wet season US$/ha 50 Wet season US$ 39,320 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A Rice Production Wet Season yield (t/ha) 1.2 Area converted (ha) for Ag 786.4 Rice price (US$/t) 178 Total yield (t) 943.7 Ag, Benefit (US$) 167,975 81 APPENDIX 5: TOURISM AND CARBON NPV ANALYSIS Tourism benefits. Tourism benefits are quantified primarily based on expected changes in visitation data of tourists at project sites, average daily spending and average length of stay over a 50-year time horizon. Visitation data to tourism sites in Pursat province, and that were used where available, averaged over the two-year collection period. Average tourism spending and average stay length were derived from tourism statistics available from Cambodia’s Ministry of Tourism and the Ministry of Environment. The “without project” scenario assumes that visitation continues at an estimated 9 percent annually over the project period, based on historical arrivals data at ecotourism sites nationally, over the preceding seven years. It assumes that spending per tourist remains constant (in real prices) in the absence of any infrastructure and capacity improvements (currently US$37.88 per day). The average length of stay continues to increase according to national historical growth rates (0.3 days increase over the project period), from a baseline of 2.5 days. The “with project” conservatively estimates a 3 percent annual visitation growth above the “without project” projections, starting in Year 3 of the project (assuming infrastructure and capacity building spending is effectively implemented). It projects an annual 5 percent increase in overall tourist spending starting in Year 3 and the average length of stay increases by 1.5 days from a baseline of 2.5 days. This analysis is intentionally conservative and expects results beyond the projections considered, given the extremely low daily spending and length of stay baselines. Furthermore, multiplier benefits of the tourists’ spending on local food, transport and lodging and the employment created are not quantified in this analysis as no value chain analysis (VCA) data were available but represent an additional stream of benefits that helps ensure that the overall conclusion is based on conservative assumptions. The number of visitors is capped at 200,000 as anything beyond this is likely to be unsustainable. Emission reduction benefits. The value of emission reductions is estimated based on an updated GHG analysis VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A using data from the Cambodia’s Initial Forest Reference Level under the UNFCCC Framework report and estimations of GHG emissions reductions were modeled using the Food and Agriculture Organization (FAO) EX-ACT model. The analysis follows the “with” and “without project” scenarios to model the reduction in rate of forest loss and degradation. To place a value on the benefits of reducing carbon emissions a conservative value of US$5 t/ CO2e was used. 82 83 VA L U I N G T H E E CO SYS T E M S E RV I C E S P R OV I D E D BY F O R E S TS I N P U R SAT BA S I N , C A M B O D I A