A NATIONAL ASSESSMENT OF ECOSYSTEM SERVICES IN ZIMBABWE AND THE POTENTIAL BENEFITS OF LARGE-SCALE LAND RESTORATION JANUARY 2024 PREFACE AND ACKNOWLEDGMENTS This study was funded by the World Bank and undertaken by Anchor Environmental Consultants, South Africa. The work will support the formulation of Zimbabwe’s County Climate and Development Report (CCDR). The report was prepared by Jane Turpie, Luke Wilson and Gwyn Letley. The team is grateful for the reviews and technical direction provided by Dominick Revell de Waal and Gibson Guvheya of the World Bank. Citation: Turpie, J.K., Wilson, L. & Letley, G. 2024. A national assessment of ecosystem services in Zimbabwe and the potential benefits of large-scale land restoration. Report prepared by Anchor Environmental Consultants for the World Bank. Cover image: Luke Wilson TABLE OF CONTENTS ACRONYMS & ABBREVIATIONS I GLOSSARY OF KEY TERMS II EXECUTIVE SUMMARY IV 1. INTRODUCTION  1 2. BIOPHYSICAL AND SOCIO-ECONOMIC CONTEXT 3 ZIMBABWE'S LANDSCAPES, ECOSYSTEMS, & BIODIVERSITY 3 ADMINISTRATIVE AREAS, LAND TENURE, & PROTECTED AREAS 6 POPULATION & RURAL LIVELIHOODS 8 WATER SUPPLY 9 3. BENCHMARKING OF LAND CONDITION  10 NET PRIMARY PRODUCTIVITY 10 NDVI TRENDS  12 CARBON STORAGE 13 4. BASELINE ASSESSMENT OF ECOSYSTEMS AND ECOSYSTEM SERVICES14 OVERVIEW14 BASELINE LAND COVER AND ECOSYSTEM CONDITION 15 PROVISIONING SERVICES 20 CULTURAL SERVICES 34 REGULATING SERVICES 38 SUMMARY OF ECOSYSTEM VALUES AND THEIR BENEFICIARIES 51 5. SCENARIO ANALYSIS  53 OVERVIEW53 DESIGN AND GENERATION OF TWO FUTURE SCENARIOS 54 IMPACTS ON ECOSYSTEM SERVICES 60 COST-BENEFIT ANALYSIS OF THE RESILIENT FUTURE SCENARIO  72 6. CONCLUSIONS 74 REFERENCES75 ACRONYMS & ABBREVIATIONS AET Actual Evapotranspiration MAP Mean Annual Precipitation AGB Aboveground Biomass MAT Mean Annual Temperature Ministry of Environment, Climate, AMD Acid Mine Drainage MECTHI Tourism and Hospitality Industry ANR Assisted Natural Regeneration Mg Megagrams (1000kg or metric tonne) Ministry of Lands, Agriculture, Fisher- ASA Advisory Services and Analytics MLAFWRR ies, Water and Rural Resettlement Normalised Difference Vegetation ASCC Annualised Social Cost of Carbon NDVI Index BAU Business-as-Usual NPP Net Primary Productivity BGB Belowground Biomass NPV Net Present Value CA Conservation Agriculture PES Payments for Ecosystem Services CAMP- Community Areas Management PUD Photo User Day FIRE Programme for Indigenous Resources CBD Convention of Biodiversity PV Present Value CSA Climate-Smart Agriculture ROI Return on Investment Climate-Smart Agriculture Investment CSAIP RUSLE Revised Universal Soil Loss Equation Plan Community-Based Natural Southern African Development CBNRM SADC Resources Management Community CN Curve Number SCC Social Cost of Carbon DEM Digital Elevation Model SDGs Sustainable Development Goals EMA Environmental Management Agency SDR Sediment Delivery Ratio ESA European Space Agency SWY Seasonal Water Yield FAO Food and Agriculture Organization Tg Teragrams (millions of metric tonnes) FSR Future Suitability Ratio TLU Tropical Livestock Unit FTLRP Fast-Track Land Reform Programme UN United Nations GCF Green Climate Fund UNICEF United Nations Children's Fund United Nations Convention to GDP Gross Domestic Product UNCCD Combat Desertification United Nations Framework GEF Global Environment Facility UNFCCC Convention on Climate Change GLW3 Gridded Livestock of the World WDPA World Database on Protected Areas GoZ Government of Zimbabwe WFP World Food Programme GRanD Global Reservoir and Dam Database WWF World Wide Fund for Nature IAP Invasive Alien Plant ZINWA Zimbabwe National Water Authority Integrated Valuation of Ecosystem Zimbabwe Parks and Wildlife InVEST ZPWMA Service Tradeoffs Management Authority LAI Leaf Area Index ZTA Zimbabwe Tourism Authority LDN Land Degradation Neutrality I GLOSSARY OF KEY TERMS The variability among living organisms and the ecological complexes of Biodiversity which they are part. This includes variation within species, the diversity of species within ecosystems, and the diversity of ecosystem types in nature. Carbon Sequestration The process of capturing and storing atmospheric carbon dioxide. An area where water is collected by the natural landscape. Precipitation that falls in a catchment runs downhill into creeks, rivers, lakes, oceans, or into Catchment built infrastructure, such as reservoirs. In this document, the terms catch- ment and watershed are used interchangeably. Climate-Smart A broad term for reforming agricultural practices to achieve a more produc- Agriculture tive, resilient, and low-emission agricultural sector. A farming system that promotes minimum soil disturbance, maintenance of Conservation Agriculture permanent soil cover, and diversification of plant species. A conceptual framework and tool used to evaluate the viability and desirabili- Cost-Benefit Analysis ty of projects or policies based on their costs and benefits over time. The interest rate used in discounted cash flow analysis to determine the pres- Discount Rate ent value of future cash flows. The benefits people obtain from the earth’s many life-support systems. The Ecosystem Services Millennium Ecosystem Assessment defines four categories of ecosystem ser- vices: provisioning, regulating, cultural, and supporting services. Water added to an aquifer through the unsaturated zone after infiltration Groundwater Recharge and percolation following any storm rainfall event. The reduction or loss in biological or economic productive capacity of the Land Degradation land resource base. A state whereby the amount and quality of land resources necessary to Land Degradation support ecosystem functions and services remain stable or increase within Neutrality (LDN) specified temporal and spatial scales and ecosystems. Normalised Difference A widely used metric for estimating the health and density of vegetation Vegetation Index (NDVI) based on the analysis of remote sensing or aerial imagery. The net long-term rate of accumulation of carbon in ecosystems, which can Net Ecosystem Carbon be approximated as NEP (see below) less of carbon losses from disturbances Balance (NECB) (e.g. fire, land cover change, grazing), harvesting and leaching of dissolved organic carbon into groundwater and rivers. A measure of the net effect of carbon taken up by ecosystems through pho- Net Ecosystem tosynthesis, and the return of carbon to the atmosphere through plant and Productivity (NEP) soil respiration. Net Primary Productivity A measure of the amount of carbon taken up by plants through photosyn- (NPP) thesis, net of carbon returned to the atmosphere through plant respiration. A calculation used to estimate the net benefit over the lifetime of a particular project. Net present value allows decision-makers to compare various alterna- Net Present Value (NPV) tives on a similar time scale by converting all options to current dollar figures. A project is deemed acceptable if the net present value is positive over the expected lifetime of the project. II A scheme where beneficiaries of ecosystem services compensate ecosystem managers (landowners or resource stewards) to change their practices, to Payments for Ecosystem secure those ecosystem services. This may involve desisting from damaging Services (PES) activities or adopting more expensive practices that are less damaging to the environment. A simple ratio of the gain from an investment relative to the amount invest- Return on Investment ed. ROI is calculated by dividing net profit (current value of investment − cost (ROI) of investment) by the cost of investment. Land occurring along watercourses and water bodies. For this study, it can be Riparian Buffer defined as the area within 30 m of the river channel. Managing the use and protection of natural resources in a way (or at a rate) Sustainable Resource which enables social, economic, and cultural well-being while ensuring these Management resources are sustained for future generations and any adverse effects on the environment are minimized. III EXECUTIVE SUMMARY INTRODUCTION This study was commissioned to inform Zimbabwe’s Country Climate and Development Report (CCDR), specifically the “deep dive” on “Resilient agriculture, landscape restoration, food and water security”. It describes the baseline and compares a future (2050) Business-as-Usual (BAU) Scenario with a future Resilient Scenario in which Vision 2030 is extended to include a range of landscape interventions designed to secure and enhance ecosystem services from natural, semi-natural and cultivated areas throughout Zimbabwe. The aim of the study was to undertake a national-scale assessment of the nature and value of ecosystem services, and to estimate how these would change under the Resilient relative to the BAU Scenario. The study also develops a baseline land cover layer incorporating vegetation types and condition and provides some benchmarking against other countries in the region. Ecosystem services considered included provisioning, regulating and cultural services. Provisioning services assessed were the ecosystem contributions to cultivated production, livestock production and harvested natural resources. The regulating services assessed were global climate regulation through sequestration and storage of carbon, water flow regulation and groundwater recharge and sediment retention. The cultural service considered was ecosystem contribution to nature-based tourism. CONTEXT Zimbabwe is a landlocked subtropical country that is richly endowed with biodiversity in a variety of arid to mesic ecological zones that are dominated by woodland and shrubland vegetation. Closed forests, grasslands and wetlands also contribute significantly to its ecological wealth. However, this has diminished considerably since 2000 due to land reform and subsequent policies, which saw a reduction in private conservation and in commercial agriculture output. Land tenure remains insecure and uncertain in many areas. The majority of Zimbabwe’s approximately 15 million inhabitants are young and depend on land and natural resources for their livelihoods. Population densities are mostly concentrated in the higher rainfall central and north-eastern parts of the country, broadly coinciding with the miombo woodland biome which is a rich source of provisioning services. Most of these households depend on groundwater, but there is also a large network of dams that supply urban areas, irrigation schemes and hydro-electric power, and which depend on healthy catchment areas for their operation. BENCHMARKING Measures of vegetative productivity (Net Primary Productivity and NDVI) and carbon retention were compared with four East and Southern African countries - Kenya, Malawi, Tanzania and Zambia.1 Net primary productivity (NPP) is a measure of plant biomass growth, expressed as the amount of carbon captured by plants through photosynthesis over a given period. Annual NPP data from 2001-2022 were obtained from MODIS for this analysis. NPP in Zimbabwe was found to be most similar to Malawi. Given Malawi’s higher rainfall, this suggests that overall ecosystem health in Zimbabwe is higher than in Malawi. Zambia and Tanzania also both have higher mean rainfall then Zimbabwe, and mean NPP is slightly higher in both these countries as would be expected. Kenya has markedly lower mean NPP than all other countries in the comparison. This 1  Further detail on the methods used for the benchmarking analysis is provided in Chapter 3 of the report. IV ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE is due at least in part to the extensive arid areas in the north and east of the country. Zimbabwe had more marked interannual oscillations in NPP than any of the other countries, which could indicate high vulnerability of the country to interannual rainfall variability given the strong links between rainfall and NPP. However, Zimbabwe was also one of the few countries that exhibited a general increase in NPP from 2001 to 2022. Zimbabwe also has a relatively high NPP/capita, second only to Zambia where this is twice as high as Zimbabwe. Given that forest and woodland area in Zimbabwe is declining, the positive NPP trend may seem somewhat surprising. A possible explanation is that there has been some passive recolonisation of grassland and shrubland on commercial farmland parcels that have fallen into disuse following the land reform programme of the early 2000s. Analysis of remotely sensed fire data also indicates there has been a significant reduction in burning across Zimbabwe since 2012, which may also have contributed to increasing NPP. NDVI provides a measure of vegetation greenness and is another useful metric for monitoring changes in vegetation health over time. Trends in NDVI between 2001 and 2021 were analysed using Trends.Earth, which corrects for the influence of rainfall variability. Zimbabwe and Zambia had the lowest proportion of land that exhibited a statistically significant decline in NDVI from 2001 to 2021. As with the NPP measures, Zambia had the highest value of aboveground carbon storage of 25.8 tC/ha, followed by Tanzania, while Zimbabwe had an average of 14.4 tC/ha. BASELINE ASSESSMENT The baseline assessment involved mapping the extent and condition of natural and human- made ecosystems and quantifying the flows of the above-mentioned ecosystem services for the year 2019. Modeling of ecosystem services requires spatial information on the extent of different ecosystem types in the landscape, and ideally an indication of the condition of these ecosystems. This was assessed on the basis of global land cover data and the ecoregions map for Zimbabwe (generating 20 land cover types), and satellite-derived Normalised Difference Vegetation Index (NDVI) data (to assess condition). Natural land cover classes cover 74% of Zimbabwe with the bulk of remaining land cover (24%) being farmland. Following the statistical analysis of NDVI data, pixels with values below the expected range of NDVI for specific ecosystem types and rainfall zones were identified as degraded and were parameterised accordingly in the models. At the catchment level, the coverage of degraded natural habitats was estimated to be highest in the Mzingwane Catchment (15% of total natural area) and lowest in the Mazowe Catchment (9% of total natural area). Further analysis of degradation levels was conducted at a sub-catchment level to better highlight degradation hotspots. These included farming areas north and west of Masvingo, the dry western half of the Mzingwane Catchment, and communal farming areas to the southwest of Gokwe where significant expansion of farmland into woodland has occurred. We assessed the ecosystem provisioning services relating to the 11 main crops, livestock and wild harvested natural resources in terms of resource rent. The contribution to crop production was estimated to be US$365.4 million/y, or US$39/ha of farmland, with 34% of this being for food crop production, while the contribution to livestock production was estimated to be US$291.7 million/y. Household harvesting of wild natural resources such as wood, thatch and wild foods was estimated to be worth at least US$576 million/y, or US$20/ha of natural land on average. Values per unit area were highest in the northeast where rural populations are highest. Zimbabwe’s natural ecosystems also make a substantial contribution to tourism. Based on tourism statistics, attraction-based tourism accounted for 65% of tourism expenditure, and based on the location of tourism photographic activity, nature-based tourism amounted to 81% of attraction-based tourism, thus almost 53% of all tourism expenditure. The ecosystem contribution, valued as resource rent from nature-based tourism, was estimated to be US$229.7 million in 2019. More than half this value came from the Gwayi catchment area, which includes Victoria Falls and Hwange National Park. Regulating services were assessed in terms of the avoided costs associated with carbon sequestration and storage, flow regulation and sediment retention. Avoided costs are the costs V Executive Summary that would be incurred if ecosystems were not performing these services. At a national level, it was estimated that ecosystems currently sequester 90.6 million tC/y (an average of 2.3 tC/ ha/year), and retain 808 million tC (equivalent to 2966 million tonnes of CO2) in above and belowground vegetation biomass (an average of 20.6 tC/ha/year). A much greater amount of carbon is stored in the country’s soils but was not included in the stock estimate as this has not been accurately mapped or quantified. These active and passive services avoid global damages worth some US$905 million/y and US$8.5 billion/y, respectively. The latter figure would be much be much larger if the soil carbon pool was also included. The mediation of groundwater recharge by current vegetation cover avoids surface water supply costs of an estimated US$496 million/y. Vegetation cover also prevents the transport of some 167 million t of sediment per year into dam catchment areas, saving on water storage replacement costs of about US$208 million/y. In summary, the total value of ecosystem services provided by baseline landscapes was estimated to be in the order of US$13 billion/y, with over US$2.17 billion of benefits accruing annually to Zimbabwe and the balance being climate regulation services benefiting the rest of the world. THE BUSINESS-AS-USUAL SCENARIO The BAU Scenario estimated future changes in the delivery of services based on the extrapolation of past changes in ecosystem extent and condition to 2050. The analysis also incorporates available information on the impacts of future climate change on key services such as crop production. Changes in land cover and condition were projected using the InVEST Scenario Generator model, under the assumption of a continuation of past trends. The resulting estimates were that the cultivated area would increase from 24% coverage to 30%, built-up areas would triple in size to 0.6% of area, and that natural land cover classes would decline from 74% to 68% of national area. Levels of degradation of natural areas would be expected to double. These changes would be contrary to Zimbabwe’s Land Degradation Neutrality (LDN) commitments, which include an ambitious goal of reforestation of 6.4 million ha of degraded forest (16.5% of national area) and the return of 215 000 ha of cropland back to forest by 2030. The projected changes highlight that significant effort will be required to reverse the BAU trajectory if Zimbabwe is to achieve its aspirational LDN commitments. While ecosystems still provide immense value to Zimbabwe under the baseline scenario, this value has been compromised by ecosystem degradation and loss. Further declines in the capacity of ecosystems to provide services are projected under the BAU scenario, as a result of the further conversion and degradation of natural habitats, in conjunction with projected climate change impacts. This includes a reduction in the availability of wild harvested resources, declining agricultural land productivity, and a reduction in carbon storage and sequestration with the loss and degradation of high biomass natural habitats. Similarly, soil erosion and export of sediments to watercourses and dams are projected to increase, while increased runoff at the expense of infiltration will reduce groundwater recharge rates. In spite of a degree of biodiversity loss, nature-based tourism would still grow due to projected tourism growth for Zimbabwe as well as the establishment of conservancies in previously unprotected areas. A RESILIENT FUTURE SCENARIO Potential landscape interventions to restore, maintain, or enhance the flow of ecosystem services from natural and cultivated lands were explored. Potentially suitable interventions that could be applied across the country’s various socio-ecological contexts were identified, and the impact that these could have on ecosystem condition and services were then estimated. The selected interventions were guided by and are consistent with Zimbabwe’s various environmental commitments as articulated in the Climate-Smart Agricultural Investment Plan (CSAIP), Revised Nationally Determined Contributions (NDCs), National Adaptation Plan (NAP), Vision 2030 and Land Degradation Neutrality targets. VI ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE A range of agricultural and natural habitat restoration interventions were proposed for the study area: climate-smart agriculture (CSA), the restoration and protection of riparian buffers, the passive restoration of degraded natural habitats through better rangeland management or wildlife-based land use and sustainable resource use. These interventions will meet several of Zimbabwe’s LDN commitments, including forest restoration, implementation of conservation agriculture and the enforcement of appropriate stocking rates. It was estimated that, relative to the BAU Scenario: ∙ CSA could increase crop production revenues by US$154.8 million/y; ∙ Restoration of riparian buffers and degraded natural habitats could increase the value of wild resource harvesting by US$13.6 million/y; ∙ Increased carbon storage and sequestration could reduce local climate change-related damages by US$33.5 million/y and damages to the rest of the world by US$1.64 billion/y, compared to the BAU scenario. The potential revenue that could be generated through carbon credits was estimated to be US$55.9 million/y. ∙ The recovery of riparian buffers and degraded natural habitats could generate additional climate regulation benefits worth US$1.73 billion/y, including US$89.4 million of benefits to Zimbabwe; ∙ The expansion of conservancies and the improvement of management and tourism facilities within the state protected area estate could generate further increased nature- based tourism benefits of US$51.9 million/y; ∙ Collectively, the interventions could result in an increase in groundwater recharge benefits of around US$124.5 million/y and avoided reservoir sedimentation costs of US$11.2 million/y; Overall, the benefits of following a resilient development pathway could to some US$445.4 million/y locally and US$2.09 billion/y globally, when compared to the BAU trajectory. The total benefits are highest in the Save Catchment, where the Resilient Scenario yields ecosystem service gains worth US$118.5 million/y locally, relative to BAU. When the global value of avoided climate change-related damages is included, the benefits of interventions in the Save Catchment increase to US$389 million/y. To further evaluate the economic rationale for the Resilient Scenario interventions, a high-level cost benefit analysis was conducted. To do so, establishment and maintenance costs were estimated for the various interventions and expressed in present value terms, using a 25-year time horizon and social discount rate of 4.56%. Ecosystem service benefits were also expressed in present value terms using the same time period and discount rate. Additionally, it was assumed that the full ecosystem service benefits would be realised gradually over time, varying from five years in the case of crop production benefits from CSA, to 25 years for the full realisation of carbon and wild resource harvesting benefits, reflecting the slower recovery and restoration of natural habitats. Overall, ecosystem service benefits from the Resilient Scenario over the next 25 years were estimated to have a present value of US$5.17 billion to Zimbabwe, relative to the BAU Scenario. The total present value cost of the Resilient Scenario interventions was estimated to be US$1.59 billion. This results in a positive net present value (NPV) of US$3.59 billion and a high ROI of US$3.3 for every dollar invested in the Resilient Scenario interventions. If the global value of avoided climate damages to the world is included in the analysis, the ROI increases to 9.4. These results highlight that well-implemented restoration and conservation interventions could generate benefits that significantly outweigh their costs. This assumption holds across each catchment, with local ROI ranging from 1.5 in the Runde Catchment to as high as 4.7 in the Save Catchment. In addition, the Save, Manyame and Gwayi Catchments were all estimated to have a high ROI of more than 3.5. Furthermore, the high global benefits of the resilient scenario lend credence to Zimbabwe's request for external financing of its climate change adaptation and mitigation investments, which could be used to help fund the interventions proposed in this study. VII Executive Summary VIII 1. INTRODUCTION A Country Climate and Development Report water security”. The deep dive is informed by (CCDR) is a diagnostic that aims to provide two policy scenarios and two climate scenarios options for how a country can integrate – a Business-as-Usual (BAU) Scenario in their development objectives with climate which historical trends are projected into action needs and commitments. Zimbabwe’s the future, and a Resilient Scenario in which CCDR will present deep dives on three Vision 2030 is extended to include a range thematic areas in light of the government’s of landscape interventions designed to Vision 2030. The objective of this report is to secure and enhance ecosystem services support the formulation of Zimbabwe’s CCDR from natural, semi-natural and cultivated by informing the deep dive on “Resilient areas throughout Zimbabwe. The outcomes agriculture, landscape restoration, food and are compared over the period to 2050. 1 Introduction This study draws on available information software to create the 2050 BAU Scenario. from government reports, published literature and remote sensing data, supported through For the Resilient Scenario, we estimate the use of geographic information systems land cover changes based on the relevant (GIS) software and the Integrated Valuation interventions planned under Zimbabwe’s of Ecosystem Services and Tradeoffs (InVEST) Vision 2030 National Development Strategy modelling tool. The study builds on earlier 1 (NDS1) plus additional landscape-scale work conducted on Zimbabwe’s Biodiversity interventions, all of which are implemented Economy (Turpie et al., 2022b) and the by 2030, and the scenario is projected to Mazowe Catchment (Turpie et al., 2022a). 2050. A number of contextual changes are also expected under the Resilient scenario The study had three objectives: relative to the BAU that also affect the value of ecosystem services, such as changes in the rate ∙ To estimate the current value of of urbanisation, improvements in electricity ecosystem services including selected supply and improvements in sectoral outputs benchmarks against regional peers; such as mining due to better governance. ∙ To estimate the value of the ecosystem services lost or gained under the two We then estimate how each of the ecosystem policy scenarios (BAU and Resilient); and services described above changes under the adjusted land covers for the BAU and Resilient ∙ To set out a prioritised list of interventions scenarios, in physical and monetary value to ameliorate the combined impact of terms, or in qualitative terms where appropriate. development and climate change. In the third section, we propose a suite of In the first section, we describe changes in priority measures to ameliorate the impact habitat quality and net primary productivity, climate change under the two development and generate national and catchment-level scenarios to off-set the potential losses in estimates of the following in physical (where ecosystem services. This is based on a clear appropriate) and monetary terms (net of description of (a) what is already being human inputs where appropriate) for: implemented (BAU), (b) what is planned ∙ Harvested wild resources; (Vision 2030) and (c) what additional ∙ Cultivated production; measures could further support Zimbabwe ∙ Livestock production; adapt to climate change and pursue a low- ∙ Sediment regulation; carbon development pathway. The additional ∙ Flow regulation: measures focus on nature-based solutions ∙ Tourism; and for mitigating and adapting to climate ∙ Carbon retention. change and enhancing ecosystem health, such as agricultural, rangeland and forest In developing these estimates, we cover interventions. The choice and location compare data for NPP, carbon retention of adaptation measures incorporates our and other readily available benchmarking understanding of the nature of climate change data for Zimbabwe with data for damages under the two scenarios, including Zambia, Tanzania, Malawi and Kenya. future projections for reduced crop outputs In the second section, we generate land cover Finally, we review the costs and potential and productivity projections under the two efficacy of potential interventions in different policy scenarios – BAU and Resilient. For the landscape/land tenure contexts and use BAU scenario, we analyse changes in land this to prioritise a suite of key interventions. cover and ecosystem productivity over the past We then estimate their impacts on the two decades and project these trends up to flows and values of ecosystem services and 2050. The changes in land cover are analysed calculate return on investment in each in GIS using available data for 1992 to 2019, and catchment, separately and in combination, taking into account recent projections of forest so as to provide clear guidance on additional cover changes for Zimbabwe. Ecosystem interventions to be undertaken in Zimbabwe. condition is estimated using satellite imagery- derived Normalised Difference Vegetation Index (NDVI) data. We project the trends in land cover and condition using InVEST 2 2. BIOPHYSICAL AND SOCIO-ECONOMIC CONTEXT This section provides an overview of the contextual backdrop to the study, helping biophysical characteristics of Zimbabwe, to understand the relationships between the land tenure, population characteristics, Zimbabwe’s people and its natural assets, livelihoods, and economy. This provides the and the current state of the environment. KEY POINTS ∙ Zimbabwe is a landlocked subtropical country that is richly endowed with biodiversity in a variety of arid to mesic ecological zones that are dominated by woodland and shrubland vegetation, and also include closed forests, grasslands and wetlands. ∙ Biodiversity has diminished considerably since 2000 due to land reform and subsequent policies, which saw a reduction in private conservation and in commercial agriculture output. Land tenure remains insecure and uncertain in many areas. ∙ The majority of Zimbabwe’s approximately 15 million inhabitants are young and depend on land and natural resources for their livelihoods. They are concentrated in the central and north-eastern parts which are more suitable for cultivation. ∙ Most rural households across the country depend on groundwater as their main drinking source, with 83% of households depending on water from wells and boreholes, but there is also a large network of dams that supply urban areas, irrigation schemes and hydro-electric power. ZIMBABWE'S LANDSCAPES, ECOSYSTEMS, & BIODIVERSITY Zimbabwe is a land-locked country, much of the south to over 1000 mm in the mountainous which is situated at a relatively high elevation Eastern Highlands (Mugandani et al., 2012; (160 to 2592 m; (GoZ, 2019). Its prominent World Bank, 2021). Rainfall is generally higher topographical feature is a central watershed in the north and east of Zimbabwe (>700 mm) which bisects the country along a southwest and decreases towards the south and west. to northeast axis (Figure 1). This and other areas above 1200 m make up the “highveld”, a gently The climate is seasonal, with a wet season undulating landscape with granite domes from November to March, and a long dry (dwalas) that are a characteristic feature. The season from April to October (Mugandani et central watershed divides the Zambezi River al., 2012). Temperatures are highest in October basin to the north and the Save and Limpopo at the end of the dry season, and lowest in River basins to the south (Moore et al., 2009). June-July (Mugandani et al., 2012). There is also high inter-annual rainfall variability, Despite its tropical location, much of the with rainfall becoming increasing unreliable country experiences moderate temperatures towards the south (Unganai & Mason, 2001). due to its altitude, with cooler temperatures associated with higher altitudes. The higher- A map of aridity index values is shown in lying central and eastern areas have a Figure 2. This is derived by dividing mean subtropical to temperate climate (Mugandani annual rainfall by potential evapotranspiration et al., 2012), with mean annual temperatures and provides an indication of water availability of 16 – 20°C, while mean temperatures in for vegetation growth (Zomer, Xu & Trabucco, the lower-lying Zambezi, Limpopo and 2022). This highlights that much of the Save River basins rise to 23°C or more. country has semi-arid conditions, with an aridity index of under 0.5. Conditions for crop Mean annual rainfall varies from 300 mm in growth are most favourable in the north 3 Ch 2: Context Figure 1. Topographic map of Zimbabwe with rivers and names of the seven major river catchments. Derived from 90 m Shuttle Radar Topography Mission (SRTM) Digital Elevation Model (DEM) and eastern parts of Zimbabwe’s central 1988). Most remaining forest is montane and watershed, where conditions are relatively sub-montane forest, with very little of the cool and wet. Conversely, the south and west original lowland forest left (Müller, 2006). of the country have low aridity index values with very low potential for rainfed agriculture. In the drier south-west, miombo woodland grades into tree and shrub savanna, The topography, soils and climate collectively dominated by Senegalia/Vachellia and determine the natural distribution of Terminalia, and which is shorter and more vegetation and of biodiversity in general. open (Whitlow, 1988). The Kalahari sands in the Vegetation is dominated by savanna woodland northwest support Baikiaea (teak) woodland, and grassland. The most widespread woodland which contains a number of prized timber type is miombo, which dominates in the species, including Pterocarpus angolensis wetter central, northern and eastern parts of (Mukwa) and Baikiaea plurijuga (Zambezi the country (Figure 3). Miombo woodlands teak) (Forestry Commission, 2011; WWF, 2016a). provide significant income to Zimbabwean Additionally, Baikiaea woodland resources households from woody and non-woody are widely used by rural communities. The forest products, and was estimated to be the value of forest products harvested from most valuable of the country’s woodland types woodland was estimated to fall between in this regard (World Bank, 2019a). Woodland miombo and mopane woodland (World Bank, vegetation on the central watershed varies 2019b), with the latter estimated to have the from open to closed canopy and is often lowest value of the three woodland types interspersed with seasonally inundated included in the World Bank study. Mopane grassland in lower-lying areas, known locally (Colophospermum mopane) woodland or as vleis or dambos. In the higher rainfall areas, scrub, which is also characterised by the woodland gives way to a mosaic of forest and presence of baobabs (Adansonia digitata), grassland, although natural forest patches are is widespread in the hot, low-lying southern mostly fairly small and confined to sheltered areas and the Zambezi River Valley. It is also slopes and valleys where they are protected worth noting that riparian areas throughout from fire (Forestry Commission, 2011; Whitlow, the country tend to be densely wooded. 4 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE Figure 2. Aridity index map for Zimbabwe, which provides an indication of moisture availability for vegetation growth. Source (Zomer et al., 2022). Figure 3: Simplified vegetation map for Zimbabwe derived from WWF Ecoregions (Olson et al., 2001). 5 Ch 2: Context ADMINISTRATIVE AREAS, LAND TENURE, & PROTECTED AREAS Zimbabwe has ten provinces (Figure 4), Almost 6 million ha of commercial farmland including the two smaller metropolitan were subdivided into small A1 units, while provinces of Bulawayo and the capital city around 3.5 million ha was allocated under Harare. Communal land is the most widespread the A2 scheme (Moyo 2011). The intensity of form of land tenure, accounting for 42% of the production of these A2 farms varies, with country’s land area, though much of this is in some having converted to smaller-scale drier parts of the country where agricultural operations (Scoones et al. 2018). Overall, potential is low (Figure 5; Scoones et al., 2011). large-scale commercial farmland decreased substantially in area from 15.5% of the country Under the Fast-track Land Reform in 1980 to 3.4% in 2010 (Scoones et al. 2011). Programme (FTLRP) of 1999–2009, many of the country’s large-scale commercial farms Land tenure remains insecure and uncertain were subdivided into “A1” family farming units in many of the areas redistributed under the or “A2” commercially-oriented medium-scale FTLRP. For example, A1 households cannot use farming units (±300 ha) (Moyo 2011; Scoones et their land as collateral security, which affects al. 2011; Sukume, Mahofa, and Mutyasira 2022).1 their access to credit (Mugabe et al. 2014). This, combined with capacity constraints, has resulted in generally low crop production (Godwin et al. 2011; Mugabe et al. 2014). 1  A1 farms have individual family farm of 6 ha plus a common livestock grazing area (ZIMSTAT 2019). Figure 4. Provinces and cities of Zimbabwe. 6 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE Figure 5: Land tenure across Zimbabwe. Source: Zimbabwe Forestry Commission, Date unknown. Figure 6. Population density map of Zimbabwe (Source: data derived from www.worldpop.org). 7 Ch 2: Context POPULATION & RURAL LIVELIHOODS Zimbabwe has a population of just over 15 about 28% have cattle and 46% have goats. million (UNFPA, 2021), of whom about two Thus it is unsurprising that the pattern of thirds (68%) are rural (ZIMSTAT, 2017). The rural population density is clearly influenced population is young, with 40% of people being by the areas most suitable for maize (Figure under the age of 15, and has a growth rate 7; Chivasaa, Mutanga & Biradarc, 2019). of 2% (ZIMSTAT, 2017). The spatial variation in population density (Figure 6) is influenced by Nevertheless, rural Zimbabweans are still climate, land tenure and road infrastructure, highly dependent on casual labour and with highest densities in the more mesic formal employment for income (Figure northern and eastern areas. These factors also 8). About three-quarters of Zimbabwe’s influence livelihood patterns across Zimbabwe. rural population receives some sort of social support, including from relatives Some 80% of rural households grow maize, which are the main source of cash income. Figure 7. Relative suitability for growing maize in Zimbabwe. Source: Chivasaa et al., (2019) 8 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE Figure 8. Most important sources of income in Zimbabwe. Source: ZimVAC (2020) WATER SUPPLY Both surface and groundwater sources are capacity significantly, with an estimated loss used across Zimbabwe. While the country has in capacity of 29 million m3 by 2003. This figure no sizeable natural lakes, it does have a dense is likely to have increased significantly since. network of at around 14 000 dams, of which 260 are classed as large dams (Murwira et al. Groundwater is also important for many 2014; FAO, 2016). This includes Lake Kariba of the country’s residents. Most rural along the country’s northwestern border, households across the country depend on the largest artificial reservoir in the world by groundwater as their main drinking water volume. However, Lake Kariba is primarily source, with 83% of households dependent used for hydropower generation rather than on boreholes and wells (ZIMSTAT and UNICEF, water supply. Zimbabwe’s internal dams 2019). Even in urban areas, a sizeable 31% supply much of the water used by irrigated of households rely on wells and springs as agriculture and urban water consumers. their main sources of water (ZIMSTAT and Including the Zimbabwean portion of Lake UNICEF, 2019), while groundwater is still Kariba, it is estimated the country has a total an important source of water for irrigation, dam capacity of 99 930 million m3 (FAO, mining, and tourism (Davis and Hirji 2014). 2016). However, siltation reduces dam storage 9 3. BENCHMARKING OF LAND CONDITION Various metrics can be used as a proxy for varied over time between 2001 and 20221. gauging the health and ability of ecosystems to NPP measures the net gain in plant biomass deliver services. In this study, a benchmarking through photosynthesis. It is expressed as the analysis was done comparing measures of net mass of carbon produced per unit area and primary productivity (NPP), the normalised time (e.g. kgC/m2/y). In contrast, NDVI is an difference vegetation index (NDVI) and index which quantifies vegetation greenness, carbon retention between Zimbabwe and and is useful for measuring changes in four East and Southern African peers (Kenya, vegetation health over time and space. Malawi, Tanzania and Zambia). This includes an assessment of how NPP and NDVI have 1  The NDVI trend analysis ran from 2001 to 2021, as 2022 data were not available through Trends.Earth at the time of writing. NET PRIMARY PRODUCTIVITY Data on annual NPP from 2001 to 2022 were downwards were evident for Kenya, Malawi and obtained from the 500 m MODIS Net Primary Tanzania. However, mean NPP in Zimbabwe Productivity Gap-Filled product, accessed via does appear to show a general increasing Google Earth Engine. This layer is generated trend from 2001 to 2022, which is suggestive from the sum of daily gross productivity data of improved ecosystem productivity. Zambia for a given year, less of carbon released during showed a similar trend of increasing NPP plant respiration. Additionally, average NPP between 2001 and 2022, though the increase per capita was calculated for each country is less pronounced than for Zimbabwe. by calculating the mean NPP between 2015 and 2022 and dividing this by each country’s The apparent increase in NPP in Zimbabwe population. A shorter time window was used may be somewhat unexpected, given that the for the per capita NPP analysis since the area of forest and woodland has declined over populations of all countries in the sample have time. One possible explanation is that there increased significantly over time. As such, it is has been some reversion of vegetation to less meaningful to relate NPP from the earlier grass and shrubland on commercial farmland years in the analysis to current population levels. parcels that have fallen into disuse following the FTLRP in the early 2000s. Indeed, a The mean annual NPP for Zimbabwe and tendency by some resettled farmers to leave the four comparison countries from 2001 to land fallow has been identified as one of 2022 is shown in Figure 9. These data suggest the main reasons behind low agricultural NPP in Zimbabwe is most similar to Malawi, productivity since the FTLRP, prompting the particularly in more recent years. Given government to start crafting a “use it or lose it” that Malawi has higher mean rainfall than policy in 2015 (Mpofu, 2018). Another possible Zimbabwe, this suggests overall ecosystem explanation is the significant decline in burned health in Zimbabwe is higher than in Malawi. area, as revealed by a recent analysis of MODIS Zambia and Tanzania also both have higher fire data (Shekede et al., 2024). This could mean rainfall then Zimbabwe, and mean NPP have facilitated increased uptake of carbon is slightly higher in both these countries as by recovering woody ecosystems that have would be expected. Kenya has markedly lower benefited from a reduction in fire frequency. mean NPP than all other countries in the comparison. This is due at least in part to the Another notable pattern is that the high and extensive arid areas in the north and east of low NPP years largely parallel each other in the country, where rainfall is much lower than Zambia and Zimbabwe. This is likely indicative in any of the other countries in the assessment. of the influence of broadly comparable rainfall variability patterns in these two All five countries exhibited a degree of neighbouring countries. Similarly, Tanzania variability in annual NPP over the assessment and Kenya show very comparable temporal period (Figure 9). No clear trends upwards or patterns in NPP over the assessment period, 10 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE even though mean NPP is much higher somewhat in the crop production data used in in Tanzania in absolute terms. Again, this this study. According to Zimbabwe’s Crop and suggests a similar rainfall signature in the NPP Livestock Assessment reports, crop production data for these two neighbouring countries. in both the 2018/2019 and 2019/2020 rainy seasons was low due to poor rainfall. In contrast, It is also notable that Zimbabwe had the most crop production was markedly higher in marked inter-annual variability in mean NPP 2020/2021 and 2021/2022. For example, maize (Figure 9). This could be indicative of high rainfall production in the 2018/2019 and 2019/2020 variability and resulting impacts on range and growing seasons was 770 000 t and 908 000 t, farmland productivity, which is borne out respectively (MoLAWFRR, 2020). This increased Figure 9. Comparison of mean annual NPP between 2001 and 2022. Figure 10. Comparison of mean annual NPP (between 2015 and 2022) expressed in per capita terms. 11 Ch 3: Benchmarking Land Condition dramatically to 2.72 million t in 2020/2021 10). Zambia’s high NPP per capita value can and 1.56 million t in 2021/2022 (MoLAWFRR, be explained by the fact that it is a relatively 2022). These trends broadly align with the high rainfall country with a low population lower mean NPP for Zimbabwe between 2018 density. For example, Zambia’s population and 2020, relative to 2021 and 2022 (Figure density is around 25 people/km2, compared 9). In conjunction with the crop production to around 40 people/km2 in Zimbabwe (the data, these data underscore the high second most sparsely populated country in vulnerability of Zimbabwe to rainfall variability. the comparison). In contrast, while Malawi also receives relatively high rainfall, it has a When NPP is estimated per capita, Zimbabwe much higher population density of around was found to have a relatively high value, 160 people/km2. This results in greater second only to Zambia where NPP per capita pressure on ecosystems and agricultural is almost twice as high as Zimbabwe (Figure land, and in turn lower NPP per capita. NDVI TRENDS To provide an additional indicator of changes period (both 11.1%). However, Zambia had an in land health and productivity over time increase in NDVI over a much larger proportion across the five countries, the Trends.Earth of its land area than Zimbabwe did. The fact analysis of NDVI trends in Zimbabwe between that areas exhibiting an increase in NDVI 2001 and 2021 was extended to the other significantly exceeded areas with declining countries in the sample. This shows which NDVI in Zambia and Zimbabwe is also in line parts of each country have experienced with the general increase in NPP in these a statistically significant change in land two countries over the assessment period, productivity (using NDVI as the proxy), after as described above. While satellite imagery accounting for the effects of rainfall variability. limitations mean that the analysis cannot be extended back to assess degradation The results of the NDVI trend analysis are shown prior to 2001, these results do suggest that in Figure 11, which depicts the percentage of degradation rates in Zimbabwe over the last land area in each country that exhibited an two decades are relatively low compared to its increase, decrease or stable NDVI between regional peers. As with the positive NPP trend, 2001 and 2021. Zimbabwe and Zambia had this finding may be somewhat surprising the lowest proportion of land that exhibited a given the general perception of ongoing statically significant decline in NDVI over this deforestation and ecosystem degradation in Figure 11. Percentage of each country’s area with declining, increasing or stable NDVI f rom 2001 to 2021. 12 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE Zimbabwe. Again, the recovery of NDVI may 26.4% of the country’s land area had a reflect the passive regrowth of grassland statistically significant decline in NDVI. This and shrubland on idle resettled farmland is in line with the fact that Malawi has the and/or the decline in burning that appears lowest NPP per capita and a much higher to have occurred in Zimbabwe since 2012. population density than the other countries in the sample, both of which indicate Degradation was greatest in Malawi, where higher pressure on land resources here. CARBON STORAGE The third metric used in the regional biomass for the benchmarking exercise. benchmarking comparison was carbon storage. For this, the focus was on the storage Zimbabwe fell around the middle of the of carbon in the aboveground biomass of regional sample in terms of mean aboveground plants, as a suitably accurate harmonised carbon storage across the country (Figure 12). data layer encompassing all five countries As with the NPP measures, Zambia had the is available (Santoro et al., 2018). Conversely, highest value by some margin, with mean global belowground carbon biomass datasets aboveground carbon storage of 25.8 tC/ha, explored for this study appear to often followed by Tanzania. Higher biomass in these overestimate BGB in the region and/or show countries would be expected, given that they some unexpected spatial patterns. As a result, have higher mean rainfall than Zimbabwe and in the estimation of the carbon retention a positive relationship exists between biomass service for Zimbabwe, BGB was instead and rainfall for woodland types like miombo derived from the AGB layer using root:shoot (Frost, 1996). By the same token, the fact that ratios from the literature. This approach Zimbabwe has higher mean carbon storage could have been extended to the other four than Malawi is again indicative of more countries. However, it would require land extensive degradation of natural habitats cover data to be obtained and prepared for and their carbon stocks in Malawi, relative to each country in order to assign appropriate Zimbabwe. Finally, Kenya had the lowest mean root:shoot ratios for different habitats. carbon storage. As with NPP, this is in large Hence, given time limitations, a decision part due to the extensive areas of low biomass, was made to focus only on aboveground arid vegetation in the north and east of Kenya. Figure 12. Mean aboveground carbon storage across the five countries, based on Santoro et al. (2018). 13 Ch 3: Benchmarking Land Condition 4. BASELINE ASSESSMENT OF ECOSYSTEMS AND ECOSYSTEM SERVICES KEY POINTS ∙ A detailed baseline land cover dataset was generated from global land cover data, vegetation type data (ecoregions) and vegetation condition information (using NDVI). ∙ Natural land cover classes cover 74% of Zimbabwe and 24% is cultivated land. ∙ Overall, the total value of ecosystem services currently provided by Zimbabwe’s ecosystems was estimated to be in the order of US$11.5 billion/y. Of this, around US$2.4 billion/year of ecosystem benefits accrue to Zimbabwe, with the balance the balance being climate regulation services benefitting the rest of the world. ∙ The value of provisioning services was estimated to be in the region of US$1.23 billion/y. This includes the ecosystem contribution to crop and livestock production, estimated to be worth US$365 million/y and US$292 million/y, respectively. An additional US$576 million is derived from the harvesting of wild natural resources by rural households. ∙ Hydrological-related ecosystem services also result in significant cost savings from the retention of soil and enhancement of groundwater recharge by vegetation cover. The value of avoided sedimentation was reservoir estimated to be US$208 million/y, while the facilitation of groundwater recharge by ecosystems was estimated to be worth US$496 million/y. and the ecosystem contribution to nature-based tourism in the region of US$230 m/y. ∙ Climate regulation through the storage and sequestration of carbon by ecosystems generates around US$9.4 billion/y in benefits globally from avoided climate change-related damages, with Zimbabwe’s share estimated to be around US$188 million/y. ∙ The ecosystem contribution to nature-based tourism was estimated to be around US$230 million/y. Box 1 Brief overview of the concept and types of ecosystem services Ecosystem services are defined as “the benefits people obtain from ecosystems” (Millennium Ecosystem Assessment 2003, 2005). These benefits depend on the nature of ecosystems and their biodiversity. Ecosystem services are typically considered to include provisioning, regulating, and cultural services. Provisioning services are the harvestable resources supplied by ecosystems. These include: ∙ Wild foods and medicines; ∙ Raw materials; 14 ∙ Ecosystem inputs to crop and livestock production; and ∙ Genetic resources. Regulating services are the functions that ecosystems and their biota perform that benefit people in surrounding or downstream areas or even distant areas. These include: ∙ Climate regulation; ∙ Flow regulation; ∙ Sediment regulation; ∙ Water quality amelioration; and ∙ Pollination. Cultural services are the ecosystem attributes (for example, beauty and species diversity) that give rise to the ‘use values’ gained through any type of activity ranging from adventure sports to birdwatching, religious or cultural ceremonies, or just passive observation or the ‘non-use values’ gained from knowing that they exist and can be enjoyed by future generations. BASELINE LAND COVER AND ECOSYSTEM CONDITION Modelling of ecosystem services requires WWF ecoregions data (Olson et al., 2001) spatial information on the extent of different and satellite-derived Normalised Difference ecosystems in the landscape, and ideally Vegetation Index (NDVI) data. The combination an indication of the condition of these of land cover data with ecoregions was used ecosystems. Together, ecosystem extent and to effectively disaggregate land cover data condition determine the physical capacity into broad vegetation types (for example of landscapes to supply ecosystem services. miombo, mopane etc.), while NDVI data For this study, current ecosystem extent was used as a proxy measure of ecosystem and condition were estimated through the condition. These steps are described in integration of three data sources, namely more detail in the following sections. global land cover data (Buchhorn et al., 2020), GENERATION OF INITIAL LAND COVER DATASET Various land cover datasets were considered ecoregions layer (Olson et al., 2001). This for use in this study. Zimbabwe’s Forestry effectively allows the land cover data to be Commission has generated land cover maps split into broad vegetation types. For example, for 1992 and 2017. However, the map layers an area of open tree cover within the miombo can only be viewed online and are not freely woodland ecoregion would be classed as available to downland. Out of freely accessible open miombo woodland in the updated land global land cover data, the 90 m Copernicus cover layer. A further modification was made land cover dataset for 2019 (Buchhorn et al., to separate plantation forest areas, largely 2020) was chosen. This dataset was considered located in Zimbabwe’s Eastern Highlands, reasonably accurate based on comparison from indigenous woodland and forest. Even with satellite imagery across various parts of though the raw spatial data could not be the country, with a relatively high number obtained, rough plantation locations were of land cover classes compared to other mapped through digitising an image of the global land cover products. For example, 2017 Forestry Commission land cover. These areas of closed and open tree cover are plantation areas were then refined based separated, which is a meaningful distinction on satellite imagery, in which plantation for ecosystem service modelling purposes. vegetation is generally distinct. These changes resulted in the generation of an Further detail was added to the land cover initial land cover layer containing 20 classes. data through combining it with the WWF 15 Ch 4: Baseline Assessment INTEGRATION OF ECOSYSTEM CONDITION Most land cover products do not incorporate of both vegetation types falls within protected ecosystem condition into their classification areas, and most of these protected areas are in schemes, requiring the integration of other turn located in the lower rainfall range of this datasets. Other remote sensing data can vegetation type. In this case, it was assumed the be used to assess various indicators of land NDVI data adequately reflected degradation degradation and ecosystem condition, without needing to control for rainfall variability. including the Normalised Difference Vegetation Index (NDVI). NDVI data have As examples, the observed relationships been widely used as an indicator of vegetation between NDVI and rainfall for mopane and health and ecosystem dynamics since the miombo woodland are shown in Figure 13. 1980s (Anyamba & Tucker, 2005). NDVI The patterns were in line with those observed measures vegetation health and density using in a regional study of East African woodland plant reflectance characteristics and is well types, where NDVI increases initially before suited to monitoring changes in ecosystem plateauing above a certain rainfall amount health over space and time (Kinyanjui, 2011). (Nicholson et al., 1990). However, in our data a subsequent sharp increase in NDVI was Sentinel-2 NDVI data covering the full extent evident above 800 mm. This is associated of Zimbabwe was downloaded for three with higher elevation hilly areas where relief consecutive years (2019 – 2021) using the rainfall results in denser, mixed vegetation Google Earth Engine code platform. To ensure communities which are not necessarily a consistent timeframe for comparison, dominated by mopane. This level of detail is not only imagery from the month of March was captured in the WWF Ecoregions layer from downloaded in each year. March is at the end which the vegetation classes were derived. of the growing season when disparities in NDVI between healthy and degraded vegetation Miombo woodland exhibited an increase in should be high. Conversely, in the dry season NDVI with rainfall until around 1200 mm (Figure both healthy and degraded vegetation would 13). However, after this point NDVI tended to have low NDVI values over much the country decline with further increases in rainfall. An given that deciduous species are widespread evaluation of the NDVI data for miombo in the in Zimbabwe. Clouds were masked out of >1200 mm rainfall band showed that a lot of the imagery and the average NDVI for each pixels exhibiting lower NDVI were areas where pixel calculated, generating a cloud-free land cover was a mixture of trees and bare layer of mean March NDVI for 2019 to 2021. rock. These areas were simply lumped as tree cover at the resolution of the 90 m land cover The NDVI values for each natural land cover data used in the study. To reduce the effect type were then analysed separately to classify of these anomalous pixels, areas with rainfall degraded and undegraded areas. There is >1225 mm were excluded from the estimation often a positive relationship between NDVI of the best-fit curve for relating NDVI and and rainfall within a given vegetation type, rainfall in miombo woodland (Figure 13). This as has been found for miombo (Nicholson, resulted in a relationship of increasing NDVI Davenport & Malo, 1990). Thus, relationships with rainfall up to around 1200 mm, after which between NDVI and rainfall were assessed for NDVI starts to plateau, again better matching each natural land cover type. In addition to the relationship between rainfall and NDVI for rainfall, other factors that could improve the miombo in East Africa (Nicholson et al., 1990). predictive model for NDVI were investigated, including altitude, mean annual temperature For each land cover type where a positive and potential evapotranspiration. However, correlation between NDVI and rainfall was all were highly collinear. Rainfall was strongly evident, the best-fit relationship between positively correlated with altitude and strongly rainfall and NDVI was used to create a layer of negatively correlated with mean annual predicted NDVI from the mean annual rainfall temperature and potential evapotranspiration raster. Actual measured NDVI was then divided Thus only rainfall was used to predict NDVI. by the predicted NDVI layer to generate a ratio of actual to predicted NDVI for each pixel. Areas Not all land cover types exhibited a clear where this ratio is significantly lower than one positive relationship with rainfall. This was are likely to be degraded, since this means notable in Baikiaea woodland and shrubland, actual NDVI is much lower than predicted where NDVI appeared to decline with rainfall. A NDVI for that land cover type and rainfall potential explanation is that a high proportion amount. A consistent threshold criterion 16 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE was then required to separate degraded and For land cover classes where no clear positive undegraded areas, which involves a degree relationship was evident between NDVI and of arbitrariness. In this case, the distribution rainfall the approach could be comparatively of actual to predicted NDVI ratio values was simpler, since it was assumed that NDVI extracted for each natural land cover type. The directly reflects condition here. In these cases, lower 15th percentile value was then used as the distribution of actual NDVI values was the threshold for identifying degradation. For examined, and the threshold value again example, for miombo woodland, pixels falling set based on the value of the lowest 15th within the lowest 15th percentile of actual percentile for actual NDVI in each habitat to predicted NDVI had a value of 0.82 or less type. For example, pixels falling within the (Figure 14). Hence, all miombo woodland pixels lowest 15th percentile of NDVI in Baikiaea where the ratio of actual to predicted NDVI woodland had an NDVI value of 0.52 or was less than 0.82 were classed as degraded. less, which was used as the threshold for identifying degradation in this land cover type. Figure 13. Example of relationships between NDVI and rainfall NDVI correlation for mopane woodland (a) and miombo woodland (b) Figure 14. Distribution of the ratio of actual to predicted NDVI values for all miombo woodland pixels. The red line indicates the lower 15th percentile threshold value that was used to classify pixels as degraded. 17 Ch 4: Baseline Assessment FINAL BASELINE LAND COVER The final current land cover dataset generated degradation of natural habitats is greater, from the combination of raw land cover an analysis of the percentage coverage of data, WWF ecoregions and NDVI data degraded habitats was conducted at a finer consisted of 30 land cover classes, including sub-catchment level, as shown in Figure 16. ten degraded natural habitats classes. A simplified version of this land cover layer Notable hotspots for habitat degradation is shown in Figure 15. Overall, natural land include the centre of Zimbabwe, from cover classes (degraded and undegraded) south of Harare to Masvingo (Figure 16). were estimated to cover 74% of Zimbabwe. This degradation hotspot is centred around the upper western reaches of the Save The bulk of remaining land cover consists of Catchment, which encompasses densely cultivation, which was estimated to cover populated communal farming areas around 24% of Zimbabwe. Cultivated land cover is Buhera and Gutu. Notably, highly degraded particularly widespread in the Mazowe, Save sub-catchments west of Masvingo overlap and Runde Catchments, with many of the with Chivi and Zvishavane Districts, which country’s densely populated communal were identified as degradation hotpots in areas falling within these catchments. Water Zimbabwe’s LDN mapping process (GoZ, 2017). accounts for an additional 1.1% of the country, plantations 0.3% and built-up areas 0.2%. High prevalence of degraded natural habitats is also evident across the western half of The coverage of degraded natural habitats the Mzingwane Catchment, to the south of at the catchment level varied from a low of Bulawayo. This includes Umzingwane and 6% of total catchment area in the Mazowe Beitbridge Districts, which were also identified Catchment to 13% in the Mzingwane as hotspots in the LDN process. The western Catchment. When expressed as a proportion Mzingwane Catchment encompasses some of total natural habitat area, the coverage of the most arid areas of Zimbabwe, with of degraded habitats increases to 9% in the high vulnerability to drought and sensitivity Mazowe Catchment and 15% in the Mzingwane to overgrazing. Another notable area of Catchment (Table 1). The proportion of degradation straddles the Gwayi and Sanyati natural habitat classed as degraded was also Catchments, southwest of Gokwe. This is relatively high in the Runde (14%), Gwayi, another relatively dry area characterised Sanyati and Save Catchments (all 13%). by a high density of farming households. Additionally, significant expansion of farmland To better highlight areas of the country where into Baikiaea woodland habitats is evident here Table 1. Coverage of natural land cover in each catchment, along with the percentage of natural land cover that was classified as degraded CATCHMENT % NATURAL LAND COVER % DEGRADED Gwayi 86% 11% Manyame 74% 8% Mazowe 62% 6% Mzingwane 86% 13% Runde 70% 10% Sanyati 70% 9% Save 59% 8% TOTAL 74% 10% 18 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE Figure 15. Baseline land cover map generated f rom a combination of Copernicus land cover data, WWF ecoregions and Sentinel-2 NDVI data. Note: The full 30 land cover classes are presented in a simplified format here for ease of interpretability. Figure 16. Percentage of natural habitat area classified as degraded at sub-catchment scale. 19 Ch 4: Baseline Assessment PROVISIONING SERVICES KEY POINTS ∙ Crop production: The net value of the top 11 crops was estimated to be US$365.4 million/y, or US$39/ha of farmland, with 34% of this being for food crop production. The overall value per ha was highest in Save and Runde catchments, while the contribution to food production was highest in Manyame catchment. ∙ Livestock production: Zimbabwe supports around 4.34 million TLUs, at a mean density of 11.1 TLUs/km2 with a net value of US$291.7 million/y. Cattle production accounts for over 90% of this value. ∙ Harvested wild resources: Household harvesting of wild natural resources such as wood, thatch and wild foods was estimated to be worth at least US$576 million/y, or US$20/ha of natural land on average. This is particularly high in areas with large rural populations such as the Mazowe and Save catchments. CROP PRODUCTION OVERVIEW OF THE SERVICE sunflower). Production of the 11th crop, sugar cane, was mapped directly to the sugar cane Agriculture is a key sector, providing a source production areas identified through satellite of livelihood for around 70% of Zimbabwe’s imagery. These are readily distinguished from population, 15-20% of GDP and 40% of export other farmland and are almost exclusively earnings (World Bank, 2019b). Nevertheless, it limited to the country’s hot southeast lowveld. is not reaching its full potential in the country. Once food self-sufficient, Zimbabwe has The resource rent was assumed to be 15% become increasingly dependent on imports of gross output, since the government sets and foreign aid (Runganga & Mhaka, 2021). The producer prices for crops with the aim of vast majority of farmers are smallholders who ensuring farmers realise a 15% profit margin. largely rely on rainfed agriculture and natural Producer prices were obtained from the resources as their livelihood (CIAT & World Grain Marketing Board, Tobacco Industry and Bank, 2017). These farmers face significant Marketing Board, and Cotton Company. Sugar declines in yields under a hotter and potentially cane was valued using a case study in Zimbabwe drier future climate (World Bank, 2019b). and Zambia (Shumba, Roberntz & Kuona, 2011). Nature’s contributions to crop production are complex and include the soil, nutrient and CROP PRODUCTION IN PHYSICAL TERMS moisture inputs that sustain plant growth. As a proxy, we use production volume as the Total annual production of the eleven crops physical measure of the ecosystem service. included in the assessment amounts to 7.95 million t, or an average of 0.85 t/ha of farmland area per year. Production is generally highest in ESTIMATION APPROACH the wetter centre and northeast of the country, where the climate is most favourable for crop The InVEST Crop Production model was used growth (Figure 17). Notably, 69% of the overall to map the production reported at provincial- physical crop production figure comes from level for 11 major crops in Zimbabwe’s Crop and sugar cane, although it is also worth noting Livestock Assessment Reports (MoLAWFRR, that sugar has a much higher wet weight than 2021, 2022), using correction factors to ensure other crops. Despite covering a relatively small the final production estimates match the area, the fact that sugar cane yields are much official figures. The crops included six major greater than any other crop means that it food crops (maize, sorghum, millet, ground and accounts for a disproportionately large amount bambara nuts, beans, and sweet potatoes) and of overall production. Sugar cane is grown on four major cash crops (tobacco, cotton, soya, and irrigated farmland, almost exclusively located 20 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE within the hot, dry southeast Lowveld, chiefly higher than production per hectare of in the Runde and Save Catchments. These farmland (Table 2). Note that for comparability are visible as the areas of high production purposes, these figures exclude sugar cane. in the far southeast of the country, which include the sugar estates around Triangle, With sugar cane excluded, the Manyame Hippo Valley and Chisumbanje (Figure 17). Catchment was estimated to have the highest crop production per hectare, followed by the Overall production of the ten main crops Mazowe Catchment (Table 2). However, in excluding sugar amounts to 2.46 million t/y absolute terms, the larger farmland area of (million tonnes per year), averaging 0.26 t/ the Mazowe Catchment means that overall ha (tonnes per ha) of farmland (Table 2). Of crop production is greater here. Both of these this, food crop production is 2.05 million t/y, catchments include much of the higher or 0.21 t/ha/y. The low production per hectare yielding commercial farmland in relatively is largely because only a fraction of the area wet north and east of the country, resulting classified as farmland is planted in any given in higher average yields than other areas. year, with the remainder being fallow or Collectively, the Mazowe, Manyame and abandoned fields, grazing land and other moderately high yielding Sanyati Catchments area such as hedgerows. While farmland account for 67% of total crop production in covers 24% of Zimbabwe according to the the country (sugar cane excluded). Average land cover, the total planted area for field production is much lower for catchments crops in 2020/2021 (a good rainfall season) in the drier south and west of the country, was just 3.5 million ha, or around 9% of the particularly the Gwayi, Runde and Mzingwane country (MoLAFWRR 2021). Hence, the actual Catchments, where average production per yield of crops per planted area is significantly hectare of farmland ranges from 0.09 – 0.17 t/y. Figure 17. Estimated total production of the 11 major crops included in the assessment per hectare of farmland. 21 Ch 4: Baseline Assessment Table 2. Total and Average crop production across Zimbabwe's major catchments (sugar cane excluded for comparability across catchments). TOTAL CROP % OF NATIONAL AVERAGE PRODUCTION CATCHMENT PRODUCTION PRODUCTION (T/HA OF FARMLAND) Gwayi 102 211 4% 0.09 Manyame 470 119 19% 0.48 Mazowe 612 588 25% 0.42 Mzingwane 147 763 6% 0.17 Runde 175 094 7% 0.14 Sanyati 579 401 24% 0.31 Save 376 706 15% 0.20 TOTAL 2 463 883 0.26 ECOSYSTEM CONTRIBUTION TO CROP PRO- However, the Manyame Catchment had a DUCTION VALUE marginally higher value for food crop production per hectare of farmland (US$23.51/ha). Total revenue from production of the eleven crops included in the assessment was The picture is slightly different when estimated to be US$2.07 billion/y, an average production of all crops (food and cash crops) value of US$220/ha of farmland. In surplus value is estimated, reflecting the high value of sugar terms, the service was estimated to be worth cane production. This results in the Save and US$365.4 million/y, or US$39/ha (Table 3). Food Runde Catchments having the highest value crop production was estimated to be worth for total crop production (US$84.5 million/y US$139.5 million/y, or 34% of the total value and US$84.2 million/y, respectively), with of the crop production service. The Mazowe each catchment accounting for 23% of the Catchment had the highest overall value for national crop production value (Table 3). These food crop production (US$34.2 million/y). catchments contain almost all of Zimbabwe’s Table 3. Approximate value of the crop production service in terms of resource rent value of the eleven crops included in the assessment. TOTAL FOOD CROP FOOD CROP TOTAL CROP ALL CROP % OF NATIONAL PRODUCTION PRODUCTION PRODUCTION PRODUCTION CROP PRODUCTION CATCHMENT (US$ MILLION /Y) (US$/HA) (US$ MILLION /Y) (US$/HA) VALUE Gwayi 5.9 5.37 6.2 5.66 2% Manyame 23.1 23.51 50.3 51.12 14% Mazowe 34.2 23.19 72.9 49.40 20% Mzingwane 10.2 11.56 13.2 14.91 4% Runde 13.6 11.23 84.2 69.71 23% Sanyati 27.8 14.82 54.1 28.84 15% Save 24.7 13.08 84.5 44.83 23% TOTAL 139.5 14.82 365.4 38.83 22 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE sugar growing areas, with the bulk of the large by the Manyame Catchment. In contrast, the commercial estates falling within the Runde Gwayi and Mzingwane Catchments account Catchment. This contributes to the Runde for very small portions of Zimbabwe’s crop Catchment having the highest estimated value production value, reflecting the low yields per unit of farmland (US$69.7/ha), followed in these generally water scarce catchments. LIVESTOCK PRODUCTION OVERVIEW OF THE SERVICE in the Crop and Livestock Assessment reports (MoLAWFRR, 2021, 2022). This was done Livestock are an important component of rural through the same procedure used for crop livelihoods. In addition to meat production, production, with livestock numbers in the they have several important roles, including GLW3 datasets summed at provincial level to exploit crop-livestock interactions, provide and adjustment factors applied to ensure a store of wealth, for draught power and to these align with the provincial totals in the pay lobola1 . Mixed crop-livestock production Crop and Livestock Assessment reports. is the most widely practiced form of small- Numbers of the three livestock species were holder agriculture in the country (Mkuhlani then standardised by conversion to TLUs. et al., 2018). In these systems, livestock are typically herded during the rainy season and The service was valued in terms of resource kept away from fields to avoid crop damage. rent as an approximation of the residual value During the dry season, they can graze on crop attributed to the environment. Valuation was residues after harvesting (Mkuhlani et al., 2018). conducted separately for commercial farmland areas on the one hand and resettlement Ecosystem inputs to livestock production areas and communal land on the other, due include fodder production and natural to significant differences in these production water sources. A suitable proxy physical systems. Livestock sale is the primary aim of measure of the service is the number of production in the commercial farming sector, tropical livestock units (TLUs) supported.2 resulting in much higher livestock offtake rates and thus sales revenues. In contrast, livestock ESTIMATION APPROACH (particularly cattle) serve a wider range of roles in small-scale farming systems, with much Livestock densities were mapped using lower offtake and sales rates here (Mukhebi et global spatial data, adjusted to match the al. 1999; Scoones 1992). Hence, the valuation of official statistics reported in Zimbabwe’s small-scale livestock production incorporated Crop and Livestock Assessment reports. The the value of manure, milk, draught power, and focus here was on cattle, sheep and goats, hides, all of which are particularly important the major free-roaming livestock species. components of the value of livestock to small- scale farmers in Zimbabwe (Scoones, 1992). The spatial distribution of cattle, sheep and goats in the landscape was modelled using the Gridded Livestock of the World (GLW3) LIVESTOCK PRODUCTION IN PHYSICAL dataset (Gilbert et al., 2018). This is derived TERMS from official government livestock estimates, Overall, it was estimated that Zimbabwe which are spatially disaggregated to a 10 km2 supports around 4.34 million TLUs, at a cell size using statistical methods with high mean density of 11.1 TLUs/km2 (Table 4). spatial resolution covariates. Since the GLW3 It was estimated that two thirds (66%) of data is dated to 2010, it was aligned with this livestock biomass is located within more recent livestock numbers as reported communal and resettlement areas, where production is largely small-scale. This is evident in the map of livestock production 1  Lobola is a cultural practice where the husband pays (Figure 18), where areas with the highest a “bride price” to the wife’s family and is customarily paid livestock densities are generally located in with cattle. communal lands, such as around Masvingo and in the far northeast of the country. 2  One TLU is equivalent to 250 kg of live animal biomass. Cattle are typically assumed to be 0.7 TLU per head, while The total number of TLUs supported was fairly sheep and goats are 0.1 TLU per head. 23 Ch 4: Baseline Assessment Table 4. Total and mean numbers of livestock across Zimbabwe's catchments, expressed in Tropical Livestock Unit (TLU) terms. TOTAL NUMBER OF MEAN NUMBER OF TLUS CATCHMENT TLUS SUPPORTED % OF NATIONAL TOTAL SUPPORTED PER KM2 Gwayi 640 605 15% 7.3 Manyame 338 872 8% 8.3 Mazowe 630 366 15% 15.7 Mzingwane 602 914 14% 9.6 Runde 708 645 16% 17.2 Sanyati 709 258 16% 10.2 Save 709 966 16% 14.4 TOTAL 4 340 628 11.1 Figure 18. Map of livestock production across Zimbabwe, expressed in terms of tropical livestock units (TLUs). 24 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE comparable across most catchments at around ECOSYSTEM CONTRIBUTION TO LIVESTOCK 600 000 to 700 000, except for Manyame PRODUCTION VALUE Catchment where the number was notably The total net value of the livestock production lower (Table 4). Mean TLU densities were more service was estimated to be US$291.7 variable, with the Runde, Mazowe and Save million/y (Table 5). Communal and small- Catchments all estimated to have relatively scale livestock were estimated to account higher TLUs/km2. All three of these catchments for 59% of this total value. Since cattle are include large areas of densely populated much more valuable than sheep and goats communal land, where livestock densities per head, they account for over 90% of the tend to be high. The Gwayi Catchment was total value of the livestock production service. estimated to have the lowest density of TLUs, which can be explained in large part by the extensive coverage of protected areas in this catchment, where livestock grazing is excluded. Table 5. Estimated value of the livestock production service in terms of approximate resource rent. COMMUNAL LIVE- COMMERCIAL LIVE- STOCK VALUE STOCK VALUE TOTAL LIVESTOCK VALUE CATCHMENT (US$ MILLION/Y) (US$ MILLION/Y) (US$ MILLION/Y) Gwayi 27.1 16.1 43.2 Manyame 10.3 13.2 23.6 Mazowe 28.2 14.4 42.6 Mzingwane 20.0 19.5 39.4 Runde 29.1 20.1 49.3 Sanyati 28.7 18.4 47.1 Save 28.2 18.4 46.6 TOTAL 171.6 120.1 291.7 HARVESTED WILD RESOURCES Harvested wild resources are essential to rural Wild harvested resources can be particularly livelihoods in Zimbabwe. Firewood remains important as a safety-net during times of the main source of energy for cooking for the crisis, such as droughts. For example, in years vast majority of households, and is still largely of low rainfall, indigenous fruits were found to harvested from natural wood stocks (ZIMSTAT contribute about 20% of the energy intake of and UNICEF 2019). Additionally, most rural wealthier farming households and 40% of the households obtain a range of other important energy intake of poor farming households in a products from natural habitats, including study of Wedza District (Woittiez et al., 2013). wood and thatching grass for construction, wild fruits and vegetables, mushrooms, honey, medicinal plants, and other products. These MODELLING OF WILD RESOURCE HARVEST- supplement rural diets, as well as providing an ING additional source of income (Woittiez et al., 2013; Harvested wild resources were modelled Chagumaira et al., 2016; Kupurai, Kugedera & using the methods described in Turpie et al. Sakadzo, 2021). In some areas of Zimbabwe, (2020) and Turpie, Weiss & Letley (2022). These wild harvested resources contribute up to 35% estimate the use of natural resources based of rural incomes (Feresu 2010, in Miller, 2012). on the capacity of the landscape to supply 25 Ch 4: Baseline Assessment Table 6. Data sources used in estimation of stocks and household demand for the selected natural resources. RESOURCE SOURCES OF STOCK ESTIMATES SOURCES OF DEMAND ESTIMATES Campbell et al., 1991, 1997; McGregor, 1991; Grundy et al., 1993; Mabugu & Chitiga, 2002; Wood Bouvet et al., 2018; Santoro et al., 2018 Chambwera & Folmer, 2007; Mudekwe, 2007; Woittiez et al., 2013; ZIMSTAT, 2017 Grundy et al., 2000; Twine et al., 2003; Mude- kwe, 2007; Woittiez et al., 2013; Chagumaira Frost, 1996; Poilecot & Gaidet, 2011; Thatching grass et al., 2016; Dowo, Kativu & de Garine-Wicha- Pritchard et al., 2018 titsky, 2018; Kupurai, Kugedera & Sakadzo, 2021 Campbell, 1987; Campbell, Vermeulen Campbell, 1987; Campbell et al., 1991; McGre- & Lynam, 1991; Campbell et al., 1997; gor, 1991; Twine et al., 2003; Mudekwe, 2007; Wild plant foods Ngulube, Hall & Maghembe, 1996; Woittiez et al., 2013; Mashapa et al., 2014; Cha- Poilecot & Gaidet, 2011; Ngadze et al., gumaira et al., 2016; Kupurai et al., 2021 2017 Campbell et al., 1991; McGregor, 1991; Mude- Degreef et al., 2020; Mlambo & Ma- Mushrooms kwe, 2007; Woittiez et al., 2013; Dowo et al., phosa, 2021 2019 Campbell et al., 1991; Shackleton & Shackle- Honey Jaffé et al., 2010; Garcia et al., 2013 ton, 2004; Mudekwe, 2007; Mahlatini et al., 2020; Kupurai et al., 2021 Kozanayi & Frost, 2002; Stack et al., 2003; Mopane worms Illgner & Nel, 2000 Twine et al., 2003; Dowo et al., 2018; Kupurai et al., 2021 different types of resources on the one hand Based on the comparison of remotely-sensed and the spatial distribution of the human biomass data (Bouvet et al., 2018; Santoro et al., demand for a given resource on the other. A 2018) between degraded land cover types and further factor considered is accessibility, with their undegraded equivalents, resource stocks resources in protected areas assumed to be were assumed to be 25% lower in degraded less available for harvesting. Varying levels habitats. However, it is possible that some of accessibility were assumed depending on resource stocks could be depleted more than protected area strictness and effectiveness. the biomass change might suggest, in which case this assumption may be conservative. Harvesting was quantified and valued for selected groups of resources based on data The demand for natural resources was availability. It was assumed that wild resources estimated based on household density, the are harvested from all natural land cover proportion of households using a particular classes (degraded and undegraded) but not resource, and the average consumption from cultivated or built-up areas. For woody per user household. Population data used resources, stocks were estimated directly from for was obtained from the Worldpop 100 the layer of aboveground biomass (Bouvet m constrained population density map et al., 2018; Santoro et al., 2018). For other (Bondarenko et al., 2020) and ZIMSTAT data resources, stocks per unit area were assigned to (ZIMSTAT, 2017). Data on the proportion of the different natural land cover classes based households using the various resources on information from the literature (Table 6). and average consumption per household 26 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE were also obtained from ZIMSTAT reports WOODY RESOURCES and numerous literature studies (Table 6). Firewood is the main cooking fuel for the vast majority (94%) of rural households in Once resource demand, stocks, and Zimbabwe (ZIMSTAT and UNICEF, 2019). accessibility had been mapped, the quantities Around 16% of urban households also rely on of resources harvested were calculated from wood as their primary cooking fuel, a figure the minimum of the estimated demand and which has increased since the 2000s as a result the estimated available stocks of resources of the country’s unreliable electricity supply within a specified distance of the demand (Dube, Musara & Chitamba, 2014). In addition to source. To estimate and map harvesting firewood, poles, withies and timber are widely at a high resolution, the running mean harvested for construction by rural households, method developed by Turpie et al. (2020) as well as wood for crafts (Grundy et al., 1993; in KwaZulu-Natal, South Africa was used. Mudekwe, 2007). The latter tends to be a fairly specialist activity. However, harvesting Harvesting was valued using estimates of local of wood for construction purposes by rural farmgate prices and expressed in current US$. households is included in these estimates. Descriptions of the use and value of fuelwood and poles, thatch, plant foods, mushrooms, Generally, harvesting of wood is lowest in honey and mopane worms harvesting are the south and west of the country where given below, followed by an overall summary. population densities and thus demand are lower (Figure 19). Higher wood harvesting Figure 19. Quantities of wood harvested for local firewood consumption and rural household construction needs. 27 Ch 4: Baseline Assessment quantities are associated with densely these catchments include extensive areas of populated rural areas in the north and east of densely populated communal land, as well as the country, particularly on communal land, as dense peri-urban populations around Harare well as densely populated-peri urban regions (in the Mazowe Catchment) and Mutare (in the surrounding towns and cities. Areas with Save Catchment). Harvesting rates are lowest negligible wood harvesting (shown in grey) in the Gwayi and Mzingwane Catchments. are either those where natural habitats have The former includes large, uninhabited been converted to cultivation and settlement protected areas such as the Hwange National (i.e. no wood stocks available for harvesting), Park and surrounding safer areas and forest or the interior of strict protected areas which reserves, while the Mzingwane Catchment have no nearby populations (i.e. no demand). has extensive sparsely populated areas of private livestock and wildlife ranching. In total, the subsistence or small-scale harvesting of woody resources was estimated to be around 12.2 million t/y at the national THATCHING GRASS scale or an average of 0.42 t/ha of natural land Thatching grass is widely used for roofing cover. This was valued at US$305 million/y, with in rural areas of the country, though its an average value of US$10.46/ha of natural usage varies spatially from virtually 0% in land cover. At a catchment level, harvesting urban provinces to 63% of households in rates are highest in the Mazowe and Save Matabeleland North (ZIMSTAT, 2017). Since Catchments, exceeding 0.5 t/ha here. Both miombo and other local woodland types Figure 20. Estimated harvested quantities of thatching grass for roofing of rural households. 28 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE WILD PLANT FOODS generally have a grassy understory, the resource can be harvested from a wide range The estimated rates of harvesting of wild plant of habitats from grassland to woodland, foods (wild fruits and vegetables) are shown though not from forests. Based on available in Figure 21. Population densities again have data, the highest harvesting of thatching grass a strong influence on the spatial pattern of was estimated to be associated with densely harvesting. Additionally, the higher densities populated rural areas and peri-urban areas of valuable fruit trees in miombo woodland surrounding cities and large towns (Figure 20). (Campbell, 1987) means that harvesting rates are highest in miombo areas in the centre, north In total, subsistence harvesting of thatching and east of Zimbabwe. Key fruit tree species grass for roofing by rural households was associated with miombo woodland include estimated to be around 120 000 t/y, or an muzhanje/mahobohobo (Uapaca kirkiana) average of 4.11 kg/ha of natural land cover. and mobola plum (Parinari curatellifolia), The harvest of thatching grass was valued at which are harvested in large quantities by local US$48.1 million/y, or US$1.65/ha of natural land communities for consumption and sale locally cover. The value of thatching grass per unit and in urban markets (Karaan et al., 2006; area was highest for mesic grassland, followed Woittiez et al., 2013; Chagumaira et al., 2016). by dry grassland and miombo shrubland, reflecting the greater grass productivity After wood, wild plant foods were the second in these more open habitats (Frost, 1996). most valuable of the harvested resources Figure 21. Estimated harvested quantities of wild plant foods. 29 Ch 4: Baseline Assessment included in this study, with a total estimated collection provides a relatively quick and value of NTFPs (US$106.5 million/y). Overall, it easy income source (Mlambo & Maphosa, was estimated that around 304 000 tonnes 2017). Wild mushrooms also provide a rich of wild plant foods are harvested each year, or source of protein, which is often lacking in an average of 10.4 kg/h of natural habitat. This diets of the rural poor (Chittaragi, Naika & gives a mean value of US$3.65/ha/y. The value Vinayaka, 2014; Mlambo & Maphosa, 2017). of wild plant food harvesting was particularly high in miombo woodland (US$26.3/ha/y), and Spatial variation in the estimated household accounts for around 80% of the total value harvesting of mushrooms is shown in Figure if wild plant food harvesting in the country. 22. Miombo woodland areas had particularly high values for mushroom harvesting due to high mushroom stocks here. This MUSHROOMS results from the dominance of tree genera (Brachystegia, Jubelnardia and Uapaca) Mushroom growth and harvesting is associated with ectomycorrhizal fungi, strongly seasonal in Zimbabwe (December- leading to a high abundance of edible fungi April) and varies significantly in response (Degreef et al., 2020; Mlambo & Maphosa, to rainfall (Mlambo & Maphosa, 2017, 2021). 2021). Additionally, the diversity of edible Mushroom flushes are sporadic and may fungi is also high in Zimbabwe’s miombo last for only a few weeks following good woodlands. Harvesting of at least 45 different rainfall events. During such times, mushroom Figure 22. Estimated harvested quantities of wild mushrooms. 30 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE mushroom species has been documented. country results in high potential for honey However, only a portion of these are attractive production (Campbell et al., 2007). The annual enough for commercial sale along roadsides harvested amount of honey was estimated to or at town stalls (Mlambo & Maphosa, 2017). be around 6.4 million litres, an average of 0.16 l/ ha of natural habitat. It was estimated that 6.8 The overall value of wild mushroom harvesting million litres of honey are harvested nationally, was estimated to be US$74.9 million/y, with an with an average harvest of 0.17 l/ha of natural average value of US$1.91/ha across all natural habitat. As the most dominant woodland land cover. Total mushroom harvesting was habitat, miombo woodland unsurprisingly estimated to be 62 400 t/y, or 1.59 kg/ha of natural accounts for the highest proportion of honey land cover. Around 55 500 t/y was estimated to harvesting. However, on a per unit area basis, be harvested from miombo woodland alone the value of harvesting is slightly higher in (worth US$66.7 million/y), accounting for 88% of indigenous forest, reflecting greater tree total wild mushroom harvesting in the country biomass and thus potential hive density. Highest values for honey harvesting HONEY are evident where woodland and forest habitats in are located close to densely Wild honey harvesting across Zimbabwe was populated rural and peri-urban areas (Figure estimated to have a total value of US$22.5 23). This is evident across much of the million, or US$0.58/ha. The widespread central and eastern regions of Zimbabwe. distribution of wooded habitats across the Figure 23. Estimated harvested quantities of honey. 31 Ch 4: Baseline Assessment MOPANE WORM the central watershed. Harvesting levels were estimated to generally be highest in Mopane worm harvesting was estimated the southeast of the country, where densely to have a total value of US$18.6 million/y. In populated communal lands intersect with modelling this resource, stocks of mopane mopane habitats. Land tenure has a marked worm were assumed to be exclusively limited impact over population densities and thus to wooded mopane habitats in the north, south harvesting. For example, in the far south of the and west of the country. Within these habitats, country, large areas with low harvesting rates mopane worm harvesting has a relatively high tend to align with more sparsely populated value per unit area of US$2.38/ha/y. Notably, the private ranching areas. Additionally, a large value per hectare of mopane worm harvesting amount of mopane habitat falls within protected in these habitats was second only to that of areas, with limited to no harvesting here. wood. The total quantity of mopane worms harvested was estimated to be 16 900 t/y, which amounts to 2.16 kg/ha/y of mopane habitat. OVERALL VALUE The spatial map of mopane worm harvesting The total value of subsistence and small- (Figure 24) differs markedly from any of the scale harvesting of the selected NTFPs was other resources included in this study due to estimated to be around US$576 million/y. the limitation of harvesting to areas within Notably, this excludes the value of medicinal the mopane ecoregion. These habitats are plants. Given that it is estimated that around associated with lower-lying areas away from 80% of Zimbabwe’s population uses traditional Figure 24. Estimated harvested quantities of mopane worms 32 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE medicines derived from indigenous biodiversity, ha/y), which was the highest mean value of any the inclusion of traditional medicines would catchment. These estimates reflect the fact likely significantly add to the total value of that both the Mazowe and Save Catchments the resources considered here (Maroyi, 2013). include large rural populations, resulting in high resource demands. Conversely, harvesting At a catchment level, the total value of wild was lowest in both total and per hectare terms resource harvesting was highest in the Save in the Gwayi and Mzingwane Catchments, Catchment (US$124.1 million/y), followed by both of which are relatively sparsely populated. the Mazowe Catchment (US$105.7 million/y; Furthermore, the Gwayi Catchment has a Table 7). However, the latter had a slightly large coverage of protected areas where higher harvesting value per hectare (US$26.4/ resource harvesting is low or non-existent.. Table 7. Total and average value of the subsistence harvesting of selected wild resources, broken down by catchment. TOTAL VALUE OF RE- MEAN VALUE OF RE- % OF NATIONAL RE- SOURCE HARVESTING SOURCE HARVESTING SOURCE HARVESTING CATCHMENT (US$ MILLION/Y) (US$/HA) VALUE Gwayi 54.3 6.17 9.4% Manyame 64.3 15.79 11.2% Mazowe 105.7 26.37 18.3% Mzingwane 45.6 7.27 7.9% Runde 77.9 18.88 13.5% Sanyati 104.2 14.99 18.1% Save 124.1 25.25 21.5% TOTAL 576.1 14.72 33 Ch 4: Baseline Assessment CULTURAL SERVICES KEY POINTS ∙ Tourism is a significant economic sector in Zimbabwe and the country draws heavily on its natural attractions in its tourism marketing strategies. ∙ In total, the value of nature-based tourism across Zimbabwe was estimated to be US$364 million in 2019 and was highest in the Gwayi catchment which accounted for 55% of this value. ∙ The nature-based tourism value for terrestrial natural land cover within protected areas was 5.5 times higher than for unprotected natural areas, highlighting that protection does make these areas significantly more attractive for tourism. NATURE-BASED TOURISM OVERVIEW OF THE SERVICE attractions located outside protected areas. Tourism is a significant economic sector Zimbabwe also ranks within the top third of in Zimbabwe, with an estimated direct countries globally on the natural resources contribution to GDP of US$1.03 billion in pillar of the World Economic Forum’s (WEF) 2019, according to the country’s first Tourism travel and tourism competitiveness index (TTCI) Satellite Account report (MECHTI, 2021). This (WEF, 2019), The country also draws heavily on amounts to 4.25% of overall GDP. Wildlife its natural attractions in its tourism marketing and natural landscapes are a key drawcard. strategies (see https:// zimbabwetourism.net/). Indeed, state protected areas managed by the Parks and Wildlife Management Although relatively volatile, the tourism Authority (PWMA) received around 530 sector in Zimbabwe has experienced steady 000 foreign visitors (ZTA, 2020), which is growth over the past two and a half decades roughly a third of the 1.7 million arrivals in (Figure 25). The COVID-19 pandemic had a 2019 (excluding transit travellers). This does significant impact during 2020-2022, and not include visitors who came to see natural tourism was still recovering during 2023. Figure 25. Total number of inbound tourists to Zimbabwe over the period 1995-2022. Data source: The World Bank, 2023 (for period 1995-2019) and (Kenya Ministry of Tourism, 2023)Zimbabwe Tourism Authority (for period 2020-2022). 34 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE ESTIMATION APPROACH of ecosystem-related content in natural landscapes, plantations, cultivated areas The available data do not provide an explicit and urban areas found in studies of photo estimate of the contribution of biodiversity content in Kenya and Uganda (Turpie, Wilson and natural attractions to the value of tourism & Letley, 2023; UBOS, 2023). This study updates in Zimbabwe. Given that tourists often have the estimate given in Turpie et al., (2022b). complex itineraries, this can be difficult to tease out, and is usually estimated using survey data. For this study, the service was valued as ECOSYSTEM CONTRIBUTION TO NA- the resource rent generated by nature-based TURE-BASED TOURISM VALUE tourism expenditure in 2019 (before the COVID Pandemic), using a combination of tourism A map presenting the spatial distribution of statistics and spatial data on tourism activity. density of tourism value, as disaggregated from PUDs, is shown in Figure 26. There is an obvious First, we isolated the expenditure on visiting concentration of photographs around larger attractions (as opposed to other reasons such urban centres. However, there are also clear as business or visiting friends and family), based clusters of higher density around a number of on WTTC data (WTTC, 2021) and information protected areas and other biodiversity-related from ZTA (2020) which splits leisure tourist attractions. Away from urban centres and the numbers by purpose of visit. Transit visitors more popular protected areas, the distribution were excluded from the analysis. Based of geotagged photographs is generally sparse, on data from South Africa, it was assumed indicating the high value of protected areas as that the percentage of tourist spending on focal points for tourism outside of major towns. attractions varied as follows for the different categories of visitors: 100% spending on Photograph densities were particularly high attraction for holiday tourists, 4% for business in Zambezi National Park and neighbouring visitors, 2% for those visiting family and Matetsi Safari area in the northwest corner relatives and 15% for other visitor categories. of the country (Figure 26). This reflects their prime location as wildlife-viewing areas According to WTTC (2021), 92% of tourism close to Victoria Falls, the country’s premier expenditure in Zimbabwe was for leisure tourist attraction. Photo densities were also purposes in 2019. Of this, the main purpose high in Hwange National Park, particularly in of visit was visiting friends and relatives its more accessible eastern reaches, and in (56%), followed by holiday tourists (33%) and Matobo National Park just south of Bulawayo. other categories (11%) (ZTA, 2020). Based on the estimated attraction-related spending Other notable tourism hotspots include ratios for the different visitor categories total Lake Kariba, Mana Pools National Park and expenditure on attraction-based tourism in neighbouring safari areas, which are popular Zimbabwe was estimated to be US$450 million areas for both wildlife and recreational fishing. in 2019, which is a third of total tourism receipts. The same can be said for Lake Kariba, which also provides a mix of fishing and wildlife- Attraction-based tourism value was then viewing opportunities. Just southwest of disaggregated based on the spatial distribution Harare, Lake Chivero is another major nature- of photo-user-days (PUDs) calculated from geo- based tourism attraction, offering fishing tagged photographs uploaded to the internet and a game park with rhino. In the east of (www.flickr.com), accessed using the InVEST the country, a concentration of photographs Recreation model.1 This method provides a is evident around the Nyanga National Park reliable proxy for visitation rates (see Wood et which offers a mixture of wildlife and hiking al., 2013), and maps value to the attractions with attractive mountainous scenery, while in rather than to overnighting locations (Turpie the southeast of the country a notable cluster et al. 2021). Because a significant proportion of photographs is evident in the Gonarezhou of geotagged photographs on Flickr are not National Park, particularly in the northern tourism-related, adjustments were made to section which is most accessible to visitors. the PUD density based on average proportions Based on this, nature-based tourism was estimated to account for 81% of attraction- based tourism, thus almost 53% of all tourism 1  The average annual photo-user-days (PUDs) were expenditure. The ecosystem contribution, obtained for each grid cell (2 km2) across the period valued as resource rent from nature-based 2005-2017. One PUD is one unique photographer who tourism, was estimated to be US$229.7 took at least one photo in the gird cell on a single day. 35 Ch 4: Baseline Assessment Figure 26. Estimated attraction-based tourism expenditure per km2, derived f rom the spatial distribution of geotagged photographs (updated f rom Turpie et al., 2022b). Table 8 The total value of the ecosystem contribution to tourism (resource rent) by catchment in 2019. RESOURCE RENT VALUE OF % OF NATIONAL PER HECTARE NATURE NATURE-BASED TOURISM NATURE-BASED BASED TOURISM VALUE CATCHMENT (US$ MILLION/Y) TOURISM VALUE (US$/HA/Y) Gwayi 125.3 55% 14.2 Manyame 37.7 16% 9.3 Mazowe 14.5 6% 3.6 Mzingwane 14.0 6% 2.2 Runde 12.9 6% 3.1 Sanyati 11.8 5% 1.7 Save 13.6 6% 2.8 TOTAL 229.7 5.9 36 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE million/y (Table 8). Gwayi Catchment highest value per hectare by some margin, as accounted for 55% of this value (Table 8). This the concentration of high densities of people catchment includes Victoria Falls, as well as into relatively small areas results in high Hwange and Zambezi National Parks and photograph densities, even after adjusting for neighbouring State Forest areas. Manyame non-nature-related content. Water had the Catchment, which includes most of Harare next highest value per hectare. This reflects as well as Lake Chivero and Mana Pools the fact that many of Zimbabwe’s larger National Park and neighbouring safari areas, dams are popular spots for various nature- had the second highest value. All remaining related tourism activities, including fishing catchments had relatively lower attraction- and wildlife viewing. Notably, the nature- based tourism value in absolute and per based tourism value for terrestrial natural hectare terms, with Sanyati Catchment having land cover within protected areas was 5.5 the lowest value. Much of this catchment is times higher than for unprotected natural relatively inaccessible, with few major roads areas, highlighting that protection does make and limited tourist facilities, except Lake Kariba. natural areas significantly more attractive for tourism. Cultivated areas had the lowest per The breakdown of nature-based tourism value hectare value for tourism, suggesting that by broad land cover types is given in Table these areas are of limited interest to visitors. 9. Urban areas were estimated to have the Table 9. The total value of the ecosystem contribution to tourism (resource rent) by broad land cover type in 2019 RESOURCE RENT VALUE OF % OF NATIONAL PER HECTARE NATURE BROAD LAND NATURE-BASED TOURISM NATURE-BASED BASED TOURISM VALUE COVER (US$ MILLION/Y) TOURISM VALUE (US$/HA/Y) Natural (inside PAs) 107.0 47% 20.0 Natural (outside PAs) 85.1 37% 3.6 Plantation 0.7 0% 5.4 Cultivation 15.4 7% 1.6 Urban 8.0 3% 84.3 Water 13.6 6% 31.0 TOTAL 229.7 5.9 37 Ch 4: Baseline Assessment REGULATING SERVICES KEY POINTS ∙ Carbon storage and sequestration: In total, aboveground and belowground biomass carbon storage across Zimbabwe was estimated to be 808 million tC or 2 966 million tCO2e. Retention of this carbon results in avoided climate change-related losses worth US$8.4 billion/y globally. This value would be even greater if the large soil carbon pool was included. However, this has unfortunately not been adequately assessed in Zimbabwe. Additionally, ecosystems were estimated to remove around 86 million tC (317 tCO2e) from the atmosphere each year, avoiding climate change-related losses worth US$905 million/y globally. ∙ Flow regulation: Through mediating infiltration, ecosystems can help reduce overall seasonal variation in flows, relative to the seasonal variation in rainfall. This potentially has an important bearing on the cost of supplying or obtaining water. Modelling of flows with and without vegetation cover did not generate a significant benefit for surface infrastructure. However, it was estimated that groundwater recharge would decline by 59% under a bare ground scenario, with a replacement cost of US$496 million/y. ∙ Sediment retention: Vegetative cover prevents erosion by stabilising soil and intercepting rainfall, thereby reducing its erosivity. It was estimated that landscapes within dam catchment areas specifically retain some 167 million tonnes of sediment per year (16 t/ha/y), relative to a hypothetical landscape where all land cover is converted to bare ground. The value of the sediment retention service within dam catchment areas was estimated to be worth US$208 million/y. CARBON STORAGE AND SEQUESTRATION OVERVIEW OF THE SERVICE provide a reliable estimate for higher biomass forest areas. For these areas, the GlobBiomass Carbon accumulates in vegetation biomass AGB layer generated by Santoro et al. (2018) through plant growth and in soils through the was used to fill the gaps to generate a production of leaf litter and partially decayed final combined AGB layer for the country. biomass. Thus, significant amounts of carbon Belowground biomass was then derived from are removed from the atmosphere and stored the aboveground layer based on root shoot in healthy ecosystems. When ecosystems ratios for similar land cover types from the are degraded or transformed, the amount of literature (Ryan, Williams & Grace, 2011; IPCC, carbon sequestered is reduced, and the stored 2019). These ranged from 0.27 for cultivated carbon is released into the atmosphere as CO2, areas to 1.58 for grassland. Biomass values were thereby contributing to global climate change. converted to the carbon equivalent (biomass x Indeed, the loss of forests is responsible for 0.5) and to the CO2 equivalent (carbon x 3.67). about 12–17% of the world’s greenhouse gas emissions (Nakakaawa, Vedeld & Aune, 2011). Due to a lack of accurate spatial data, This assessment estimated the active service soil organic carbon was not mapped and of carbon sequestration, as well as the passive included in this assessment. Studies in the service (benefit) of the retention of carbon. miombo biome have found that soil carbon stocks are highly variable, but often exceed ESTIMATION APPROACH carbon storage in vegetation biomass (Ryan et al., 2011). Hence, the stock estimates in Carbon storage was quantified in terms this study are very conservative, as they of aboveground and belowground carbon are limited to biomass carbon storage only. biomass. Carbon storage in aboveground biomass (AGB) was based primarily on the Carbon sequestration was mapped in physical AGB map of African savannas and woodlands terms as the net uptake of carbon per hectare (Bouvet et al. 2018). However, it does not per year. Net primary productivity (NPP) data 38 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE from the 500 m MODIS NPP Gap-Filled product, year into perpetuity, estimates of the SVC are accessed via Google Earth Engine, was used as usually expressed in terms of the net present the initial data input. NPP is considered to be value of damages avoided. For comparability a good measure of the carbon sequestration with the annual values of other ecosystem (carbon sink) function of ecosystems, as it services, these values were annualised. For measures the amount of carbon removed from this study, we used the World Bank’s median the atmosphere by plants, less of respiration estimate of the social value of carbon for (Sha et al., 2022). The mean NPP value for 2022 (US$62/tCO2e; World Bank 2017). 2015 to 2022 was obtained, to account for data gaps and inter-annual variability. Additionally, NPP was clipped to natural ecosystems only, CARBON STORAGE AND SEQUESTRATION as per SEEA EA guidelines, since cultivated The overall average rate of carbon sequestration biomass is typically harvested each year and across Zimbabwe was estimated to be 2.2 thus does not result in a net removal of carbon. tC/ha/y, ranging from 2.0 tC/ha/y in the Mzingwane and Runde Catchments to 2.5 tC/ However, without adjustment, NPP ha/y in the Save Catchment (Table 10). Total overestimates net carbon uptake. It does not sequestration by ecosystems was estimated account for carbon returned to the atmosphere to be around 86 million tC/y, or 317 million by heterotrophic respiration, that is, CO2 tCO2e/y. It should be noted that this includes released from the soil during the decomposition uptake of carbon into both vegetation of organic matter by soil microbes and fauna. biomass and the soil organic carbon pool. To account for this, NPP was converted to Net Ecosystem Productivity (NEP), which excludes The total aboveground and belowground heterotrophic respiration losses (Turner et al., storage of carbon across the country was 2004). The conversion was made using the estimated to be 808 million tC, or 2966 million NPP/NEP relationship estimated by Pregitzer tCO2e (Table 10). This amounts to an average & Euskirchen (2004), which generally results in storage of 20.6 tC/ha, or 75.8 tCO2e/ha. Of this, NEP estimates that are around 40-50% of NPP. aboveground biomass accounts for 14.7 tC/ha and belowground biomass the remaining 6.0 To obtain a more realistic estimate of net carbon tC/ha. At the catchment-scale, average carbon uptake, annual carbon removals from wood biomass per hectare was broadly comparable, harvesting (as estimated in this study above) ranging from 18.6 tC/ha in Runde Catchment and fires were then deducted from the NEP to 23.0 tC/ha in Mazowe Catchment. These estimate. To estimate the latter, annual burned carbon storage estimates would likely at least area was obtained from the ESA MODIS Fire_cci double with the inclusion of the soil carbon pool. Burned Area Pixel product (Padilla Parellada, 2018), and the amount of carbon removed by Indigenous forest had the highest average these burns was subtracted from the NEP biomass per hectare (57.4 tC/ha; Table 11). estimate. It was assumed that a 22% removal However, plantation forest had the highest of aboveground carbon occurs in burned sequestration rate (7.6 tC/ha/year), reflecting areas, based on studies of biomass removals the faster growth of exotic plantation species. by fire in Zambian miombo woodland (Hoffa In contrast, cultivation had lower mean et al., 1999). This may still overestimate the net biomass (9.0 tC/ha) than any other land cover ecosystem carbon balance (NECB), as some type, except for bare, with a biomass over 3 processes such as carbon removals by livestock times less than for most woodland types. This and wildlife, commercial wood harvesting and underscores the significant decline in carbon the long-term natural removals of carbon from biomass storage that occurs when natural ecosystems (such as through leaching and the land cover types are converted to cultivation. export of dissolved organic carbon by rivers) were not estimated within the limitations of The fairly high storage and sequestration this study. However, the method used should estimates for built-up areas may be somewhat still provide a reasonable approximation surprising (Table 11). This reflects the extensive of the carbon sequestration service. presence of large urban trees across many of Zimbabwe’s cities. At the resolution of the 100 Carbon storage and sequestration were both m land cover used in this study, these tend to be valued in terms of the “social value of carbon” classified as built-up area. Similarly, the small (SVC), which is the avoided damage caused allocation of carbon in waterbodies is due to by an additional tonne of carbon dioxide in overlaps and the resolution of the spatial data, the atmosphere. Because each tonne added for example where a waterbody pixel partially to the atmosphere leads to damages in every 39 Ch 4: Baseline Assessment Table 10. Average and total amount of carbon sequestered and stored in natural ecosystem biomass across Zimbabwe’s major catchments, and the annualised values of these active and passive services, respectively. ACTIVE SERVICE: PASSIVE SERVICE: CARBON SEQUESTRATION CARBON RETENTION GLOBAL GLOBAL AVERAGE TOTAL VALUE AVERAGE TOTAL VALUE CATCHMENT (T/HA/Y) (MILLION T/Y) (US$ MILLION/Y) (T/HA/Y) (MILLION T/Y) (US$ MILLION/Y) Gwayi 2.4 20.8 218.3 21.2 186.8 1 959.6 Manyame 2.2 9.0 94.3 22.1 90.0 944.1 Mazowe 2.3 9.3 97.9 23.0 92.2 966.8 Mzingwane 2.0 12.4 129.9 19.2 120.0 1 258.6 Runde 2.0 8.2 85.6 18.6 76.9 806.3 Sanyati 2.1 14.5 152.6 20.1 139.5 1 462.9 Save 2.5 12.1 126.5 20.9 102.7 1 076.6 TOTAL 2.2 86.3 905.0 20.6 808.1 8 474.9 includes riverine or shoreline vegetation. border. Elsewhere, pockets of high biomass are associated with higher rainfall hilly areas The spatial distribution of carbon biomass with dense woodland vegetation, such as in across Zimbabwe is shown in Figure 27. the hills around Masvingo and Zvishavane. Notable areas here where high biomass woody natural habitats remain include most of the Areas of low biomass are largely associated Zambezi Valley, which exhibits particularly with cultivation, with the lowest values in high biomass between Victoria falls and the densely-cultivated communal areas. These southern end of Lake Kariba. Patches of high align with many of the large patches of biomass are also associated with indigenous low biomass spread across the centre, forest, plantations and dense woodlands southeast and north of the country. along and around Zimbabwe’s eastern 40 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE VALUE OF CARBON STORAGE AND SEQUES- to about $905 million/y. About 2% of these TRATION passive and active benefits would be felt in Zimbabwe. The combined global value of At the global level, climate change-related the carbon storage and sequestration service damages avoided by the retention of carbon was thus estimated to be around US$9.4 stocks in Zimbabwe were estimated to be billion/y, with a local value to Zimbabwe worth US$183.9 billion in present value terms, of around US$188 million/y. The potential equivalent to an annual value of US$8.47 to generate income from carbon credits billion/y (Table 10). The benefit of damages through enhancing the carbon storage and avoided globally from the active service of sequestration service is explored in Chapter 5. carbon sequestration was estimated to be $19.6 billion in present value terms, equivalent Table 11. Total and average carbon storage and sequestration by land cover type. Note that some land cover classes have been aggregated to ease interpretability. BIOMASS BIOMASS CARBON STORAGE CARBON STORAGE CARBON SEQ. CARBON SEQ. LAND COVER (MILLION T) (T/HA) (MILLION T/YEAR) (T/HA/YEAR) Indigenous Forest 5.6 57.4 0.5 5.3 Plantation Forest 5.3 41.5 1.0 7.6 Miombo woodland 270.6 30.4 29.7 3.3 Mopane woodland 132.2 29.9 10.6 2.4 Baikiaea woodland 60.5 25.9 6.6 2.8 Acacia-Terminalia 82.4 28.4 9.0 3.1 savanna Miombo shrubland 45.2 15.4 8.3 2.8 Mopane shrubland 60.0 17.5 6.3 1.9 Baikiaea shrubland 15.0 14.1 2.7 2.5 Acacia-Terminalia 33.7 13.9 6.8 2.8 shrubland Mesic grassland 6.1 15.8 1.3 3.4 Dry grassland 3.2 16.8 0.4 2.1 Cultivation 84.4 9.0 2.5 0.3 Bare 0.0 6.0 0.0 0.7 Waterbodies 1.1 2.4 0.3 0.6 Built-up 2.7 28.7 0.2 2.4 TOTAL 808.1 20.6 86.3 2.2 41 Ch 4: Baseline Assessment Figure 27: Carbon biomass (above and below ground) across Zimbabwe derived f rom remote sensing datasets. Source: Based on Bouvet et al. (2018) and Santoro et al. (2018) FLOW REGULATION Land cover plays a significant part in to some extent by ecosystems. Vegetation regulating the timing and quantity of surface, cover reduces the proportion of rainfall that subsurface and groundwater flows, which can runs off the surface during rainfall events, impact the availability of water to users as which helps to retard flood flows and allow well as exposure to hazards like floods. During more opportunity for precipitation to infiltrate rainfall events, some water soaks into the into the soil. This can result in enhanced ground, while the balance runs off the surface baseflows and groundwater recharge with (herein referred to as ‘quickflow’). Some of healthy vegetation cover, though this benefit the former is lost through evaporation from may be negated where vegetation has high the soil or evapotranspiration by plants. Of evapotranspiration rates (Owuor et al., 2016). the remainder (herein referred to as the ‘net infiltration’), some emerges at springs to Through mediating infiltration, healthy join streams and rivers (termed ‘baseflows’), ecosystems can contribute to more sustained while some replenishes groundwater or river flows between rainfall events and reduce aquifers (termed ‘groundwater recharge’). overall seasonal variation in flows. This can affect the cost of surface or groundwater The balance between quickflow and infiltration supply by water utilities and/or the cost of varies across the landscape and is mediated collecting water (for households not supplied 42 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE by infrastructure). As runoff variability The InVEST SWY model is relatively simple increases, a larger built storage capacity is compared to many other hydrological generally required to meet water demands models. While this does result in certain during low flow seasons (for small dams) limitations, it means the model is well or drier years (for large dams, Guswa et al. suited to modelling over large scales and in 2017; Vogel et al. 1999, 2007). However, the data-poor environments. The InVEST SWY extent to which ecosystems play a role in model estimates the pixel-level contribution smoothing surface flow variability and/or of the landscape to both quickflow and contributing to groundwater replenishment infiltration at the resolution of the digital depends on a range of context-specific elevation model input (90 m in this case). factors such as slope, geology, soil type, evapotranspiration, groundwater depth, etc. The model first estimates quickflow generation on each pixel using a curve number (CN)- The groundwater recharge mediated by based approach. These were estimated ecosystems was valued in terms of the from available CN values for comparable additional expense that would be incurred to land cover class and hydrological soil group meet current demands under a no-ecosystem combinations (Baker and Miller 2013; Beatty (bare-ground) scenario. To do so, groundwater et al. 2018; SANRAL 2013; Descheemaeker et recharge was re-modelled for a scenario when al. 2008; Wischmeier and Smith 1978). Loss of all ecosystems are converted to bare ground water to evapotranspiration is modelled using and compared to current groundwater information on reference evapotranspiration recharge rates. It was assumed that the and the monthly water requirements of percentage decrease in net infiltration would different land cover types. The latter is have a similar impact on the combined yield estimated using the plant evapotranspiration from existing borehole and well infrastructure coefficient (Kc) parameter. Initial monthly in each catchment, and the deficit would Kc values for different land cover types were be made up through investment in surface guided by MODIS-derived leaf area index water infrastructure. To estimate the cost of data for different land cover types as well replacement surface water infrastructure, as by estimating the ratio of potential to we assumed a storage-yield ratio of 3.21, a actual evapotranspiration from available capital expenditure (CAPEX) of US$0.84/m3 data. The Kc estimates were then refined by and annual operating costs (OPEX) of 0.5% of comparing modelled evapotranspiration in CAPEX (based on the National Water Resources initial model runs to actual evapotranspiration masterplan), and a 15% cost of capital. derived from remote-sensing data. Information on groundwater demand in each catchment was obtained from the estimates Precipitation that does not run off as quickflow of borehole abstraction rates from the Water or get lost through evapotranspiration can Resources Master Plan (MoLAWRR, 2020a). infiltrate the soil and contribute to groundwater recharge and baseflows. However, the model does not partition infiltrating water between MODELLING FLOWS AND GROUNDWATER long-term aquifer recharge and water that RECHARGE is discharged into streams to form baseflow. This requires more detailed hydrological A rapid-level estimate of the effects of study. In general, groundwater dynamics ecosystems on flows and groundwater remain poorly studied in Zimbabwe (Davis recharge was made using the InVEST & Hirji, 2014). However, the sectoral report Seasonal Water Yield (SWY) model. This on groundwater from Zimbabwe’s National model provides an indication of which parts Water Resources Master Plan does provide of the landscape contribute the most to the estimates of overall groundwater recharge various flow components, and can be used and groundwater discharges as baseflows for to evaluate changes in flow dynamics and each of the seven catchments (MoLAWRR, groundwater under future land use scenarios. 2020a). The mean of the low and high baseflow estimates was first calculated from the MoLAWFRR (2020a) figures. The total groundwater recharge estimate for each 1  The yield of a dam is usually much less than its catchment was then divided by the mean capacity. A ratio of 2 means that the annual yield figure for baseflow discharges to estimate is half of the storage volume. The ratio of 3.2 was baseflow as a proportion of total recharge. estimated based on information in the Water Resources This varies significantly at a catchment level, Masterplan 2020-2040. ranging from just 8% of annual recharge 43 Ch 4: Baseline Assessment in the Gwayi Catchment to 58% of annual discharge estimates from MoLAWRR (2020a). recharge in the Mzingwane Catchment. These proportions were then applied to the InVEST A map showing the contribution of different model estimates to partition infiltrating parts of the country to baseflows and water between long-term groundwater groundwater recharge is shown in Figure 28 recharge and baseflow discharges. . This tends to be higher in the wetter north and east of the country, while baseflow and groundwater recharge is generally low in the ECOSYSTEM EFFECTS ON FLOWS west and south. Due to low rainfall and high evapotranspiration, many pixels in the south It was estimated that total quickflow (surface of the country were estimated to use up all runoff during or shortly after rainfall events) available water and thus make no contribution nationally is 19 466 Mm3/y, or an average value of to groundwater recharge or baseflows 50 mm/y (Table 12). Net infiltration of rainfall to (shown as pale brown areas in the map). recharge groundwater and contribute to river baseflows was estimated to be an additional 28 Estimates of quickflow, baseflow and 577 Mm3/y, or an average value of 73 mm. Based groundwater recharge across the seven on the partitioning of recharge derived from catchments are shown in Table 12. Note MoLAWRR (2020a), discharge as baseflows that the groundwater recharge estimate was estimated to be 6504 Mm3. Combined excludes the baseflow component, which with quickflow, this gives a total mean was instead added to the total flow estimate. annual runoff (MAR) estimate of 25 969 Mm3. Mean quickflow was estimated to be fairly similar across the five catchments draining The model results align well with the runoff the central, northern and eastern parts of and groundwater recharge estimates in Zimbabwe (Manyame, Mazowe, Save, Runde Zimbabwe’s Water Resources Master Plan and Sanyati), ranging from 50.7 mm/y to 79.3 reports. The total annual runoff (quickflow and mm/y (Table 12). Quickflow was notably lower baseflow) estimated here is comparable to the in the very dry Mzingwane Catchment, as current runoff estimate of 23 921 Mm3 from well as in the Gwayi Catchment. The latter has the Water Resources Master Plan (MoLAWRR, moderate to low rainfall, generally flat terrain 2020b). The total modelled estimate for and sandy soils, resulting in low quickflow. recharge (28 577 Mm3) is also very similar to the total annual recharge estimate of 29 500 Baseflow and groundwater recharge estimates Mm3 from MoLAWFRR (2020). Lastly, the varied more across catchments (Table 12). Very estimated discharge of recharge as baseflows low baseflow in the Gwayi Catchment reflects (6504 Mm3) falls within the range of the low the prevalence of deep permeable sediments in (4899 Mm3) and high (9281 Mm3) baseflow this basin, which promotes deep groundwater Table 12: Average and total quickflow, baseflow and groundwater recharge across Zimbabwe’s seven catchments estimated by the InVEST SWY model. MEAN MEAN MEAN GROUND- TOTAL GROUND- QUICKFLOW BASEFLOW WATER RECHARGE TOTAL RUNOFF WATER RECHARGE CATCHMENT (MM/Y) (MM/Y) (MM/Y) (MILLION M3/Y) (MILLION M3/Y) Gwayi 22.1 4.6 55.8 2345 4913 Manyame 64.2 23.7 132.5 3581 5398 Mazowe 79.3 46.6 84.9 5046 3403 Mzingwane 27.3 8.9 6.5 2267 408 Runde 63.8 7.0 24.7 2920 1021 Sanyati 50.7 11.3 75.5 4310 5248 Save 78.5 33.2 34.2 5498 1684 TOTAL 49.7 17.5 55.5 25 969 22 074 44 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE Figure 28. Estimated contribution to baseflow and groundwater recharge across Zimbabwe. recharge and results in low discharge of natural vegetation types (forest, dense groundwater as baseflows (MoLAWRR, 2020a). woodland) despite their location in higher Baseflow was estimated to be highest in the rainfall areas. This reflects the fact that Save and Mazowe Catchments, reflecting interception by the plant canopy and dense both relatively high rainfall and relatively ground cover result in low runoff generation high values for baseflow discharges as a in these habitats. Groundwater recharge and proportion of total recharge (MoLAWRR, baseflow generation were generally highest 2020a). Groundwater recharge rates (after for moderately dense to open natural habitats subtracting baseflow discharges) were highest located in moderate to high rainfall areas, in the relatively wetter Manyame and Mazowe such as miombo woodland and shrubland Catchments, in line with the generally strong and mesic grassland. These habitats have positive relationship between precipitation higher mean groundwater recharge and and baseflow/groundwater recharge reported baseflow than dense woodland and forest, as by other hydrological studies of Zimbabwe the denser vegetation of the latter habitats (Farquharson & Bullock, 1992; Mazvimavi et results in greater losses of water through al., 2005; World Bank Water, 2014). Conversely, transpiration by plants. Degradation of groundwater recharge was estimated to be natural habitats generally resulted in higher very low in the dry Mzingwane Catchment. quickflow and lower contributions to baseflow and groundwater recharge, as the reduction Across land cover types, quickflow was in vegetation cover results in higher runoff generally estimated to be low for dense and less opportunity for rainfall to infiltrate. 45 Ch 4: Baseline Assessment For the same reason, cultivated areas mostly particularly in the Mzingwane Catchment. exhibit much higher quickflow and lower recharge than surrounding natural habitats. Based on estimates of groundwater demand and the cost of constructing replacement surface water infrastructure, the overall value of VALUE OF THE FLOW REGULATION SERVICE the contribution of ecosystems to groundwater recharge was estimated to be US$496.3 It was estimated that groundwater recharge million/y. The value of the service at a catchment would decline significantly by 59% under a no- level reflects the combination of the absolute ecosystem (bare ground) scenario (Table 13), change in recharge under a no-ecosystem due to much higher runoff in the absence of scenario and the amount of groundwater vegetation cover. In drier catchments in the use in each catchment, ranging from US$43.1 south and west of the country, it was estimated million/y in the Manyame Catchment to that groundwater recharge would decline to US$184.4 million/y in the Save Catchment. almost zero in the absence of vegetation cover, Table 13. Difference in recharge between current land cover and a no-ecosystem (bare ground scenario). The contribution of ecosystems to groundwater recharge was valued in terms of the cost of replacement surface water inf rastructure. MEAN GROUND- % CHANGE IN WATER RECHARGE MEAN RECHARGE RECHARGE WITH VALUE OF ECOSYSTEM CURRENT BARE NO ECOSYSTEM CONTRIBUTION CATCHMENT (MM/Y) (MM/Y) CONTRIBUTION (US$ MILLION/Y) Gwayi 55.8 11.0 -80.2% 75.1 Manyame 132.5 74.1 -44.1% 43.1 Mazowe 84.9 43.8 -48.3% 53.5 Mzingwane 6.5 0.0 -99.6% 45.2 Runde 24.7 2.0 -92.0% 51.7 Sanyati 75.5 37.7 -50.1% 43.2 Save 34.2 12.4 -63.9% 184.4 TOTAL 55.5 22.6 -59.0% 496.3 EROSION CONTROL AND SEDIMENT RETENTION Vegetative cover prevents erosion by and related infrastructure. The reduction of stabilising soil and intercepting rainfall, natural vegetation cover, whether through thereby reducing its erosivity. Vegetated building roads and settlements, mining, areas may also capture the sediments that resource harvesting, grazing, agriculture, or are eroded from upstream agricultural and burning, results in elevated levels of erosion degraded lands and transported in surface and subsequent increases in sediment loads flows, preventing them from entering streams carried downstream. Globally, anthropogenic and rivers (Blumenfeld et al., 2009). Thus, sedimentation accounts for about 37% of vegetation protects downstream areas from the annual costs of reservoirs (Basson 2009). the impacts of sedimentation, which can include impacts on water storage capacity, Soil erosion has been a serious concern in hydropower generation and navigability of Zimbabwe for some time (Whitlow 1988). High rivers (Pimentel et al. 2008). While some level of soil erosion rates reduce topsoil depth as well as sedimentation of reservoirs is expected under reducing soil water content, soil organic carbon natural conditions, and planned for, elevated and removing nutrients (Roose 2008). This catchment erosion either incurs dredging costs imposes costs on farmers, who must increase or shortens the projected lifespan of reservoirs fertilizer application to replace lost nutrients. In 46 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE extreme cases, soils may become too shallow Mazowe Catchment, Zimbabwe (Figure 29). to support crop growth (Whitlow 1988). In addition to affecting agricultural productivity, the export of eroded soil to watercourses MODELLING SEDIMENT RETENTION results in siltation, which can affect river flows Soil erosion and sedimentation rates were and reduce the storage capacity of reservoirs. modelled using the InVEST Sediment Delivery Ratio (SDR) model, which estimates both The resulting sedimentation of rivers and the potential amount of erosion from the reservoirs is also a serious issue in Zimbabwe, landscape, and of this how much is exported (Makwara and Gamira 2012; Tundu, Tumbare, to watercourses as sediment. In this study, and Onema 2018). For example, there we focus on the value of sediment retention has been a 39% reduction in capacity of by ecosystems for reducing reservoir Chimhanda Dam (Tundu, Tumbare, and sedimentation. The sediment retention service Onema 2018) and a 67% loss in storage was quantified by comparing sediment export capacity for Chesa Causeway Dam (Godwin from the current landscape to one where no et al. 2011). These smaller, low-capacity dams sediment is being retained or trapped by in communal areas are particularly prone vegetation. The service was valued based on the to high siltation rates (Whitlow 1988). It cost of dredging dams to recover lost storage. has been estimated that such dams have an effective lifespan of just 15 years before The InVEST SDR model first estimates soil being filled with sediment (Magadza 1984). loss using the Revised Universal Soil Loss Notably, the impact of vegetation on reducing Equation (RUSLE). It then models sediment sedimentation rates has been demonstrated export using the sediment delivery ratio, in Zimbabwe. For example, Murwira et al. which calculates the proportion of soil loss (2014) found a strong negative relationship that ends up actually reaching a watercourse. between the degree of vegetative cover and The RUSLE component estimates erosion risk the suspended sediment loads of rivers in the based on soil erodibility (K-factor), the erosive Figure 29: Relationship between vegetation density (as indicated by NDVI) and the concentration of suspended solids in rivers in the Mazowe Catchment, Zimbabwe. Source: Murwira et al. (2014). 47 Ch 4: Baseline Assessment power of rainfall (R-factor), topography (LS- all small farm dams with small catchment factor), the extent to which land cover reduces areas, situated in the upper reaches of river erosion risk (C-factor) and the presence basins. The vast majority of these dams lie of erosion control practices (P-factor). within the mapped catchment areas of larger downstream dams, and thus the sediment The soil erodibility (K-factor) layer was retention service would still be captured. generated from soil property layers obtained from the Africa SoilGrids dataset (Hengl Once dam locations had been accurately et al. 2015). For rainfall erosivity (R), the mapped, catchment areas for all dams were Global Rainfall Erosivity Database (GloREDa) delineated using the 90 m DEM through raster was used (Panagos et al., 2015). The the InVEST DelineateIT watershed creation topographic (LS) factor is calculated internally tool. This area represents the “serviceshed” by the model using the DEM input. The land over which the sediment retention service cover management (C) component of the is demanded. Lake Kariba was excluded, RUSLE equation accounts for how different as its size means sedimentation has a land cover types reduce soil erosion, relative much lower impact on storage capacity to bare fallow areas (Wischmeier and Smith than for smaller dams. Additionally, most 1978). A C-factor value was assigned to each of the water and sediment inputs to the land cover class based on literature estimates lake originate from outside of Zimbabwe. for comparable land cover types (Wischmeier & Smith, 1978a; Angima et al., 2003; Panagos et The service was only valued within the al., 2015; Fenta et al., 2020). C-factor estimation catchment areas of existing dams. Options was also informed by the assessment of mean for valuation include estimating the cost NDVI across the different land cover classes, of preventing sedimentation of dams by given that other work has found a relationship constructing sediment check dams or between NDVI and C-factor values (Knijff et al., estimating the replacement cost of lost 1999; Almagro et al., 2019; Li et al., 2020). C-factors storage capacity through building additional were also guided by existing data, including water storage. For this study we used an a large survey of siltation in dams across the estimated cost of check dam construction, country undertaken in the 1980s (Van Den obtained from Mekonnen et al. (2015). The Wall Bake, 1985), and some additional studies volume of sediment was estimated from of specific dams (Kabell, 1984; Dalu et al., 2013). sediment mass using a density of 1.2 t/m3. Finally, the supporting practice (P) factor is used to account for erosion control measures CURRENT EROSION AND SEDIMENT EXPORT in agricultural land. Given the fairly widespread Total soil erosion across Zimbabwe was use of practices like contour ploughing and estimated to be around 367 million t/y, hedgerows on farmland in Zimbabwe, a P-factor averaging 9.4 t/ha/y. Tolerable soil loss rates value of 0.85 was used for cultivated land. vary significantly, but generally range from 1 to 12 t/ha/y across different land cover Since reservoir sedimentation is one of the types, and around 10 t/ha/y for agricultural major negative impacts of sediment export soils (Roose 1996). Modelled erosion rates to watercourses, mapping of dams and their exceeded tolerance limits in certain landcover catchment areas was conducted to identify types, particularly on bare areas (33.0 t/ the area over which the sediment retention ha/y) and cultivated lands (16.1 t/ha/y), as service is most demanded. While the GOOD2 well as degraded natural land cover types dam database (Mulligan, van Soesbergen, (for example, 20.4 t/ha/y in degraded dry and Sáenz 2020) does map the location and grasslands and 13.1 t/ha/y in degraded miombo catchment areas of 38,000 dams globally, it was shrubland). High rates of soil erosion on found that dam wall locations were not always farmland impose costs on farmers who must correct, resulting in incorrect catchment replace lost nutrients with fertilizers, while in areas. Additionally, certain dams were not extreme cases, soils may become too shallow captured in the GOOD2 dataset. Hence, dam to support crop growth (Whitlow 1988). wall locations were updated and missing dams added using a combination of GOOD2, Of the soil eroded in the country, around 43.5 HydroLAKES data (Messager et al. 2016), and million t/y were estimated to be exported to satellite imagery. The final layer included 1147 watercourses as sediment, with the remainder dams, of which 67 were manually added. While being deposited across the landscape before there are many more dams in Zimbabwe, reaching streams or reservoirs. This gives the dams that were not captured are almost 48 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE an average sediment export rate of 1.1 t/ SEDIMENT RETENTION ha/y. These results aligned well with the To calculate the contribution of current estimates of sediment exports from various land cover to reducing sediment export, the Zimbabwean catchments in the literature SDR model estimates the contribution of used for calibration (Kabell, 1984; Van Den vegetation to reducing erosion in situ on each Wall Bake, 1985; Dalu et al., 2013). Reported pixel. Additionally, it estimates the amount sediment yields reported across these study of sediment originating upslope that gets catchments ranged from 0.1 to 3.5 t/ha/y, with a trapped by downslope vegetation, before mean value of 1.6 t/ha/y across all catchments. it can reach a watercourse. Through these In comparison, mean modelled sediment yield two processes, it is estimated that current across the same catchments was 1.7 t/ha/y, vegetation cover across the country reduces which aligns well with the literature estimates. sediment export to watercourses by around 692 million t/y, an average of 17.7 t/ha/y. This Sediment export across all dam catchments figure represents the amount of sediment that was estimated to be 11.4 million t/y (1.1 t/ha/y). would enter watercourses if vegetation was not Cultivated areas are associated with higher performing any sediment retention service. sediment export rates (mean of 2.1 t/ha/y) relative to intact natural habitats, with mean sediment The amount of sediment export avoided export generally less than 1 t/ha/y in the latter. by current land cover is shown spatially in Figure 30. Areas that do not fall inside a dam Figure 30. Sediment retained and trapped by current land cover before it can reach watercourses. Note: Areas that do not drain into dams have been slightly greyed out on this map. 49 Ch 4: Baseline Assessment catchment are greyed out to highlight the Within dam catchment areas specifically, it areas where the sediment retention service is was estimated that an additional 167 million demanded (i.e. parts of the country that do fall t of sediment would be exported to dams within a dam catchment). As seen on the map, each year if no sediment retention service dams are mostly located towards the centre was being provided by the landscape (Table and north of the country where most of the 14). This is an average sediment retention commercial farmland is located. Topography rate of 16 t/ha/y. Dense woody habitats that is a major factor determining the potential typically occur in high rainfall, hilly areas are for sediment export, with the highest values particularly important for retaining sediment. for sediment retention associated with hilly For example, indigenous forest and dense areas under natural land cover. Sediment miombo woodland reduce sediment export export from these areas would be very high in to dams by some 40 t/ha/y on average. the absence of vegetation cover. Within dam catchment areas, this is evident for densely wooded hilly terrain south of Masvingo and VALUE OF SEDIMENT RETENTION the Matobo Hills south of Bulawayo, as well Based on the cost of dam dredging, the as the Great Dyke to the northwest of Harare value of the sediment retention service (Figure 30). Sediment retention is also high within dam catchment areas was estimated across much of the northeast of the country. to be worth US$208.3 million/y or US$20.0/ However, only a relatively small portion of the ha (Table 14). The total value of the service landscape falls within dam catchments here. In is highest in the Runde Catchment where contrast, sediment retention is relatively low in it was estimated that ecosystems result in flat areas in the west and south of the country, cost savings of US$52.8 million/y. However, as well as the wide Zambezi Valley floor in the value per unit area was highest in the far north. These areas also have low agricultural Mazowe Catchment (US$37.2/ha/y), which potential and limited dam development, has relatively high rainfall and hilly terrain, resulting in low demand for the service resulting in high sediment retention rates. from a reservoir sedimentation perspective. Table 14. Total sediment retained by the landscape within dam catchment areas, and value of the sediment retention service based on avoided check dam construction costs. TOTAL SEDIMENT MEAN SEDIMENT RETENTION RETENTION AVOIDED COSTS AVOIDED COSTS CATCHMENT (MILLION T/Y) (T/HA/Y) (US$ MILLION/Y) (US$/HA/Y) Gwayi 5.5 6.2 6.9 7.7 Manyame 22.0 19.0 27.4 23.8 Mazowe 26.8 29.8 33.5 37.2 Mzingwane 39.1 14.0 48.9 17.5 Runde 42.2 21.3 52.8 26.6 Sanyati 12.4 6.4 15.5 8.1 Save 18.5 24.5 23.2 30.6 TOTAL 166.6 16.0 208.3 20.0 50 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE SUMMARY OF ECOSYSTEM VALUES AND THEIR BENEFICIARIES The services provided by ecosystems across benefit from the opportunities for recreational, Zimbabwe benefit a range of stakeholders, as cultural, or spiritual fulfilment offered by the summarised in Table 15. The overall value of country’s natural assets. Tourism associated the selected ecosystem services to Zimbabwe with visiting and experiencing natural was estimated to be around US$2.4 billion/y. assets was estimated to be worth US$229.7 When the global value of avoided climate million/y. This value is linked to the extent change-related damages to the rest of the of road infrastructure and tourism facilities world is included, the value generated by as well as attractive scenery and wildlife. Zimbabwe’s ecosystems increases to around US$11.5 billion/y. The way in which these The water sector is also a major beneficiary. benefits are distributed among the different In this study, it was estimated that sediment stakeholders is determined by the use of the retention by ecosystems generates cost landscape, the resulting balance between savings of US$208.3 million/y from avoided natural and transformed land cover types, dam sedimentation risk alone. Further and the condition of those land cover types. benefits may be obtained where water is treated for use. Additionally, it was estimated Commercial farmers and rural households, that the mediation of groundwater recharge who make up most of Zimbabwe’s by ecosystems has a value in the region of population, enjoy the greatest aggregate US$496.3 million/y, based on the cost of benefits from ecosystems. These include constructing replacement surface water agriculture and livestock production, worth infrastructure as an alternative water source. an estimated US$365.4 and US$291.7 million/y, respectively (Table 15). Rural households Finally, there are global benefits from the also harvest a value of at least US$576.1 storage and sequestration of carbon in the million/y from the use of natural resources landscape, which helps avoid further climate provided by ecosystems. Since only a change damages, potentially worth billions selection of resources were included in the of dollars. Zimbabwe accounts for a small assessment, the total, value of wild resource share of this value, with local benefits of the harvesting is likely to be significantly higher. carbon storage and sequestration service estimated to be around US$187.8 million/y. In addition, all the inhabitants of Zimbabwe 51 Ch 4: Baseline Assessment Table 15: Summary of the current values of selected ecosystem services assessed in this study in US$ millions per year, and the actors who benefit from each service. Note that value to the rest of the world has some potential to be captured by Zimbabwe. VALUE PER YEAR TYPES OF SERVICES EXPLANATION VALUE TO WHOM (US$, MILLIONS) Cultivated produc- Commercial farmers Production value net of human inputs 365.4 tion and rural households Communal farmers 171.6 Livestock production Production value net of human inputs Commercial farmers 120.1 Value of wild harvested foods, fuel, and Wild resources Rural households 576.1 raw materials net of human inputs Cost savings due to vegetation capacity Water utilities and Sediment regulation to hold soil in place or trap eroded soils 208.3 private dam owners before entering streams Flow regulation (con- Cost savings in water resources infra- Water utilities and/or tribution to baseflows structure due to facilitation of recharge 496.3 direct water users and groundwater) by vegetation Nature-based Net income generated as a result of Tourism sector 229.7 tourism tourism to ecosystem-based attractions Avoided climate-change damages as a Zimbabwe 187.8 Carbon sequestration result of CO2 uptake and avoided CO2 and retention emissions Rest of world* 9192.4 52 5. SCENARIO ANALYSIS OVERVIEW The previous chapter of the report quantified Scenario, based on the extrapolation of past key ecosystem services provided by landscapes changes in ecosystem extent and condition, across Zimbabwe in their current state. and a Resilient Future Scenario, in which Although the current value of these services landscape interventions to restore, maintain, is substantial, ecosystem modification and or enhance the flow of ecosystem services degradation mean that ecosystem services from natural and cultivated lands are fall well short of their full potential value. implemented. The section concludes with a high-level cost benefit analysis to evaluate This chapter estimated future changes in the potential return on investment (ROI) the delivery of services under two alternative from the Resilient Scenario interventions. future scenarios: a Business As Usual (BAU) KEY POINTS ∙ Under a BAU Scenario, cultivation expands significantly, increasing from 24% coverage to 30% by 2050. Built-up areas triple in area to cover 0.6% of the country’s area at this point. Natural land cover classes decline from 74% currently to 68% under BAU with the coverage of degraded natural habitats increasing significantly from 9.6% to 19.5%. ∙ The BAU scenario sees significant negative impacts on ecosystem services. ∙ Under the Resilient Scenario, a range of agricultural and natural habitat restoration interventions are implemented: climate-smart agriculture (CSA), restoration of riparian buffers that have been lost to cultivation, the passive restoration of degraded natural habitats, management of resource harvesting pressures, and the improvement of community conservation and state protected areas. ∙ CSA could increase crop revenues from by US$154.8 million/y compared to the BAU scenario, despite a significant reduction in farmland area under the Resilient Scenario. ∙ Restoration of riparian buffers and degraded natural habitats could increase the value of wild resource harvesting by US$13.6 million/y. ∙ Increased carbon storage and sequestration could reduce local climate change-related damages by US$33.5 million/y and damages to the rest of the world by US$1.64 billion/y, compared to the BAU scenario. The potential revenue that could be generated through carbon credits was estimated to be US$55.9 million/y. ∙ The expansion of conservancies and the improvement of management and tourism facilities within the protected area estate could increase the value of nature-based tourism by US$52 million/y compared to the BAU scenario. ∙ Collectively, the interventions could result in an increase in groundwater recharge worth around US$169.0 million/y and avoided reservoir sedimentation costs of US$11.2 million/y compared to the BAU scenario. ∙ Overall, the benefits of following a resilient development pathway could amount to US$445 million/year locally and some US$2.09 billion/y globally, compared to BAU. ∙ Based on a high-level cost-benefit analysis, it was estimated that every dollar invested in the Resilient Scenario would return US$3.3 in local ecosystem service benefits, and US$9.4 in global benefits. Local ROI is positive in all catchments, ranging from 1.5 in the Runde Catchment to 4.7 in the Mazowe Catchment. 53 Ch 5: Scenario Analysis DESIGN AND GENERATION OF TWO FUTURE SCENARIOS THE BUSINESS AS USUAL (BAU) SCENARIO The impact of a BAU trajectory on ecosystem this is a highly unlikely land cover transition. service values was modelled by the extrapolation of land cover change and natural In addition to modelling expansion of habitat degradation trends (as measured by settlement and cultivation into natural NDVI), projected to 2050. Changes in land habitats, the expansion of degradation cover extent were obtained from Zimbabwe’s in remaining natural habitats was also national land cover assessments in 1992 and estimated. Rates of degradation were first 2017 produced by the Forestry Commission. estimated using Trends.Earth (Conservation While the raw GIS data is not available for International, 2018), which was used to public use, figures giving the extent of the identify areas of the country which exhibited various land cover categories in both years are a significant decline in NDVI between 2001 publicly available. The proportional increase and 2021, after accounting for the effects of in cultivated and built-up area between rainfall variability. The extent of degradation 1992 and 2017 was thus obtained from in this period was then used to project the these data and projected forward to 2050. extent of degradation by 2050, assuming a similar rate of degradation in the future. The The estimated expansion of cultivation and estimated expansion of degraded natural settlement was then modelled spatially habitats was then modelled spatially, again using the InVEST Scenario Generator model using the InVEST Scenario Generator model. in a stepwise manner. Urban expansion was modelled first, as it was assumed that settlement A map highlighting areas of change derived can expand into both natural and cultivated from the BAU trajectory is shown in Figure land cover. Cultivation was then expanded 31, depicting areas of new degradation, into remaining natural land cover but was not cultivation and urban expansion relative to permitted to expand into settlement, since current land cover. The percentage coverage of Table 16. Comparison of the percentage coverage of simplified land cover classes between current land cover and the BAU Scenario. LAND COVER CURRENT BAU Indigenous forest 0.2% 0.2% Plantations 0.3% 0.3% Dense woodland 3.8% 3.6% Woodland 37.6% 29.1% Shrubland 21.8% 14.5% Grassland 1.3% 1.0% Degraded shrubland 3.4% 7.2% Degraded woodland 6.0% 12.0% Degraded grassland 0.2% 0.4% Cultivation 24.0% 29.9% Bare 0.0% 0.0% Open water 1.1% 1.1% Built-up 0.2% 0.6% 54 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE different land cover classes is given in Table 16. Hence an overall decline in the extent of natural habitats was projected, coupled Cultivation expands significantly under with a large increase in the proportion BAU, increasing from current coverage of of natural habitat in a degraded state. 24% of the country to 30% by 2050. Built- up areas expand rapidly to triple in size. These changes would be contrary to However, they still account for a small Zimbabwe’s Land Degradation Neutrality 0.6% of the country’s area under BAU. (LDN) commitments, which include an ambitious goal of reforestation of 6.4 million The expansion of cultivation and settlement ha of degraded forest (16.5% of national area) results in a decline in the overall coverage of and the return of 215 000 ha of cropland back natural land cover classes from 74% currently to forest by 2030. The projected changes to 68% under BAU. Out of this, the coverage highlight that significant effort will be required of degraded natural habitats increases to reverse the BAU trajectory if Zimbabwe is significantly from 9.6% to 19.5% under BAU. to achieve its aspirational LDN commitments. Figure 31. Map highlighting the projected expansion of degraded natural habitats and new areas under cultivation and built-up land cover under BAU by 2050. 55 Ch 5: Scenario Analysis THE RESILIENT FUTURE SCENARIO OVERVIEW capital. For example, CSA practices and irrigation should increase agricultural The Resilient Scenario assumes the productivity per unit area, thus reducing the implementation of a suite of landscape pressure to convert natural areas to farmland. interventions that restore and/or maintain If such practices were applied well, the rate of vegetation cover, soil retention, biodiversity, loss of natural ecosystems within Zimbabwe and agricultural productivity. These could be reduced. The recovery of rangelands mutually supportive interventions include (a) would ensure the provision of ecosystem supporting, regulating, and/or incentivising benefits into the longer term, including during climate-smart agriculture (CSA) practices and times of economic shocks or climate stress. the rehabilitation and sustainable expansion The restoration of natural areas would also of irrigation infrastructure; (b) limiting grazing help reduce erosion and sedimentation rates, and the use of wild resources to sustainable while potentially enhancing infiltration and levels, to maintain their productivity and groundwater recharge. Ecosystem restoration flows of other services; and (c) restoring and interventions also provide alternative income protecting key natural areas important for opportunities, such as from biodiversity biodiversity and ecosystem services (Figure tourism, carbon credits, and payments for 32). The selected interventions were guided by hydrological services, some of which could be and are consistent with Zimbabwe’s various reinvested in land and resource management. environmental commitments as articulated in the Climate-Smart Agricultural Investment A number of policy and supporting measures Plan (CSAIP), Revised Nationally Determined would need to be implemented to achieve these Contributions (NDCs), National Adaptation broad ecosystem management objectives Plan (NAP), Vision 2030 and LDN targets. (Figure 32). CSA measures are generally a win-win solution for both farmers (through These actions would work synergistically increased productivity) and ecosystems and towards maintaining and enhancing the can be self-sustaining after initial input. Other benefits derived from Zimbabwe’s natural measures, such as reducing unsustainable Figure 32: The three broad goals of interventions to achieve sustainable use of landscapes in order to derive maximal benefit, and some of the measures that can be used to achieve them (Turpie et al., 2022a) 56 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE resource use may require more complex and drought-tolerant varieties were obtained longer-term interventions. The interventions from the CSAIP, ranging from 3% for soya to are discussed at length in Turpie et al., 2022a. 19% for tobacco. These represent the yield gains, relative to using non-drought-tolerant The resilient scenario involved modelling varieties, under future climate conditions. the effects of a range of interventions to improve land and resource management It was assumed that additional yield practices. This includes enhancing and improvements relative to BAU would also upscaling management actions that are occur with increased adoption of reduced/no already being implemented or planned, till and other CA practices. Estimates of the e.g. the government’s Pfumvudza/Intwasa benefits of CA vary widely across Zimbabwean programme, efforts to rehabilitate and expand studies. For example, within 16 project districts, irrigated agriculture and the establishment average maize yields in Pfumvudza plots (4.19 of community conservancies for wildlife. The t/ha) were almost twice as high as those from assumptions are outlined below, including non-Pfumvudza plots (2.27 t/ha) (IAPRI & FAO, the establishment and maintenance 2021). These yield gains are likely due in part to costs used in the cost-benefit analysis. improved extension services and the provision of inputs to Pfumvudza plots, as well as to the CA practices themselves. Other studies CLIMATE-SMART AGRICULTURE report more moderate yield benefits from CA, ranging from 12% to 59% when averaged The adoption of Climate-smart agriculture spatially and temporally (Marongwe et al., 2011; (CSA) has the potential to address both Thierfelder, Cheesman & Rusinamhodzi, 2012). low productivity and land degradation, Based on the average of yield gains reported in particularly in the face of declining crop yields the literature, it was estimated that adoption predicted under climate change. Recognising of CA would increase yields by 34%. To account this, a detailed Climate Smart Agriculture for the fact that some farmers already practice Investment Plan (CSAIP) has already been CA and the fact that universal adoption is prepared for Zimbabwe (World Bank 2019). unlikely, it was assumed that CA adoption CSA is a broad term which encompasses would increase by 50% in calculating the a wide range of practices, including overall yield benefit, relative to BAU. This results conservation agriculture (CA) (which reduces in an aggregate yield improvement of 17%. soil and water losses), agroforestry, sustainable expansion of irrigation improved livestock Irrigation has the potential to significantly fodder production, rainwater harvesting, increase crop yields, particularly under future and soil conservation infrastructure. climate conditions. According to the CSAIP, yield improvement factors with irrigation under For this study, it was assumed that CSA will climate change range from 1.86 for tobacco to include the use of drought-tolerant crop as high as 3.69 for cotton (World Bank, 2019b). varieties and the increased adoption of Based on Zimbabwe’s Vision 2030 NDS1, it was reduced/no tillage approaches and other assumed that around 140 000 ha of farmland conservation agriculture practices. Zimbabwe would benefit from the rehabilitation and already officially promotes CA practices expansion of irrigation infrastructure under through the Pfumvudza/Intwasa programme, the Resilient Scenario. This roughly doubles and has committed to the implementation of the current area under functional, formal CA on 360,000 ha of cropland and 1.1 million ha irrigation (CIAT & World Bank, 2017). However, of degraded arable land under the country’s this still amounts to just 8.4% of total planted LDN commitments. In addition, it was assumed area (over 3 million ha) of the eleven main that Zimbabwe’s plans to rehabilitate and crops as reported in the Crop and Livestock extend irrigation infrastructure, as detailed Assessment Reports (MoLAWFRR, 2020, 2021, under the Vision 2030 National Development 2022). Yield improvement factors under the Strategy 1 (NDS1), are fully realised. Resilient Scenario were thus derived from the proportional increase in future irrigated According to Zimbabwe’s CSAIP, switching area and the irrigation yield improvement to drought-tolerant crop varieties will help factors estimated in the CSAIP (World Bank, reduce yield losses under future climatic 2019b). Given that the irrigation still covers conditions, but will not be sufficient on its a relatively small area under the Resilient own to avoid overall yield losses relative Scenario, aggregate yield benefits were to current conditions (World Bank, 2019b). estimated to range from 4% - 12% for each crop. Improvement factors with the adoption of 57 Ch 5: Scenario Analysis The same initial impacts of climate change precise level of payment varying based on on crop yields as per the BAU scenario were the current levels of wood harvesting and used for the Resilient Scenario, based on value of agriculture in each catchment. the CSAIP projections (World Bank, 2019b). Additionally, it was assumed that the rollout of The estimated yield improvement factors subsidised efficient cookstove technologies to for each crop through the use of drought- further reduce pressure on woody resources tolerant varieties, increased adoption of would have a cost of US$15/household. CA and expansion of irrigation were then applied to these initial future yield estimates. CONSERVANCIES AND IMPROVED PRO- For the purposes of the cost-benefit TECTED AREA MANAGEMENT AND TOURISM analysis, the establishment cost was for CSA FACILITIES interventions was derived from the estimate Opportunities for community conservation for a similar suite of interventions in the CSAIP, development were identified. These included which was given as US$95 – 125 per household the plans to establish and improve community (World Bank, 2019b). Based on the estimated conservancies in the middle Zambezi Valley, number of participating households and adjacent to Gonarezhou NP, and in the intervention extent, this was translated to a Zambezi Valley between Victoria Falls and cost of US$100/ha of farmland planted under Lake Kariba. This scenario also includes the CSA. This includes the costs of extension and improved management and upgrading training and costs of equipment and inputs of tourism facilities in state protected required for CSA (for example, planting areas, as envisaged under the Vision 2030. equipment to reduce labor burdens of reduced and no tillage farming). Once they Costs for these interventions were derived are established and with suitable equipment, from Vision 2030, which set aside US$2.4 most of the CSA practices should be largely million for the establishment of community self-sustaining. Hence, maintenance costs conservancies and US$22 million for upgrading were assumed to be modest at around US$15/ infrastructure and tourist facilities in state ha. This covers the cost of ongoing improved parks. This amounts to an establishment extension support, as well as some incentives cost of around US$6.2/ha for community for farmers to not expand their fields. conservancies and US$4.4/ha for improving state protected areas. Once established, it RECOVERY AND SUSTAINABLE USE OF is assumed that communities will cover the RANGELANDS AND HARVESTED RESOURC- costs of conservancy management out of ES the revenues (rent and royalty) received from joint-venture partners. Nevertheless, ongoing Degraded habitats throughout the country are support would be required. Based on figures restored to a more natural and productive state from Namibia, this was assumed to be around through improved rangeland management US$0.4/ha/year (MEFT/NASCO 2021). Ongoing and resource harvesting controls. This is costs for maintaining improved visitor facilities incentivised through provision of support for in state protected areas were assumed to be the formation of community conservancies and 10% of establishment costs, thus also resulting through the implementation of a payments in an ongoing cost of around US$4/ha/year. for ecosystem services scheme. These measures could be accompanied by measures to reduce demand for woody resources, such RIPARIAN BUFFERS as through the rollout of subsidised efficient The legal protection of the 30 m riparian cookstove technologies to rural households. buffer zone is enforced along all streams and rivers throughout the study area. This The cost of setting up such a payments for involves the cessation of existing cultivation ecosystem services scheme was estimated and mining activities within these areas. at US$10/ha (Bond et al. 2010). The aim of Degraded and cultivated areas are allowed to the scheme would be to partially offset the regenerate to riparian woodland vegetation. opportunity costs associated with reducing firewood harvesting and grazing, as well It was estimated that a mix of passive and active as the opportunity costs associated with restoration in the riparian zone would have an not converting these areas to farmland. establishment cost of around US$200/ha in the The average annual payment required was first year or US$1,200/km of river (Brancalion estimated to be around US$16/ha, with the et al. 2019; Dugan 2011). Costs would gradually 58 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE decline over the next two years to reach a long- buffers are allowed to recover and is based on term maintenance cost of US$30/ha/year or cost estimates for guarding and patrolling US$180/km of river. This captures the costs of nature reserves (Chardonnet 2019; Frazee monitoring and enforcement to ensure riparian et al. 2003; James, Green, and Paine 1999). 59 Ch 5: Scenario Analysis IMPACTS ON ECOSYSTEM SERVICES CROP PRODUCTION BAU SCENARIO that sugar cane production would be similar under BAU despite the increase in suitability. Both the expansion of farmland and predicted impacts of climate change on yields were Using the available information on predicted considered in the estimation of the crop crop yields under the “medium” future climate production service under BAU. Modelling of scenario (World Bank, 2019b), total crop crop yields under a “medium” future climate production in physical terms was predicted scenario undertaken as part of Zimbabwe’s to decline by 3.8% under BAU. This is despite Climate Smart Agriculture Investment Plan the projected 14% increase in farmland area, (CSAIP) predicted that yields of nine of suggesting a shift towards ever more extensive, the ten crops included in the study would low productivity agriculture. Furthermore, this decline significantly under climate change figure includes sugar cane production, which (World Bank, 2019b). This includes a 33% was assumed to remain stable under BAU yield decline for maize, the staple food crop, and has a disproportionate impact on total under a “medium” climate scenario.1 In crop production estimates in physical terms. contrast, the CSAIP indicated that warmer When sugar cane is excluded, total production conditions will increase suitability for sugar across the ten remaining crops declines cane, as cold stress is a major limiting factor by 12% under BAU. Since crop production (World Bank, 2019b). However, since sugar declines under BAU despite the expansion cane is reliant on extensive irrigation, the of farmland area, production decreases are CSAIP also notes that additional irrigation much higher when expressed in terms of infrastructure would be required to take production per hectare. When analysed in per advantage of the increase in suitability. Given hectare terms, crop production was projected this and the fact that sugar cane production to decline by 29.5% under BAU (Table 17). has generally declined since 2000 (Madzokere, Mutambara & Zirenga, 2018), it was assumed Notably, food crop production was projected to decline by 12.7% relative to current. These estimates raise serious concerns for 1  This was the intermediate scenario in the Climate future food security, given that Zimbabwe’s Smart Agriculture Investment Plan, lying between the population continues to grow at a rate of more extreme hot and dry future scenario and a more around 2% per annum (ZIMSTAT, 2015) and favourable wet future scenario. Table 17. Projected changes in the value of the crop production service under BAU, relative to current. CROP PRODUCTION CROP PRODUCTION MEAN CROP MEAN CROP CURRENT BAU PRODUCTION CURRENT PRODUCTION BAU CATCHMENT (US$ MILLION /YEAR) (US$ MILLION /YEAR) (US$/HA) (US$/HA) Gwayi 6.2 5.9 5.7 4.0 Manyame 50.3 43.7 51.1 36.7 Mazowe 72.9 65.2 49.4 35.6 Mzingwane 13.2 11.9 14.9 10.8 Runde 84.2 82.3 69.7 56.4 Sanyati 54.1 49.1 28.8 20.5 Save 84.5 78.6 44.8 34.7 TOTAL 365.4 336.7 38.8 28.8 60 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE the country already experiences food scarcity occurs under the Resilient Scenario relative challenges (World Bank, 2019b; Runganga to current land cover, since no further & Mhaka, 2021). Food production per capita conversion of natural habitats occurs and will thus be drastically lower under BAU. existing farmland within riparian buffers is (World Bank, 2019b; Runganga & Mhaka, 2021). restored to natural vegetation. This results in the conversion of around 306 000 ha of In monetary terms, it was estimated that the existing cultivation to natural riparian buffers. value of the crop production service would Relative to the BAU scenario, farmland area decline by 8% under BAU, representing a is around 20% lower in the Resilient Scenario. loss in revenue of US$191 million/year relative to current. The resource rent value of crop Despite the much smaller farmland area, production (i.e. the ecosystem contribution) was the sustainable expansion of irrigation and estimated to decline by of US$28.7 million/year application of CSA practices results in a 2.6% relative to the current value (Table 17). Notably, increase in total crop production by weight the value of crop production per hectare of in the Resilient Scenario, relative to BAU. farmland declines quite substantially by 26%. When the disproportionate effect of sugar cane is excluded (assumed to remain at a similar level of production in both scenarios), RESILIENT SCENARIO production across the ten remaining crops was estimated to increase by 9.2% under the An overall reduction in farmland area Resilient Scenario, relative to BAU. Notably, Figure 33. Changes in the value of the crop production service under the Resilient Scenario, relative to BAU. 61 Ch 5: Scenario Analysis crop production per unit area was estimated million/y under the Resilient Scenario. Again, to increase by 31.8% relative to BAU, indicating the increase in value per hectare of farmland much more efficient use of agricultural land. is much greater, rising by 37.3% to US$39.5/ At the same time, this will reduce the pressure ha/y under the Resilient Scenario, compared to convert natural ecosystems to agriculture, to US$28.8/ha/y under the BAU Scenario. and thus mitigate the degradation of the ecosystem services they provide relative to BAU. The highest increases in the value of the crop production service under the Resilient Scenario The overall value of the crop production are associated with catchments in the north service was estimated to increase by 6.9% and east of the country (Figure 33), where under the Resilient Scenario, relative to agricultural productivity is generally higher to BAU. This represents an increase in crop begin with. The largest benefit occurs in the production of US$154.8 million/y in gross Mazowe Catchment, where it was estimated revenue, while the ecosystem contribution to that revenue from crop production would crop production was estimated to increase increase by US$43.9 million/y, relative to BAU. by US$23.2 million/y to reach US$359.9 HARVESTED WILD RESOURCES BAU SCENARIO BAU, given their physical supply constraints. As a result of further loss and degradation Declines in value of resource harvesting of natural habitats, the total value of the six ranged from US$0.4 million/y in groups of harvested resources considered in Mzingwane Catchment to US$2.5 million/y this study was estimated to decline by US$10.3 in Manyame Catchment (Table 18). million/y (Table 18). This represents a loss in value of 2% relative to current value, which, given projected population growth, represents RESILIENT SCENARIO a significant loss in terms of value per capita. The availability of harvested wild resources The extent to which harvesting was predicted was higher under the Resilient Scenario due to decline varied by resource. For example, to the recovery of degraded natural habitats, thatching grass stocks generally exceeded the the reduction in agricultural expansion and amounts demanded in most areas, resulting the sustainable harvesting of resources from in minimal change in value. However, more riparian buffers. Through these measures, it was significant declines were predicted for estimated that harvesting of the selected wild resources like mushrooms and honey under resources would increase by US$13.6 million/y Table 18. Changes in the value of wild resource harvesting under the BAU Scenario. CURRENT VALUE VALUE UNDER BAU CHANGE IN VALUE UNDER BAU CATCHMENT (US$ MILLION /Y) (US$ MILLION /YEAR) (US$ MILLION /Y) Gwayi 54.3 53.6 -0.7 Manyame 64.3 61.8 -2.5 Mazowe 105.7 104.1 -1.6 Mzingwane 45.6 45.2 -0.4 Runde 77.9 76.7 -1.2 Sanyati 104.2 102.3 -1.9 Save 124.1 122.0 -2.1 TOTAL 576.1 565.8 -10.3 62 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE under the Resilient Scenario, relative to BAU. populated Mzingwane and Gwayi Catchments (less than US$1 million/y), while the highest At a catchment level, the gains in the value of the increase was estimated for the Manyame service are relatively low in the more sparsely Catchment (US$3.9 million/y; Figure 34). Figure 34. Changes in the value of wild resources under the Resilient Scenario, relative to BAU. CARBON STORAGE AND SEQUESTRATION BAU SCENARIO capacity. This would result in the release of 168 million tCO2e into the atmosphere, while Carbon storage and sequestration were reducing the amount of carbon sequestered predicted to decline under BAU. This reflects each year by Zimbabwean ecosystems the release of carbon to the atmosphere with by 29 million tCO2e. These reductions will the clearance and degradation of natural contribute to climate change impacts habitats (thus reducing carbon storage) both in Zimbabwe and the world at large. and the reduced capacity of remaining ecosystems to uptake carbon (thus reducing The largest losses of carbon biomass were carbon sequestration). Relative to current projected for the Sanyati and Save Catchments land cover, carbon storage in vegetation (declines of 7.1% and 7.0%, respectively). The biomass (above and belowground) across Gwayi Catchment was projected to have the the country was predicted to decline by 5.7% lowest decline in carbon biomass (3.8% loss), under BAU (Table 19), accompanied by a reflecting the fact that a significant portion 9.3% decline in annual carbon sequestration 63 Ch 5: Scenario Analysis Table 19. Changes in carbon storage and sequestration between current land cover and the BAU Scenario. BIOMASS CARBON BIOMASS CARBON SEQ. CARBON SEQ. BASELINE CARBON BAU % CHANGE BASELINE BAU % CHANGE CATCHMENT (MILLION T) (MILLION T) STOCK (MILLION T/YEAR) (MILLION T/YEAR) SEQ. Gwayi 186.8 179.7 -3,8% 20.8 19.6 -5.6% Manyame 90.0 85.2 -5.3% 9.0 8.2 -8.5% Mazowe 92.2 86.0 -6.7% 9.3 7.9 -15.3% Mzingwane 120.0 113.4 -5.5% 12.4 11.6 -6.3% Runde 76.9 72.9 -5.2% 8.2 7.4 -9.6% Sanyati 139.5 129.6 -7.1% 14.5 13.0 -10.7% Save 102.7 95.5 -7.0% 12.1 10.5 -12.9% TOTAL 808.1 762.3 -5.7% 86.3 78.3 -9.3% of the catchment falls within protected areas, in avoided local climate-change related resulting in less conversion of natural habitats. damages by 2050, and US$330.5 million/y in Declines in sequestration were estimated to be avoided damages for the rest of the world. highest in the Mazowe (15.3%) and Save (12.9%) Catchments, with the smallest reduction By helping to reduce global climate change again being in the Gwayi Catchment (5.6%). impacts, the Resilient Scenario interventions also have the potential to generate significant revenue through carbon credit sales. This RESILIENT SCENARIO provides a measure of the potential direct financial gain that could accrue to Zimbabwe Carbon storage and sequestration are from providing these global benefits. increased in the Resilient Scenario by the restoration of degraded natural habitats, Using a relatively low estimate of US$4.5 per reduction in natural habitat loss and tCO2e (Ecosystem Marketplace 2021), it was conversion of cultivation in riparian zones estimated that enhanced carbon storage and to riparian woodland. It was estimated that sequestration under the Resilient Scenario carbon storage would increase by 10% and relative to BAU could generate a total of sequestration by 12% by 2050, relative to the US$1.40 billion in carbon credits for Zimbabwe BAU Scenario (Table 20). This would increase at current prices. Given that studies have total carbon storage by 75.4 million tC, and the shown full recovery of biomass in miombo amount of carbon sequestered by ecosystems ecosystems takes around 25 years (Kalaba et by 9.2 million tC/year (Table 20), relative to BAU. al. 2013; Williams et al. 2008), the potential annual value of carbon credits generated by Through these increases in carbon storage the recovery of natural vegetation would be and sequestration, it was estimated that around US$55.9 million/y. Notably, the value around US$1.68 billion/y in global climate that could be generated through carbon change-related damages would be avoided credit sales would increase significantly with under the Resilient Scenario by 2050, based likely future increases in carbon credit pricing. on the values projected by World Bank (2017). Of this, the value of local avoided The potential value that could be generated climate change damages to Zimbabwe through carbon credit sales is shown at the would be around US$33.5 million/y, with the catchment level in Figure 30. This exceeds remainder of US$1.64 billion/y accruing to US$5 million/year in all catchments, with the the rest of the world (Table 20). The largest highest potential value in the Sanyati and value gains occur in the Sanyati Catchment, Save Catchments, where it was estimated that where enhanced carbon storage and restoration could generate US$11.3 million/y and sequestration would generate US$6.7 million/y US$9.2 million/y in carbon credits, respectively. 64 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE Table 20. Changes in carbon storage and sequestration by 2050 under the Resilient Scenario relative to BAU, and the annual value of this carbon in terms of avoided climate-change related losses to Zimbabwe and to the rest of the world (ROW). INCREASE LOCAL VALUE OF IN CARBON INCREASE IN AVOIDED CLIMATE- CARBON SEQ. CARBON SEQ. STORAGE CARBON SEQ. CHANGE RELATED LOSSES BASELINE BAU CATCHMENT (MILLION T) (MILLION T/YEAR) (US$ MILLION/Y) (MILLION T/YEAR) (MILLION T/YEAR) Gwayi 12.0 1.4 5.3 261.2 6.8% Manyame 8.3 0.8 3.6 176.6 9.7% Mazowe 8.7 1.5 4.0 198.4 10.9% Mzingwane 11.6 1.1 5.0 246.4 10.2% Runde 7.2 0.9 3.2 157.8 10.1% Sanyati 15.3 1.7 6.7 330.5 12.0% Save 12.2 1.8 5.5 270.5 13.2% TOTAL 75.4 9.92 33.5 1 641.6 10.1% Figure 35. Changes in the value of carbon storage under the Resilient Scenario relative to BAU, valued in terms of carbon credits for Zimbabwe 65 Ch 5: Scenario Analysis FLOW REGULATION BAU SCENARIO However, given the simplicity of the model and lack of detailed data for calibration, this The further conversion of natural habitats should be interpreted as a rough estimate. to agriculture and settlement, along with increased thinning and degradation of remaining natural habitats, results in a 20.1% RESILIENT SCENARIO increase in quickflow under the BAU Scenario, relative to current conditions (Table 21). While Restoring natural habitats and increasing higher quickflow runoff increases short-term soil and water conservation techniques on river flows, this is not necessarily beneficial, as it farmland through CSA practices contribute means river flows are flashier and less sustained to an increase in groundwater recharge of between rainfall events. It also increases 12.4% under the Resilient Scenario, relative to flood risks. At the catchment level, projected BAU. This amounts to an estimated 2600 Mm3 increases in quickflow under BAU were increase in annual groundwater recharge, fairly similar, ranging from 16.5% in the Save though the simplicity of the model and Catchment to 27.9% in the Gwayi Catchment. limited data validation means this should be regarded as a rough estimate of the restoration Conversely, baseflow and groundwater benefit. At a catchment level, the increases recharge were predicted to decline by in groundwater recharge are quite variable, 4.9% (Table 21), since increased quickflow ranging from 5% in the Gwayi Catchment runoff means less water is available for to as high as 23% in the Runde Catchment. infiltration. However, increased quickflow is partially compensated for by lower Based on levels of groundwater demand evapotranspiration rates when natural and the cost of constructing surface water vegetation is cleared or degraded. As a result, infrastructure, enhanced groundwater the decline in baseflow and groundwater recharge under the Resilient Scenario recharge under BAU is proportionally was estimated to have a value of US$124.5 smaller than the increase in quickflow. million/y. The benefits are highest for catchments in the north and east of the At a catchment level, declines in baseflow country (Figure 36), with the highest value and groundwater recharge under BAU of US$62.1 million/y for the Save Catchment. ranged from 3.5% in the Gwayi and Manyame This reflects both the significant increase in Catchments to 8.1% in the Runde Catchment. recharge under the Resilient Scenario (22%) In absolute terms, it was estimated that this for this catchment as well as relatively high would result in a decline in groundwater demand for groundwater in this region. recharge nationally by around 1052 Mm3/y. 66 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE Table 21. Relative changes in quickflow, baseflows and groundwater recharge under the BAU scenario. % CHANGE IN QUICK- % CHANGE IN BASEFLOW RECHARGE LOSS CATCHMENT FLOW AND RECHARGE (MILLION M3/Y) Gwayi 27.9% -3.5% -173.5 Manyame 19.6% -3.5% -189.8 Mazowe 19.0% -6.1% -208.7 Mzingwane 21.9% -4.6% -18.7 Runde 16.7% -8.1% -82.3 Sanyati 22.7% -5.1% -266.2 Save 16.5% -6.7% -112.6 TOTAL 20.1% -4.9% -1 051.7 Figure 36. Changes in the value of groundwater recharge under the Resilient Scenario, relative to BAU, valued in terms of replacement surface water inf rastructure costs. 67 Ch 5: Scenario Analysis EROSION CONTROL AND SEDIMENT RETENTION BAU SCENARIO to 37% in the Save Catchment. Overall, land cover changes under the BAU scenario result The overall reduction in vegetation cover in the additional export of 3.8 million t of under the BAU Scenario results in greater sediment to the country’s dams each year. exposure of the soil to rainfall, leading to increases in both erosion and sedimentation rates. Mean erosion rates across Zimbabwe RESILIENT SCENARIO were estimated to increase by 19% from 9.4 t/ ha/y to 11.2 t/ha/y. The export of sediment to In dam catchment areas, total sediment export watercourses was estimated to increase by is reduced by 59% (from 1.47 to 0.61 t/ha/y), or 8.95 31%, rising from a current mean of 1.1 t/ha/y million t relative to BAU. This has an estimated to 1.5 t/ha/y under BAU. The proportionally cost saving of US$11.2 million/y. Notably, the greater increase in sediment export relative reduction in sediment export from small-scale to erosion reflects the fact that the ability of farmland accounts for over half of the overall vegetation to reduce erosion in situ and its sediment reduction in dam catchment areas. ability to trap eroded sediment originating from upslope areas both decline under BAU. The overall reduction in sediment export to dams at the catchment level is shown in Figure Within dam catchment areas specifically, 37. The spatial patterns shown are closely tied to sediment export was predicted to increase by the number of dams in each catchment, with 33% under BAU (Table 22). At a catchment level, all avoided sediment export generally higher in major river basins were predicted to experience the east of the country. The catchments with the a significant increase in sediment export, highest reduction in sediment export are the ranging from 25% in the Gwayi Catchment Mzingwane, Runde and Mazowe catchments. Table 22. Changes in sediment export to dams between current land cover and the BAU Scenario. Additional sedimentation under BAU was costed based on the price of check dam construction. SEDIMENT EXPORT SEDIMENT EXPORT TO DAMS CURRENT TO DAMS BAU CATCHMENT (MILLION T/Y) (MILLION T/Y) % INCREASE Gwayi 27.9% -3.5% -173.5 Manyame 19.6% -3.5% -189.8 Mazowe 19.0% -6.1% -208.7 Mzingwane 21.9% -4.6% -18.7 Runde 16.7% -8.1% -82.3 Sanyati 22.7% -5.1% -266.2 Save 16.5% -6.7% -112.6 TOTAL 20.1% -4.9% -1 051.7 68 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE Figure 37. Value of enhanced sediment retention under the Resilient Scenario relative to BAU, valued in terms of avoided check dam construction costs. NATURE-BASED TOURISM BAU SCENARIO the average annual growth rate was 1.9%. No reasonable long-term forecasts to 2050 Future tourism value under the BAU Scenario could be found. Thus, tourism values were was adjusted in response to projected grown to 2050 using the more conservative changes in land cover and protection status. estimate of 1.9% average annual increase International growth in tourism demand based on the projected growth for Zimbabwe. was factored into the analysis. According to the UNWTO long-term forecast, the number Tourism values in 2050 under BAU were of international tourist arrivals worldwide adjusted in proportion to changes in land were expected to increase by an average cover extent as well as in response to the of 3.3% a year over the period 2010 to 2030 ongoing establishment of conservancies (UNWTO, 2014). Over this period the rate of in previously unprotected areas. This led to growth was expected to gradually decline the overall resource rent value of nature- from 3.8% in 2012 to 2.9% in 2030. The long- based tourism in Zimbabwe under the BAU term forecast for southern Africa was higher to be around US$283 million in 2050, with at 4.3% per annum (UNWTO, 2014). However, an average value of US$7.2/ha/y (Table 23). based on actual data of international visitors to Zimbabwe over the period 1995-2019 69 Ch 5: Scenario Analysis Table 23. Projected nature-based tourism value (resource rent) by 2050 under the BAU Scenario across the seven catchments in Zimbabwe. RESOURCE RENT VALUE OF % OF NATIONAL PER HECTARE NATURE NATURE-BASED TOURISM NATURE-BASED BASED TOURISM VALUE CATCHMENT (US$ MILLION/Y) TOURISM VALUE (US$/HA/Y) Gwayi 154.3 55% 17.5 Manyame 49.4 17% 12.1 Mazowe 18.1 6% 4.5 Mzingwane 16.5 6% 2.6 Runde 15.6 6% 3.8 Sanyati 13.7 5% 2.0 Save 15.2 5% 3.1 TOTAL 282.8 7.2 RESILIENT SCENARIO tourism revenue. The expansion of conservancy areas and strengthening protected area Tourism values in 2050 under a Resilient management was assumed to increase Scenario were adjusted in proportion to tourism inside these areas by 40% over the 27- changes in land cover extent and changes year period to 2050 as habitat degradation is in protection status with the establishment halted and wildlife numbers increase further. of conservancies in previously unprotected areas and the assumed improvement The overall value of nature-based tourism in management and tourism facilities in Zimbabwe was projected to be higher of protected environments as a result of under the Resilient Scenario relative to interventions implemented under the Resilient the BAU by 18% due to the expansion and Scenario envisaged under the Vision 2030. improvement of community conservation Financial and technical support for protected areas as part of an enhanced protected area areas through collaborative management estate across the country. This amounts to partnerships has shown to deliver positive an increase in the annual value of nature- ecological, economic and social outcomes based tourism by just over US$52 million (Lindsey et al., 2021a). For example, the by 2050, reflecting a significant increase in establishment of a collaborative management revenue relative to the current trajectory. partnership in Gonarezhou National Park in the south-east of Zimbabwe with Frankfurt The increase in the tourism value is a result Zoological Society (FZS) in 2017 has increased of the expansion of conservancy area under funding levels 20-fold, increased tourism by the Resilient Scenario in combination 40% relative to pre-FZS levels and improved with improved capacity to enhance community conservation efforts surrounding protected area performance. This value the park (Lindsey et al., 2021b). This shows that could make a meaningful contribution to investment in park management and facilities diversifying livelihoods across Zimbabwe. can have a significant positive impact on SUMMARY OF ECOSYSTEM SERVICE BENEFITS The above sections show that the interventions relative to BAU (Table 25). When the global proposed under the Resilient Scenario can value of the enhanced carbon storage service generate significant benefits across a range is included, the annual ecosystem service of ecosystem services, relative to the BAU benefits would be some US$2.09 billion/y Scenario. Overall, the interventions proposed greater than under the BAU Scenario. under the Resilient Scenario would generate local benefits worth US$445.4 million/y, At a catchment level, the aggregate greatest 70 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE ecosystem service benefits to Zimbabwe occur relative to BAU in each catchment (Table 25). in the Save Catchment, where the Resilient The elevated value of the Sanyati Catchment in Scenario yields ecosystem service gains worth global benefit terms reflects the fact that the US$118.5 million/y, relative to BAU (Table 25). largest carbon storage gains under the Resilient Scenario occur here. In per hectare terms, When the global value of the enhanced the largest increases in both local and global carbon storage service is included, the Save benefits under the Resilient Scenario occurred and Sanyati Catchments yield the greatest in the Save Catchment, with local benefits of ecosystem service benefits in aggregate US$24/ha and global benefits of US$79/ha. terms, with a gain in value of US$369 million/y Table 24. Projected nature-based tourism value by 2050 (resource rent) under the Resilient Scenario across the seven catchments in Zimbabwe and the gain in this value relative to the BAU Scenario. RESOURCE RENT PER HECTARE GAIN IN NATURE BASED VALUE OF NATURE- % OF NATIONAL NATURE BASED TOURISM VALUE BASED TOURISM NATURE-BASED TOURISM VALUE RELATIVE TO BAU CATCHMENT (US$ MILLION/Y) TOURISM VALUE (US$/HA/Y) (US$ MILLION/Y) Gwayi 187.7 66% 21.3 33.4 Manyame 55.9 20% 13.7 6.6 Mazowe 19.2 7% 4.8 1.1 Mzingwane 19.2 7% 3.1 2.8 Runde 18.7 7% 4.5 3.1 Sanyati 15.9 6% 2.3 2.3 Save 17.9 6% 3.6 2.7 TOTAL 334.7 8.5 51.9 Table 25. The combined value of ecosystem service benefits under the Resilient Scenario, relative to BAU. Local benefits refer to benefits that accrue nationally to Zimbabwe, while the value of global benefits includes benefits to the rest of the world f rom enhanced carbon storage under the Resilient Scenario. GAIN IN LOCAL GAIN IN LOCAL GAIN IN GLOBAL GAIN IN GLOBAL BENEFITS BENEFITS BENEFITS BENEFITS CATCHMENT (US$ MILLION/Y) (US$/HA) (US$ MILLION/Y) (US$/HA/Y) Gwayi 54.3 6.2 315.5 35.8 Manyame 69.4 17.0 246.1 60.4 Mazowe 79.5 19.8 277.9 69.3 Mzingwane 26.7 4.3 273.1 43.6 Runde 38.5 9.3 196.3 47.6 Sanyati 58.4 8.4 388.9 56.0 Save 118.5 24.1 389.0 79.1 TOTAL 445.4 11.4 2086.9 53.3 71 Ch 5: Scenario Analysis COST-BENEFIT ANALYSIS OF THE RESILIENT FUTURE SCENARIO The potential for the Resilient Future Scenario to increase more sharply in the first five years, to generate a positive ROI was explored reflecting the spread of CSA. After year five, through a high-level cost-benefit analysis. the value of these services was assumed to This analysis was performed for the country increase more gradually reflecting the slower as a whole and individually for each river recovery of natural habitats, with the full catchment. Costs and benefits were converted ecosystem service benefit realised after around to present value using a time horizon of 25 20 years. Finally, tourism benefits from the years and a social rate of discount of 4.56%. establishment of community conservancies and improvement of visitor facilities in state The ecosystem service benefits described protected areas were assumed to increase above were assumed to be realised gradually linearly, with the full benefit realised by year ten. over time. Crop production benefits from CSA practices began in year two and were fully Overall, ecosystem service benefits to realised by year five. Benefits from natural Zimbabwe from the Resilient Scenario over habitat restoration and recovery grew more the next 25 years were estimated to have slowly. For carbon and harvested resources, a present value of US$5.16 billion, relative a linear increase was assumed, with benefits to the BAU Scenario (Table 26). The total starting in year two and these services reaching cost of the Resilient Scenario interventions their maximum value by year 25. This was based (both establishment and ongoing costs) was on studies of the rates of miombo woodland estimated to be US$1.59 billion in present recovery after disturbance (Kalaba et al. 2013; value terms, resulting in a positive net present Williams et al. 2008). Sediment retention and value (NPV) of US$3.59 billion. This results in a groundwater recharge benefits were assumed ROI of 3.3. In other words, each dollar invested Table 26. Summary of present value costs, benefits and return on investment for the Resilient Scenario at national-level, relative to the BAU Scenario. PRESENT VALUE (US$ MILLION) COSTS Restore degraded natural habitats 883.5 Establish community conservancies and improve state PAs 55.0 Implement CSA (50% adoption) 466.3 Restore riparian buffers 182.3 Total value of costs 1 587.1 BENEFITS Crop production 1 927.2 Wild resource harvesting 81.0 Carbon storage and sequestration 1 169.2 Groundwater recharge 1 316.0 Sediment retention 118.3 Nature-based tourism 563.0 Total value of benefits 5 174.8 NET PRESENT VALUE 3 587.7 RETURN ON INVESTMENT 3.3 72 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE in the Resilient Scenario interventions would At a catchment level, the local ROI of the generate a return of US$3.3 in ecosystem Resilient Scenario varied from 1.5 in the service benefits. Notably, this includes only the Runde Catchment to as high as 4.7 in the value of local benefits that accrue to Zimbabwe. Mazowe Catchment (Table 27). This shows If the global value of avoided climate damages that the benefits of restoration and improved to the world is included in the analysis, the ecosystem management exceed the costs ROI increases to 9.4. These results highlight in every catchment of Zimbabwe, though that well-implemented restoration and the return on investment is significantly conservation interventions could generate higher in certain catchments. In addition benefits that significantly outweigh their costs, to the Mazowe Catchment, the Save, thus providing a strong economic rationale Manyame and Gwayi Catchments were all for the Resilient Scenario interventions. estimated to have a high ROI of 3.5 or more. Table 27. Summary of the cost-benefit analysis of the Resilient Scenario interventions at a catchment level. PRESENT VALUE OF COSTS PRESENT VALUE OF BENEFITS NET PRESENT VALUE LOCAL CATCHMENT (US$ MILLION/Y) (US$/HA) (US$ MILLION/Y) ROI Gwayi 170.5 616.5 446.0 3.6 Manyame 193.8 815.5 621.7 4.2 Mazowe 198.7 939.9 741.2 4.7 Mzingwane 161.8 323.2 161.4 2.0 Runde 291.6 441.1 149.5 1.5 Sanyati 282.0 702.2 420.2 2.5 Save 288.7 1 336.4 1 047.7 4.6 TOTAL 1 587.1 5 174.8 3 587.7 3.3 73 6. CONCLUSIONS The majority of Zimbabwe’s approximately natural habitats through better rangeland 15 million inhabitants are young and depend management or wildlife-based land use and on land and natural resources for their sustainable resource use. These would not livelihoods. Benchmarking of satellite-based only meet the country’s LDN commitments measures of land health and productivity but would generate much greater benefits suggest that the country’s landscapes are to society than a BAU scenario, by increasing still in a relatively good state. Indeed, natural rather than decreasing the supply of land cover classes still cover 74% of Zimbabwe. ecosystem services. Overall, the benefits of following a resilient development pathway However, the analysis of past trends shows could amount to some US$ 445.4 million that there has been a fair amount of ecological locally and US$2.09 billion/y globally, when degradation over the past 20-30 years, and that compared to the BAU trajectory. This does not without directed interventions, a business- include all ecosystem services and benefits as-usual scenario would result in significant that would result from these investments. losses of ecosystem integrity and capacity to supply ecosystem services. Currently, The economic rationale for following a cultivated lands contribute a resource rent resilient development pathway was further of some US$365.4 million/y, while natural highlighted through a high-level cost-benefit landscapes contribute resource rents of analysis, which showed that local ecosystem US$291.7 million/y from livestock, US$576 service benefits arising from the Resilient million/y from wild harvested resources, and Scenario interventions would generate a ROI US$229.7 from nature-based tourism. They of 3.3, relative to following the BAU trajectory. If also avoid costs of US$496 million/y, US$208 global benefits from avoided climate change million/y and US$9.4 billion/y as a result of damages are included, the ROI increases to groundwater recharge, sediment retention 9.4. These high ROI estimates highlight the and climate regulating services, respectively, fact that the benefits of investing in improved the last being a global value. Collectively, ecosystem management and protection the total value of the selected ecosystem significantly outweigh the costs of doing so. services was estimated to be US$2.4 billion/y Furthermore, the steep increase in ROI when to Zimbabwe and US$11.5 billion/y globally. the global benefits of the Resilient Scenario are included underscores the potential for Under BAU, natural land cover would decline Zimbabwe to benefit from international by six percentage points to 68% of national carbon markets in support of the proposed area, and levels of degradation would increase, interventions. Ongoing efforts to appropriately reducing capacity to supply ecosystem services, structure the policy regulatory framework However, if the country follows a more resilient for Zimbabwe, together with international growth path, it would invest in the restoration reforms to guide carbon markets under the and maintenance of key natural areas and the auspices of the United Nations Framework enhancement of cultivated lands, through Convention on Climate Change (UNFCCC) climate-smart conservation agriculture (CSA), will help realise this funding potential if the restoration and protection of riparian Zimbabwe makes the requisite commitments buffers, the passive restoration of degraded and investments in the Resilient Scenario. 74 REFERENCES Almagro, A., Thomé, T.C., Colman, C.B., Pereira, R.B., Marcato Junior, J., Rodrigues, D.B.B. & Oliveira, P.T.S. (2019). Improving cover and management factor (C-factor) estimation using remote sensing approaches for tropical regions. Int. Soil Water Conserv. Res. 7, 325–334. Angima, S.D., Stott, D.E., O’Neill, M.K., Ong, C.K. & Weesies, G.A. (2003). Soil erosion prediction using RUSLE for central Kenyan highland conditions. Agric. Ecosyst. Environ. 97, 295–308. Anyamba, A. & Tucker, C.J. (2005). Analysis of Sahelian vegetation dynamics using NOAA-AVHRR NDVI data from 1981-2003. J. Arid Environ. 63, 596–614. Baker, T.J. & Miller, S.N. (2013). Using the Soil and Water Assessment Tool (SWAT) to assess land use impact on water resources in an East African watershed. J. Hydrol. 486, 100–111. Beatty, C., Raes, L., Vogl, A.L., Hawthorne, P.L., Moraes, M., Saborio, J.L. & Meza Prado, K. (2018). Landscapes, at your service: applications of the Restoration Opportunities Optimization Tool (ROOT). Landscapes, your Serv. Appl. Restor. Oppor. Optim. Tool. Gland, Switzerland: IUCN. Blumenfeld, S., Lu, C., Christophersen, T. & Coates, D. (2009). Water, Wetlands and Forests: A Review of Ecological, Economic and Policy Linkages. Montreal and Gland: Secretariat of the Convention on Biological Diversity and Secretariat of the Ramsare Convention on Wetlands. Bond, I., Chambwera, M., Jones, B., Chundama, M. & Nhantumbo, I. (2010). REDD+ in dryland forests: Issues and prospects for pro-poor REDD in the miombo woodlands of southern Africa. International Institute for Environment and Development, London. Bondarenko, M., Kerr, D., Sorichetta, A., Tatem, A. & WorldPop. (2020). Census/projection-disaggregated gridded population datasets, adjusted to match the corresponding UNPD 2020 estimates, for 51 countries across sub-Saharan Africa using building footprints. Bouvet, A., Mermoz, S., Le Toan, T., Villard, L., Mathieu, R., Naidoo, L. & Asner, G.P. (2018). An above-ground biomass map of African savannahs and woodlands at 25 m resolution derived from ALOS PALSAR. Remote Sens. Environ. 206, 156–173. Brancalion, P.H.S., Meli, P., Tymus, J.R.C., Lenti, F.E.B., M. Benini, R., Silva, A.P.M., Isernhagen, I. & Holl, K.D. (2019). What makes ecosystem restoration expensive? A systematic cost assessment of projects in Brazil. Biol. Conserv. 240. Buchhorn, M., Smets, B., Bertels, L., De Roo, B., Lesiv, M., Tsendbazar, N., Herold, M. & Fritz, S. (2020). Copernicus Global Land Service: Land Cover 100m: collection 3: epoch 2019. Campbell, B.M. (1987). The use of wild fruits in Zimbabwe. Econ. Bot. 41, 375–385. Campbell, B.M., Angelsen, A., Cunningham, A., Katerere, Y., Sitoe, A. & Wunder, S. (2007). Miombo woodlands – opportunities and barriers to sustainable forest management. ISRN Biodivers. Center for International Forestry Research, Bogor, Indonesia. Campbell, B.M., Luckert, M. & Scoones, I. (1997). Local-level valuation of savanna resources: A case study from Zimbabwe. Econ. Bot. 51, 59–77. Campbell, B.M., Vermeulen, S.J. & Lynam, T. (1991). Value of trees in the small-scale farming sector of Zimbabwe. International Development Research Centre. Chagumaira, C., Rurinda, J., Nezomba, H., Mtambanengwe, F. & Mapfumo, P. (2016). Use patterns of natural resources supporting livelihoods of smallholder communities and implications for climate change adaptation in Zimbabwe. Environ. Dev. Sustain. 18, 237–255. 75 References Chambwera, M. & Folmer, H. (2007). Fuel switching in Harare: An almost ideal demand system approach. Energy Policy 35, 2538–2548. Chardonnet, B. (2019). Lets Talk about it! Reconfiguring the Protected Areas in Africa. IUCN Programme on African Protected Areas & Consservation (PAPACO). Chittaragi, A., Naika, R. & Vinayaka, K.S. (2014). Nutritive value of few wild mushrooms from the western ghats of Shivamogga district, Karnataka, India. Asian J. Pharm. Clin. Res. 7, 50–53. Chivasaa, W., Mutanga, O. & Biradarc, C. (2019). Mapping land suitability for maize (Zea mays L.) production using GIS and AHP technique in Zimbabwe. South African J. Geomatics 8, 265–281. CIAT & World Bank. (2017). Climate-Smart Agriculture in Zimbabwe. CSA Country Profiles for Africa Series. Washington D.C.: International Center for Tropical Agriculture. Conservation International. (2018). Trends.Earth. Dalu, T., Tambara, E.M., Clegg, B., Chari, L.D. & Nhiwatiwa, T. (2013). Modeling sedimentation rates of Malilangwe reservoir in the south-eastern lowveld of Zimbabwe. Appl. Water Sci. 3, 133–144. Davis, R. & Hirji, R. (2014). Climate Change and Water Resources Planning, Development, and Management in Zimbabwe. Clim. Chang. Water Resour. South Asia. World Bank. Degreef, J., Kasongo, B., Niyongabo, E. & De Kesel, A. (2020). Edible mushrooms, a vulnerable ecosystem service from african miombo woodlands. Biotechnol. Agron. Soc. Environ. 24, 70–80. Descheemaeker, K., Poesen, J., Borselli, L., Nyssen, J., Raes, D., Haile, M., Muys, B. & Deckers, J. (2008). Runoff curve numbers for steep hillslopes with natural vegetation in semi-arid tropical highlands, northern Ethiopia. Hydrol. Process. 22, 4097–4105. Dowo, G.M., Kativu, S. & de Garine-Wichatitsky, M. (2018). Local perceptions of tree diversity, resource utilisation and ecosystem services provision at the periphery of Gonarezhou National Park, Zimbabwe. For. Trees Livelihoods 27, 1–21. Dube, P., Musara, C. & Chitamba, J. (2014). Extinction Threat to Tree Species from Firewood Use in the Wake of Electric Power Cuts : A Case Study of Bulawayo , Zimbabwe. Resour. Environ. 4, 260–267. Dugan, P. (2011). Cost comparison analysis of ANR compared to conventional reforestation. In Forests Beneath the grass - Proceedings of the regional workshop on advancing the application of assisted natural regenration for effective low-cost forest restoration. Bohol, Philippines, 19-22 May 2009.: 9–14. Durst, P.B., Sajise, P. & Leslie, R.N. (Eds.). Ecosystem Marketplace. (2021). Markets in Motion: State of the Voluntary Carbon Markets 2021. FAO. (2016). AQUASTAT Country profile – Zimbabwe. Rome, Italy: Food and Agriculture Organization of the United Nations. Farquharson, F.A.K. & Bullock, A. (1992). The hydrology of basement complex regions of Africa with particular reference to southern Africa. Geol. Soc. Spec. Publ. 66, 59–76. Fenta, A.A., Tsunekawa, A., Haregeweyn, N., Poesen, J., Tsubo, M., Borrelli, P., Panagos, P., Vanmaercke, M., Broeckx, J., Yasuda, H., Kawai, T. & Kurosaki, Y. (2020). Land susceptibility to water and wind erosion risks in the East Africa region. Sci. Total Environ. 703, 135016. Frazee, S.R., Cowling, R.M., Pressey, R.L., Turpie, J.K. & Lindenberg, N. (2003). Estimating the costs of conserving a biodiversity hotspot: A case-study of the Cape Floristic Region, South Africa. Biol. Conserv. 112, 275–290. Frost, P. (1996). The Ecology of Miombo Woodlands. In The Miombo in Transition: Woodlands and Welfare in Africa: 11–57. Campbell, B.M. (Ed.). Bogor, Indonesia: Center for International Forestry Research. 76 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE Garcia, R.C., De Oliveira, N.T.E., Camargo, S.C., Pires, B.G., De Oliveira, C.A.L., Teixeira, R. de A. & Pickler, M.A. (2013). Honey and propolis production, hygiene and defense behaviors of two generations of Africanized honey bees. Sci. Agric. 70, 74–81. Gilbert, M., Nicolas, G., Cinardi, G. & et al. (2018). Global distribution data for cattle, buffaloes, horses, sheep, goats, pigs, chickens and ducks in 2010. Sci Data 5, 180227. Godwin, M.A., Soropa, G., Makurira, H. & Dzvairo, W. (2011). Sedimentation impacts on reservoir as a result of land use on a selected catchment in Zimbabwe. Int. J. Eng. Sci. Technol. 3, 6599–6609. GoZ. (2017). Land Degradation Neutrality National Report. Government of Zimbabwe, Ministry of Environment, Water and Climate. GoZ. (2019). Zimbabwe’s Sixth National Report to the Convention on Biodiversity. Government of Zimbabwe, Ministry of Environment, Water and Climate. Grundy, I., Turpie, J., Jagger, P., Witkowski, E., Guambe, I., Semwayo, D. & Solomon, A. (2000). Implications of co-management for benefits from natural resources for rural households in north-western Zimbabwe. Ecol. Econ. 33, 369–381. Grundy, I.M., Campbell, B.M., Balebereho, S., Cunliffe, R., Tafangenyasha, C., Fergusson, R. & Parry, D. (1993). Availability and use of trees in Mutanda Resettlement Area, Zimbabwe. For. Ecol. Manage. 56, 243–266. Hengl, T., Heuvelink, G.B.M., Kempen, B., Leenaars, J.G.B., Walsh, M.G., Shepherd, K.D., Sila, A., MacMillan, R.A., De Jesus, J.M., Tamene, L. & Tondoh, J.E. (2015). Mapping soil properties of Africa at 250 m resolution: Random forests significantly improve current predictions. PLoS One 10. Hoffa, E.A., Ward, D.E., Hao, W.M., Susott, R.A. & Wakimoto, R.H. (1999). Seasonality of carbon emissions from biomass burning in a Zambian savanna. J. Geophys. Res. Atmos. 104, 13841–13853. IAPRI & FAO. (2021). Impact of Pfumvudza Farming Concept on Maize Productivity in Zimbabwe. Lusaka, Zambia: Indaba Agricultural Policy Research Institute (IAPRI). Illgner, P. & Nel, E. (2000). The geography of edible insects in Sub-Saharan Africa: A study of the mopane caterpillar. Geogr. J. 166, 336–351. IPCC. (2019). 2019 Refinement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories (2019 Refinement). 2019 Refinement to 2006 IPCC Guidel. Natl. Greenh. Gas Invent. IPCC Task Force on National Greenhouse Gas Inventories. Jaffé, R., Dietemann, V., Allsopp, M.H., Costa, C., Crewe, R.M., Dall’olio, R., De La Rúa, P., El-Niweiri, M.A.A., Fries, I., Kezic, N., Meusel, M.S., Paxton, R.J., Shaibi, T., Stolle, E. & Moritz, R.F.A. (2010). Estimating the Density of Honeybee Colonies across Their Natural Range to Fill the Gap in Pollinator Decline Censuses. Conserv. Biol. 24, 583–593. James, A.N., Green, M.J.B.B. & Paine, J.R. (1999). A Global Review of Protected Area Budgets and Staff. Cambridge, UK: World Conservation Press. Kabell, T.C. (1984). Sediment storage requirements for reservoirs. Harare Symposium, challenges in African hydrology and water resources. IAHS 144, 569–576. Kalaba, F.K., Quinn, C.H., Dougill, A.J. & Vinya, R. (2013). Floristic composition, species diversity and carbon storage in charcoal and agriculture fallows and management implications in Miombo woodlands of Zambia. For. Ecol. Manage. 304, 99–109. Karaan, M., Ham, C., Akinnifesi, F., Moombe, K., Jordaan, D., Franzel, S. & Aithal January, A. (2006). Baseline Marketing Surveys and Supply Chain Studies for Indigenous Fruit Markets in Tanzania, Zimbabwe and Zambia. World Agroforestry Centre and CPWild Research Alliance. 77 References Kenya Ministry of Tourism, W. and H. (2023). Tourism performance statistics. https://www.tourism.go.ke/. Accessed 20 April 2023. Kinyanjui, M.J. (2011). NDVI-based vegetation monitoring in Mau forest complex, Kenya. Afr. J. Ecol. 49, 165– 174. Knijff, J. Van Der, Jones, R.R.J. a., Montanarella, L. & Van der Knijff, J.M. (1999). Soil erosion risk assessment in Italy. European Commission, European Soil Bureau. Kozanayi, W. & Frost, P. (2002). Marketing of Mopane Worm in Southern Zimbabwe. Processing 1–31. Kupurai, P., Kugedera, A.T. & Sakadzo, N. (2021). Evaluating the potential contribution of non-timber forest products (NTFPs) to smallholder farmers in semi-arid and arid regions: A case of Chivi, Zimbabwe. Res. Ecol. 3, 22–30. Li, C., Lin, L., Hao, Z., Post, C.J., Chen, Z., Liu, J. & Yu, K. (2020). Developing a USLE cover and management factor (C) for forested regions of southern China. Front. Earth Sci. 14, 660–672. Lindsey, P., Baghai, M., Bigurube, G., Cunliffe, S., Dickman, A., Fitzgerald, K., Flyman, M., Gandiwa, P., Kumchedwa, B., Madope, A., Morjan, M., Parker, A., Steiner, K., Tumenta, P., Uiseb, K. & Robson, A. (2021a). Attracting investment for Africa’s protected areas by creating enabling environments for collaborative management partnerships. Biol. Conserv. 255, 108979. Lindsey, P., Baghai, M., Bigurube, G., Cunliffe, S., Dickman, A., Fitzgerald, K., Flyman, M., Gandiwa, P., Kumchedwa, B., Madope, A., Morjan, M., Parker, A., Steiner, K., Tumenta, P., Uiseb, K. & Robson, A. (2021b). Attracting investment for Africa’s protected areas by creating enabling environments for collaborative management partnerships. Biol. Conserv. 255, 108979. Mabugu, R. & Chitiga, M. (2002). Accounting for Forest Resources in Zimbabwe. CEEPA. University of Pretoria, Pretoria, South Africa. Madzokere, F., Mutambara, J. & Zirenga, V.L. (2018). An Investigation into the Factors Affecting Outgrower Sugarcane Farming In the South- East Lowveld of Zimbabwe. Bus. Soc. Sci. J. 3, 1–11. Magadza, C.H.D. (1984). An analysis of siltation rates in Zimbabwe. Zimbabwe Sci. News 18, 1–5. Mahlatini, P., Hove, A., Maguma, L.F. & Chemura, A. (2020). Using direct use values for economic valuation of wetland ecosystem services: a case of Songore wetland, Zimbabwe. GeoJournal 85, 41–51. Makwara, E.C. & Gamira, D. (2012). About to Lose all the Soil in Zaka’s Ward 5, Zimbabwe: Rewards of Unsustainable Land Use. Eur. J. Sustain. Dev. 1, 457. Marongwe, L.S., Kwazira, K., Jenrich, M., Thierfelder, C., Kassam, A. & Friedrich, T. (2011). An African success: The case of conservation agriculture in Zimbabwe. Int. J. Agric. Sustain. 9, 153–161. Maroyi, A. (2013). Traditional use of medicinal plants in south-central Zimbabwe: Review and perspectives. J. Ethnobiol. Ethnomed. 9. Mashapa, C., Gandiwa, E., Mhuriro-Mashapa, P. & Zisadza-Gandiwa, P. (2014). Increasing demand on natural forest products in urban and peri-urban areas of Mutare, eastern Zimbabwe: implications for sustainable natural resources management. Nat. Faune 28, 42–48. Mazvimavi, D., Meijerink, A.M.J., Savenije, H.H.G. & Stein, A. (2005). Prediction of flow characteristics using multiple regression and neural networks: A case study in Zimbabwe. Phys. Chem. Earth 30, 639–647. McGregor, J. (1991). Woodland resources: ecology, policy and ideology an historical case study of communal area, Zimbabwe. Loughbrgh. Univ. Loughborough University. MECHTI. (2021). Tourism Satellite Account for Zimbabwe. Ministry of Environment, Climate, Tourism and Hospitality Industry Zimbabwe. 78 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE MEFT/NASCO. (2021). The State of Community Conservation in Namibia (Annual Report 2020). Ministry of Environment, Forestry and Tourism (MEFT) and the Namibian Association of CBNRM Support Organisations (NACSO). Mekonnen, M., Keesstra, S.D., Baartman, J.E., Ritsema, C.J. & Melesse, A.M. (2015). Evaluating sediment storage dams: structural off-site sediment trapping measures in northwest Ethiopia. Geogr. Res. Lett. 41, 7–22. Messager, M.L., Lehner, B., Grill, G., Nedeva, I. & Schmitt, O. (2016). Estimating the volume and age of water stored in global lakes using a geo-statistical approach. Nat. Commun. 7, 1–11. Millennium Ecosystem Assessment. (2003). Ecosystems and Human Well-being: a Framework for Assessment. Island Press, Washington. Millennium Ecosystem Assessment. (2005). Ecosystems and Human Well-being: Synthesis. Island Press, Washington. Mkuhlani, S., Mupangwa, W., Macleod, N., Gwiriri, L., Nyagumbo, I., Manyawu, G. & Chigede, N. (2018). Crop- livestock integration in smallholder farming systems of Goromonzi and Murehwa, Zimbabwe. Renew. Agric. Food Syst. 35, 249–260. Mlambo, A. & Maphosa, M. (2017). Miombo Woodland Mushrooms of Commercial Food Value: A Survey of Central Districts of Zimbabwe. J. Food Secur. 5, 51–57. Mlambo, A. & Maphosa, M. (2021). Ectomycorrhizal mushroom yield association with woody species and leaf litter in amiombo woodland , central Zimbabwe. African J. Ecol. Ecosyst. 8, 1–10. MoLAFWRR. (2021). Zimbabwe First Round Crop and Livestock Assessment Report 2020/2021 Season. Ministry of Lands, Agriculture, Fisheries, Water and Rural Resettlement. MoLAWFRR. (2020). Zimbabwe Second Round Crop and Livestock Assessment Report 2019/2020 Season. Republic of Zimbabwe, Ministry of Lands, Agriculture, Water, Fisheries and Rural Resettlement. MoLAWFRR. (2021). Zimbabwe Second Round Crop and Livestock Assessment Report 2020/2021 Season. Republic of Zimbabwe, Ministry of Lands, Agriculture, Water, Fisheries and Rural Resettlement. MoLAWFRR. (2022). Zimbabwe Second Round Crop and Livestock Assessment Report 2021/2022 Season. Republic of Zimbabwe, Ministry of Lands, Agriculture, Water, Fisheries and Rural Resettlement. MoLAWRR. (2020a). National Water Resources Master Plan 2020 - 2040. Sectoral Report B2: Groundwater Resources Management and Development Volume 3. Republic of Zimbabwe, Ministry of Lands, Agriculture, Water and Rural Resettlement. MoLAWRR. (2020b). National Water Resources Master Plan 2020 - 2040. Main Report of the Master Plan Volume 2. Republic of Zimbabwe, Ministry of Lands, Agriculture, Water and Rural Resettlement. Moore, A.E., Cotterill, F.P.D., Broderick, T. & Plowes, D. (2009). Landscape evolution in Zimbabwe from the permian to present, with implications for kimberlite prospecting. South African J. Geol. 112, 65–88. Moyo, S. (2011). Three decades of agrarian reform in Zimbabwe. J. Peasant Stud. 38, 493–531. Mpofu, B. (2018). The Land Question, Agriculture, Industrialization and the Economy in Zimbabwe: A Critical Reflection. In The Development of Africa: Issues, Diagnoses and Prognoses: 115–132. Akanle, O. & Adésìnà, J.O. (Eds.). Springer International Publishing. Mudekwe, J. (2007). The Impact of Subsistence Use of Forest Products and the Dynamics of Harvested Woody Species Populations in a Protected Forest Reserve in Western Zimbabwe. University of Stellenbosch. 79 References Mugabe, D., Chipunza, N., Mupaso, N., Munyati, V.T. & Makarudze, F.V. (2014). Estimation of Optimal Land Use Allocation Among Small Holder (A1) Farmer Households in Zimbabwe. A Case Study of Long Croft Farm, in Mazowe District. J. Agric. Sci. 6, 170–181. Mugandani, R., Wuta, M., Makarau, A. & Chipindu, B. (2012). Re-Classification of Agro-Ecological Regions of Zimbabwe in Conformity With Climate Variability and Change. African Crop Sci. J. 20, 361–369. Mukhebi, A.W., Chamboko, T., O’Callaghan, C.J., Peter, T.F., Kruska, R.L., Medley, G.F., Mahan, S.M. & Perry, B.D. (1999). An assessment of the economic impact of heartwater (Cowdria ruminantium infection) and its control in Zimbabwe. Prev. Vet. Med. 39, 173–189. Mulligan, M., van Soesbergen, A. & Sáenz, L. (2020). GOODD, a global dataset of more than 38,000 georeferenced dams. Sci. Data 7, 1–31. Murwira, A., Masocha, M., Magadza, C.H.D., Owen, R. & Nhiwatiwa, T. (2014). Zimbabwe-Strategy for Managing Water Quality and Protecting Water Sources 1–97. Nakakaawa, C.A., Vedeld, P.O. & Aune, J.B. (2011). Spatial and temporal land use and carbon stock changes in Uganda: Implications for a future REDD strategy. Mitig. Adapt. Strateg. Glob. Chang. 16, 25–62. Ngadze, R.T., Linnemann, A.R., Nyanga, L.K., Fogliano, V. & Verkerk, R. (2017). Local processing and nutritional composition of indigenous fruits: The case of monkey orange (Strychnos spp.) from Southern Africa. Food Rev. Int. 33, 123–142. Ngulube, M.R., Hall, J.B. & Maghembe, J.A. (1996). A review of the silviculture and resource potential of a miombo fruit tree: Uapaca kirkiana (Euphorbiaceae). J. Trop. For. Sci. 8, 395–411. Nicholson, S.E., Davenport, M.L. & Malo, A.R. (1990). A comparison of the vegetation response to rainfall in the Sahel and East Africa, using normalized difference vegetation index from NOAA AVHRR. Clim. Change 17, 209–241. Olson, D.M., Dinerstein, E., Wikramanayake, E.D., Burgess, N.D., Powell, G.V.N., Underwood, E.C., D’Amico, J.A., Itoua, I., Strand, H.E., Morrison, J.C., Loucks, C.J., Allnutt, T.F., Ricketts, T.H., Kura, Y., Lamoreux, J.F., Wettengel, W.W., Hedao, P. & Kassem, K.R. (2001). Terrestrial Ecoregions of the World: A New Map of Life on Earth. Bioscience 51, 933–938. Owuor, S.O., Butterbach-Bahl, K., Guzha, A.C., Rufino, M.C., Pelster, D.E., Díaz-Pinés, E. & Breuer, L. (2016). Groundwater recharge rates and surface runoff response to land use and land cover changes in semi- arid environments. Ecol. Process. 5, 1–21. Padilla Parellada, M. (2018). ESA Fire Climate Change Initiative (Fire_cci): MODIS Fire_cci Burned Area Pixel product, version 5.1[dataset]. Cent. Environ. Data Anal. Panagos, P., Borrelli, P., Meusburger, K., Alewell, C., Lugato, E. & Montanarella, L. (2015). Estimating the soil erosion cover-management factor at the European scale. Land use policy 48, 38–50. Pimentel, D., Wilson, C., Mccullum, C., Huang, R., Dwen, P., Flack, J., Tran, Q., Saltman, T. & Cliff, B. (2008). Economic and Environment Benefits of Biodiversity The annual economic and environmental in the United States total approximately $ 300 billion. Sci. York 47, 747–757. Poilecot, P. & Gaidet, N. (2011). A quantitative study of the grass and woody layers of a Mopane (Colophospermum mopane) savannah in the mid-Zambezi Valley, Zimbabwe. Afr. J. Ecol. 49, 150–164. Pregitzer, K.S. & Euskirchen, E.S. (2004). Carbon cycling and storage in world forests: Biome patterns related to forest age. Glob. Chang. Biol. 10, 2052–2077. Pritchard, R., Ryan, C.M., Grundy, I. & van der Horst, D. (2018). Human Appropriation of Net Primary Productivity and Rural Livelihoods: Findings From Six Villages in Zimbabwe. Ecol. Econ. 146, 115–124. 80 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE Roose, E. (1996). Land husbandry - Components and strategy. FAO Soils Bull. 70. Rome: Food and Agriculture Organization. Roose, E. (2008). Soil erosion, conservation and restoration: a few lessons from 50 years of research in Africa. In 5th International Congress of the European Society for Soil Conservation “Soils changing in a changing World: the soils of tomorrow.” Reiskirchen, Germany: Catena Verlag. Runganga, R. & Mhaka, S. (2021). Impact of Agricultural Production on Economic Growth in Zimbabwe. Munich Pers. RePEc Arch. 1–21. Ryan, C.M., Williams, M. & Grace, J. (2011). Above- and belowground carbon stocks in a miombo woodland landscape of Mozambique. Biotropica 43, 423–432. Santoro, M., Cartus, O., Mermoz, S., Bouvet, A., Le Toan, T., Carvalhais, N., Rozendaal, D., Herold, M., Avitabile, V., Quegan, S., Carreiras, J., Rauste, Y., Balzter, H., Schmullius, C. & Seifert, F.M. (2018). GlobBiomass global above-ground biomass and growing stock volume datasets. Scoones, I. (1992). The economic value of livestock in the communal areas of southern Zimbabwe. Agric. Syst. 39, 339–359. Scoones, I., Marongwe, N., Mavedzenge, B., Murimbarimba, F., Mahenehene, J. & Sukume, C. (2011). Zimbabwe’s Land Reform: A summary of findings. Brighton, UK: Institute of Development Studies. Scoones, I., Mavedzenge, B., Murimbarimba, F. & Sukume, C. (2018). Tobacco, contract farming, and agrarian change in Zimbabwe. J. Agrar. Chang. 18, 22–42. Sha, Z., Bai, Y., Li, R., Lan, H., Zhang, X., Li, J., Liu, X., Chang, S. & Xie, Y. (2022). The global carbon sink potential of terrestrial vegetation can be increased substantially by optimal land management. Commun. Earth Environ. 3, 1–10. Shackleton, C. & Shackleton, S. (2004). The importance of non-timber forest products in rural livelihood security and as safety nets: A review of evidence from South Africa. S. Afr. J. Sci. 100, 658–664. Shekede, D.M., Kusangaya, S., Chavava, C.B., Gwitira, I. & Chemura, A. (2024). A two-decade analysis of the spatial and temporal variations in burnt areas across ecosystems of Zimbabwe. PLOS Clim. 3, 1–20. Shumba, E., Roberntz, P. & Kuona, M. (2011). Assessment of Sugarcane Outgrower Schemes for Bio-Fuel Production in Zambia and Zimbabwe. Harare: World Wide Fund For Nature. Stack, J., Dorward, A., Gondo, T., Frost, P., Taylor, F., Kurebgaseka, N., Gwavuya, S., Tawedzegwa, M., Rutamaba, W., Tlotlego, S. & Zhou, R. (2003). Mopane worm utilisation and rural livelihoods in Southern Africa. In The International Conference on Rural Livelihoods, Forests and Biodiversity: 1–38. Sukume, C., Mahofa, G. & Mutyasira, V. (2022). Effects of commercialisation on seasonal hunger: Evidence from smallholder resettlement areas, Mazowe District, Zimbabwe. Agricultrual Policy Research in Africa. The World Bank. (2023). Data Bank: International tourism, number of arrivals - Kenya. https://data. worldbank.org/indicator/ST.INT.ARVL?locations=KE. Accessed 20 April 2023. Thierfelder, C., Cheesman, S. & Rusinamhodzi, L. (2012). A comparative analysis of conservation agriculture systems: Benefits and challenges of rotations and intercropping in Zimbabwe. F. Crop. Res. 137, 237– 250. Tundu, C., Tumbare, M.J. & Onema, J.M.K. (2018). Sedimentation and its impacts/effects on river system and reservoir water quality: Case study of Mazowe catchment, Zimbabwe. Proc. Int. Assoc. Hydrol. Sci. 377, 57–66. Turner, D.P., Guzy, M., Lefsky, M.A., Ritts, W.D., Van Tuyl, S. & Law, B.E. (2004). Monitoring forest carbon sequestration with remote sensing and carbon cycle modeling. Environ. Manage. 33, 457–466. 81 References Turpie, J., Wilson, L. & Letley, G. (2023). Kenya CCDR Note: Potential changes in biodiversity and effects on cultural ecosystem services. Turpie, J., Wilson, L., Letley, G., Weiss, J., Leon, J. & Vogl, A. (2022a). Mapping and valuing ecosystem services and analysis of potential landscape interventions in Mazowe Catchment, Zimbabwe. Report submitted by Anchor Environmental Consultants to the World Bank. August 2022. 125 pp. Turpie, J., Wilson, L., Shone, F., Ngwenya, N., N., N., Machena, C. & Zamasiya, B. (2022b). Zimbabwe’s biodiversity economy: Status report, investment blueprint and framework for accounting. Turpie, J.K., Letley, G., Schmidt, K., Weiss, J., O’Farrell, P. & Jewitt, D. (2020). Towards a method for accounting for ecosystem services and asset value : Pilot accounts for KwaZulu-Natal, South Africa, 2005-2011. Anchor Environmental Consultants. Turpie, J.K., Weiss, J. & Letley, G.K. (2022c). Spatial modelling to determine the use and sustainability of wild resources harvesting at landscape scale for ecosystem asset valuation and accounting. Rev. Twine, W., Moshe, D., Netshiluvhi, T. & Siphugu, V. (2003). Consumption and direct-use values of savanna bio- resources used by rural households in Mametja, a semi-arid area of Limpopo province, South Africa. S. Afr. J. Sci. 99, 467–473. UBOS. (2023). Physical and Monetary Ecosystem Services and Asset Accounts for Uganda 1990-2015. Uganda Bureau of Statistics. UNFPA. (2021). World Population Dashboard Zimbabwe. United Nations Population Fund. Unganai, L.S. & Mason, S.J. (2001). Spatial characterization of Zimbabwe summer rainfall during the period 1920-1996. S. Afr. J. Sci. UNWTO. (2014). Tourism Highlights 2014 Edition. World Tourism Organization (UNWTO), Madrid Spain. Van Den Wall Bake, G.. (1985). Siltation and Soil Erosion Survey in Zimbabwe. WEF. (2019). The Travel and Tourism Competitiveness Report 2019. Geneva: World Economic Forum. Whitlow, R. (1988). Land Degradation in Zimbabwe: A Geographical Study. Department of Natural Resources,. Williams, M., Ryan, C.M., Rees, R.M., Sambane, E., Fernando, J. & Grace, J. (2008). Carbon sequestration and biodiversity of re-growing miombo woodlands in Mozambique. For. Ecol. Manage. 254, 145–155. Wischmeier, W. & Smith, D. (1978a). Predicting rainfall erosion losses—a guide to conservation planning. United States Department of Agriculture. Wischmeier, W. & Smith, D. (1978b). Predicting rainfall erosion losses—a guide to conservation planning. Woittiez, L.S., Rufino, M.C., Giller, K.E. & Mapfumo, P. (2013). The use of woodland products to cope with climate variability in communal areas in Zimbabwe. Ecol. Soc. 18, 24. Wood, S.A., Guerry, A.D., Silver, J.M. & Lacayo, M. (2013). Using social media to quantify nature-based tourism and recreation. Sci. Rep. 3. World Bank. (2017). Shadow price of carbon in economic analysis: Guidance Note. : Strategic Actions World Bank. (2019a). Zimbabwe Climate Change Technical Assistance Program  Supporting the Forest Sector. World Bank. (2019b). Zimbabwe Climate Smart Agriculture Investment Plan. International Bank for Reconstruction and Development / The World Bank. 82 ECOSYSTEM SERVICES ASSESSMENT OF ZIMBABWE World Bank. (2021). Climate Risk Profile: Zimbabwe. Washington D.C.: The World Bank Group. World Bank Water. (2014). Climate Change and Water Resources Planning, Development, and Management in Zimbabwe. The Zimbabwe Water Forum. WTTC. (2021). Zimbabwe 2021 Annual Research: Key Highlights. World Travel & Tourism Council. ZIMSTAT. (2015). Population projections thematic report. Zimbabwe National Statistics Agency. ZIMSTAT. (2017). Inter-censal Demographic Survey 2017. Zimbabwe National Statistics Agency. ZIMSTAT. (2019). Zimbabwe Poverty Report 2017. Zimbabwe National Statistics Agency. ZIMSTAT and UNICEF. (2019). MICS 2019 Survey Findings Report. Zimbabwe National Statistics Agency and the United Nations Children’s Fund. Zomer, R.J., Xu, J. & Trabucco, A. (2022). Version 3 of the Global Aridity Index and Potential Evapotranspiration Database. Sci. Data 9, 1–15. ZTA. (2020). Tourism Trends and Statistics Report 2019. Zimbabwe Tourism Authority. 83