Avoiding ecosystem and social impacts of hydropower, wind, and solar in Southern Africa's low-carbon electricity system

The study addresses how Southern Africa can meet rapidly growing electricity demand and climate goals while minimizing negative ecosystem and social impacts from energy infrastructure, especially large hydropower. Although hydropower is often promoted as low-cost, low-carbon, its social and environmental harms have been underestimated in planning. Wind, solar, and storage offer alternatives but entail variability, land-use, and siting conflicts. Southern Africa, with high reliance on hydropower in several countries and vast proposed hydro projects across major river basins, also hosts globally important biodiversity and culturally significant landscapes. The research question is how excluding socially and environmentally sensitive areas from wind, solar, and hydropower development affects optimal electricity pathways, hydropower selection, emissions, and system costs, particularly under a regional carbon-reduction target consistent with Paris Agreement ambitions.

Prior sustainable hydropower research has emphasized minimizing rather than avoiding impacts by optimizing dam portfolios against ecological and social criteria, typically assuming a fixed hydropower contribution. Such approaches often decouple dam siting from power system planning and thus cannot test substitution by other technologies. Some recent work assesses replacing hydropower with solar/wind on an annual energy basis, but it overlooks hydropower’s dispatchability and typically aggregates hydropower as a fleet rather than modeling individual projects with realistic temporal dynamics. Recent capacity expansion studies in Africa suggest about half of proposed dams are economic, but they do not screen for socio-environmental impacts. Meanwhile, utility-scale wind and solar have encountered opposition tied to environmental and social concerns, with such factors a leading cause of project failure in some regions. This body of literature highlights the need to jointly consider socio-environmental siting and system-level economics and reliability to identify low-impact, feasible low-carbon portfolios.

The analysis integrates spatial resource assessment, environmental and social screening, hydrological modeling, and capacity expansion planning for the Southern African Power Pool (SAPP) through 2040. - Scenarios: Seven candidate resource scenarios were constructed: Base; Legal (excludes IUCN I–II and national parks); Social (excludes legally protected and livelihood/cultural use areas; screens out dams displacing >5000 people); Environmental (excludes IUCN III–VI, additional conservation areas, Key Biodiversity Areas; screens dams on or significantly altering large free-flowing rivers); Environmental and Landscape (adds dense forests and wetlands exclusions); All Exclusions (combines Legal, Social, Environmental and Landscape); All Exclusions No New Hydropower (as All Exclusions with a hydropower moratorium). - Wind and solar resource mapping: Using the MapRE framework, candidate sites were identified at 500 m resolution and aggregated into project areas (up to 25 km² for PV ~1 GW; 100 km² for wind ~300 MW). Hourly 2018 time series were derived: wind from ERA5 adjusted to Global Wind Atlas with a Vestas 2 MW/90 m power curve via NREL SAM; solar from NSRDB GHI converted to fixed-tilt capacity factors (tilt = latitude) via SAM. Project economics include spur line and road costs based on proximity to infrastructure. - Hydropower characterization and screening: Existing and planned/proposed dams (about 41 GW planned/proposed) were mapped across eight major basins. Daily runoff and reservoir operations were simulated using VIC and VIC-Res with dam-specific rule curves to yield average daily energy budgets; GridPath then dispatches hourly constrained by daily energy limits. Reservoir extents for 34 large planned/proposed projects (~20 GW) were modeled using 90 m DEM, dam heights, and watershed delineation to estimate inundation footprints and volumes, validated against literature and imagery. Projects were screened per scenario using overlap thresholds with protected/conservation/social datasets, forest land-use efficiency thresholds, displacement (>5000 people), and free-flowing river criteria (including downstream Degree of Regulation thresholds). - Power system modeling: GridPath-SAPP (mixed-integer program) optimizes generation, storage, and transmission investments and operations for 12 country load zones over investment years 2020, 2025, 2030, 2035, 2040 (plus 2045 for end effects). Demand doubles from 2020 to 2040 (based on SAPP Plan) with hourly profiles extrapolated from 2018. A 15% planning reserve margin is imposed; only dispatchables and 10% of wind contribute. Technology costs follow SAPP Plan with NREL ATB mid-case trajectories; fuels from SAPP Plan; emissions factors from EIA. Scenarios include cases with no carbon target (emissions capped to Base trajectory) and a low-carbon target (linear path to 50% below 2020 emissions by 2040). Outputs include capacities, generation, dispatch, trade, costs, and CO2 emissions.

- Resource impacts of exclusions: Under All Exclusions, remaining potential is about 17% of Base for solar and 72% for wind. Planned hydropower capacity available falls to ~58% (~25 GW) of total planned/proposed potential; about 12% overlaps legally protected areas and 17% overlaps high conservation value areas or large free-flowing rivers. - Capacity build and mix: Without socio-environmental exclusions, the least-cost system adds 176 GW by 2040 (about +130% over 2020), with wind and solar providing roughly half of new capacity and their generation share rising from 4% to 55%. Adding protections shifts portfolios toward more solar, batteries, and some gas, and less hydropower; differences across Legal to All Exclusions are <10% of Base new capacity. A hydropower moratorium requires markedly more wind, solar, storage, and gas. - Effects of protections on technologies: Protections reduce selected wind capacity (exclusion of high-quality sites) and increase solar capacity. In no-carbon-target cases, new gas capacity needs vary non-monotonically across intermediate protections; All Exclusions adds only ~1.5 GW gas. - Carbon target interactions: Meeting a 50% emissions reduction by 2040 (vs. Base without target) requires ~50 GW additional renewables including hydro. With strong protections plus the carbon cap (All Exclusions), capacities increase notably: wind (+7.5%), solar (+29%), batteries (+23%), and hydropower decreases (−31%) relative to the low-carbon Base, resulting in a 43% capacity increase by 2040 vs. Base without carbon target. No extra gas capacity is selected under the carbon cap as protections increase (except with a hydro moratorium), though gas generation and batteries help balance higher solar. - Hydropower selection economics: Of ~41 GW planned/proposed hydro, only ~18 GW (with carbon target) and ~13 GW (without) are needed in 2040 in Base. Protections further reduce selected hydro by ~5.5 GW (with carbon target) and ~3 GW (without), so only ~10–12 GW (≈25% of planned/proposed) is cost-competitive by 2040. Some cost-competitive projects are excluded in protective scenarios due to socio-environmental impacts. High-capacity-factor sites in Kwanza and Zambezi basins dominate selections; some Congo basin projects (e.g., Inga 3) are generally too expensive unless others are unavailable. - Emissions and costs: Without the low-carbon cap, emissions slightly rise to 2035 then drop >10% by 2040 (coal retirements, cheaper VRE). The low-carbon target avoids ~100 Mt CO2/year in 2040 and ~350 Mt cumulatively (2020–2040). System cost impacts of protections (no carbon cap): +0.4% (Legal), +1.8% (Environmental), up to +2.3% (All Exclusions). Adding the carbon target increases costs by ~3% (Base) and ~4% (All Exclusions); jointly pursuing all protections and the carbon target raises costs by ~6.8% vs. Base with no carbon target. A hydropower moratorium yields the highest cost increases: ~+3.4% (no carbon cap) and ~+6% (with carbon cap). - Avoided impacts: In unconstrained Base, selected dams could inundate ~230–400 km² of legally protected areas and >1000–1500 km² of conservation areas and displace >150,000 people. Environmental and social screens reduce all assessed impacts by more than half; with social screens, displacement falls to <20,000. - Operations: Wind complements hydropower seasonally (dry season contributions), and hydropower provides evening ramping; with All Exclusions and no carbon cap, gas generation rises by ~23% to compensate for reduced hydro/wind and increased solar.

The results show that Southern Africa can meet growing electricity demand and achieve deep emissions reductions while largely avoiding development in areas with high social, terrestrial, and freshwater ecosystem value. A significant share of proposed hydropower capacity presents substantial socio-environmental risks, whereas low-impact wind and solar resources remain abundant even after stringent screens. System-level planning that explicitly co-optimizes siting and technology choice reveals that only a quarter of planned/proposed hydropower is economically needed by 2040, and much of it can be substituted by wind, solar, and storage with modest cost premiums. While protections and a low-carbon target increase costs, these premiums are small relative to typical cost overruns and delays for large hydropower, and to uncounted mitigation/compensation costs. Geographic differences in hydro selection slightly shift optimal transmission needs but do not require major cross-border buildouts beyond modeled corridors. The analysis underscores the importance of integrating conservation and community priorities into power system planning to reduce conflict risk, expedite development, and preserve critical freshwater ecosystems and livelihoods.

This study provides an integrated, spatially explicit framework for jointly evaluating socio-environmental siting constraints and least-cost power system development in Southern Africa. It demonstrates that: (1) strong land and freshwater protections eliminate many high-impact hydropower projects while leaving ample low-impact wind and solar potential; (2) only about 25% of planned/proposed hydropower is economically needed by 2040; and (3) meeting a 50% emissions reduction target by 2040 while applying comprehensive protections is feasible with modest system cost increases (~4–7%). Future work should more fully integrate river basin process metrics (e.g., sediment transport, aquatic biodiversity, fragmentation) and lifecycle GHGs from reservoirs, explore uncertainty in technology costs and hydrology, and assess country-level cost distribution and equity impacts. Enhancing regional transmission and power trade, and mobilizing international support to cover incremental costs, can further enable low-impact, low-carbon pathways.

The study does not quantify several important dam-related environmental impacts (e.g., sediment transport, reservoir GHG emissions, fish biodiversity, river fragmentation), partly due to data limitations and portfolio dependence. Cost and timeline overruns, mitigation/compensation expenses, and project-level permitting risks are not incorporated into system cost projections. Demand is treated as inelastic, technology cost trajectories use mid-case assumptions, and temporal operations are represented with selected days per month. Hydropower reservoir modeling was conducted for 34 large planned/proposed projects with detailed data; other projects were represented without explicit reservoir footprints. Country-level distributional cost impacts are not reported. These factors may affect the precision and generalizability of results, though the main conclusions are robust to scenario variations examined.