Global transcontinental power pools for low-carbon electricity

The study addresses how to reliably and cost-effectively meet global electricity demand with near-100% renewable energy given the spatial and temporal variability of resources. Electricity accounts for about 40% of global energy-related CO2 emissions, and electrification of transport and industry increases the importance of decarbonizing power. While global technical potentials of solar, wind, and hydropower far exceed projected demand, temporal variability and uneven geographic distribution of resources create mismatches between supply and demand and lead to large cross-country cost disparities. The paper investigates whether forming transcontinental power pools (large-scale regional electricity trade and transmission) can ensure reliability and reduce costs in 2050 under different land-availability constraints.

Prior work has assessed benefits of international electricity trade. Guo et al. and Zhao et al. examined intercontinental trade and found higher renewable shares with increasing trade but did not model near-100% clean systems. Studies from Lappeenranta University modeled 100% renewable systems with trade within continental-scale pools (e.g., Europe, Sub-Saharan Africa, Northeast Asia, MENA), but without country-level spatial resolution. Earlier analyses did not explicitly assess how land-use constraints limiting renewable siting affect international trade needs and system costs. This study fills these gaps by modeling country-level investments and operations in 2050 with hourly resolution, high-resolution resource mapping, and explicit land-availability scenarios.

• Scope and scenarios: Modeled 100% renewable electricity systems across 211 countries/administrative areas for 2050, evaluating two scenarios: (1) Country scenario (no cross-border transmission), and (2) Transcontinental scenario (six regional power pools enabling international trade within regions: Sub-Saharan Africa; East Asia & Russia; Europe & MENA; North America; South America; Southeast Asia & Oceania). Two land-availability cases: (a) all suitable sites; (b) only the global top 10% suitable sites (excluding rooftop PV) based on a composite siting index. A sensitivity with top 25% sites and with higher global demand (76 PWh) was also run.
• Resource assessment: Global potentials developed at 0.01° × 0.01° resolution. A Development Potential Index (DPI) combined resource yield (capacity factors), distances to grid/roads/urban areas/rail or ports, land cover, and inverse population density with AHP-derived weights. Offshore wind potential assessed via levelized cost and filtered to within 200 km of coasts, excluding protected areas and sea ice, requiring wind speed ≥ 8 m/s. Rooftop PV potential assigned very high in urban areas with a land-use factor derived from rooftop area share and suitability.
• Capacity calculations: For each pixel, the least-LCOE technology was chosen; capacity per pixel equals land-use factor times pixel area. Country capacities sum pixel capacities. Generation potential computed from capacity factors and hours. Annual average capacity factors for technologies derived from Global Solar Atlas (PV GTI for r_PV), direct normal irradiation for CSP via empirical relation, Global Wind Atlas (IEC turbine classes), and hydropower from Hoes et al. Land-use factors from NREL; PV AC/DC adjusted by 1.17. Pumped hydro generation profiles monthly from IEA data; assumed CF 0.5 where lacking data (notably in Africa).
• System planning model: Used GridPath (open-source) for least-cost capacity expansion and dispatch. First optimized capacity investments with 3-hour time steps over all 365 days; then fixed capacities and simulated hourly (8760 h) operations to assess reliability. Storage: existing pumped hydro (10 h duration); battery storage allowed up to 24 h discharge and only used intra-day balancing. Planning reserve margin 15% of peak load per country. Transmission losses assumed 1.6% per 1000 km. Existing inter-country lines in Europe included for relevant comparisons. HVDC line lengths approximated by distances between population centroids.
• Costs and penalties: Penalty for loss of load $100 million/MWh; any modeled loss of load is assumed to be met by fossil generation. Levelized system cost includes renewables, fossil generation, and social cost of carbon (SCC). SCC $81/tCO2 in 2050 (3% discount rate). Fossil LCOEs: coal $95/MWh, gas $90/MWh; CO2 intensities from NREL ATB; coal-gas mix from World Bank. Technology capital/O&M cost projections from IRENA and NREL ATB; HVDC costs and losses from literature.
• Demand and profiles: 2018 electricity demand from EIA; growth 2030–2050 per IEA Sustainable Development (SDG) scenario in the main case; Net Zero 2050 (NZE) for sensitivity. Hourly wind/solar and load profiles for 42 major countries and 23 subregions from Tong et al. (2018 data), extended regionally where missing; hydropower monthly profiles from IEA (2015–2021), with assumptions where data absent.
• Metrics and accounting: Unmet demand in country scenario equals load minus renewable production; unmet demand assumed met by fossil fuels incurring LCOE and SCC. System costs computed per unit of electricity demand by country. In the transcontinental scenario, total costs include renewable, fossil, and transmission costs; country LCOE is demand-weighted average within the power pool. Costs assumed shared across member countries proportional to demand for benefit/cost incidence analysis.

• Reliability and adequacy:
  • All suitable sites: In the country (no-trade) scenario, renewables fall short by 0.8 PWh (2%) of global annual demand in 2050; with transcontinental power pools, unmet demand falls to 0.
  • Top 10% land constraint: Country scenario is 5.2 PWh (12%) short after accounting for temporal mismatch; pooling within six transcontinental regions reduces unmet demand to 0, enabling 100% renewable supply despite land constraints.
  • Country-level reliability challenges arise even where annual potential exceeds demand due to seasonal variability (e.g., Switzerland) and resource scarcity (e.g., South Korea, India under constraints). Existing interconnections modestly reduce shortages in some regions.
• Cost impacts:
  • All suitable sites: Transcontinental power pools reduce system LCOE by 5–52% across regions vs country-only systems; largest average reduction in Europe & MENA (52%, ~$47/MWh), with substantial decreases in PV ($24/MWh), onshore wind ($10/MWh), and storage ($8/MWh). North America sees smaller reductions (~5%) due to ample domestic resources.
  • Top 10% sites: Pooling reduces system costs notably in Europe & MENA (23% without prior interconnections; 21% with existing interconnections), South America (23%), and Sub-Saharan Africa (14%). Benefits are modest in North America (6%). In East Asia & Russia and Southeast Asia & Oceania, system costs change by +3% and +0.7% respectively due to higher transmission investment offsetting reductions in fossil and renewable costs.
  • Under land constraints, country-only systems can exceed $70/MWh in many Southeast Asian countries; regions with abundant high-quality sites (South and North America) maintain lower costs. In some cases (Europe & MENA) average costs appear lower with 10% site availability than with all sites due to reduced overbuild and some fossil backfill at assumed costs.
• Trade patterns:
  • All suitable sites: Annual traded electricity equals ~16% of global demand; Europe & MENA trades ~40% of its demand; Sub-Saharan Africa and Southeast Asia & Oceania >20%; East Asia & Russia, North America, South America ~10%. Largest net exporters include Syria and Oman (>0.2 PWh each); Germany is the largest net importer (~0.5 PWh). Globally, China is the largest net exporter (~0.9 PWh) and South Korea the largest net importer (~0.7 PWh).
  • Top 10% sites: Trade rises to ~30% of global demand. Southeast Asia & Oceania trades ~75% of regional demand (~2 PWh), with Australia exporting primarily to Indonesia (0.5 PWh), Vietnam (0.5 PWh), Thailand (0.4 PWh), and Malaysia (0.4 PWh). East Asia & Russia trades ~5 PWh (~25% of regional demand); net importers include South Korea (1 PWh), India (1 PWh), Japan (0.6 PWh), while net exporters include Pakistan (1.5 PWh), Iran (1.3 PWh), Afghanistan (0.8 PWh), and Kazakhstan (0.3 PWh). Europe & MENA imports >40% (3 PWh), with Syria, UK, Morocco, and Libya jointly exporting ~1.5 PWh; Germany and Italy are major importers. In North America, Canada exports ~0.3 PWh and the USA imports ~0.2 PWh.
• Distribution of benefits and costs within pools:
  • About one quarter of ~80 net exporters see >$10/MWh reductions due to reduced curtailment; benefits >$20/MWh are common among European exporters. About half of exporters face cost increases as they add capacity and share costs with importers (e.g., >$20/MWh increases in Pakistan and Australia under 10% sites).
  • Roughly three-quarters of net importers see ≥$5/MWh cost reductions; European importers often achieve >$20/MWh reductions.
• Robustness: With top 25% sites and with higher 2050 demand (~76 PWh), pooling still meets nearly 100% of demand and lowers overall costs in most regions; exceptions occur where increased renewable and transmission investments outweigh fossil cost reductions under tight land constraints (10% sites) in East Asia & Russia and Southeast Asia & Oceania.
• Modeling fidelity: Using 3-hour resolution over a full year in capacity expansion yields zero curtailment-driven loss of load in subsequent hourly operations; representative-day approaches with reserve margins can lead to ~1% curtailment, indicating the value of full-year modeling for reliability and cost quantification.

The findings demonstrate that forming transcontinental power pools is a robust strategy to tackle the spatiotemporal mismatch between renewable generation and demand while minimizing system costs. By aggregating diverse, high-quality renewable resources across countries and leveraging complementary temporal profiles, pooled systems reduce the need for overbuilding solar/wind and large-scale battery storage, substantially lowering LCOE and eliminating unmet demand, even under stringent land-availability constraints. The study highlights that international trade enables countries with poor renewable endowments (e.g., Japan, South Korea) to access low-cost clean power, accelerating decarbonization and mitigating high domestic costs. Exporters benefit through reduced curtailment and improved utilization of abundant resources, though cost-sharing arrangements can raise some exporters’ average system costs; still, total pool costs decline. The results underscore the importance of transmission expansion and coordinated market design. Practical implementation requires addressing geopolitical, regulatory, and market governance challenges, including ownership, financing, and equitable cost/revenue allocation. While long-term storage (e.g., hydrogen, synthetic methane) could further help balance seasonal variability, it cannot solve absolute resource shortfalls under land constraints without trade; even without long-duration storage, transcontinental pooling achieves reliability and cost-effectiveness in highly renewable systems.

This paper shows that continent-scale power pools can enable 100% renewable electricity supply reliably and at lower cost by sharing resources and expanding HVDC transmission. Under both unconstrained and land-constrained siting, pooling eliminates unmet demand and reduces system costs substantially, particularly in Europe & MENA and the Americas, with modest benefits in North America where resources are abundant domestically. The work advances prior literature by incorporating country-level resolution, hourly operations, and explicit land-constraint scenarios. Future research should: (1) develop and evaluate market pricing and cost-allocation mechanisms to equitably distribute benefits and costs among importers and exporters; (2) incorporate regional variations in technology costs, financing, and labor; (3) include long-duration storage technologies and assess their interactions with transmission expansion under land constraints; (4) refine land-use data and consider co-location of wind and solar to expand siting options; and (5) improve transmission cost estimates and explore learning-driven cost reductions for HVDC and related infrastructure.

• Technology cost assumptions are uniform across countries, not reflecting differences in cost of capital, access to technology, and labor.
• Land-use assumptions carry significant uncertainty; competing demands (agriculture, conservation) and alternative land-use factors could materially change siting potential. Co-location of technologies (e.g., wind with PV) was not allowed at a pixel, potentially understating practical potential.
• Long-duration storage options (e.g., hydrogen, synthetic methane) were not included in the core optimization; while discussed qualitatively, their exclusion may affect seasonal balancing outcomes and cost trade-offs.
• Transmission costs may be overestimated (conservative assumptions and limited learning considered); actual costs could be lower, making pooling more favorable.
• Load and renewable profiles rely on 2018 data scaled to 2050 and regional proxies where data are missing, introducing profile uncertainty.
• Battery storage constrained to intra-day operation (≤24 h discharge) may overstate the role of transmission relative to seasonal storage in some contexts.