Comparative assessment of United States coastal hubs for large scale electrochemical marine carbon dioxide removal

Spatial Analysis of Coastal Facilities

An assessment of the geographical distribution of the facilities revealed important insights about their locations along the U.S. coastline. Most facilities are located near major metropolitan areas, as proximity to these cities is crucial for their operations. For example, on the West Coast, facilities are concentrated around the Greater San Francisco and Los Angeles Areas, while along the Gulf Coast, they are located in or near the Greater Houston area. Similarly, on the East Coast, facilities are situated around major cities (Fig. 2A).

A Geographical distribution of 38 coastal facilities identified as potential colocation sites for e-mCDR deployment along the U.S. coastline. Facilities are distributed relatively uniformly, with clusters forming near major metropolitan areas. The three facility types included power plants (•), desalination plants (▴), and LNG terminals (■). B Hierarchical clustering dendrogram illustrating the classification of facilities into five distinct clusters, each representing a potential e-mCDR hub based on geographical proximity. C Identified e-mCDR hubs and facility composition. The five hubs—Northeast, Southeast, South, West, and Northwest—are delineated using the Convex Hull algorithm. Each hub’s composition is represented by a pie chart indicating the percentage distribution of power plants, desalination plants, and LNG terminals within the hub.

A comparison of facility types across different geographic regions revealed distinct patterns. The East predominantly includes power plants, the South hosts a higher number of LNG terminals, and the West features more seawater desalination plants⁶⁵. This distribution aligns with the unique needs and characteristics of each region. For instance, the East’s higher population density necessitates a greater number of power plants to meet electricity demands⁶⁶. The South’s abundance of LNG terminals is attributed to being located on the coast of the Gulf of Mexico, which is favorable for LNG export/import terminals^67,68. In contrast, the West requires desalination plants to address the region’s limited freshwater resources⁶⁹. These regional variations highlight the functional specialization of facilities based on local demands and resources.

Hub Analysis

Using the HC method based on the geolocation of facilities, five distinct clusters were identified through the dendrogram, shown in Fig. 2B. These clusters correspond to five coastal regions that vary significantly in terms of economic conditions, population, and climate⁷⁰. These regions are designated as the Northeast, Southeast, South, West, and Northwest hubs, each representing a potential e-mCDR hub. These identified hubs were illustrated on the US map using the Convex Hull method (Fig. 2C).

An initial analysis of the facility distribution within each hub reveals notable differences. The Northeast and Southeast hubs are characterized by water intake capacity predominantly from power plants, accounting for 98% and 97% of their total intake, respectively. In the South hub, power plants contribute 61% of the water intake capacity, followed by LNG terminals at 35%. Despite housing several desalination plants, power plants’ water intake capacity dominates the West hub, representing 96.5% of its total water intake capacity. In contrast, the Northwest hub is unique, with 94% of its water intake capacity attributed to LNG terminals (Fig. 2C). Overall, the majority of water intake capacity across all hubs is associated with power plants, highlighting their significant water consumption compared to LNG terminals and desalination plants. Although the South and West hubs have fewer power plants, this facility type still accounts for the largest share of water intake capacity. The Northwest hub is an exception, as the lack of large fossil fuel-based power plants results in LNG terminals being the primary contributors to water intake capacity. Among the identified hubs, the South stands out as the most diverse, with a balanced mix of facility types.

Comparative Analysis of Hub Criteria

To gain deeper insights into the suitability of each hub for e-mCDR implementation, a comparative analysis was conducted based on the seven evaluation criteria previously established. Figure 3 provides a visual representation of the hubs’ performance across these criteria, facilitating a comprehensive understanding of their strengths and limitations. The Northeast hub demonstrated relatively high CO₂ removal capacity, primarily attributed to the high water intake of the facilities located in this hub. The hub also offers high grid emissions efficiency, making it a viable option within the hybrid pathway to achieve net carbon negative outcomes. However, the Northeast hub ranked low in facility diversity, with power plants being the predominant colocation facilities. Considering its strengths and weaknesses, the Northeast hub is best suited for e-mCDR projects that focus on the hybrid pathway, prioritizing total CO₂ removal capacity with a net carbon negative.

Comparison of the five e-mCDR hubs across the seven hub-level criteria, highlighting their relative strengths and limitations. These criteria included CO₂ removal capacity (ktonne CO₂/day), removal affordability (kg CO₂/$), grid emissions efficiency (kWh/kg CO₂), social vulnerability index (-), local carbon footprint (kg CO₂/year/capita), facility diversity index (-), and hydrogen management infrastructure (-).

The Southeast hub demonstrates high removal affordability and a favorable social vulnerability index. The relatively low electricity costs in states, such as Florida and Georgia are primarily driven by the region’s reliance on cost-effective natural gas and nuclear power generation⁷¹. Additionally, the region exhibits a high social vulnerability index due to factors, such as higher population densities in coastal areas, economic disparities, and exposure to climate-related hazards. The hub possesses average local carbon footprint values, as the region does not host a significant number of CO₂-emitting industries. Despite these advantages, the Southeast hub faces limitations in grid emissions efficiency, with relatively low integration of clean energy sources into the grid. Furthermore, the region lacks significant hydrogen management infrastructure, which could pose challenges for implementing hydrogen-focused e-mCDR pathways.

The South hub demonstrated the highest performance across five of the evaluated criteria, including removal affordability, social vulnerability index, local carbon footprint, facility diversity index, and hydrogen management infrastructure. Similar to the Southeast hub, the South benefits from cost-effective electricity, primarily due to the widespread availability of natural gas, wind, and nuclear power generation in states, such as Texas and Louisiana⁷¹. The social vulnerability index is also notably high in this region, largely because many of the facilities are situated in areas designated under the Justice40 initiative⁷², which aims to address environmental and economic disparities in disadvantaged communities. The South hub’s high local carbon footprint is attributed to the presence of some of the largest CO₂-emitting industries in the country, primarily within the oil and gas sector⁷³. Texas and Louisiana are major energy-producing states, housing numerous refineries, petrochemical plants, and processing facilities that contribute significantly to national CO₂ emissions. One of the South hub’s key strengths is its high facility diversity, including power plants, desalination plants, and LNG terminals. This diversity is crucial from a logistical perspective, as it enhances the hub’s adaptability to potential regulatory changes and technical challenges. For instance, if stringent environmental regulations were imposed on a specific facility type, such as desalination plants, or if retrofitting e-mCDR systems into certain process flow diagrams proved challenging, the hub’s diversity would provide flexibility in selecting suitable alternative facilities for deployment, thereby mitigating logistical and technical constraints. Additionally, the South hub includes robust hydrogen management infrastructure, with Texas and Louisiana being home to some of the largest H₂ storage facilities, as well as extensive pipeline networks and transportation systems^74,75,76. The region hosts several large hydrogen hubs and pipelines to meet the growing demand for H₂ as an alternative energy source⁷⁷. The generated H₂ can also serve as valuable feedstocks for numerous industrial units located within or near the hubs, further enhancing the economic feasibility of e-mCDR deployment.

The South hub also has certain limitations that may impact its potential for e-mCDR implementation. The hub demonstrated the lowest CO₂ removal capacity among all evaluated regions. This limitation is primarily attributed to the low overall seawater intake capacity. Despite achieving a low score in grid emissions efficiency, the South hub has a sufficient score for the hybrid pathway to provide a net carbon negative result. This performance can be attributed to the region’s substantial wind energy generation capacity, with Texas being the largest wind electricity producer in the United States⁷⁸. The state has made significant investments in wind power infrastructure, contributing to a relatively higher share of renewables compared to other industrialized regions. However, fossil fuels still dominate the energy mix, with natural gas and coal remaining primary energy sources. Louisiana continues to heavily depend on fossil fuels, with minimal integration of renewable energy sources into its grid⁷¹.

Considering all evaluated criteria, the South hub emerges as an attractive location for implementing e-mCDR technologies. Its strong removal affordability, extensive hydrogen management infrastructure, and high facility diversity collectively position it as a strategic choice for large-scale deployment. The availability of well-established industrial networks and extensive transportation infrastructure further enhances the hub’s attractiveness for integrating e-mCDR with existing processes. While the Southeast hub excels in criteria, such as removal affordability and social vulnerability index, the South hub demonstrates even greater strengths in these areas. This superior performance, coupled with the geographic proximity of the two hubs, suggests that the South hub is likely to attract e-mCDR projects that would otherwise consider the Southeast.

The Western hub exhibited the highest CO₂ removal capacity among all evaluated hubs. This superior performance is attributed to the high water intake capacity of this hub. Although the cost of electricity is relatively high in this region, which lowers its removal affordability, it offers a very high clean energy capacity, mostly located in Southern and Central California, introducing substantial renewable energy sources into the grid⁷⁹. The Western hub also benefits from well-developed hydrogen management infrastructure, providing access to existing storage facilities and transportation networks that could support e-mCDR deployment. The development of H₂ infrastructure is driven by environmental policies⁸⁰, with H₂ infrastructure being particularly focused on its use in transportation⁸¹. One of the primary limitations of the Western hub is its low facility diversity index, as the majority of its water intake capacity comes from power plants. This lack of diversity may introduce logistical challenges, as previously discussed, potentially limiting the hub’s adaptability to regulatory changes or technical constraints associated with integrating e-mCDR systems into desalination operations.

The Northwest hub achieved a moderate score in grid emissions efficiency, with most facilities located in Northern California and Oregon, regions that have some of the highest renewable energy integration in the country^71,82. The availability of renewable energy resources, from hydroelectric, wind, and solar power, enhances the region’s potential for low-carbon e-mCDR operations. Similar to most of the hubs, the Northwest hub also ranked high in social vulnerability index, largely due to the presence of several Justice40-designated communities that can benefit from sustainable climate initiatives⁷². The hub demonstrated moderate performance in other criteria, including removal affordability, local carbon footprint, and facility diversity index. A key limitation of the Northwest hub is its underdeveloped hydrogen management infrastructure, which could hinder the integration of hydrogen-focused e-mCDR pathways and limit opportunities for utilizing captured H₂ as a feedstock for local industries, if new facilities (e.g., for Sustainable Aviation Fuels, SAFs) are not built. Given its moderate overall score, the Northwest hub remains an attractive option for implementing e-mCDR systems that prioritize low-carbon operations and sustainability goals, while further development of H₂ infrastructure would unlock additional incentives for e-mCDR deployment.

Ranking of Hubs Based on Weighted Criteria

The criteria weights were determined using the AHP method to reflect their relative importance in evaluating e-mCDR hub suitability. The results, presented in Fig. 4A, indicated that CO₂ removal capacity holds the highest weight at 24%, followed by removal affordability at 23%. Green energy share contributed 23%, while local carbon footprint contributed 12%. Hydrogen management infrastructure was weighed at 9%. The facility diversity index and social vulnerability index carried the lowest weights at 6% and 4%, respectively.

A Criteria weights assigned using the Analytic Hierarchy Process (AHP), with CO₂ removal capacity receiving the highest weight, followed by removal affordability, and grid emissions efficiency. B Hub rankings across individual criteria, determined using normalized weighted scores. C Overall hub rankings derived from the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) method, indicating that the South hub achieved the highest relative closeness score (C = 0.60), closely followed by the West and Northeast (C = 0.52).

This weight distribution highlights the central importance of CO₂ removal capacity. Removal affordability closely follows in importance, given its critical impact on the economic feasibility of e-mCDR projects. The significant weighing of grid emissions efficiency highlights the growing emphasis on sustainable energy sources in carbon removal initiatives. If the grid emissions efficiency is low (i.e., energy mix used to power these mCDR approaches is largely dominated by fossil fuels), then the overall process may not be carbon negative. However, as the share of renewable energy increases over the years, the net carbon removal potential of these hubs is expected to improve. Local carbon footprint and hydrogen management infrastructure, while lower in weight, still contribute meaningfully to the overall assessment. The relatively lower weights assigned to the facility diversity index and social vulnerability index suggest that while these factors are considered, they play a less critical role in determining the suitability of e-mCDR hubs for the DOC pathway.

Analyzing the top- and lowest-performing hubs for each weighted criterion is essential for understanding the factors influencing overall hub rankings. While Fig. 3 presented the absolute scores of each criterion for the hubs, Fig. 4B illustrates the hub rankings based on normalized weighted scores determined through AHP. Regarding CO₂ removal capacity, the West and Northeast hubs outperformed the others, primarily due to their high water intake capacities. The remaining hubs exhibited relatively lower performance in this criterion. Removal affordability demonstrated substantial variability, with the South and Southeast hubs exhibiting the higher values, indicating lower electricity costs. In contrast, the grid emissions efficiency criterion showed relatively minor variation across the hubs, with the Northeast and West Coast hubs (i.e., West and Northwest) achieving slightly higher scores due to the widespread integration of renewable energy into their regional grids^83,84,85. It is worth to note that all the hubs meet the grid emissions efficiency threshold for net carbon negativity which makes them suitable for e-mCDR integration⁴⁵. In terms of the local carbon footprint, the South hub exhibited the highest score, indicating its substantial contributions to national CO₂ emissions⁸⁶. This highlights the potential for e-mCDR implementation to achieve meaningful emission reductions in this region, whereas other hubs exhibit relatively lower emissions and may offer less immediate impact in this regard. The hydrogen management infrastructure criterion was led by the South hub, followed by the West hub, reflecting the extensive availability of H₂ transportation and storage facilities in these regions^80,81,87,88. Additionally, the South hub stood out in the facility diversity index criterion, highlighting its broad mix of facility types, which enhances flexibility and adaptability for e-mCDR deployment. Finally, for the social vulnerability index, all hubs except the Northeast ranked high, suggesting that e-mCDR implementation in these regions could provide substantial community benefits. This further reinforces the importance of deploying such technologies in socially vulnerable areas to address environmental and economic disparities.

All weighted criteria were incorporated to rank the hubs using the TOPSIS method. The analysis revealed that the South hub achieved the highest overall relative closeness score (C = 0.60), closely followed by the West and Northeast hubs (C = 0.52). The Southeast (C = 0.43) and Northwest (C = 0.24) hubs ranked fourth and fifth, respectively (Fig. 4C). The South hub’s top ranking is primarily attributed to its strong performance across multiple criteria, particularly in removal affordability, hydrogen management infrastructure, local carbon footprint, and facility diversity index. The West hub ranked highly due to its superior CO₂ removal capacity, grid emissions efficiency, and well-established hydrogen management infrastructure. Given the varied performance across the hubs, categorizing them based on their overall suitability provides a more practical perspective. Instead of a linear ranking from first to fifth, the hubs can be grouped into high potential (West, South, and Northeast), moderate potential (Southeast), and low potential (Northwest) hubs.

Sensitivity Analysis

The narrow margin between the top three hubs indicates that slight changes in the criteria’s weights could influence the overall rankings. Therefore, a sensitivity analysis of the weight assignments was conducted, and the adjusted weights were subsequently applied in the TOPSIS method to re-rank the hubs. The sensitivity analysis was conducted on all seven evaluation criteria, and the results are presented in Fig. 5, highlighting the four most influential criteria impacting each hub’s ranking. For the South hub, CO₂ removal capacity emerged as the most influential criterion, exerting the greatest impact on its closeness score (C). It was followed by removal affordability, local carbon footprint, and hydrogen management infrastructure, indicating the hub’s reliance on these factors for maintaining its top position. Similarly, in the case of the West hub, its ranking was found to be most sensitive to the weight assigned to CO₂ removal capacity. This criterion was followed in importance by removal affordability, local carbon footprint, and facility diversity index. For the Northeast hub, the ranking was primarily influenced by CO₂ removal capacity, with local carbon footprint, removal affordability, and facility diversity index also playing significant roles. In contrast, the ranking of the Southeast hub was found to be relatively stable, with no single criterion exerting a substantial impact on their overall ranking positions. The relative closeness score for the Northwest hub is mainly impacted by the removal affordability, and CO₂ removal capacity.

Sensitivity analysis was performed on the Analytic Hierarchy Process (AHP)-derived weights to assess the impact of subjectivity in the pairwise comparison process. The analysis involved systematically adjusting the weight of each criterion by a factor of 2, both increasing (upper bound) and decreasing (lower bound), followed by recalculating the weights using AHP and re-ranking the hubs using the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). The sensitivity analysis was conducted on all seven hub-level criteria, and the results highlighted the four most influential criteria affecting each hub’s ranking.

For the high-potential hubs, West, South, and Northwest, which performed well across most criteria, their rankings were primarily influenced by their most weighted criterion—CO₂ removal capacity. This criterion represented the primary determinant in their performance, making their rankings highly sensitive to any changes in this criterion. As a result, an improvement or deterioration in CO₂ removal capacity could significantly impact these hubs’ overall standing. For instance, if the weight of CO₂ removal capacity, which reflects the relative importance of the criterion, is decreased to its lower bound, the South hub would become the single dominant hub with a C of 0.71, increasing its margin over the West (C = 0.38) and the Northeast (C = 0.36) hubs. However, if the weight is increased to its upper bound, the South hub would rank third (C = 0.49), after the Northeast hub (C = 0.64), while the West hub slightly surpasses to take first place (C = 0.65). The next most influential criteria were those that differentiated the most between the high-potential hubs. For instance, the South hub excelled in removal affordability, local carbon footprint, hydrogen management infrastructure, and facility diversity index, while the West and Northeast hubs struggled the most in all of these criteria. This suggests that further improvements in these criteria reinforce the South hub’s competitive advantage and contribute to enhancing its overall ranking. Conversely, these criteria represent key limitations on the West and Northeast hubs’ performance, acting as bottlenecks that limit their overall potential. Any improvement in these areas yields significant gains in the hub’s overall ranking. For the moderate and low potential hubs, which generally perform strongly in one or two specific criteria, their rankings were primarily influenced by those criteria. For example, the rankings of the Southeast hub were most impacted by the removal affordability, as it represents its strongest criterion.

Beyond the weight-based sensitivity analysis, an additional round of sensitivity analysis was conducted to evaluate the influence of specific facility types and infrastructure characteristics on hub rankings. This analysis aimed to assess how the elimination of key facilities within each hub affects the overall ranking outcomes, providing further insight into the structural dependencies of e-mCDR deployment via the hybrid pathway. Three distinct scenarios were examined. In the first scenario (Fig. 6A), the facility with the highest water intake was removed from each hub to determine the extent to which hub rankings rely on a single high-capacity facility. Since water intake capacity directly influences CO₂ removal potential, this analysis evaluates the resilience of each hub when its largest seawater-processing facility is excluded, for example, due to logistical constraints in implementation, challenges in retrofitting, facility-specific regulatory restrictions, or potential operational disruptions. In the second scenario (Fig. 6B), all power plants were removed from each hub. Power plants generally exhibit the highest water intake capacity among the three facility types. Eliminating power plants simulates a scenario in which e-mCDR deployment is limited to smaller-scale systems that are better suited for facilities with lower water intake capacities, such as desalination plants and LNG terminals. This analysis provides insight into how hub rankings would shift under a deployment model focused on lower-capacity e-mCDR implementations rather than large-scale systems co-located with high-intake power plants. In the third scenario (Fig. 6C), all LNG terminals were removed from each hub to investigate their role in supporting e-mCDR deployment. Unlike power plants and desalination facilities, LNG terminals do not operate continuously at full capacity, and their long-term availability may be uncertain due to shifts in energy markets and policy changes⁸⁹. This analysis assesses the extent to which hub rankings are influenced by the presence of LNG infrastructure and evaluates the potential challenges of e-mCDR deployment in hubs without LNG terminal support.

Sensitivity analysis was conducted to evaluate the impact of removing specific facility types on hub rankings, providing insight into the structural dependencies of e-mCDR deployment. Closeness scores (C) were recalculated using the TOPSIS method for each of the following scenarios. A Scenario I: Removal of the highest water intake facility from each hub to assess the resilience of hub rankings when a single high-capacity facility is unavailable due to logistical constraints or retrofitting challenges. B Scenario II: Removal of all power plants from each hub to simulate a scenario where e-mCDR deployment relies solely on smaller-scale seawater-processing facilities, such as desalination plants and LNG terminals, offering insights into hub suitability for lower-capacity e-mCDR systems. C Scenario III: Removal of all LNG terminals from each hub to examine their influence on hub rankings, considering that LNG terminals do not operate continuously at full capacity and their long-term availability is uncertain due to shifts in energy markets and policy changes. The labels in each panel represent the two most impacted hubs for each scenario, with values indicating the percentage change in C relative to the base scenario.

The Northwest hub exhibited the greatest sensitivity to the removal of its highest water intake facility, with a 20% decrease in its closeness score (C) compared to the base scenario, which was represented in Fig. 4C. This was followed by the South hub, which showed a 14% reduction in C. These findings suggest that the Northwest hub’s ranking is disproportionately reliant on a single high-capacity facility, making its suitability for e-mCDR deployment particularly vulnerable to logistical constraints or facility-specific limitations. In contrast, the relatively lower impact observed for other hubs indicates a more distributed reliance on multiple facilities, reducing the risk associated with the unavailability of any single high-capacity site (Fig. 6; Scenario I).

The removal of power plants had a profound effect on hub rankings, with the Northeast hub experiencing the most significant decline, a 71% decrease in C, followed by the West hub, which showed a 46% reduction. This substantial drop highlights the Northeast hub’s strong dependence on power plants, which account for 98% of its total water intake. Conversely, the South hub emerged as the highest-ranked hub under this scenario, with a C of 0.79, indicating that this region presents the most favorable conditions for lower-capacity e-mCDR implementations that rely on smaller seawater-processing facilities, such as desalination plants and LNG terminals (Fig. 6; Scenario II).

The impact of removing LNG terminals was less pronounced than the removal of power plants, as LNG facilities generally contribute a smaller fraction of total water intake across hubs. The Northwest hub experienced the greatest reduction in ranking, with a 20% decrease in C, followed by the South hub, which showed a 7% decline. In this scenario, the high potential hubs (South, West, and Northeast hubs) obtained similar high ranking (C = 0.57), indicating that it is the most reliable for e-mCDR deployment, as LNG terminals operate intermittently and face long-term uncertainty due to shifts in energy markets and policy changes (Fig. 6; Scenario III).

The results of these scenarios provide a deeper understanding of the sensitivity of hub rankings to facility-level variations, informing future site selection and infrastructure planning for e-mCDR implementation. In general, the Northwest hub is the most susceptible to disruptions, as it exhibited the highest sensitivity across multiple scenarios, making it a relatively uncertain candidate for future e-mCDR deployment.