The Striking Impact of Natural Hazard Risk on Global Green Hydrogen Cost

Abstract: Due to climate change, natural hazards that affect energy infrastructure will become more frequent in the future. However, to incorporate natural hazard risk into infrastructure investment decisions, we develop an approach to translate this risk into discount rates. Thus, our newly developed discount rate approach incorporates both economic risk and natural hazard risk. To illustrate the impact of including the risk of natural hazards, we apply country-specific discount rates for hydrogen production costs. The country-specific relative difference in hydrogen generation cost ranges from a 96% surplus in the Philippines to a -63% cost reduction in Kyrgyzstan compared to a discount rate that only consists of economic risks. The inclusion of natural hazard risk changes the cost ranking of technologies as outcome of energy system models and thus policy recommendations. The derived discount rates for 254 countries worldwide are published in this publication for further use.

• The paper presents a dual-risk framework that integrates economic and natural hazard risks into discount rate calculations, leading to revised LCOH estimates.
• It employs a system-level energy model with hourly resolution and regional differentiation to quantify how risk factors affect cost sensitivity, showing significant surpluses and reductions.
• The analysis reveals that using uniform discount rates mischaracterizes country cost competitiveness, offering vital policy insights for resilient renewable infrastructure investments.

The Impact of Natural Hazard Risk on Global Green Hydrogen Cost: A Technical Analysis

Introduction and Motivation

This paper presents a rigorous methodology to assess how natural hazard risk, alongside economic risk, influences the levelized cost of hydrogen (LCOH) from renewables on a country-specific basis. While prior LCOH research often applies uniform discount rates, this contribution introduces a framework where both economic and natural hazard risks are explicitly quantified and incorporated in the derivation of the discount rate used for infrastructure investment calculations. This approach allows for a more granular and realistic estimation of LCOH, capturing spatial heterogeneity in risk exposure due to geography and governance.

Methodology

The core analytical innovation is the decomposition and recombination of discount rates. The country-specific final discount rate, $i_c$, is the sum (or a weighted sum) of an economic risk component (sourced primarily from Damodaran's country risk premia and credit default swaps) and a natural hazard risk component (quantified using the World Risk Report and related indices):

$$i_c = a \cdot i_{e,c} + b \cdot i_{n,c},\quad a + b = 1$$

where $i_{e,c}$ is the economic discount rate and $i_{n,c}$ is the natural hazard risk-derived discount rate, with $a$ and $b$ weighting their influence. The method ensures that the final $i_c$ does not exceed empirically established maxima, guarding against unrealistic cost inflation in highly risky contexts.

LCOH is then calculated through a system-level energy model (ETHOS.FINE), with multi-technology capacity optimization and hourly temporal resolution, incorporating power generation from onshore wind and PV, electrolysis, and hydrogen storage. All techno-economic parameters are regionally differentiated based on up-to-date IEA projections for 2050, and cost trajectories are referenced to 2023 USD.

Results and Numerical Findings

The analysis quantifies the sensitivity of LCOH to alternative compositions of risk in the discount rate. Key numerical findings include:

• The relative impact of incorporating natural hazard risk in the discount rate leads to LCOH surpluses up to 96% in the Philippines and reductions as steep as -63% in Kyrgyzstan, compared to models using only economic risk in the discount rate.
• The low correlation (Pearson $r = 0.31$) between economic and natural hazard discount rates at the country level underpins these large differentials and demonstrates the independent influence of both risk classes.
• Uniform discount rate approaches – typical in many major studies (e.g., a flat 8%) – are shown to mischaracterize country cost competitiveness, often underestimating costs for some high-risk countries and overestimating costs elsewhere by as much as 86% (e.g., in Somalia).
• The ten-year average of economic risk premia (rather than a single year's data) is shown to yield more stable, less volatile investment cost estimations, with annual fluctuations in Damodaran's rate introducing LCOH shifts of up to $5.64/kg$ hydrogen.

Notably, the inclusion of explicit natural hazard risk can invert the LCOH cost rankings among countries, impacting both system optimization recommendations and investment decision-making.

Theoretical and Practical Implications

The explicit modeling of natural hazard risk in discount rates reshapes optimal infrastructure allocation and associated policy. Countries previously considered low-cost locations for green hydrogen production may lose this status when their heightened exposure to natural hazards is considered. Policy recommendations and global trade patterns for H2 could shift, as underlying financial models adapt to more accurately reflect risk-adjusted capital requirements.

Practically, the methodology outlined can be generalized to any long-lived energy infrastructure where climate-induced risk is non-trivial. This is particularly salient for siting decisions in grids, storage, and renewables deployments as climate extremes become more frequent and severe.

The approach delineates between immutable factors (e.g., geolocation-driven hazard exposure) and modifiable ones (societal vulnerability, governance, adaptation measures), clarifying where national policy can meaningfully reduce LCOH via targeted risk reduction.

Future AI and Modeling Directions

Advances in AI-based geospatial hazard analytics, probabilistic risk assessment, and data-driven socio-economic modeling could enhance the resolution and dynamism of such discount rate formulations. Forward-looking models may incorporate real-time disaster frequency projections, infrastructure adaptation dynamics, or probabilistic Bayesian updating driven by continuously refreshed hazard data.

Further research may explore endogenizing the evolution of natural hazard risk (e.g., via climate models), copula-based dependency structures between economic and environmental risk, or dynamic discount rates wherein risk adjusts with ongoing mitigation investments.

Conclusion

This work demonstrates that global green hydrogen cost assessments are fundamentally contingent on a dual-risk framework integrating both economic and natural hazard exposures into discount rates. The resulting country-specific LCOH landscapes invalidate many of the homogeneous risk assumptions prevalent in energy systems modeling. The proposed methodology yields a robust scaffold for incorporating compound risk in capital cost planning, paving the way for more resilient, context-sensitive infrastructure investment strategies in the decarbonization transition.

The provision of country-specific discount rate data for 254 countries enables direct uptake by both modeling communities and policy analysts, likely impacting future studies spanning hydrogen, renewables, and resilient infrastructure finance.