The timing is significant. Exploration is moving toward ultra-deep formations more than 8,000 meters beneath the surface and into increasingly complex deepwater environments, while unconventional resources account for much of the growth in reserves and production. At such depths, familiar assumptions begin to fail. Pressures can exceed 140 megapascals and temperatures can rise above 200 degrees Celsius, creating conditions in which rocks deform unpredictably, fluids behave in unfamiliar ways, and conventional seismic imaging loses accuracy. Meanwhile, hydrocarbons must be studied alongside carbon capture, utilization and storage, hydrogen, geothermal energy, artificial intelligence, and the emerging economic value of subsurface data.

The challenge list was launched in 2021 by the journal’s editorial board, which invited contributions from researchers and drew on strategic reports from organizations including the International Energy Agency, the American Association of Petroleum Geologists, and the Society of Petroleum Engineers. Candidate questions were assessed through expert workshops, external review, and screening criteria that included originality, urgency, transformative potential, interdisciplinarity, long-term influence, and balance across the field. Researchers from China University of Petroleum ultimately assembled a portfolio of 100 fundamental questions spanning exploration and development, storage and pipeline networks, refining and petrochemicals, petroleum materials, carbon management and emerging energy systems, and energy economics and digital transformation.

The authors classify the problems into three broad scientific patterns, revealing that petroleum research is no longer focused solely on finding and extracting oil and gas. Sixty-eight challenges are mechanism-oriented, seeking theories that explain why complex processes occur across multiple scales. These include the coupled effects controlling hydrocarbon accumulation in ultra-deep formations and the behavior of gas, liquid, and solid interfaces during multiphase catalysis. Thirteen are technology-oriented, addressing engineering limits such as high-precision geophysical prediction and drilling-fluid rheology under extreme pressure and temperature. The remaining 19 are system-level problems, involving the optimization of entire energy chains, the resilience of global supply networks, and the ownership and pricing of industrial data.

Among the ten challenges highlighted as particularly disruptive is the search for a unified theory of hydrocarbon accumulation and preservation below 8,000 meters. Solving that problem could clarify how temperature, pressure, mineral reactions, fluid migration, and rock deformation interact in the deepest reservoirs, potentially extending effective exploration toward 10,000 meters. Another question concerns the survival limit of liquid hydrocarbons in ultra-deep formations. If researchers can determine how organic molecules remain stable—or transform—at temperatures approaching 250 degrees Celsius, they could revise long-standing assumptions in organic geochemistry about where liquid petroleum can exist.

Several of the highlighted problems connect petroleum engineering directly to climate technology. One concerns water-free fracturing using supercritical carbon dioxide, a fluid state reached above its critical temperature and pressure in which carbon dioxide combines gas-like mobility with liquid-like density. Researchers must understand how this unusual phase changes during injection, how it transports proppant into fractures, and how it interacts with rock and formation fluids. A successful approach could reduce freshwater use in hydraulic fracturing while creating a pathway for carbon dioxide storage. The authors suggest that, if technical and economic barriers are overcome, such systems could help shift carbon capture and storage from a costly obligation toward a revenue-generating industrial process.

Another proposed breakthrough is the development of reservoir nanotracers capable of revealing the distribution of remaining oil at pore scale. These engineered particles would need carefully designed surfaces so they can travel through complex porous networks, resist chemical degradation, and produce detectable signals without becoming trapped prematurely. In principle, their movement could provide a form of “underground CT” imaging, allowing engineers to map fluid pathways that conventional well measurements cannot resolve. The resulting information could improve recovery strategies by showing where oil remains, how it is connected, and which microscopic channels control its movement through the reservoir.

The list also treats the subsurface as a potential platform for renewable energy and large-scale storage. In enhanced geothermal systems, engineers inject fluid into hot rock to create or reactivate fractures through which heat can be recovered. A central challenge is understanding the dynamic interaction between artificial fractures created by stimulation and natural fracture networks already present in the rock. Better control of these interactions could improve heat extraction while reducing the risk of induced seismicity, a major social and regulatory concern. For underground hydrogen storage, researchers must determine how hydrogen reacts geochemically and biologically with depleted reservoirs. Microorganisms may consume hydrogen, while minerals and formation fluids can alter its composition; controlling these losses could raise the working-gas ratio from below 50 percent to above 80 percent.

Ultra-long gravity heat pipes represent another proposed route to extracting heat from hot dry rock. These devices use phase changes in a working fluid to transport thermal energy over long distances, but their performance depends on evaporation, condensation, flow resistance, material stability, and pressure management at extreme temperatures. Identifying the limits of heat transfer and selecting suitable working fluids could lower the cost of geothermal electricity, with the authors estimating a possible levelized cost of 5 to 8 U.S. cents per kilowatt-hour. Such a target would place geothermal power closer to the cost range of other competitive low-carbon energy sources, although substantial engineering validation would still be required.

Artificial intelligence appears throughout the roadmap not simply as a tool for faster computation, but as a force changing how petroleum science is practiced. More than 30 of the challenges depend heavily on data, algorithms, or intelligent systems. Physics-informed neural networks can combine governing equations with observations, while differentiable programming allows models to be optimized through machine-learning methods. Neural operators can learn relationships between fields, such as pressure, temperature, and saturation, across changing geological conditions. Together, these techniques could produce “grey-box” models that retain the flexibility of data-driven systems while preserving some physical interpretability. Digital twins would extend this approach by continuously updating virtual representations of reservoirs as new measurements arrive, potentially turning reservoir management from periodic, experience-based intervention into real-time adaptive control.

The roadmap also identifies a rapidly expanding frontier in energy economics and global infrastructure. Satellite observations, vessel tracking, and machine-learning forecasts could enable daily monitoring of tanker capacity and freight rates, offering an independent view of oil and gas transportation markets. At the same time, subsurface data—once treated mainly as technical records—could become a factor of production with measurable ownership rights, economic value, and pricing mechanisms. The authors argue that future research must establish how geological information is created, shared, protected, and monetized. More broadly, the 100 questions reflect a transition from petroleum as a discipline of resource extraction to what the authors call “smart integrated energy science,” combining subsurface storage, oil, gas, hydrogen, electricity, heat, carbon-cycle management, and intelligent decision-making. The list is deliberately open-ended: as researchers solve existing problems, new uncertainties will emerge. Its central message is that defining what remains unknown may be the first step toward transforming the energy system.

Subject of Research: Not applicable

Article Title: One hundred grand challenges in petroleum science.

Web References: https://doi.org/10.1016/j.petsci.2026.06.013

Image Credits: Xiao, L. Z., Jin, Y., Zhang, L. B., et al. (journal cover)

Keywords: petroleum science, ultra-deep reservoirs, artificial intelligence, digital twins, carbon capture and storage, geothermal energy, underground hydrogen storage, supercritical carbon dioxide fracturing, reservoir nanotracers, energy transition

A groundbreaking study published recently in Nature Energy leverages cutting-edge machine-learning techniques to uncover the most potent predictors of public support for energy and climate mitigation policies among informed citizens across Europe. This research not only rigorously assesses a vast array of potential variables but validates its insights by accurately forecasting the outcome of a real-world referendum in Switzerland focused on renewable energy. Importantly, the study extends its scope to verify the generalizability of its findings across six European countries, scrutinizing public support for an array of climate mitigation strategies.

The urgency of decarbonizing the energy sector cannot be overstated. Transitioning from fossil-fuel dependence to renewable energy sources is central to meeting international climate goals, reducing greenhouse gas emissions, and combating global warming’s escalating impacts. Yet, while technology and economics often dominate the conversation, the role of informed public opinion is just as critical. Policymakers require reliable insights into the psychosocial and perceptual factors that shape support or opposition to complex policy instruments. This study bridges that knowledge gap through an innovative analytical framework.

Employing sophisticated machine-learning models, the research team sifted through extensive survey data collected from informed citizen cohorts, parsing out the most meaningful predictors of policy endorsement. These algorithms, designed to handle complex, multidimensional datasets, excelled at identifying patterns and ranking variables by their predictive power, surpassing traditional statistical methods in both accuracy and granularity. This methodological advance enabled the researchers to ascertain the relative weight of variables across diverse energy policy contexts.

One of the pivotal findings was the outsized influence of affective responses—emotional reactions underpinning individual attitudes towards energy policies. Unlike purely cognitive or rational evaluations of policy merit, affective responses tap into deeper feelings such as hope, fear, and moral conviction. These emotional dimensions were found to directly impact support levels, highlighting the importance of addressing public sentiment alongside factual information in policy communication strategies.

In addition to emotions, the study identified societal and environmental policy-impact beliefs as strong predictors. These beliefs reflect how citizens perceive potential benefits and trade-offs of mitigation measures not only for the environment — such as pollution reduction or biodiversity protection — but also for society at large, including economic opportunities and health improvements. Notably, support increased when policies were seen to generate equitable societal benefits, underscoring the role of fairness perceptions in shaping energy policy preferences.

Fairness perceptions emerged as a critical dimension, reinforcing the idea that public endorsement hinges on trust that policies distribute benefits and burdens justly. Equity concerns span socioeconomic factors, geographic considerations, and intergenerational justice, and this study shows that perceived discrepancies can dampen support. Hence, transparent communication about policy impacts and inclusive policymaking that addresses fairness directly will be vital to sustaining support.

The innovative aspect of this research lies not only in identifying individual predictors but also in integrating perceived trends in collective public support over time. The sense that a policy is gaining momentum, winning broader acceptance, or becoming a social norm was shown to significantly boost individual endorsement. This social dynamic provides policymakers with a psychological lever: framing mitigation efforts as part of an irreversible, widely embraced movement could mobilize fence-sitters and hesitant constituents.

Validity and real-world applicability of the machine-learning model were demonstrated through its deployment to forecast the outcome of a landmark renewable energy referendum in Switzerland. The model achieved remarkable accuracy, confirming that the key predictors identified do not just exist theoretically but have practical explanatory and predictive power. This case study underscores the potential to anticipate societal responses to policy proposals before implementation, enabling proactive strategy adjustments.

Extending beyond Switzerland, the team tested the robustness of their model across a broader European context, incorporating data on public support for various mitigation policies in six countries. The model successfully generalized, confirming the universality of the core predictors—affective responses, fairness perceptions, impact beliefs, and social trend awareness—as foundational elements driving energy policy support across diverse national landscapes. This European-wide validation signals that despite cultural and political differences, the psychological mechanics of policy endorsement show consistent patterns.

By illuminating these intricately intertwined predictors, this research injects fresh rigor into the design and implementation of energy policies. It stresses the necessity for policymakers to craft narratives and strategies that resonate emotionally while demonstrating tangible benefits and fairness. Moreover, the study advocates for continuous public engagement that nurtures perception of positive social dynamics to amplify collective buy-in. These insights could guide communication campaigns, stakeholder dialogues, and legislative frameworks to better align with public values.

From the scientific perspective, the application of machine learning in social science research represents a paradigm shift. It allows handling extensive, multi-faceted datasets with enhanced objectivity and predictive accuracy. This study exemplifies how advanced computational methods can unpack complex human attitudes toward climate action, breaking free from reductive approaches. Such integrative exploration of psychological, social, and environmental dimensions provides a richer, more actionable understanding, vital in the multidimensional challenge of climate policy.

This research also brings to light important considerations regarding public knowledge and information. The focus on informed citizens emphasizes that depth of understanding enhances the discernment of policy impacts and the reliability of expressed preferences. Consequently, education, transparent information provision, and combating misinformation remain indispensable priorities to foster an informed electorate capable of making decisions aligned with long-term sustainability.

Looking forward, the implications of this study are profound. Governments aiming to implement ambitious climate policies can leverage these predictive insights to tailor strategies that maximize public acceptance. This could reduce political resistance, accelerate policy deployment, and ultimately hasten the transition towards a low-carbon future. The approach also signals potential for adaptive policymaking informed by real-time sentiment tracking assisted by machine-learning analytics.

However, challenges remain. Emotions and fairness are subjective and may evolve rapidly in response to external events, media framing, or political rhetoric. Maintaining continuous engagement and updating models to reflect shifting public moods will be essential to preserve predictive relevance. Additionally, the interplay between local contexts and broader societal trends requires nuanced understanding to avoid one-size-fits-all policy messaging.

In conclusion, this pioneering research delineates a pathway for harmonizing science, policy, and society in the quest to combat climate change. Through harnessing advanced analytical tools and centering psychological and social predictors, it lays the groundwork for more effective, citizen-aligned energy policies. The promise of accelerating Europe’s—and potentially the world’s—energy transition hinges not only on technology and economics but fundamentally on decoding and integrating the human factors that shape democratic support for transformative change.

Subject of Research: Predictors of informed energy policy support and public attitudes towards climate mitigation measures across Europe

Article Title: Predictors of informed energy policy support across Europe

Article References:

Krainz, M., Sorgato, V., Vallaeys Mora, I. et al. Predictors of informed energy policy support across Europe.
Nat Energy (2026). https://doi.org/10.1038/s41560-026-02050-5

Image Credits: AI Generated

DOI: https://doi.org/10.1038/s41560-026-02050-5

Keywords: Energy policy support, climate mitigation, decarbonization, machine learning, public opinion, affective responses, fairness perceptions, environmental beliefs, social trends, Europe