INTELLIGENCE

Explore Intelligence, Econometrics, News, Science, and Standards that translate climate evidence into defensible infrastructure investment, valuation, risk, and resilience decisions.

INDUSTRIES

Foundational Research Gaps And Future Directions For Digital Twins

Future weather and climate conditions — safer, more efficient engineering processes

“A digital twin is a set of virtual information constructs that mimics the structure, context, and behavior of a natural, engineered, or social system (or system-of-systems), is dynamically updated with data from its physical twin, has a predictive capability, and informs decisions that realize value. The bidirectional interaction between the virtual and the physical is central to the digital twin.”

The Committee on Foundational Research Gaps and Future Directions for Digital Twins adopts a modified definition of a digital twin from that initially formulated by the AIAA Digital Engineering Integration Committee in 2020. According to this refined interpretation, a digital twin consists of virtual informational replicas that accurately reflect systems’ structure, context, and dynamics, whether natural, engineered, or social, and systems composed of interlinked subsystems.

This digital representation is continually synchronized with its physical counterpart through data updates, encompasses the ability to forecast future states, and enables decision-making processes that generate value. The essence of a digital twin lies in its two-way communicative link, allowing for continuous data exchange from the physical entity to its digital counterpart and vice versa, facilitating informed decision-making processes, whether automated or involving human intervention.

The committee’s elaboration further broadens the scope of what constitutes digital twins’ physical counterparts, spanning artificial constructs, organic life forms, and societal constructs. It underlines the importance of predictive capabilities in digital twins, stressing the role of enabling proactive decision-making based on insights extended beyond current data. In summary, this enhanced definition strongly emphasizes reciprocal interaction as a pivotal feature.

Abstract

“Across multiple domains of science, engineering, and medicine, excitement is growing about the potential of digital twins to transform scientific research, industrial practices, and many aspects of daily life. A digital twin couples computational models with a physical counterpart to create a system that is dynamically updated through bidirectional data flows as conditions change. Going beyond traditional simulation and modeling, digital twins could enable improved medical decision-making at the individual patient level, predictions of future weather and climate conditions over longer timescales, and safer, more efficient engineering processes. However, many challenges remain before these applications can be realized.”

“This report identifies the foundational research and resources needed to support the development of digital twin technologies. The report presents critical future research priorities and an interdisciplinary research agenda for the field, including how federal agencies and researchers across domains can best collaborate.”

Contributor(s): National Academy of Engineering; National Academies of Sciences, Engineering, and Medicine; Division on Engineering and Physical Sciences; Division on Earth and Life Studies; Board on Mathematical Sciences and Analytics; Committee on Applied and Theoretical Statistics; Computer Science and Telecommunications Board; Board on Life Sciences; Board on Atmospheric Sciences and Climate; Committee on Foundational Research Gaps and Future Directions for Digital Twins.

(Source: National Academies of Sciences, Engineering, and Medicine. 2024. Foundational Research Gaps and Future Directions for Digital Twins. Washington, DC: The National Academies Press. https://doi.org/10.17226/26894.)

© 2024 National Academy of Sciences. All rights reserved.

About ClimaTwin®

ClimaTwin® is the leading climate risk intelligence solution for infrastructure assets and the built environment. To learn more, please visit www.climatwin.com today.

© 2024 ClimaTwin Corp. ClimaTwin® is a registered trademark of ClimaTwin Corp. The ClimaTwin logos, ClimaTwin Solutions™, and Future-proofing assets today for tomorrow’s climate extremes™ are trademarks of ClimaTwin Corp. All rights reserved.

###

by | May 13, 2024 | Intelligence

CONTINUE THE DECISION PATH

ClimaTwin Expands Flood Intelligence and Asset Analysis for Climate Business Decisions

FOR IMMEDIATE RELEASE September platform update adds flood depth by return period, expands hazard mapping and strengthens reporting and AI-assisted asset analysis BOSTON, Oct. 5, 2026 — ClimaTwin introduces new flood analysis, mapping and asset management capabilities in its September 2026 platform update. The enhancements strengthen ClimaTwin’s Climate Business Intelligence™ platform, helping infrastructure owners, asset managers and institutional investors bring more detailed climate information into operational reviews, capital planning and investment discussions. The update combines...

read more

Decision Systems and Model Governance: Implementing Reproducible Econometrics and Model Governance

Direct Answer Reproducible climate econometrics requires versioned data, code, assumptions, model artifacts, and decision outputs that another qualified analyst can reconstruct. Model governance adds clear ownership, independent validation, documented limitations, approval thresholds, change control, performance monitoring, and retirement procedures. Each KPI needs to trace from source data through transformation and model version to the reported number. Governance needs to match decision materiality: experimental analyses are labeled and sandboxed; capital-allocation or disclosure models...

read more

Decision Systems and Model Governance: Running Portfolio Stress Tests and Optimizing Capital Allocation

Direct Answer A portfolio climate stress test applies coherent hazard, vulnerability, valuation, financing, and recovery assumptions to estimate losses and performance under adverse but plausible conditions. Capital allocation then compares adaptation, acquisition, disposition, insurance, and liquidity choices subject to budget, risk, return, and concentration constraints. Optimization uses stress-test outputs as uncertain inputs, not exact forecasts. Robust or scenario-based decision rules identify actions that perform acceptably across several futures. The objective function, constraints,...

read more

Decision Systems and Model Governance: Estimating Heterogeneous Effects With Causal Machine Learning

Direct Answer Causal machine learning estimates treatment effects by using flexible algorithms to model complex relationships and heterogeneity. Double or debiased machine learning uses sample splitting and orthogonal scores to reduce regularization bias in the target causal parameter. Causal forests estimate how treatment effects vary with observed asset, hazard, and market characteristics. These methods do not establish identification: treatment, outcome, timing, confounding, and overlap assumptions must be credible. Validate heterogeneous effects on held-out data, summarize it into stable...

read more

ClimaTwin Shortlisted for 2026 Global Tech Insider Awards in GreenTech & Climate Tech

FOR IMMEDIATE RELEASE Climate Business Intelligence™ platform connects asset-level climate insights with financial analysis, operational priorities, and investment planning BOSTON, Sept. 28, 2026 — ClimaTwin announces its inclusion on the shortlist for the 2026 Global Tech Insider Awards in the GreenTech & Climate Tech category. The announcement highlights ClimaTwin’s work connecting climate information with business and financial decisions across infrastructure assets, capital programs, and investment portfolios. The Global Tech Insider Awards recognize companies and individuals across...

read more

Model Risk, Uncertainty, and Validation: Propagating Uncertainty and Conducting Hindcasts

Direct Answer Uncertainty propagation carries variability from exposures, hazard projections, vulnerability, damage functions, financial forecasts, and parameter estimates into final risk outputs. Hindcasting tests whether a model, using only information available at an earlier date, reproduces subsequently observed conditions or outcomes. Together, the methods address different questions: propagation quantifies conditional uncertainty, and hindcasting evaluates historical predictive performance. Validation uses time-aware data splits, preserves real-time model versions when possible, and...

read more

Model Risk, Uncertainty, and Validation: Combining Scenario Ensembles and Alternative Models

Direct Answer Scenario ensembles combine results across climate pathways, general circulation models, hazard models, damage functions, or econometric specifications. A simple average is appropriate only when members are comparable and the weighting rule is defensible. Scenarios are conditional narratives rather than automatically calibrated probabilities; do not create expected values by assigning arbitrary probabilities without disclosure. Analysts can report ranges, robust rankings, scenario-specific decisions, or model-averaged estimates when weights are evidence-based. Dependence among...

read more

Model Risk, Uncertainty, and Validation: Decomposing Portfolio Risk and Financial Performance

Direct Answer Portfolio decomposition attributes total climate risk or climate-adjusted performance to assets, hazards, locations, and model components. For additive expected loss, contribution is straightforward: each asset's weighted loss sums to the portfolio total. For nonlinear measures such as tail loss, probability of negative growth, or diversification benefit, marginal and component contributions depend on the selected risk functional. Analysts need to reconcile contributions to the portfolio measure, distinguish standalone from marginal risk, and avoid interpreting a large...

read more

Model Risk, Uncertainty, and Validation: Building Bayesian Hierarchical Climate Models

Direct Answer Bayesian hierarchical models combine a multilevel data structure with probability distributions for unknown parameters and latent climate-risk quantities. The models partially pool sparse asset or region estimates, incorporate prior engineering or empirical evidence, and propagate parameter uncertainty into posterior predictions. Priors are transparent, calibrated to plausible scales, and tested for sensitivity. Posterior intervals describe uncertainty conditional on the model, data, and priors — not protect against omitted hazards, misspecified damage functions, or poorly...

read more