Dynamical seasonal forecast models are improving with time but tend to underestimate the amplitude of atmospheric circulation variability and to have lower skill in predicting summer variability than in winter. Here, we construct Nonlinear AutoRegressive Moving Average models with eXogenous inputs (NARMAX) to develop the analysis of drivers of North Atlantic atmospheric circulation and jet‐stream variability, focusing on the East Atlantic (EA) and Scandinavian (SCA) patterns as well as the North Atlantic Oscillation (NAO) index. New time series of these indices are developed from empirical orthogonal function (EOF) analysis. Geopotential height data from the ERA5 reanalysis are used to generate the EOFs. Sets of predictors with known associations with these drivers are developed and used to formulate a sliding‐window NARMAX model. This model demonstrates a high degree of predictive accuracy, as indicated by its average correlation coefficients over the testing period (2006–2021): 0.78 for NAO, 0.83 for EA and 0.68 for SCA. In comparison, the SEAS5 and GloSea5 dynamical forecast models exhibit lower correlations with observed circulation changes: for NAO, the correlation coefficients are 0.51 for SEAS5 and 0.34 for GloSea5, for EA they are 0.15 and 0.09, respectively, and for SCA, they are 0.28 and 0.24, respectively. Comparison of NARMAX predictions with forecasts and hindcasts from the SEAS5 and GloSea5 models highlights areas where NARMAX can be used to help improve seasonal forecast skill and inform the development of dynamical models, especially in the case of summer.
The building sector, responsible for 40% of global energy consumption, faces increasing demands for sustainability and energy efficiency. Accurate energy consumption forecasting is essential to optimise performance and reduce environmental impact. This study introduces a hybrid machine learning framework grounded in Sparse, Interpretable, and Transparent (SIT) modelling to enhance building energy management. Leveraging the REFIT Smart Home Dataset, the framework integrates occupancy pattern analysis, appliance-level energy prediction, and probabilistic uncertainty quantification. The framework clusters occupancy-driven energy usage patterns using K-means and Gaussian Mixture Models, identifying three distinct household profiles: high-energy frequent occupancy, moderate-energy variable occupancy, and low-energy irregular occupancy. A Random Forest classifier is employed to pinpoint key appliances influencing occupancy, with a drop-in accuracy analysis verifying their predictive power. Uncertainty analysis quantifies classification confidence, revealing ambiguous periods linked to irregular appliance usage patterns. Additionally, time-series decomposition and appliance-level predictions are contextualised with seasonal and occupancy dynamics, enhancing interpretability. Comparative evaluations demonstrate the framework’s superior predictive accuracy and transparency over traditional single machine learning models, including Support Vector Machines (SVM) and XGBoost in Matlab 2024b and Python 3.10. By capturing occupancy-driven energy behaviours and accounting for inherent uncertainties, this research provides actionable insights for adaptive energy management. The proposed SIT hybrid model can contribute to sustainable and resilient smart energy systems, paving the way for efficient building energy management strategies.
Industrialised construction increasingly relies on digital building models to encode component geometry, assembly interfaces, and spatial constraints. However, Building Information Models (BIM) and Industry Foundation Classes (IFC) encode geometry and semantics but do not encode executable task knowledge or action-level recovery semantics required for robot-executable task reasoning. This paper presents a knowledge-driven engineering informatics framework that bridges IFC-based building information and task-level robotic assembly planning through an Action Knowledge Graph (AKG). The AKG encodes empirically grounded relationships among robot-executable action primitives—including precedence, refinement, and failure-conditioned recovery edges—derived from systematic analysis of established embodied manipulation datasets. Based on this representation, assembly processes are expressed as task graphs whose feasibility is evaluated against BIM-derived geometry and IFC-based semantic information extracted from the digital twin environment. Three controlled experiments on an IFC-modelled door-articulation construction scenario demonstrate that the framework: (i) generates geometrically consistent task graphs from IFC-derived semantics; (ii) differentiates between recoverable and irrecoverable feasibility failures based on AKG edge structure; and (iii) restores task feasibility through recovery graph traversal without modifying the underlying AKG structure. A comparative experiment further isolates the functional contribution of recovery semantics, demonstrating that task completion under a 52.5 mm hinge offset is achievable only when AKG recovery edges are activated. These results demonstrate that knowledge-driven task representations provide a structured, interpretable, and transferable mechanism for extending digital twin environments into reasoning-enabled robotic planning systems, with direct implications for semantic interoperability and engineering informatics research on industrialised construction robotics.
Material choices fixed early in design determine much of a building’s embodied carbon and how much of its material can later be recovered, but the data needed to weigh the two are usually held in separate tools. To bring them together, a deterministic process was implemented that reads Industry Foundation Classes (IFC) data, enumerates whole-building material scenarios, filters them for Pareto dominance, answers controlled semantic queries, and packages the results as prompt text for optional downstream interpretation. One buildingSMART reference IFC model yielded seven valid components, and 150 wall-floor-roof material configurations were evaluated on that basis. Twelve configurations were non-dominated: six low-carbon-oriented configurations, in which timber systems predominate, and six circularity-oriented configurations built on aluminium-roof systems. The same candidates were returned under all five Building Circularity Score (BCS) weighting schemes (12/12 overlap with the equal-weight baseline), and the classification of 0 balanced, 6 low-carbon-oriented, and 6 circularity-oriented scenarios held across nine threshold combinations. In a scenario-based Monte Carlo check, every one of the 12 baseline candidates stayed non-dominated in at least 83.2% of 1000 sampled runs under the stated assumptions. The full 150-configuration case took approximately 0.79 s, of which pairwise Pareto filtering accounted for approximately 0.023 s. Within this single test case, IFC-derived quantities were sufficient for a reproducible environmental ranking, and the outputs are candidate scenarios for prototype-stage decision support, not complete design prescriptions. External validation, additional IFC models, engineering feasibility checks, and deployment beyond the prototype are left to future work.
Recent advances in machine learning (ML) have enabled significant progress in geoscience by capturing complex relationships and enhancing predictive skills. However, the success of many ML algorithms in data-rich settings does not seamlessly transfer to climate and atmospheric applications, where observational datasets are often limited. This underscores the need for methods that deliver high predictive accuracy under data-scarce conditions while retaining interpretability.Here, we compare various ML approaches with the Nonlinear AutoRegressive Moving Average model with eXogenous inputs (NARMAX) in typical small-data climate applications, such as seasonal weather forecasting and Greenland Blocking Index (GBI) prediction. NARMAX, a transparent, interpretable, parsimonious and simulatable (TIPS) framework, demonstrates robust performance and avoids common pitfalls such as overfitting and unstable predictions when data are scarce. Notably, it achieves superior or competitive forecast accuracy for small or limited data conditions, underscoring its practical value in operational climate science. By adopting a sparse system identification approach, NARMAX yields model structures that readily reveal key predictors and their relative contributions, providing valuable physical and statistical insights into climate variability.Our findings illustrate how NARMAX bridges the gap between purely data-driven modelling (focusing on prediction) and mechanistic modelling (focusing on physical insights), offering a clear pathway for refining model strategies and deepening our understanding of climate dynamics. We propose that NARMAX and similar methods play an inherently powerful role for both small and large data modelling problems and meanwhile serve as potent components to potentially improve the explainability of ML methods. By showcasing both interpretability and predictive efficacy, this work encourages the adoption of machine learning methods that best meet the needs for specific data modelling tasks in climate science and beyond.
Modular construction robotics requires reasoning that links building information to dependencies, action knowledge and recovery. We propose the Knowledge-Driven Construction Robotics Framework (KDCRF), a BIM-linked decision-support ap-proach for modular assembly planning. An assembly is represented as A={C,I,D,T,R}; hierarchical task graphs are instantiated from an action-knowledge-graph primitive library. A rule-based layer populates this representation directlyfrom IFC: the sample models declare no IfcRelConnectsElements relationships,yet connection semantics are recovered from aggregation, accessory typing andproperty sets under a geometric contact constraint, yielding a 109-node task graphordered bottom-up by elevation. Recovery is graph rewiring admitted by an explicitadmissibility rule; on nine controlled and 102 synthetic perturbations, retrievalwithout it wrongly declares 41% of proposed plans executable. Reasoning cost islinear in task-graph size from 25 to 1601 nodes. It is a decision-support prototype,not a claim of physical deployment.