Abstract Focusing on typical highway structures, this study introduces a novel approach to developing structure-specific parametric seismic fragility curves. Combining the capacity spectrum method (CSM) with a structural reliability framework, the proposed method differs significantly from existing methods for fragility analysis by relying solely on cyclic or backbone pushover performance. In particular, a fragility model is shown to be generally governed by seismic capacity and spectral modification factors. These two high-level parameters are then related to geometry, material, and spectral characteristics following common seismic design and evaluation procedures. As a result, the new approach aligns naturally with deterministic or semiprobabilistic seismic analyses of structures. Additionally, it can reduce to the HAZUS fragility curves (a popular class fragility model) based on assumptions clearly laid out in this study. When the HAZUS assumptions are refined or loosened, the proposed method yields structure-specific fragility models comparable to those obtained from more sophisticated techniques for fragility curve development. The new method also enables rapid development of fragility curves for retrofitted or deteriorated structures. Finally, the approach is used to demonstrate that (1) the widely used HAZUS fragility curves may introduce considerable model uncertainty for a specific in-class structure, and (2) due to this high in-class variation, it can be challenging to use class-level baseline models to incorporate deterioration or retrofitting effects. In contrast, the proposed method is capable of capturing these effects and distinguishing between structures of the same class, supporting more targeted and effective retrofit decision-making.
Fundamental challenges for improving supervised machine learning (ML) models for predicting geothermal resources include: 1) available training data being mostly positive (e.g., hot springs or operating power plants), so negative training sites must be selected, and in particular, negative training sites must represent the full range of conditions where no resources exist; and 2) input features potentially containing signals of several geologic processes, some of which may not be directly informative for modeling the availability of geothermal resources. We consider the artificial neural network (ANN) developed by the Nevada Machine Learning project (NVML), propose new selection strategies for negatives, and use signal separation to enhance the information content of input features. Using these strategies, we develop ML models using arguably simpler algorithmic approaches (Logistic Regression and eXtreme Gradient Boosting [i.e., XGBoost]) and compare the new results with the original ANN results.In particular, we evaluate the influence of selection strategy for negative training sites by comparing two training strategies: 1) using the original negative sites defined by NVML (NVML negatives); and 2) randomly selecting negatives from outside ellipsoids defined by NVML as having favorable structural settings, under the assumption that areas far from favorable settings are likely negative. Next, we explore the added value of splitting the topography input feature into two features to separate signals into regional elevation trend and local relative elevation, under the assumption that regional trends are correlated to regional geology (e.g., crustal thickness) and that local relative elevation (e.g., basin and range topography) is correlated to local geology (e.g., faulting). All new models using logistic regression and XGBoost were run 60 times each using Monte Carlo cross-validation to evaluate predictive uncertainty.Models using the randomly selected negatives outperform models using the negatives selected for the NVML ANN; models using XGBoost outperform models using logistic regression; and models using separated elevation features outperform models using the original elevation data. The results showing improved performance with random sampling of negatives demonstrates the value of sampling the entire input feature space when fitting a model (i.e., providing the ML workflow with all types of negatives and positives). The results showing improved performance with separated topography demonstrates that ML models cannot always find important patterns when they are obscured by larger trends, so feature engineering that enhances these patterns can be beneficial.
HfS2 is a layered semiconductor for electronic and optoelectronic applications, but humidity-assisted native oxidation limits its ambient stability. Raman spectroscopy probes HfS2 degradation but is time-consuming and condition-sensitive. Here, we revisit a previously reported HfS2 aging dataset to develop a Raman-calibrated image-based framework for rapid, non-destructive screening of CVD-grown HfS2 films on c-plane sapphire. Photographs collected under a documented configuration yielded predefined color, histogram, texture, and spatial descriptors. A simple, interpretable regression using only CIE Lab b* with physical-range clipping achieved a within-library, query-excluded leave-one-day-out RMSE of 6.78 days for 21 queries from three optically evolving conditions: native HfS2 at 35% and 70% RH and PMMA-capped HfS2 at 70% RH. The b*-based model performed comparably to kNN (7.70 days) and significantly better than a no-information pool-mean baseline (9.84 days, p = 0.0030), indicating that more complex estimators provided no clear benefit in the present dataset. Under leave-one-pair-out evaluation of 20 image-Raman pairs, the combined image-derived estimate showed a trend-level relationship with measured normalized A1g intensity (R2 = 0.600), while Delta E gave the best individual result (R2 = 0.701). The framework identified its limits: severe degradation of native HfS2 at 70% RH was underestimated, whereas Al2O3-encapsulated HfS2 showed insufficient optical evolution for reliable chronological estimation. Accordingly, this material-and configuration-specific method is a proof-of-concept screening framework, not a universal dating model or a substitute for Raman spectroscopy. Within this scope, controlled photography provides interpretable pre-screening of broad aging states and prioritizes samples for Raman or chemical analysis.
This article investigates the role of cultural theory (CT) in explaining attitudes towards solar radiation management (SRM) in the US. Partisanship traditionally provides a clear lens through which to view favourability to climate change policies; however, the lack of partisan attention to geoengineering policies, such as SRM, suggests other belief systems may be better explainers of attitudes. Previously, scholars have found climate attitudes to vary based upon cultural types, which are a function of how individuals see the role of authority in policy making and the extent to which individuals see themselves as part of a social unit. We find that partisanship is a weak predictor of SRM attitudes, while cultural affinities provide strong influences over SRM attitudes. Specifically, we find strong group affinities (particularly egalitarians) to lead to more favourable impressions of SRM research and deployment in the case of climate emergencies. Attitudes towards hierarchy, specifically cultural affinities that disfavour hierarchy, explain SRM scepticism. We contrast these results with a separate analysis of more widely known mitigation policies: carbon taxes, regulations on carbon emissions, taxes on higher income earners for green investment, and carbon cap and trade. For these better-known policy positions, partisanship is far more important. Our work contributes to the application of CT to policy attitudes as well as to explaining the shifting landscape of attitudes towards geoengineering and a variety of climate mitigation policies.
Marine paleoclimate records for the last 1,000 years are scarce in the southwest Pacific, limiting our understanding of complex environmental changes that may have affected MaAori seascapes and fisheries. We seek to begin filling this knowledge gap by studying stable oxygen (delta 18O) and carbon (delta 13C) isotopes in archaeological and modern otoliths from ta mure (Australasian snapper, Chrysophrys auratus), which provide information about water temperature, salinity, and fish diet and metabolism. Our results show that fourteenth and fifteenth century ta mure otoliths recorded environmental conditions that are comparable to twentieth century temperatures, with some evidence for anomalously warmer seas and/or higher precipitation during the fifteenth century. These findings are concordant with previous reconstructions of terrestrial climatic conditions in northern Aotearoa and of central west Pacific Ocean sea surface temperatures, providing additional evidence that MaAori experienced a warm climatic period during their first centuries of habitation and fishing in the North Island.
