arXiv subjects

Maria J. Molina

Publications and source records attributed to Maria J. Molina.

13 recordsLinked to original sources

arXiv2026

Tracing the space-time causal origins of Earth system extremes

Identifying the causes of Earth's extremes is challenging because counterfactual experiments are not possible in the observed world. Data-driven causal discovery complements computationally expensive and potentially biased numerical experiments, but existing methods can struggle with undersampled, high-dimensional data and fail to recover multi-timestep, multivariate pathways leading to specific events. We introduce Tracer of Causal Evolutions in Space and Time (TraCE-ST), a probabilistic Lagrangian approach that produces event-conditioned causal trajectories in multivariate gridded data. TraCE-ST recovers known causal drivers and estimates their relative contributions in synthetic experiments and real-world extremes, including the 1991 Mount Pinatubo eruption. TraCE-ST also highlights less-studied drivers, such as orography-driven vorticity for Tropical Storm Debby (2006) and anomalous ocean-surface fluxes for the 2021 Pacific Northwest heatwave. Here, we propose causal tracking as an efficient data-driven framework for synthesizing causal evidence and generating testable hypotheses, complementing association analyses and numerical modeling while accelerating the study of high-impact events.

physics.ao-ph↗

arXiv2026

Knowledge-guided machine learning for disentangling Pacific sea surface temperature variability across timescales

Global weather and climate patterns are strongly influenced by dominant modes of anomalous Pacific sea surface temperature (SST) variability, including the El Niño-Southern Oscillation (ENSO), Pacific Meridional Mode (PMM), and Pacific Decadal Oscillation (PDO). However, disentangling these modes of variability remains challenging due to their spatial overlap and nonlinear coupling, which violate the assumptions of traditional linear methods. We develop a Knowledge-Guided AutoEncoder (KGAE) that uses spatiotemporal constraints and a gradient-based sparsity incentive to identify physically interpretable modes of detrended SST variability, each defined by a single broad characteristic timescale, without the need for predefined temporal filters or thresholds. The KGAE separates ENSO-like modes on 2- and 3-7-year timescales, as well as a decadal mode with characteristics reminiscent of the PDO and PMM, each with distinct spatial patterns. We perturb each latent dimension and use finite differences to characterize the state-independent and state-dependent sensitivities. We demonstrate that the decadal mode modulates ENSO diversity (central versus eastern Pacific) through both interference and nonlinear state dependence, and that a quasibiennial mode interacts with the interannual mode to characterize ENSO onset and decay. We assess the robustness of the KGAEs to random initialization and training stochasticity, to sampling bias induced by a time-stratified cross-validation scheme, and to unseen data (generalization). When applied to climate model output, KGAEs reveal model-specific biases in ENSO diversity and seasonal timing. Our results highlight how machine learning can uncover physically meaningful modes of Earth system variability and characterize their complex interactions across models and timescales.

physics.ao-ph↗

arXiv2026

Exploring coupled tropical Pacific variability within a Multi-branch $β$-Variational Autoencoder

This study explores what is encoded in the latent space of a multi-branch $β$-variational autoencoder ($β$-VAE) trained on coupled tropical Pacific climate fields. We assess the reconstruction skill and physical interpretability of the latent space trained on monthly sea surface temperature, ocean heat content, and outgoing longwave radiation across the tropical Pacific from a 500-year preindustrial control simulation. The model generalizes well, with only modest degradation from training to test performance, and preserves the dominant basin-scale structure of all three fields. Latent-space diagnostics show that variability is organized unevenly across dimensions: sea surface temperature is concentrated in a smaller subset of latent dimensions, whereas ocean heat content and outgoing longwave radiation are more broadly distributed across multiple dimensions. Comparisons with conventional tropical Pacific diagnostics further show that several latent dimensions align with known El Niño and La Niña variability, while others capture related coupled ocean-atmosphere variability on decadal or longer timescales. Sensitivity experiments and latent traversals identify dimensions associated with eastern-Pacific-like, central-Pacific-like, coastal, subsurface-dominant, and atmosphere-dominant variability. Together, these results show that the multi-branch $β$-variational autoencoder yields a skillful and physically informative reduced representation of coupled tropical Pacific variability.

physics.ao-ph↗

arXiv2026

AIMIP Phase 1: systematic evaluations of AI weather and climate models

We present the AI weather and climate model intercomparison project (AIMIP), phase 1. Drawing from the rich tradition of intercomparisons in climate model development, we specify a common experiment, output data format, and training constraints (namely, training against historical reanalysis data) for AIMIP Phase 1 models. We aim to identify differences in modeling frameworks and AI architectural choices that influence model behavior, and build trust in AI weather and climate models through open data and evaluation. AIMIP Phase 1 models must simulate the atmosphere given specified historical sea surface temperatures over 1979-2024. We evaluate the models' performance using five major evaluation criteria: biases, trends, response to El Niño-related sea surface temperature anomalies, temporal variability, and out-of-sample generalization tests. We find that the AI models are able to simulate the historical climate and response to forcing as well as a conventional physically-based model, but some AI models underestimate historical warming trends, and their predictions diverge in the out-of-sample generalization tests. We describe the AIMIP Phase 1 dataset that is publicly available for additional evaluations.

physics.ao-ph↗

arXiv2026

Urban Heat MiniCubes: An AI-Ready dataset for urban heat research

Urban heat is amplified by impermeable surfaces and heterogeneous built environments, yet street-level variability remains difficult to quantify because multi-sensor observations are rarely available in consistent, analysis-ready form at the necessary spatiotemporal scales. We present "Urban Heat MiniCubes," a publicly available, FAIR-oriented dataset designed for machine learning applications in urban heat research. The dataset provides harmonized 90 x 90 km gridded data cubes for 48 cities in the Western Hemisphere spanning 2022-2023, with variables reprojected and collocated to a common grid to reduce preprocessing (e.g., reprojection, resampling, and spatiotemporal alignment). Urban Heat MiniCubes includes two complementary modalities: (i) higher-spatial-resolution, lower-frequency observations from Landsat 8/9 (e.g., surface reflectances) and Sentinel-1 (e.g., synthetic aperture radar backscatter), and (ii) higher-temporal-frequency, coarser observations from GOES-R (e.g., longwave infrared brightness temperatures) and a microwave land surface temperature product. We document variables and metadata and provide technical assessment using inter-variable analyses and autoencoder-based reconstruction-error summaries across pixel classes (e.g., water and cloud). Potential use cases and limitations are also discussed.

physics.ao-ph↗

arXiv2026

Monthly Diffusion v0.9: A Latent Diffusion Model for the First AI-MIP

Here, we describe Monthly Diffusion at 1.5-degree grid spacing (MD-1.5 version 0.9), a climate emulator that leverages a spherical Fourier neural operator (SFNO)-inspired Conditional Variational Auto-Encoder (CVAE) architecture to model the evolution of low-frequency internal atmospheric variability using latent diffusion. MDv0.9 was designed to forward-step at monthly mean timesteps in a data-sparse regime, using modest computational requirements. This work describes the motivation behind the architecture design, the MDv0.9 training procedure, and initial results.

cs.LG↗

arXiv2025

Setting the Standard: Recommended Practices for Data Preprocessing in Data-Driven Climate Prediction

Artificial intelligence (AI) - and specifically machine learning (ML) - applications for climate prediction across timescales are proliferating quickly. The emergence of these methods prompts a revisit to the impact of data preprocessing, a topic familiar to the climate community, as more traditional statistical models work with relatively small sample sizes. Indeed, the skill and confidence in the forecasts produced by data-driven models are directly influenced by the quality of the datasets and how they are treated during model development, thus yielding the colloquialism, "garbage in, garbage out." As such, this article establishes protocols for the proper preprocessing of input data for AI/ML models designed for climate prediction (i.e., subseasonal to decadal and longer). The three aims are to: (1) educate researchers, developers, and end users on the effects that preprocessing has on climate predictions; (2) provide recommended practices for data preprocessing for such applications; and (3) empower end users to decipher whether the models they are using are properly designed for their objectives. Specific topics covered in this article include the creation of (standardized) anomalies, dealing with non-stationarity and the spatiotemporally correlated nature of climate data, and handling of extreme values and variables with potentially complex distributions. Case studies will illustrate how using different preprocessing techniques can produce different predictions from the same model, which can create confusion and decrease confidence in the overall process. Ultimately, implementing the recommended practices set forth in this article will enhance the robustness and transparency of AI/ML in climate prediction studies.

physics.data-an↗

arXiv2025

Using Generative Artificial Intelligence Creatively in the Classroom and Research: Examples and Lessons Learned

Although generative artificial intelligence (AI) is not new, recent technological breakthroughs have transformed its capabilities across many domains. These changes necessitate new attention from educators and specialized training within the atmospheric and related sciences. Enabling students to use generative AI effectively, responsibly, and ethically is crucial for their academic and professional development. Educators can also use generative AI to develop engaging classroom activities, such as active learning modules and games; however, they must be aware of potential pitfalls and biases. There are also ethical implications in using tools that lack transparency and have a considerable carbon footprint, as well as equity concerns for students who lack access to more sophisticated paid versions of generative AI tools and have deficiencies in prior educational training. This article is written for students and educators alike, particularly those interested in learning more about generative AI in education and research, including its use cases, ethical concerns, and a brief history of its emergence. Sample user prompts are also provided across numerous applications in education and the atmospheric and related sciences. Current solutions addressing broader ethical concerns regarding the use of generative AI in education remain limited; however, this work aims to foster a discussion that could galvanize the education community around shared goals and values.

cs.HC↗

arXiv2025

An Earth-System-Oriented View of the S2S Predictability of North American Weather Regimes

It is largely agreed that subseasonal-to-seasonal (S2S) predictability arises from the atmospheric initial state during early lead times and from the land and ocean during intermediate and late lead times. We test this hypothesis for the large-scale mid-latitude atmosphere by training numerous XGBoost models to predict weather regimes (WRs) over North America at 1-to-8-week lead times. Each model uses a different predictor from one Earth system component (atmosphere, ocean, or land) sourced from reanalysis. According to the models, the atmosphere provides more predictability during the first two forecast weeks, and the three components performed similarly afterward. However, the skill and sources of predictability are highly dependent on the season and target WR. Our results show greater WR predictability in fall and winter, particularly for the Pacific Trough and Pacific Ridge regimes, driven primarily by the ocean (e.g., El Niño-Southern Oscillation and sea ice). For the Pacific Ridge in winter, the stratosphere also contributes significantly to predictability across most S2S lead times. Additionally, the initial large-scale tropospheric structure (encompassing the tropics and extra-tropics, e.g., Madden-Julian Oscillation) and soil conditions play a relevant role-most notably for the Greenland High regime in winter. This study highlights previously identified sources of predictability for the large-scale atmosphere and gives insight into new sources for future study. Given how closely linked WRs are to surface precipitation and temperature anomalies, storm tracks, and extreme events, the study results contribute to improving S2S prediction of surface weather.

physics.ao-ph↗

arXiv2025

Recommendations for Comprehensive and Independent Evaluation of Machine Learning-Based Earth System Models

Machine learning (ML) is a revolutionary technology with demonstrable applications across multiple disciplines. Within the Earth science community, ML has been most visible for weather forecasting, producing forecasts that rival modern physics-based models. Given the importance of deepening our understanding and improving predictions of the Earth system on all time scales, efforts are now underway to develop forecasting models into Earth-system models (ESMs), capable of representing all components of the coupled Earth system (or their aggregated behavior) and their response to external changes. Modeling the Earth system is a much more difficult problem than weather forecasting, not least because the model must represent the alternate (e.g., future) coupled states of the system for which there are no historical observations. Given that the physical principles that enable predictions about the response of the Earth system are often not explicitly coded in these ML-based models, demonstrating the credibility of ML-based ESMs thus requires us to build evidence of their consistency with the physical system. To this end, this paper puts forward five recommendations to enhance comprehensive, standardized, and independent evaluation of ML-based ESMs to strengthen their credibility and promote their wider use.

cs.LG↗

arXiv2024

Estimating Epistemic and Aleatoric Uncertainty with a Single Model

Estimating and disentangling epistemic uncertainty, uncertainty that is reducible with more training data, and aleatoric uncertainty, uncertainty that is inherent to the task at hand, is critically important when applying machine learning to high-stakes applications such as medical imaging and weather forecasting. Conditional diffusion models' breakthrough ability to accurately and efficiently sample from the posterior distribution of a dataset now makes uncertainty estimation conceptually straightforward: One need only train and sample from a large ensemble of diffusion models. Unfortunately, training such an ensemble becomes computationally intractable as the complexity of the model architecture grows. In this work we introduce a new approach to ensembling, hyper-diffusion models (HyperDM), which allows one to accurately estimate both epistemic and aleatoric uncertainty with a single model. Unlike existing single-model uncertainty methods like Monte-Carlo dropout and Bayesian neural networks, HyperDM offers prediction accuracy on par with, and in some cases superior to, multi-model ensembles. Furthermore, our proposed approach scales to modern network architectures such as Attention U-Net and yields more accurate uncertainty estimates compared to existing methods. We validate our method on two distinct real-world tasks: x-ray computed tomography reconstruction and weather temperature forecasting.

cs.LG↗

arXiv2024

Can Transfer Learning be Used to Identify Tropical State-Dependent Bias Relevant to Midlatitude Subseasonal Predictability?

Previous research has demonstrated that specific states of the climate system can lead to enhanced subseasonal predictability (i.e., state-dependent predictability). However, biases in Earth system models can affect the representation of these states and their subsequent evolution. Here, we present a machine learning framework to identify state-dependent biases in Earth system models. In particular, we investigate the utility of transfer learning with explainable neural networks to identify tropical state-dependent biases in historical simulations of the Energy Exascale Earth System Model version 2 (E3SMv2) relevant for midlatitude subseasonal predictability. Using a perfect model framework, we find transfer learning may require substantially more data than provided by present-day reanalysis datasets to update neural network weights, imparting a cautionary tale for future transfer learning approaches focused on subseasonal modes of variability.

physics.ao-ph↗

arXiv2024

Evidential Deep Learning: Enhancing Predictive Uncertainty Estimation for Earth System Science Applications

Robust quantification of predictive uncertainty is critical for understanding factors that drive weather and climate outcomes. Ensembles provide predictive uncertainty estimates and can be decomposed physically, but both physics and machine learning ensembles are computationally expensive. Parametric deep learning can estimate uncertainty with one model by predicting the parameters of a probability distribution but do not account for epistemic uncertainty.. Evidential deep learning, a technique that extends parametric deep learning to higher-order distributions, can account for both aleatoric and epistemic uncertainty with one model. This study compares the uncertainty derived from evidential neural networks to those obtained from ensembles. Through applications of classification of winter precipitation type and regression of surface layer fluxes, we show evidential deep learning models attaining predictive accuracy rivaling standard methods, while robustly quantifying both sources of uncertainty. We evaluate the uncertainty in terms of how well the predictions are calibrated and how well the uncertainty correlates with prediction error. Analyses of uncertainty in the context of the inputs reveal sensitivities to underlying meteorological processes, facilitating interpretation of the models. The conceptual simplicity, interpretability, and computational efficiency of evidential neural networks make them highly extensible, offering a promising approach for reliable and practical uncertainty quantification in Earth system science modeling. In order to encourage broader adoption of evidential deep learning in Earth System Science, we have developed a new Python package, MILES-GUESS (https://github.com/ai2es/miles-guess), that enables users to train and evaluate both evidential and ensemble deep learning.

cs.LG↗