More Like this

1. Artificial Intelligence in Hydrogeology: Applications and an Open, Reproducible Machine Learning Course

   Samalavičius, Vytautas; Zaslavsky, Ilya (January 2026, Baltica)

   Artificial intelligence (AI) is increasingly used in hydrogeological analysis for forecasting groundwater levels and properties, discovering hydrochemical facies and anomalies, and modelling spatio-temporal signals. Current practice emphasizes spatio-temporal validation, interpretability, and uncertainty quantification, while generative AI (GenAI) is emerging to accelerate data curation, documentation, augmentation, and code prototyping. This report summarizes representative applications of AI and machine learning (ML) in hydrogeology and describes an open, reproducible course developed for university students and professional hydrogeologists. This continuously updated online course is one of the outcomes of the GRANDE-U “Groundwater Resilience Assessment through Integrated Data Exploration for Ukraine” project. It covers Python fundamentals, environment setup, data engineering, and documented case studies in time-series and spatio-temporal modeling. Application examples are drawn from recent teaching materials and GRANDE-U and related studies and include predicting terrestrial water-storage anomalies, groundwater-level forecasting, isotope estimation from routine chemistry, risk or prospectivity mapping, and automated lineament/facies mapping. 

   more » « less
2. Data-Driven Time Series Forecasting for Social Studies Using Spatio-Temporal Graph Neural Networks

   https://doi.org/10.1145/3462203.3475929  

   Li, Yi-Fan; Dong, Bo; Khan, Latifur; Thuraisingham, Bhavani; Brandt, Patrick T.; D'Orazio, Vito J. (September 2021, GoodIT '21: Proceedings of the Conference on Information Technology for Social Good)

   Time series forecasting with additional spatial information has attracted a tremendous amount of attention in recent research, due to its importance in various real-world applications on social studies, such as conflict prediction and pandemic forecasting. Conventional machine learning methods either consider temporal dependencies only, or treat spatial and temporal relations as two separate autoregressive models, namely, space-time autoregressive models. Such methods suffer when it comes to long-term forecasting or predictions for large-scale areas, due to the high nonlinearity and complexity of spatio-temporal data. In this paper, we propose to address these challenges using spatio-temporal graph neural networks. Empirical results on Violence Early Warning System (ViEWS) dataset and U.S. Covid-19 dataset indicate that our method significantly improved performance over the baseline approaches. 

   more » « less
3. (Vision Paper) A Vision for Spatio-Causal Situation Awareness, Forecasting, and Planning

   https://doi.org/10.1145/3672556  

   Azad, Fahim Tasneema; Candan, K Selçuk; Kapkiç, Ahmet; Li, Mao-Lin; Liu, Huan; Mandal, Pratanu; Sheth, Paras; Arslan, Bilgehan; Chowell-Puente, Gerardo; Sabo, John; et al (June 2024, ACM Transactions on Spatial Algorithms and Systems)

   Successfully tackling many urgent challenges in socio-economically critical domains, such as public health and sustainability, requires a deeper understanding of causal relationships and interactions among a diverse spectrum of spatio-temporally distributed entities. In these applications, the ability to leverage spatio-temporal data to obtain causally based situational awareness and to develop informed forecasts to provide resilience at different scales is critical. While the promise of a causally grounded approach to these challenges is apparent, the core data technologies needed to achieve these are in the early stages and lack a framework to help realize their potential. In this article, we argue that there is an urgent need for a novel paradigm of spatio-causal research built on computational advances in spatio-temporal data and model integration, causal learning and discovery, large scale data- and model-driven simulations, emulations, and forecasting, as well as spatio-temporal data-driven and model-centric operational recommendations, and effective causally driven visualization and explanation. We thus provide a vision, and a road map, for spatio-causal situation awareness, forecasting, and planning. 

   more » « less
4. Bayesian Recurrent Neural Network Models for Forecasting and Quantifying Uncertainty in Spatial-Temporal Data

   https://doi.org/10.3390/e21020184  

   McDermott, Patrick; Wikle, Christopher (February 2019, Entropy)

   Recurrent neural networks (RNNs) are nonlinear dynamical models commonly used in the machine learning and dynamical systems literature to represent complex dynamical or sequential relationships between variables. Recently, as deep learning models have become more common, RNNs have been used to forecast increasingly complicated systems. Dynamical spatio-temporal processes represent a class of complex systems that can potentially benefit from these types of models. Although the RNN literature is expansive and highly developed, uncertainty quantification is often ignored. Even when considered, the uncertainty is generally quantified without the use of a rigorous framework, such as a fully Bayesian setting. Here we attempt to quantify uncertainty in a more formal framework while maintaining the forecast accuracy that makes these models appealing, by presenting a Bayesian RNN model for nonlinear spatio-temporal forecasting. Additionally, we make simple modifications to the basic RNN to help accommodate the unique nature of nonlinear spatio-temporal data. The proposed model is applied to a Lorenz simulation and two real-world nonlinear spatio-temporal forecasting applications. 

   more » « less
5. Bayesian spatio-temporal small area modelling: a case study investigating the late-stage melanoma incidence in Texas

   https://doi.org/10.1093/jrsssc/qlaf054  

   Cao, Jiahao; Zhang, Kehe; Hu, Guanyu; Nelson, Kelly; Bauer, Cici (November 2025, Journal of the Royal Statistical Society Series C: Applied Statistics)

   Abstract Melanoma is projected to become the second most diagnosed cancer in the United States by 2040, emphasizing the importance of early detection for improving survival outcomes. Accurately estimating late-stage melanoma incidence at granular geographical levels is challenging due to its rarity and data sparsity. Instead of relying on aggregated reported health statistics on melanoma cases, which are commonly used in cancer surveillance research, we utilize the incidence-based individual case database, and propose a Bayesian Heterogeneous Spatio-Temporal Logistic Regression with Nonlinear Demographic Effect (BHSTLR-NDE) model to provide county-level annual estimates of late-stage melanoma incidence in Texas from 2000 to 2018, using data obtained through a request to the Texas Cancer Registry. The BHSTLR-NDE integrates county-level covariates, flexible case-level demographic effects, and a reduced-rank spatial modelling approach to effectively address data sparsity, spatio-temporal non-stationarity, and demographic heterogeneity. The superiority of the proposed BHSTLR-NDE over competing models is demonstrated through extensive simulation studies and an application to melanoma registry data. Results indicate that BHSTLR-NDE effectively borrows information across spatial, temporal, and demographic dimensions, improving estimation accuracy and robustness. These findings highlight the utility of the proposed modelling approach in enhancing cancer surveillance and guiding targeted interventions. 

   more » « less