Prediction of dynamic environmental variables in unmonitored sites remains a long-standing challenge for water resources science. The majority of the world’s freshwater resources have inadequate monitoring of critical environmental variables needed for management. Yet, the need to have widespread predictions of hydrological variables such as river flow and water quality has become increasingly urgent due to climate and land use change over the past decades, and their associated impacts on water resources. Modern machine learning methods increasingly outperform their process-based and empirical model counterparts for hydrologic time series prediction with their ability to extract information from large, diverse data sets. We review relevant state-of-the art applications of machine learning for streamflow, water quality, and other water resources prediction and discuss opportunities to improve the use of machine learning with emerging methods for incorporating watershed characteristics and process knowledge into classical, deep learning, and transfer learning methodologies. The analysis here suggests most prior efforts have been focused on deep learning frameworks built on many sites for predictions at daily time scales in the United States, but that comparisons between different classes of machine learning methods are few and inadequate. We identify several open questions for time series predictions in unmonitored sites that include incorporating dynamic inputs and site characteristics, mechanistic understanding and spatial context, and explainable AI techniques in modern machine learning frameworks.

Vibration-based structural health monitoring (SHM) of (large) infrastructure through operational modal analysis (OMA) is a commonly adopted strategy. This is typically a four-step process, comprising estimation, tracking, data normalization, and decision-making. These steps are essential to ensure structural modes are correctly identified, and results are normalized for environmental and operational variability (EOV). Other challenges, such as nonstructural modes in the OMA, for example, rotor harmonics in (offshore) wind turbines (OWTs), further complicate the process. Typically, these four steps are considered independently, making the method simple and robust, but rather limited in challenging applications, such as OWTs. Therefore, this study aims to combine tracking, data normalization, and decision-making through a single machine learning (ML) model. The presented SHM framework starts by identifying a “healthy” training dataset, representative of all relevant EOV, for all structural modes. Subsequently, operational and weather data are used for feature selection and a comparative analysis of ML models, leading to the selection of tree-based learners for natural frequency prediction. Uncertainty quantification (UQ) is introduced to identify out-of-distribution instances, crucial to guarantee low modeling error and ensure only high-fidelity structural modes are tracked. This study uses virtual ensembles for UQ through the variance between multiple truncated submodel predictions. Practical application to monopile-supported OWT data demonstrates the tracking abilities, separating structural modes from rotor dynamics. Control charts show improved decision-making compared to traditional reference-based methods. A synthetic dataset further confirms the approach’s robustness in identifying relevant natural frequency shifts. This study presents a comprehensive data-driven approach for vibration-based SHM.

Displacement continues to increase at a global scale and is increasingly happening in complex, multicrisis settings, leading to more complex and deeper humanitarian needs. Humanitarian needs are therefore increasingly outgrowing the available humanitarian funding. Thus, responding to vulnerabilities before disaster strikes is crucial but anticipatory action is contingent on the ability to accurately forecast what will happen in the future. Forecasting and contingency planning are not new in the humanitarian sector, where scenario-building continues to be an exercise conducted in most humanitarian operations to strategically plan for coming events. However, the accuracy of these exercises remains limited. To address this challenge and work with the objective of providing the humanitarian sector with more accurate forecasts to enhance the protection of vulnerable groups, the Danish Refugee Council has already developed several machine learning models. The Anticipatory Humanitarian Action for Displacement uses machine learning to forecast displacement in subdistricts in the Liptako-Gourma region in Sahel, covering Burkina Faso, Mali, and Niger. The model is mainly built on data related to conflict, food insecurity, vegetation health, and the prevalence of underweight to forecast displacement. In this article, we will detail how the model works, the accuracy and limitations of the model, and how we are translating the forecasts into action by using them for anticipatory action in South Sudan and Burkina Faso, including concrete examples of activities that can be implemented ahead of displacement in the place of origin, along routes and in place of destination.

We propose a framework for identifying discrete behavioural types in experimental data. We re-analyse data from six previous studies of public goods voluntary contribution games. Using hierarchical clustering analysis, we construct a typology of behaviour based on a similarity measure between strategies. We identify four types with distinct stereotypical behaviours, which together account for about 90% of participants. Compared to the previous approaches, our method produces a classification in which different types are more clearly distinguished in terms of strategic behaviour and the resulting economic implications.

Detecting and removing hate speech content in a timely manner remains a challenge for social media platforms. Automated techniques such as deep learning models offer solutions which can keep up with the volume and velocity of user content production. Research in this area has mainly focused on either binary classification or on classifying tweets into generalised categories such as hateful, offensive, or neither. Less attention has been given to multiclass classification of online hate speech into the type of hate or group at which it is directed. By aggregating and re-annotating several relevant hate speech datasets, this study presents a dataset and evaluates several models for classifying tweets into the categories ethnicity, gender, religion, sexuality, and non-hate. We evaluate the dataset by training several models: logistic regression, LSTM, BERT, and GPT-2. For the LSTM model, we assess a range of NLP features using a multi-classification LSTM model, and conclude that the highest performing feature combination consists of word $n$-grams, character $n$-grams, and dependency tuples. We show that while more recent larger models can achieve a slightly higher performance, increased model complexity alone is not sufficient to achieve significantly improved models. We also compare this approach with a binary classification approach and evaluate the effect of dataset size on model performance.

Machine learning models have been used extensively in hydrology, but issues persist with regard to their transparency, and there is currently no identifiable best practice for forcing variables in streamflow or flood modeling. In this paper, using data from the Centre for Ecology & Hydrology’s National River Flow Archive and from the European Centre for Medium-Range Weather Forecasts, we present a study that focuses on the input variable set for a neural network streamflow model to demonstrate how certain variables can be internalized, leading to a compressed feature set. By highlighting this capability to learn effectively using proxy variables, we demonstrate a more transferable framework that minimizes sensing requirements and that enables a route toward generalizing models.

Environmental data science for spatial extremes has traditionally relied heavily on max-stable processes. Even though the popularity of these models has perhaps peaked with statisticians, they are still perceived and considered as the “state of the art” in many applied fields. However, while the asymptotic theory supporting the use of max-stable processes is mathematically rigorous and comprehensive, we think that it has also been overused, if not misused, in environmental applications, to the detriment of more purposeful and meticulously validated models. In this article, we review the main limitations of max-stable process models, and strongly argue against their systematic use in environmental studies. Alternative solutions based on more flexible frameworks using the exceedances of variables above appropriately chosen high thresholds are discussed, and an outlook on future research is given. We consider the opportunities offered by hybridizing machine learning with extreme-value statistics, highlighting seven key recommendations moving forward.

Herbicide-resistant weeds are fast becoming a substantial global problem, causing significant crop losses and food insecurity. Late detection of resistant weeds leads to increasing economic losses. Traditionally, genetic sequencing and herbicide dose-response studies are used to detect herbicide-resistant weeds, but these are expensive and slow processes. To address this problem, an artificial intelligence (AI)-based herbicide-resistant weed identifier program (HRIP) was developed to quickly and accurately distinguish acetolactate synthetase inhibitor (ALS)-resistant from -susceptible common chickweed plants. A regular camera was converted to capture light wavelengths from 300 to 1,100 nm. Full spectrum images from a two-year experiment were used to develop a hyperparameter-tuned convolutional neural network (CNN) model utilizing a “train from scratch” approach. This novel approach exploits the subtle differences in the spectral signature of ALS-resistant and -susceptible common chickweed plants as they react differently to the ALS herbicide treatments. The HRIP was able to identify ALS-resistant common chickweed as early as 72 hours after treatment at an accuracy of 88%. It has broad applicability due to its ability to distinguish ALS-resistant from -susceptible common chickweed plants regardless of the type of ALS herbicide or dose used. Utilizing tools such as the HRIP will allow farmers to make timely interventions to prevent the herbicide-escape plants from completing their life cycle and adding to the weed seedbank.

Super-resolution of turbulence is a term used to describe the prediction of high-resolution snapshots of a flow from coarse-grained observations. This is typically accomplished with a deep neural network and training usually requires a dataset of high-resolution images. An approach is presented here in which robust super-resolution can be performed without access to high-resolution reference data, as might be expected in an experiment. The training procedure is similar to data assimilation, wherein the model learns to predict an initial condition that leads to accurate coarse-grained predictions at later times, while only being shown coarse-grained observations. Implementation of the approach requires the use of a fully differentiable flow solver in the training loop to allow for time-marching of predictions. A range of models are trained on data generated from forced, two-dimensional turbulence. The networks have reconstruction errors which are similar to those obtained with ‘standard’ super-resolution approaches using high-resolution data. Furthermore, the methods are comparable to the performance of standard data assimilation for state estimation on individual trajectories, outperforming these variational approaches at initial time and remaining robust when unrolled in time where performance of the standard data-assimilation algorithm improves.

Rapid urbanization poses several challenges, especially when faced with an uncontrolled urban development plan. Therefore, it often leads to anarchic occupation and expansion of cities, resulting in the phenomenon of urban sprawl (US). To support sustainable decision–making in urban planning and policy development, a more effective approach to addressing this issue through US simulation and prediction is essential. Despite the work published in the literature on the use of deep learning (DL) methods to simulate US indicators, almost no work has been published to assess what has already been done, the potential, the issues, and the challenges ahead. By synthesising existing research, we aim to assess the current landscape of the use of DL in modelling US. This article elucidates the complexities of US, focusing on its multifaceted challenges and implications. Through an examination of DL methodologies, we aim to highlight their effectiveness in capturing the complex spatial patterns and relationships associated with US. This work begins by demystifying US, highlighting its multifaceted challenges. In addition, the article examines the synergy between DL and conventional methods, highlighting the advantages and disadvantages. It emerges that the use of DL in the simulation and forecasting of US indicators is increasing, and its potential is very promising for guiding strategic decisions to control and mitigate this phenomenon. Of course, this is not without major challenges, both in terms of data and models and in terms of strategic city planning policies.

Risk-based surveillance is now a well-established paradigm in epidemiology, involving the distribution of sampling efforts differentially in time, space, and within populations, based on multiple risk factors. To assess and map the risk of the presence of the bacterium Xylella fastidiosa, we have compiled a dataset that includes factors influencing plant development and thus the spread of such harmful organism. To this end, we have collected, preprocessed, and gathered information and data related to land types, soil compositions, and climatic conditions to predict and assess the probability of risk associated with X. fastidiosa in relation to environmental features. This resource can be of interest to researchers conducting analyses on X. fastidiosa and, more generally, to researchers working on geospatial modeling of risk related to plant infectious diseases.

Both energy performance certificates (EPCs) and thermal infrared (TIR) images play key roles in mapping the energy performance of the urban building stock. In this paper, we developed parametric building archetypes using an EPC database and conducted temperature clustering on TIR images acquired from drones and satellite datasets. We evaluated 1,725 EPCs of existing building stock in Cambridge, UK, to generate energy consumption profiles. Drone-based TIR images of individual buildings in two Cambridge University colleges were processed using a machine learning pipeline for thermal anomaly detection and investigated the influence of two specific factors that affect the reliability of TIR for energy management applications: ground sample distance (GSD) and angle of view (AOV). The EPC results suggest that the construction year of the buildings influences their energy consumption. For example, modern buildings were over 30% more energy-efficient than older ones. In parallel, older buildings were found to show almost double the energy savings potential through retrofitting compared to newly constructed buildings. TIR imaging results showed that thermal anomalies can only be properly identified in images with a GSD of 1 m/pixel or less. A GSD of 1-6 m/pixel can detect hot areas of building surfaces. We found that a GSD > 6 m/pixel cannot characterize individual buildings but does help identify urban heat island effects. Additional sensitivity analysis showed that building thermal anomaly detection is more sensitive to AOV than to GSD. Our study informs newer approaches to building energy diagnostics using thermography and supports decision-making for large-scale retrofitting.

Machine learning (ML) techniques have emerged as a powerful tool for predicting weather and climate systems. However, much of the progress to date focuses on predicting the short-term evolution of the atmosphere. Here, we look at the potential for ML methodology to predict the evolution of the ocean. The presence of land in the domain is a key difference between ocean modeling and previous work looking at atmospheric modeling. Here, we look to train a convolutional neural network (CNN) to emulate a process-based General Circulation Model (GCM) of the ocean, in a configuration which contains land. We assess performance on predictions over the entire domain and near to the land (coastal points). Our results show that the CNN replicates the underlying GCM well when assessed over the entire domain. RMS errors over the test dataset are low in comparison to the signal being predicted, and the CNN model gives an order of magnitude improvement over a persistence forecast. When we partition the domain into near land and the ocean interior and assess performance over these two regions, we see that the model performs notably worse over the near land region. Near land, RMS scores are comparable to those from a simple persistence forecast. Our results indicate that ocean interaction with land is something the network struggles with and highlight that this is may be an area where advanced ML techniques specifically designed for, or adapted for, the geosciences could bring further benefits.

Atmospheric models used for weather and climate prediction are traditionally formulated in a deterministic manner. In other words, given a particular state of the resolved scale variables, the most likely forcing from the subgrid scale processes is estimated and used to predict the evolution of the large-scale flow. However, the lack of scale separation in the atmosphere means that this approach is a large source of error in forecasts. Over recent years, an alternative paradigm has developed: the use of stochastic techniques to characterize uncertainty in small-scale processes. These techniques are now widely used across weather, subseasonal, seasonal, and climate timescales. In parallel, recent years have also seen significant progress in replacing parametrization schemes using machine learning (ML). This has the potential to both speed up and improve our numerical models. However, the focus to date has largely been on deterministic approaches. In this position paper, we bring together these two key developments and discuss the potential for data-driven approaches for stochastic parametrization. We highlight early studies in this area and draw attention to the novel challenges that remain.

Climate models are biased with respect to real-world observations. They usually need to be adjusted before being used in impact studies. The suite of statistical methods that enable such adjustments is called bias correction (BC). However, BC methods currently struggle to adjust temporal biases. Because they mostly disregard the dependence between consecutive time points. As a result, climate statistics with long-range temporal properties, such as the number of heatwaves and their frequency, cannot be corrected accurately. This makes it more difficult to produce reliable impact studies on such climate statistics. This article offers a novel BC methodology to correct temporal biases. This is made possible by rethinking the philosophy behind BC. We will introduce BC as a time-indexed regression task with stochastic outputs. Rethinking BC enables us to adapt state-of-the-art machine learning (ML) attention models and thereby learn different types of biases, including temporal asynchronicities. With a case study of number of heatwaves in Abuja, Nigeria and Tokyo, Japan, we show more accurate results than current climate model outputs and alternative BC methods.

This paper presents a machine learning approach to multidimensional item response theory (MIRT), a class of latent factor models that can be used to model and predict student performance from observed assessment data. Inspired by collaborative filtering, we define a general class of models that includes many MIRT models. We discuss the use of penalized joint maximum likelihood to estimate individual models and cross-validation to select the best performing model. This model evaluation process can be optimized using batching techniques, such that even sparse large-scale data can be analyzed efficiently. We illustrate our approach with simulated and real data, including an example from a massive open online course. The high-dimensional model fit to this large and sparse dataset does not lend itself well to traditional methods of factor interpretation. By analogy to recommender-system applications, we propose an alternative “validation” of the factor model, using auxiliary information about the popularity of items consulted during an open-book examination in the course.

Utilizing technology for automated item generation is not a new idea. However, test items used in commercial testing programs or in research are still predominantly written by humans, in most cases by content experts or professional item writers. Human experts are a limited resource and testing agencies incur high costs in the process of continuous renewal of item banks to sustain testing programs. Using algorithms instead holds the promise of providing unlimited resources for this crucial part of assessment development. The approach presented here deviates in several ways from previous attempts to solve this problem. In the past, automatic item generation relied either on generating clones of narrowly defined item types such as those found in language free intelligence tests (e.g., Raven’s progressive matrices) or on an extensive analysis of task components and derivation of schemata to produce items with pre-specified variability that are hoped to have predictable levels of difficulty. It is somewhat unlikely that researchers utilizing these previous approaches would look at the proposed approach with favor; however, recent applications of machine learning show success in solving tasks that seemed impossible for machines not too long ago. The proposed approach uses deep learning to implement probabilistic language models, not unlike what Google brain and Amazon Alexa use for language processing and generation.

A Bayes estimation procedure is introduced that allows the nature and strength of prior beliefs to be easily specified and modal posterior estimates to be obtained as easily as maximum likelihood estimates. The procedure is based on constructing posterior distributions that are formally identical to likelihoods, but are based on sampled data as well as artificial data reflecting prior information. Improvements in performance of modal Bayes procedures relative to maximum likelihood estimation are illustrated for Rasch-type models. Improvements range from modest to dramatic, depending on the model and the number of items being considered.

Word embeddings are now a vital resource for social science research. However, obtaining high-quality training data for non-English languages can be difficult, and fitting embeddings therein may be computationally expensive. In addition, social scientists typically want to make statistical comparisons and do hypothesis tests on embeddings, yet this is nontrivial with current approaches. We provide three new data resources designed to ameliorate the union of these issues: (1) a new version of fastText model embeddings, (2) a multilanguage “a la carte” (ALC) embedding version of the fastText model, and (3) a multilanguage ALC embedding version of the well-known GloVe model. All three are fit to Wikipedia corpora. These materials are aimed at “low-resource” settings where the analysts lack access to large corpora in their language of interest or to the computational resources required to produce high-quality vector representations. We make these resources available for 40 languages, along with a code pipeline for another 117 languages available from Wikipedia corpora. We extensively validate the materials via reconstruction tests and other proofs-of-concept. We also conduct human crowdworker tests for our embeddings for Arabic, French, (traditional Mandarin) Chinese, Japanese, Korean, Russian, and Spanish. Finally, we offer some advice to practitioners using our resources.