Given the growing use of Artificial intelligence (AI) and machine learning (ML) methods across all aspects of environmental sciences, it is imperative that we initiate a discussion about the ethical and responsible use of AI. In fact, much can be learned from other domains where AI was introduced, often with the best of intentions, yet often led to unintended societal consequences, such as hard coding racial bias in the criminal justice system or increasing economic inequality through the financial system. A common misconception is that the environmental sciences are immune to such unintended consequences when AI is being used, as most data come from observations, and AI algorithms are based on mathematical formulas, which are often seen as objective. In this article, we argue the opposite can be the case. Using specific examples, we demonstrate many ways in which the use of AI can introduce similar consequences in the environmental sciences. This article will stimulate discussion and research efforts in this direction. As a community, we should avoid repeating any foreseeable mistakes made in other domains through the introduction of AI. In fact, with proper precautions, AI can be a great tool to help reduce climate and environmental injustice. We primarily focus on weather and climate examples but the conclusions apply broadly across the environmental sciences.

Despite the increasingly successful application of neural networks to many problems in the geosciences, their complex and nonlinear structure makes the interpretation of their predictions difficult, which limits model trust and does not allow scientists to gain physical insights about the problem at hand. Many different methods have been introduced in the emerging field of eXplainable Artificial Intelligence (XAI), which aims at attributing the network’s prediction to specific features in the input domain. XAI methods are usually assessed by using benchmark datasets (such as MNIST or ImageNet for image classification). However, an objective, theoretically derived ground truth for the attribution is lacking for most of these datasets, making the assessment of XAI in many cases subjective. Also, benchmark datasets specifically designed for problems in geosciences are rare. Here, we provide a framework, based on the use of additively separable functions, to generate attribution benchmark datasets for regression problems for which the ground truth of the attribution is known a priori. We generate a large benchmark dataset and train a fully connected network to learn the underlying function that was used for simulation. We then compare estimated heatmaps from different XAI methods to the ground truth in order to identify examples where specific XAI methods perform well or poorly. We believe that attribution benchmarks as the ones introduced herein are of great importance for further application of neural networks in the geosciences, and for more objective assessment and accurate implementation of XAI methods, which will increase model trust and assist in discovering new science.

Modeling dependencies between climate extremes is important for climate risk assessment, for instance when allocating emergency management funds. In statistics, multivariate extreme value theory is often used to model spatial extremes. However, most commonly used approaches require strong assumptions and are either too simplistic or over-parameterized. From a machine learning perspective, generative adversarial networks (GANs) are a powerful tool to model dependencies in high-dimensional spaces. Yet in the standard setting, GANs do not well represent dependencies in the extremes. Here we combine GANs with extreme value theory (evtGAN) to model spatial dependencies in summer maxima of temperature and winter maxima in precipitation over a large part of western Europe. We use data from a stationary 2000-year climate model simulation to validate the approach and explore its sensitivity to small sample sizes. Our results show that evtGAN outperforms classical GANs and standard statistical approaches to model spatial extremes. Already with about 50 years of data, which corresponds to commonly available climate records, we obtain reasonably good performance. In general, dependencies between temperature extremes are better captured than dependencies between precipitation extremes due to the high spatial coherence in temperature fields. Our approach can be applied to other climate variables and can be used to emulate climate models when running very long simulations to determine dependencies in the extremes is deemed infeasible.

The growth of machine learning (ML) in environmental science can be divided into a slow phase lasting till the mid-2010s and a fast phase thereafter. The rapid transition was brought about by the emergence of powerful new ML methods, allowing ML to successfully tackle many problems where numerical models and statistical models have been hampered. Deep convolutional neural network models greatly advanced the use of ML on 2D or 3D data. Transfer learning has allowed ML to progress in climate science, where data records are generally short for ML. ML and physics are also merging in new areas, for example: (a) using ML for general circulation model parametrization, (b) adding physics constraints in ML models, and (c) using ML in data assimilation.

In the twenty-first century, machine learning and deep learning have been successfully used to find hidden information from coarse-grained data in various domains. In Computer Vision, scientists have used neural networks to identify hidden pixel-level information from low-resolution (LR) image data. This approach of estimating high-resolution (HR) information from LR data is called the super-resolution (SR) approach. This approach has been borrowed by climate scientists to downscale coarse-level measurements of climate variables to obtain their local-scale projections. Climate variables are spatial in nature and can be represented as images where each pixel denotes a grid point where the variables can be measured. We can apply the deep learning-based SR techniques on such “images” for statistical downscaling of such variables. This approach of downscaling can be termed as deep downscaling. In this work, we have tried to make HR projection of the Indian summer monsoon rainfall by using a novel deep residual network called ResDeepD. The aim is to downscale the 10 × 10 low LR precipitation data to get the values at 0.250 × 0.250 resolution. The proposed model uses a series of skip connections across residual blocks to give better results as compared to the existing models like super-resolution convolutional neural network, DeepSD, and Nest-UNet that have been used previously for this task. We have also examined the model’s performance for downscaling rainfall during some extreme climatic events like cyclonic storms and deep depression and found that the model performs better than the existing models.

The physical and socioeconomic environments in which we live are intrinsically linked over a wide range of time and space scales. On monthly intervals, the price of many commodities produced predominantly in tropical regions covary with the dominant mode of climate variability in this region, namely the El Niño Southern Oscillation (ENSO). Here, for the spot prices returns of vegetable oils produced in Asia, we develop autoregressive (AR) models with exogenous ENSO indices, where for the first time these indices are generated by a purpose-built state-of-the-art general circulation model (GCM) climate forecasting system. The GCM is a numerical simulation which couples together the atmosphere, ocean, and sea ice, with the initial conditions tailored to maximize the climate forecast skill at multiyear timescales in the tropics. To serve as additional benchmarks, we also test commodity forecasts using: (a) no ENSO information as a lower bound; (b) perfect future ENSO knowledge as a reference upper bound; and (c) an econometric AR model of ENSO. All models adopting ENSO factors outperform those that do not, indicating the value here of incorporating climate knowledge into investment decision-making. Commodity forecasts adopting perfect ENSO factors have statistically significant skill out to 2 years. When adopting the GCM-ENSO factors, there is predictive power of the commodity beyond 1 year in the best case, which consistently outperforms commodity forecasts adopting an AR econometric model of ENSO.

The increasing intensity and frequency of floods is one of the many consequences of our changing climate. In this work, we explore ML techniques that improve the flood detection module of an operational early flood warning system. Our method exploits an unlabeled dataset of paired multi-spectral and synthetic aperture radar (SAR) imagery to reduce the labeling requirements of a purely supervised learning method. Prior works have used unlabeled data by creating weak labels out of them. However, from our experiments, we noticed that such a model still ends up learning the label mistakes in those weak labels. Motivated by knowledge distillation and semi-supervised learning, we explore the use of a teacher to train a student with the help of a small hand-labeled dataset and a large unlabeled dataset. Unlike the conventional self-distillation setup, we propose a cross-modal distillation framework that transfers supervision from a teacher trained on richer modality (multi-spectral images) to a student model trained on SAR imagery. The trained models are then tested on the Sen1Floods11 dataset. Our model outperforms the Sen1Floods11 baseline model trained on the weak-labeled SAR imagery by an absolute margin of $ 6.53\% $ intersection over union (IoU) on the test split.

Aerosol particles play an important role in the climate system by absorbing and scattering radiation and influencing cloud properties. They are also one of the biggest sources of uncertainty for climate modeling. Many climate models do not include aerosols in sufficient detail due to computational constraints. To represent key processes, aerosol microphysical properties and processes have to be accounted for. This is done in the ECHAM-HAM (European Center for Medium-Range Weather Forecast-Hamburg-Hamburg) global climate aerosol model using the M7 microphysics, but high computational costs make it very expensive to run with finer resolution or for a longer time. We aim to use machine learning to emulate the microphysics model at sufficient accuracy and reduce the computational cost by being fast at inference time. The original M7 model is used to generate data of input–output pairs to train a neural network (NN) on it. We are able to learn the variables’ tendencies achieving an average $ {R}^2 $ score of 77.1%. We further explore methods to inform and constrain the NN with physical knowledge to reduce mass violation and enforce mass positivity. On a Graphics processing unit (GPU), we achieve a speed-up of up to over 64 times faster when compared to the original model.

A common task in Earth Sciences is to infer climate information at local and regional scales from global climate models. Dynamical downscaling requires running expensive numerical models at high resolution, which can be prohibitive due to long model runtimes. On the other hand, statistical downscaling techniques present an alternative approach for learning links between the large- and local-scale climate in a more efficient way. A large number of deep neural network-based approaches for statistical downscaling have been proposed in recent years, mostly based on convolutional architectures developed for computer vision and super-resolution tasks. This paper presents deep learning for empirical downscaling (DL4DS), a python library that implements a wide variety of state-of-the-art and novel algorithms for downscaling gridded Earth Science data with deep neural networks. DL4DS has been designed with the goal of providing a general framework for training convolutional neural networks with configurable architectures and learning strategies to facilitate the conduction of comparative and ablation studies in a robust way. We showcase the capabilities of DL4DS on air quality Copernicus Atmosphere Monitoring Service (CAMS) data over the western Mediterranean area. The DL4DS library can be found in this repository: https://github.com/carlos-gg/dl4ds

Energy consumption in buildings, both residential and commercial, accounts for approximately 40% of all energy usage in the United States, and similar numbers are being reported from countries around the world. This significant amount of energy is used to maintain a comfortable, secure, and productive environment for the occupants. So, it is crucial that energy consumption in buildings must be optimized, all the while maintaining satisfactory levels of occupant comfort, health, and safety. Machine learning (ML) has been proven to be an invaluable tool in deriving important insights from data and optimizing various systems. In this work, we review some of the most promising ways in which ML has been leveraged to make buildings smart and energy-efficient. For the convenience of readers, we provide a brief introduction to the relevant ML paradigms and the components and functioning of each smart building system we cover. Finally, we discuss the challenges faced while implementing machine learning algorithms in smart buildings and provide future avenues for research in this field.

Tree-ring chronologies encode interannual variability in forest growth rates over long time periods from decades to centuries or even millennia. However, each chronology is a highly localized measurement describing conditions at specific sites where wood samples have been collected. The question whether these local growth variabilites are representative for large geographical regions remains an open issue. To overcome the limitations of interpreting a sparse network of sites, we propose an upscaling approach for annual tree-ring indices that approximate forest growth variability and compute gridded data products that generalize the available information for multiple tree genera. Using regression approaches from machine learning, we predict tree-ring indices in space and time based on climate variables, but considering also species range maps as constraints for the upscaling. We compare various prediction strategies in cross-validation experiments to identify the best performing setup. Our estimated maps of tree-ring indices are the first data products that provide a dense view on forest growth variability at the continental level with 0.5° and 0.0083° spatial resolution covering the years 1902–2013. Furthermore, we find that different genera show very variable spatial patterns of anomalies. We have selected Europe as study region and focused on the six most prominent tree genera, but our approach is very generic and can easily be applied elsewhere. Overall, the study shows perspectives but also limitations for reconstructing spatiotemporal dynamics of complex biological processes. The data products are available at https://www.doi.org/10.17871/BACI.248.

When one or several classes are much less prevalent than another class (unbalanced data), class error rates and variable importances of the machine learning algorithm random forest can be biased, particularly when sample sizes are smaller, imbalance levels higher, and effect sizes of important variables smaller. Using simulated data varying in size, imbalance level, number of true variables, their effect sizes, and the strength of multicollinearity between covariates, we evaluated how eight versions of random forest ranked and selected true variables out of a large number of covariates despite class imbalance. The version that calculated variable importance based on the area under the curve (AUC) was least adversely affected by class imbalance. For the same number of true variables, effect sizes, and multicollinearity between covariates, the AUC variable importance ranked true variables still highly at the lower sample sizes and higher imbalance levels at which the other seven versions no longer achieved high ranks for true variables. Conversely, using the Hellinger distance to split trees or downsampling the majority class already ranked true variables lower and more variably at the larger sample sizes and lower imbalance levels at which the other algorithms still ranked true variables highly. In variable selection, a higher proportion of true variables were identified when covariates were ranked by AUC importances and the proportion increased further when the AUC was used as the criterion in forward variable selection. In three case studies, known species–habitat relationships and their spatial scales were identified despite unbalanced data.

Environmental sensors are crucial for monitoring weather conditions and the impacts of climate change. However, it is challenging to place sensors in a way that maximises the informativeness of their measurements, particularly in remote regions like Antarctica. Probabilistic machine learning models can suggest informative sensor placements by finding sites that maximally reduce prediction uncertainty. Gaussian process (GP) models are widely used for this purpose, but they struggle with capturing complex non-stationary behaviour and scaling to large datasets. This paper proposes using a convolutional Gaussian neural process (ConvGNP) to address these issues. A ConvGNP uses neural networks to parameterise a joint Gaussian distribution at arbitrary target locations, enabling flexibility and scalability. Using simulated surface air temperature anomaly over Antarctica as training data, the ConvGNP learns spatial and seasonal non-stationarities, outperforming a non-stationary GP baseline. In a simulated sensor placement experiment, the ConvGNP better predicts the performance boost obtained from new observations than GP baselines, leading to more informative sensor placements. We contrast our approach with physics-based sensor placement methods and propose future steps towards an operational sensor placement recommendation system. Our work could help to realise environmental digital twins that actively direct measurement sampling to improve the digital representation of reality.

We demonstrate the use of a probabilistic machine-learning technique to develop stochastic parameterizations of atmospheric column physics. After suitable preprocessing of NASA’s Modern-Era Retrospective analysis for Research and Applications, version 2 (MERRA2) data to minimize the effects of high-frequency, high-wavenumber component of MERRA2 estimate of vertical velocity, we use generative adversarial networks to learn the probability distribution of vertical profiles of diabatic sources conditioned on vertical profiles of temperature and humidity. This may be viewed as an improvement over previous similar but deterministic approaches that seek to alleviate both, shortcomings of human-designed physics parameterizations, and the computational demand of the “physics” step in climate models.

Measuring and attributing greenhouse gas (GHG) emissions remains a challenging problem as the world strives toward meeting emissions reductions targets. As a significant portion of total global emissions, the road transportation sector represents an enormous challenge for estimating and tracking emissions at a global scale. To meet this challenge, we have developed a hybrid approach for estimating road transportation emissions that combines the strengths of machine learning and satellite imagery with localized emissions factors data to create an accurate, globally scalable, and easily configurable GHG monitoring framework.

Understanding the meteorological drivers of extreme impacts in social or environmental systems is important to better quantify current and project future climate risks. Impacts are typically an aggregated response to many different interacting drivers at various temporal scales, rendering such driver identification a challenging task. Machine learning–based approaches, such as deep neural networks, may be able to address this task but require large training datasets. Here, we explore the ability of Convolutional Neural Networks (CNNs) to predict years with extremely low gross primary production (GPP) from daily weather data in three different vegetation types. To circumvent data limitations in observations, we simulate 100,000 years of daily weather with a weather generator for three different geographical sites and subsequently simulate vegetation dynamics with a complex vegetation model. For each resulting vegetation distribution, we then train two different CNNs to classify daily weather data (temperature, precipitation, and radiation) into years with extremely low GPP and normal years. Overall, prediction accuracy is very good if the monthly or yearly GPP values are used as an intermediate training target (area under the precision-recall curve AUC $ \ge $ 0.9). The best prediction accuracy is found in tropical forests, with temperate grasslands and boreal forests leading to comparable results. Prediction accuracy is strongly reduced when binary classification is used directly. Furthermore, using daily GPP during training does not improve the predictive power. We conclude that CNNs are able to predict extreme impacts from complex meteorological drivers if sufficient data are available.

Climate emulators are a powerful instrument for climate modeling, especially in terms of reducing the computational load for simulating spatiotemporal processes associated with climate systems. The most important type of emulators are statistical emulators trained on the output of an ensemble of simulations from various climate models. However, such emulators oftentimes fail to capture the “physics” of a system that can be detrimental for unveiling critical processes that lead to climate tipping points. Historically, statistical mechanics emerged as a tool to resolve the constraints on physics using statistics. We discuss how climate emulators rooted in statistical mechanics and machine learning can give rise to new climate models that are more reliable and require less observational and computational resources. Our goal is to stimulate discussion on how statistical climate emulators can further be improved with the help of statistical mechanics which, in turn, may reignite the interest of statistical community in statistical mechanics of complex systems.

Exposure to ground-level ozone is a concern for both humans and vegetation, so accurate prediction of ozone time series is of great importance. However, conventional as well as emerging methods have deficiencies in predicting time series when a superposition of differently pronounced oscillations on various time scales is present. In this paper, we propose a meteorologically motivated filtering method of time series data, which can separate oscillation patterns, in combination with different multibranch neural networks. To avoid phase shifts introduced by using a causal filter, we combine past observation data with a climatological estimate about the future to be able to apply a noncausal filter in a forecast setting. In addition, the forecast in the form of the expected climatology provides some a priori information that can support the neural network to focus not merely on learning a climatological statistic. We apply this method to hourly data obtained from over 50 different monitoring stations in northern Germany situated in rural or suburban surroundings to generate a prediction for the daily maximum 8-hr average values of ground-level ozone 4 days into the future. The data preprocessing with time filters enables simpler neural networks such as fully connected networks as well as more sophisticated approaches such as convolutional and recurrent neural networks to better recognize long-term and short-term oscillation patterns like the seasonal cycle and thus leads to an improvement in the forecast skill, especially for a lead time of more than 48 hr, compared to persistence, climatological reference, and other reference models.

The oceans have a very important role in climate regulation due to their massive heat storage capacity. Thus, for the past decades, oceans have been observed by satellites to better understand their dynamics. Satellites retrieve several data with various spatial resolutions. For instance, sea surface height (SSH) is a low-resolution data field where sea surface temperature (SST) can be retrieved in a much higher one. These two physical parameters are linked by a physical link that can be learned by a super-resolution machine-learning algorithm. In this work, we present a subpixel convolutional deep learning model that takes advantage of the higher resolution SST field to guide the downscaling of the SSH one. The data fields that we use are simulated by a physic-based ocean model at a higher sampling rate than the satellites provide. We compared our approach with a convolutional neural network model. Our architecture generalized well with validation performances of 3.94 cm root mean squared error (RMSE) and training performances of 2.65 cm RMSE.

Teleconnections that link climate processes at widely separated spatial locations form a key component of the climate system. Their analysis has traditionally been based on means, climatologies, correlations, or spectral properties, which cannot always reveal the dynamical mechanisms between different climatological processes. More recently, causal discovery methods based either on time series at grid locations or on modes of variability, estimated through dimension-reduction methods, have been introduced. A major challenge in the development of such analysis methods is a lack of ground truth benchmark datasets that have facilitated improvements in many parts of machine learning. Here, we present a simplified stochastic climate model that outputs gridded data and represents climate modes and their teleconnections through a spatially aggregated vector-autoregressive model. The model is used to construct benchmarks and evaluate a range of analysis methods. The results highlight that the model can be successfully used to benchmark different causal discovery methods for spatiotemporal data and show their strengths and weaknesses. Furthermore, we introduce a novel causal discovery method at the grid level and demonstrate that it has orders of magnitude better performance than the current approaches. Improved causal analysis tools for spatiotemporal climate data are pivotal to advance process-based understanding and climate model evaluation.