2024
Plésiat, É., Dunn, R.J.H., Donat, M.G. et al. Artificial intelligence reveals past climate extremes by reconstructing historical records. Nat Commun 15, 9191 (2024)
Abstract
he understanding of recent climate extremes and the characterization of climate risk require examining these extremes within a historical context. However, the existing datasets of observed extremes generally exhibit spatial gaps and inaccuracies due to inadequate spatial extrapolation. This problem arises from traditional statistical methods used to account for the lack of measurements, particularly prevalent before the mid-20th century. In this work, we use artificial intelligence to reconstruct observations of European climate extremes (warm and cold days and nights) by leveraging Earth system model data from CMIP6 through transfer learning. Our method surpasses conventional statistical techniques and diffusion models, showcasing its ability to reconstruct past extreme events and reveal spatial trends across an extensive time span (1901-2018) that is not covered by most reanalysis datasets. Providing our dataset to the climate community will improve the characterization of climate extremes, resulting in better risk management and policies.
Guido Ascenso, Giulio Palcic, Enrico Scoccimarro, Andrea Castelletti (2024) A Systematic Framework for Data Augmentation for Tropical Cyclone Intensity Estimation Using Deep Learning Machine Learning and Computation, Volume1, Issue3
Abstract
Given the exponential relationship between the intensity and the expected damage of tropical cyclones, accurately estimating their intensity from satellite images is a crucially important area of research. Yet, the imbalance and limited size of available data sets significantly hinder model training and generalization capabilities, especially for state-of-the-art deep learning models. Therefore, it is standard in this field to use data augmentation, a family of transformations to increase the size and variety of a given data set. However, the principles behind the usage of these techniques for the estimation of the intensity of tropical cyclones has been largely unexamined. In this paper, we introduce a framework for establishing how much augmentation to apply and which techniques to use. To determine the ideal amount of augmentation, we use a modified Gini coefficient to understand how the augmentation will affect the imbalance in the data set, and we find that there is an upper bound on how much augmentation should be done (for our data set, 50 times per image). Our results also indicate that data augmentation works best when used to reduce the amount of imbalance in a data set rather than uniformly over the entire data set, as is typically done in the tropical cyclone intensity estimation literature. We then devise a backward elimination feature selection algorithm to determine which augmentation techniques work best. Our findings suggest that all augmentation techniques are effective for the estimation of tropical cyclones intensity, including random erasing, which we were the first to implement in this context.
Stefano Materia, Lluís Palma García, Chiem van Straaten, Sungmin O, Antonios Mamalakis, Leone Cavicchia, Dim Coumou, Paolo de Luca, Marlene Kretschmer, Markus Donat (2024) Artificial intelligence for climate prediction of extremes: State of the art, challenges, and future perspectives. WIREs Climate Change, e914
Abstract
Extreme events such as heat waves and cold spells, droughts, heavy rain, and storms are particularly challenging to predict accurately due to their rarity and chaotic nature, and because of model limitations. However, recent studies have shown that there might be systemic predictability that is not being leveraged, whose exploitation could meet the need for reliable predictions of aggregated extreme weather measures on timescales from weeks to decades ahead. Recently, numerous studies have been devoted to the use of artificial intelligence (AI) to study predictability and make climate predictions. AI techniques have shown great potential to improve the prediction of extreme events and uncover their links to large-scale and local drivers. Machine and deep learning have been explored to enhance prediction, while causal discovery and explainable AI have been tested to improve our understanding of the processes underlying predictability. Hybrid predictions combining AI, which can reveal unknown spatiotemporal connections from data, with climate models that provide the theoretical foundation and interpretability of the physical world, have shown that improving prediction skills of extremes on climate-relevant timescales is possible. However, numerous challenges persist in various aspects, including data curation, model uncertainty, generalizability, reproducibility of methods, and workflows. This review aims at overviewing achievements and challenges in the use of AI techniques to improve the prediction of extremes at the subseasonal to decadal timescale. A few best practices are identified to increase trust in these novel techniques, and future perspectives are envisaged for further scientific development.
Anastasiya Shyrokaya, Gabriele Messori, Ilias Pechlivanidis, Florian Pappenberger, Hannah L Cloke, and Giuliano Di Baldassarre (2024) Significant relationships between drought indicators and impacts for the 2018–2019 drought in Germany Environmental Research Letters 19 014037
Abstract
Despite the scientific progress in drought detection and forecasting, it remains challenging to accurately predict the corresponding impact of a drought event. This is due to the complex relationships between (multiple) drought indicators and adverse impacts across different places/hydroclimatic conditions, sectors, and spatiotemporal scales. In this study, we explored these relationships by analyzing the impacts of the severe 2018–2019 central European drought event in Germany. We first computed the standardized precipitation index (SPI), the standardized precipitation evaporation index (SPEI), the standardized soil moisture index (SSMI) and the standardized streamflow index (SSFI) over various accumulation periods, and then related these indicators to sectorial losses from the European drought impact report inventory (EDII) and media sources. To cope with the uncertainty associated with both drought indicators and impact data, we developed a fuzzy method to categorize them. Lastly, we applied the method at the region level (EU NUTS1) by correlating monthly time series. Our findings revealed strong and significant relationships between drought indicators and impacts over different accumulation periods, albeit in some cases region-specific and time-variant. Furthermore, our analysis established the interconnectedness between various sectors, which displayed systematically co-occurring impacts. As such, our work provides a new framework to explore drought indicators-impacts dependencies across space, time, sectors, and scales. In addition, it emphasizes the need to leverage available impact data to better forecast drought impacts.
Salcedo-Sanz, S., Pérez-Aracil, J., Ascenso, G., Del Ser, J., Casillas-Pèrez, D., Kadow, C., Fister, D., Barriopedro, D., Garcìa-Herrera, R., Giuliani, M., & Castelletti, A.(2024) Analysis, characterization, prediction, and attribution of extreme atmospheric events with machine learning and deep learning techniques: a review Theoretical and Applied Climatology, 155
Abstract
Atmospheric extreme events cause severe damage to human societies and ecosystems. The frequency and intensity of extremes and other associated events are continuously increasing due to climate change and global warming. The accurate prediction, characterization, and attribution of atmospheric extreme events is, therefore, a key research field in which many groups are currently working by applying different methodologies and computational tools. Machine learning and deep learning methods have arisen in the last years as powerful techniques to tackle many of the problems related to atmospheric extreme events. This paper reviews machine learning and deep learning approaches applied to the analysis, characterization, prediction, and attribution of the most important atmospheric extremes. A summary of the most used machine learning and deep learning techniques in this area, and a comprehensive critical review of literature related to ML in EEs, are provided. The critical literature review has been extended to extreme events related to rainfall and floods, heatwaves and extreme temperatures, droughts, severe weather events and fog, and low-visibility episodes. A case study focused on the analysis of extreme atmospheric temperature prediction with ML and DL techniques is also presented in the paper. Conclusions, perspectives, and outlooks on the field are finally drawn.
Pérez-Aracil, J., D. Fister, C. Marina, C. Peláez-Rodriguez, L. Cornejo-Bueno, P. Gutiérrez, M. Giuliani, A. Castelletti, and S. Salcedo-Sanz (2024) Long-term temperature prediction with hybrid autoencoder algorithms. Applied Computing and Geosciences, 23
Abstract
This paper proposes two hybrid approaches based on Autoencoders (AEs) for long-term temperature prediction. The first algorithm comprises an AE trained to learn temperature patterns, which is then linked to a second AE, used to detect possible anomalies and provide a final temperature prediction. The second proposed approach involves training an AE and then using the resulting latent space as input of a neural network, which will provide the final prediction output. Both approaches are tested in long-term air temperature prediction in European cities: seven European locations where major heat waves occurred have been considered. The long-term temperature prediction for the entire year of the heatwave events has been analysed. Results show that the proposed approaches can obtain accurate long-term (up to 4 weeks) temperature prediction, improving Persistence and Climatology in the benchmark models compared. In heatwave periods, where the persistence of the temperature is extremely high, our approach beat the persistence operator in three locations and works similarly in the rest of the cases, showing the potential of this AE-based method for long-term temperature prediction.
Bonetti, P., Metelli, A.M. & Restelli, M. (2024) Interpretable linear dimensionality reduction based on bias-variance analysis. Data Min Knowl Disc 38, 1713–1781
Abstract
One of the central issues of several machine learning applications on real data is the choice of the input features. Ideally, the designer should select a small number of the relevant, nonredundant features to preserve the complete information contained in the original dataset, with little collinearity among features. This procedure helps mitigate problems like overfitting and the curse of dimensionality, which arise when dealing with high-dimensional problems. On the other hand, it is not desirable to simply discard some features, since they may still contain information that can be exploited to improve results. Instead, dimensionality reduction techniques are designed to limit the number of features in a dataset by projecting them into a lower dimensional space, possibly considering all the original features. However, the projected features resulting from the application of dimensionality reduction techniques are usually difficult to interpret. In this paper, we seek to design a principled dimensionality reduction approach that maintains the interpretability of the resulting features. Specifically, we propose a bias-variance analysis for linear models and we leverage these theoretical results to design an algorithm, Linear Correlated Features Aggregation (LinCFA), which aggregates groups of continuous features with their average if their correlation is “sufficiently large”. In this way, all features are considered, the dimensionality is reduced and the interpretability is preserved. Finally, we provide numerical validations of the proposed algorithm both on synthetic datasets to confirm the theoretical results and on real datasets to show some promising applications.
Enrico Scoccimarro, Paolo Lanteri, Leone Cavicchia (2024) Freddy: breaking record for tropical cyclone precipitation? Environmental Research Letters (19 064013)
Abstract
Depending on the location on the Earth, the amount of precipitation associated with tropical cyclones (TCs) can reach 20% of the total yearly precipitation over land and up to 40% over some ocean regions. TC induced freshwater flooding has been suggested to be the largest threat to human lives due to TCs. Therefore, a reliable quantification of the precipitation amount associated with each past TC is important for a better definition of the TC fingerprint on the climate. The temporal and horizontal resolution of state-of-the-art observational datasets and atmospheric reanalysis gives the possibility to quantify precipitation associated with TCs globally following the observed TC tracks. In this work we compare the TC-related precipitation in various observational and reanalysis datasets. A particular focus is given to the record-breaking TC Freddy (Southern Indian Ocean, 2023). Here we show that the time-varying bias in TC associated precipitation, due to the positive trend in assimilated observations, makes it difficult to assess long-term trend investigation based on reanalysis. To this aim we need to build on state-of-the-art general circulation models, free to evolve under historical radiative forcing.
Klehmet Katharina, Peter Berg, Denica Bozhinova, Louise Crochemore, Yiheng Du, Ilias Pechlivanidis, Christiana Photiadou, Wei Yang (2024) Robustness of hydrometeorological extremes in surrogated seasonal forecasts, International Journal of Climatology 1-14
Abstract
Water and disaster risk management require accurate information about hydrometeorological extremes. However, estimation of rare events using extreme value analysis is hampered by short observational records, with large resulting uncertainties. Here, we present a surrogate world setup that makes use of data samples from meteorological and hydrological seasonal re-forecasts to explore extremes for long return periods. The surrogate timeseries allow us to pool the re-forecasts into 1000-year-long timeseries. We can then calculate return values of extremes and explore how they are affected by the size of sub-samples as method for estimating the uncertainty. The approach relies on the fact that probabilistic seasonal re-forecasts, initialized with perturbed initial conditions, have limited predictive skill with increasing lead time. At long lead times re-forecasts will diverge into independent samples. The meteorological seasonal re-forecasts are taken from the SEAS5 system, and hydrological re-forecasts are generated with the E-HYPE process-based model for the pan-European domain. Extreme value analysis is applied to annual maxima of precipitation and streamflow for return periods of 100 years. The analysis clearly demonstrates the large uncertainty in long return period estimates with typical available samples of only few decades. The uncertainty is somewhat reduced for 100-year samples, but several 100 years seem to be necessary to have robust estimates. The bootstrap with replacement approach is applied to shorter timeseries, and is shown to well reproduce the uncertainty range of the longer samples. However, the main estimate of the return value can be significantly offset. Although the method is model based, with the associated uncertainties and bias compared to the real world, the surrogate approach is likely useful to explore rare and compounding extremes.
Verónica Torralba, Stefano Materia, Leone Cavicchia, M. Carmen Álvarez Castro, Chloé Prodhomme, Ronan McAdam, Enrico Scoccimarro and Silvio Gualdi (2024) Nighttime heat waves in the Euro-Mediterranean region: definition, characterisation, and seasonal prediction, Environ. Res. Lett. 19 034001
Abstract
The combined effect of above-normal nighttime temperatures with high humidity poses a high risk to human health by impeding the body’s recovery from daytime heat exposure. Seasonal predictions of nighttime heat waves (NHWs) can help to better anticipate these episodes and reduce their social and economic impacts. However, the ability of the seasonal forecast systems to predict NHWs has not been explored yet. This work investigates the potential of four seasonal forecasting systems and a multi-model (MM) ensemble to provide useful information on the frequency and magnitude of the NHWs in the Euro-Mediterranean region during the boreal summer season. The analysis employs a modified version of the heat wave magnitude index (HWMI) to evaluate the NHWs. Our results demonstrate for the first time that this index is an optimal choice for the seasonal prediction analysis as it is invariant to the mean biases and provides an integrated view of the NHWs for the entire season. In addition, the percentage of days in a season with temperatures exceeding the 90th percentile (NDQ90) has been used to assess the NHWs’ seasonal frequency. Different proxies for the assessment of NHWs have been considered: apparent temperature at night (ATn, computed from temperature and humidity at night), mean temperature at night, and daily minimum temperature. All these proxies are valid for the assessment of the NHWs, but ATn is more informative about the stress on human health since it includes the impact of humidity. This work has revealed that state-of-the-art seasonal forecast systems can represent the interannual variability of both HWMI and NDQ90 in Southern Europe, Eastern Europe, and the Middle East, but they show limitations in Northern Europe. The predictive capabilities of the seasonal forecasts in specific regions demonstrate the potential of these predictions for the effective management of the risks associated with summer NHWs.
2023
Paolo Bonetti, Alberto Maria Metelli, Marcello Restelli (2023) Causal Feature Selection via Transfer Entropy. arXiv:2310.11059
Abstract
Machine learning algorithms are designed to capture complex relationships between features. In this context, the high dimensionality of data often results in poor model performance, with the risk of overfitting. Feature selection, the process of selecting a subset of relevant and non-redundant features, is, therefore, an essential step to mitigate these issues. However, classical feature selection approaches do not inspect the causal relationship between selected features and target, which can lead to misleading results in real-world applications. Causal discovery, instead, aims to identify causal relationships between features with observational data. In this paper, we propose a novel methodology at the intersection between feature selection and causal discovery, focusing on time series. We introduce a new causal feature selection approach that relies on the forward and backward feature selection procedures and leverages transfer entropy to estimate the causal flow of information from the features to the target in time series. Our approach enables the selection of features not only in terms of mere model performance but also captures the causal information flow. In this context, we provide theoretical guarantees on the regression and classification errors for both the exact and the finite-sample cases. Finally, we present numerical validations on synthetic and real-world regression problems, showing results competitive w.r.t. the considered baselines.
Guido Ascenso, Leone Cavicchia, Enrico Scoccimarro and Andrea Castelletti (2023) Optimisation-based refinement of genesis indices for tropical cyclones. Environmental Research Communications, Volume 5, Number 2
Abstract
Tropical cyclone genesis indices are valuable tools for studying the relationship between large-scale environmental fields and the genesis of tropical cyclones, supporting the identification of future trends of cyclone genesis. However, their formulation is generally derived from simple statistical models (e.g., multiple linear regression) and are not optimised globally. In this paper, we present a simple framework for optimising genesis indexes given a user-specified trade-off between two performance metrics, which measure how well an index captures the spatial and interannual variability of tropical cyclone genesis. We apply the proposed framework to the popular Emanuel and Nolan Genesis Potential Index, yielding new, optimised formulas that correspond to different trade-offs between spatial and interannual variability. Result show that our refined indexes can improve the performance of the Emanuel and Nolan index up to 8% for spatial variability and 16%–22% for interannual variability; this improvement was found to be statistically significant (p < 0.01). Lastly, by analysing the formulas found, we give some insights into the role of the different inputs of the index in maximising one metric or the other.
Jerez, S., Barriopedro, D., García-López,A., Lorente-Plazas, R., Somoza,A. M., Turco, M., et al. (2023) An Action-Oriented Approach to Make the Most of the Wind and Solar Power Complementarity. (2023) Earth’s Future,11, e2022EF003332
Abstract
Solar and wind power are called to play a main role in the transition toward decarbonized electricity systems. However, their integration in the energy mix is highly compromised due to the intermittency of their production caused by weather and climate variability. To face the challenge, here we present research about actionable strategies for wind and solar photovoltaic facilities deployment that exploit their complementarity in order to minimize the volatility of their combined production while guaranteeing a certain supply. The developed methodology has been implemented in an open-access step-wise model called CLIMAX. It first identifies regions with homogeneous temporal variability of the resources, and then determines the optimal shares of each technology over such regions. In the simplistic application performed here, we customize the model to narrow the monthly deviations of the total wind-plus-solar electricity production from a given curve (here, the mean annual cycle of the total production) across five European domains. For the current shares of both technologies, the results show that an optimal siting of the power units would reduce the standard deviation of the monthly anomalies of the total wind-plus-solar power generation by up to 20% without loss in the mean capacity factor as compared to a baseline scenario with an evenly spatial distribution of the installations. This result further improves (up to 60% in specific regions) if the total shares of each technology are also optimized, thus encouraging the use of CLIMAX for practical guidance of next-generation renewable energy scenarios.
L. Cavicchia, E. Scoccimarro, G. Ascenso, A. Castelletti, M. Giuliani, S. Gualdi (2023) Tropical Cyclone Genesis Potential Indices in a New High-Resolution Climate Models Ensemble: Limitations and Way Forward. Geophysical Research Letters, Volume 50, Issue 11
Abstract
Genesis Potential Indices (GPIs) link the occurrence of Tropical Cyclones (TCs) to large-scale environmental conditions favorable for TC development. In the last few decades, they have been routinely used as a way to overcome the limitations of climate models (GCM), whose resolution is too coarse to produce realistic TCs. Recently, the first GCM ensemble with high enough horizontal resolution to realistically reproduce TCs was made available. Here, we address the questions of whether GPIs are still relevant in the era of TC-permitting climate model ensembles, and whether they have sufficient predictive skills. The predictive skills of GPIs are assessed against the TCs directly simulated in a climate model ensemble. We found that GPIs have poor skill in two key metrics: inter-annual variability and multi-decadal trends. We discuss possible ways to improve the understanding of the predictive skill of GPIs and therefore enhance their applicability in the era of TC-permitting GCMs.
Andrea Castelletti, Andrea Ficchì, Andrea Cominola, Pablo Segovia, Matteo Giuliani, Wenyan Wu, Sergio Lucia, Carlos Ocampo-Martinez, Bart De Schutter, José María Maestre (2024) Model Predictive Control of water resources systems: A review and research agenda. Annual Reviews in Control, Volume 55, Pages 442-465
Abstract
Model Predictive Control (MPC) has recently gained increasing interest in the adaptive management of water resources systems due to its capability of incorporating disturbance forecasts into real-time optimal control problems. Yet, related literature is scattered with heterogeneous applications, case-specific problem settings, and results that are hardly generalized and transferable across systems. Here, we systematically review 149 peer-reviewed journal articles published over the last 25 years on MPC applied to water reservoirs, open channels, and urban water networks to identify common trends and open challenges in research and practice. The three water systems we consider are inter-connected, multi-purpose and multi-scale dynamical systems affected by multiple hydro-climatic uncertainties and evolving socioeconomic factors. Our review first identifies four main challenges currently limiting most MPC applications in the water domain: (i) lack of systematic benchmarking of MPC with respect to other control methods; (ii) lack of assessment of the impact of uncertainties on the model-based control; (iii) limited analysis of the impact of diverse forecast types, resolutions, and prediction horizons; (iv) under-consideration of the multi-objective nature of most water resources systems. We then argue that future MPC applications in water resources systems should focus on addressing these four challenges as key priorities for future developments.
Anastasiya Shyrokaya, Florian Pappenberger, Ilias Pechlivanidis, Gabriele Messori, Sina Khatami, Maurizio Mazzoleni & Giuliano Di Baldassarre (2023) Advances and gaps in the science and practice of impact-based forecasting of droughts Water, e1698.
Abstract
Advances in impact modeling and numerical weather forecasting have allowed accurate drought monitoring and skilful forecasts that can drive decisions at the regional scale. State-of-the-art drought early-warning systems are currently based on statistical drought indicators, which do not account for dynamic regional vulnerabilities, and hence neglect the socio-economic impact for initiating actions. The transition from conventional physical forecasts of droughts toward impact-based forecasting (IbF) is a recent paradigm shift in early warning services, to ultimately bridge the gap between science and action. The demand to generate predictions of “what the weather will do” underpins the rising interest in drought IbF across all weather-sensitive sectors. Despite the large expected socio-economic benefits, migrating to this new paradigm presents myriad challenges. In this article, we provide a comprehensive overview of drought IbF, outlining the progress made in the field. Additionally, we present a road map highlighting current challenges and limitations in the science and practice of drought IbF and possible ways forward. We identify seven scientific and practical challenges/limitations: the contextual challenge (inadequate accounting for the spatio-sectoral dynamics of vulnerability and exposure), the human-water feedbacks challenge (neglecting how human activities influence the propagation of drought), the typology challenge (oversimplifying drought typology to meteorological), the model challenge (reliance on mainstream machine learning models), and the data challenge (mainly textual) with the linked sectoral and geographical limitations. Our vision is to facilitate the progress of drought IbF and its use in making informed and timely decisions on mitigation measures, thus minimizing the drought impacts globally.
Paolo Bonetti, Alberto Maria Metelli, Marcello Restelli (2023) Interpretable Linear Dimensionality Reduction based on Bias-Variance Analysis https://doi.org/10.48550/arXiv.2303.14734
Abstract
One of the central issues of several machine learning applications on real data is the choice of the input features. Ideally, the designer should select only the relevant, non-redundant features to preserve the complete information contained in the original dataset, with little collinearity among features and a smaller dimension. This procedure helps mitigate problems like overfitting and the curse of dimensionality, which arise when dealing with high-dimensional problems. On the other hand, it is not desirable to simply discard some features, since they may still contain information that can be exploited to improve results. Instead, dimensionality reduction techniques are designed to limit the number of features in a dataset by projecting them into a lower-dimensional space, possibly considering all the original features. However, the projected features resulting from the application of dimensionality reduction techniques are usually difficult to interpret. In this paper, we seek to design a principled dimensionality reduction approach that maintains the interpretability of the resulting features. Specifically, we propose a bias-variance analysis for linear models and we leverage these theoretical results to design an algorithm, Linear Correlated Features Aggregation (LinCFA), which aggregates groups of continuous features with their average if their correlation is “sufficiently large”. In this way, all features are considered, the dimensionality is reduced and the interpretability is preserved. Finally, we provide numerical validations of the proposed algorithm both on synthetic datasets to confirm the theoretical results and on real datasets to show some promising applications.
Barriopedro, D., García-Herrera, R., Ordóñez, C., Miralles, D. G., & Salcedo-Sanz, S. (2023). Heat waves: Physical understanding and scientific challenges. Reviews of Geophysics, 61, e2022RG000780.
Abstract
Heat waves (HWs) can cause large socioeconomic and environmental impacts. The observed increases in their frequency, intensity and duration are projected to continue with global warming. This review synthesizes the state of knowledge and scientific challenges. It discusses different aspects related to the definition, triggering mechanisms, observed changes and future projections of HWs, as well as emerging research lines on subseasonal forecasts and specific types of HWs. We also identify gaps that limit progress and delineate priorities for future research. Overall, the physical drivers of HWs are not well understood, partly due to difficulties in the quantification of their interactions and responses to climate change. Influential factors convey processes at different spatio-temporal scales, from global warming and the large-scale atmospheric circulation to regional and local factors in the affected area and upwind regions. Although some thermodynamic processes have been identified, there is a lack of understanding of dynamical aspects, regional forcings and feedbacks, and their future changes. This hampers the attribution of regional trends and individual events, and reduces the ability to provide accurate forecasts and regional projections. Sustained observational networks, models of diverse complexity, narrative-based methodological approaches and artificial intelligence offer new opportunities toward process-based understanding and interdisciplinary research.
Monique M Kuglitsch, Arif Albayrak, Jürg Luterbacher, Allison Craddock, Andrea Toreti, Jackie Ma, Paula Padrino Vilela, Elena Xoplaki, Rui Kotani, Dominique Berod, Jon Cox and Ivanka Pelivan (2023) Environ. Res. Lett. 18 093004
Abstract
Earth observations (EOs) have successfully been used to train artificial intelligence (AI)-based models in the field of disaster risk reduction (DRR) contributing to tools such as disaster early warning systems. Given the number of in situ and remote (e.g.radiosonde/satellite) monitoring devices, there is a common perception that there are no limits to the availability of EO for immediate use in such AI-based models. However, a mere fraction of EO is actually being used in this way. This topical review draws on use cases, workshop presentations, literature, and consultation with experts from key institutes to explore reasons for this discrepancy. Specifically, it is the main characteristics, possible challenges, and innovative solutions for EO. Finally, it suggests ways to make EO more user-ready and to facilitate its uptake in AI for DRR and beyond.
Rousi, E., Fink, A. H., Andersen, L. S., Becker, F. N., Beobide-Arsuaga, G., Breil, M., Cozzi, G., Heinke, J., Jach, L., Niermann, D., Petrovic, D., Richling, A., Riebold, J., Steidl, S., Suarez-Gutierrez, L., Tradowsky, J. S., Coumou, D., Düsterhus, A., Ellsäßer, F., Fragkoulidis, G., Gliksman, D., Handorf, D., Haustein, K., Kornhuber, K., Kunstmann, H., Pinto, J. G., Warrach-Sagi, K., and Xoplaki, E.: The extremely hot and dry 2018 summer in central and northern Europe from a multi-faceted weather and climate perspective, Nat. Hazards Earth Syst. Sci., 23, 1699–1718,
Abstract
The summer of 2018 was an extraordinary season in climatological terms for northern and central Europe, bringing simultaneous, widespread, and concurrent heat and drought extremes in large parts of the continent with extensive impacts on agriculture, forests, water supply, and the socio-economic sector. Here, we present a comprehensive, multi-faceted analysis of the 2018 extreme summer in terms of heat and drought in central and northern Europe, with a particular focus on Germany. The heatwave first affected Scandinavia in mid-July and shifted towards central Europe in late July, while Iberia was primarily affected in early August. The atmospheric circulation was characterized by strongly positive blocking anomalies over Europe, in combination with a positive summer North Atlantic Oscillation and a double jet stream configuration before the initiation of the heatwave. In terms of possible precursors common to previous European heatwaves, the Eurasian double-jet structure and a tripolar sea surface temperature anomaly over the North Atlantic were already identified in spring. While in the early stages over Scandinavia the air masses at mid and upper levels were often of a remote, maritime origin, at later stages over Iberia the air masses primarily had a local-to-regional origin. The drought affected Germany the most, starting with warmer than average conditions in spring, associated with enhanced latent heat release that initiated a severe depletion of soil moisture. During summer, a continued precipitation deficit exacerbated the problem, leading to hydrological and agricultural drought. A probabilistic attribution assessment of the heatwave in Germany showed that such events of prolonged heat have become more likely due to anthropogenic global warming. Regarding future projections, an extreme summer such as that of 2018 is expected to occur every 2 out of 3 years in Europe in a +1.5 ∘C warmer world and virtually every single year in a +2 ∘C warmer world. With such large-scale and impactful extreme events becoming more frequent and intense under anthropogenic climate change, comprehensive and multi-faceted studies like the one presented here quantify the multitude of their effects and provide valuable information as a basis for adaptation and mitigation strategies.
Enrico Scoccimarro, Oreste Cattaneo, Silvio Gualdi, Francesco Mattion, Alexandre Bizeul, Arnau Martin Risquez & Roberta Quadrelli (2023) Communications Earth & Environment
Abstract
Cooling degree days provide a simple indicator to represent how temperature drives energy demand for cooling. We investigate, at country level, the changes in cooling degree days worldwide in a recent twenty-one-year period starting in 2000. A new database, jointly generated by CMCC and IEA based on ERA5 reanalysis’ global gridded data, is used for the analysis. In contrast to the existent literature, the factors of population-weighting and humidity are considered, which affect the magnitude and the spatial distribution of these changes. Annual tendencies show a general increase of cooling degree days over the different countries, fostering more energy consumption for cooling demand, as confirmed by some regional studies. We also focus on the temporal clustering, to measure if peaks occur evenly random or tend to cluster in shorter periods. We stress that including humidity is important both for general tendencies and clustering. India, Cambodia, Thailand and Vietnam represent the emerging countries where this effect is stronger.
Pérez-Aracil, J., Camacho-Gómez, C., Lorente-Ramos, E., Marina, C. M., Cornejo-Bueno, L. M., & Salcedo-Sanz, S. (2023) New probabilistic, dynamic multi-method ensembles for optimization based on the CRO-SL Mathematics, 11(7)
Abstract
In this paper, new probabilistic and dynamic (adaptive) strategies for creating multi-method ensembles based on the coral reef optimization with substrate layers (CRO-SL) algorithm are proposed. CRO-SL is an evolutionary-based ensemble approach that is able to combine different search procedures for a single population. In this work, two different probabilistic strategies to improve the algorithm are analyzed. First, the probabilistic CRO-SL (PCRO-SL) is presented, which substitutes the substrates in the CRO-SL population with tags associated with each individual. Each tag represents a different operator which will modify the individual in the reproduction phase. In each generation of the algorithm, the tags are randomly assigned to the individuals with similar probabilities, obtaining this way an ensemble that sees more intense changes with the application of different operators to a given individual than CRO-SL. Second, the dynamic probabilistic CRO-SL (DPCRO-SL) is presented, in which the probability of tag assignment is modified during the evolution of the algorithm, depending on the quality of the solutions generated in each substrate. Thus, the best substrates in the search process will be assigned higher probabilities than those which showed worse performance during the search. The performances of the proposed probabilistic and dynamic ensembles were tested for different optimization problems, including benchmark functions and a real application of wind-turbine-layout optimization, comparing the results obtained with those of existing algorithms in the literature.
Yiheng Du, Ilaria Clemenzi and Ilias G Pechlivanidis (2023) Hydrological regimes explain the seasonal predictability of streamflow extremes. Environmental Research Letters 18 094060
Abstract
Advances in hydrological modeling and numerical weather forecasting have allowed hydro-climate services to provide accurate impact simulations and skillful forecasts that can drive decisions at the local scale. To enhance early warnings and long-term risk reduction actions, it is imperative to better understand the hydrological extremes and explore the drivers for their predictability. Here, we investigate the seasonal forecast skill of streamflow extremes over the pan-European domain, and further attribute the discrepancy in their predictability to the local river system memory as described by the hydrological regimes. Streamflow forecasts at about 35 400 basins, generated from the E-HYPE hydrological model driven with bias-adjusted ECMWF SEAS5 meteorological forcing input, are explored. Overall the results show adequate predictability for both hydrological extremes over Europe, despite the spatial variability in skill. The skill of high streamflow extreme deteriorates faster as a function of lead time than that of low extreme, with a positive skill persisting up to 12 and 20 weeks ahead for high and low extremes, respectively. A strong link between the predictability of extremes and the underlying local hydrological regime is identified through comparative analysis, indicating that systems of analogous river memory, e.g. fast or slow response to rainfall, can similarly predict the high and low streamflow extremes. The results improve our understanding of the geographical areas and periods, where the seasonal forecasts can timely provide information on very high and low streamflow conditions, including the drivers controlling their predictability. This consequently benefits regional and national organizations to embrace seasonal prediction systems and improve the capacity to act in order to reduce disaster risk and support climate adaptation.
2022
Sancho Salcedo-Sanz, Jorge Pérez-Aracil, Guido Ascenso, Javier Del Ser, David Casillas-Pérez, Christopher Kadow, Dušan Fister, David Barriopedro, Ricardo García-Herrera, Matteo Giuliani & Andrea Castelletti
Abstract
Atmospheric extreme events cause severe damage to human societies and ecosystems. The frequency and intensity of extremes and other associated events are continuously increasing due to climate change and global warming. The accurate prediction, characterization, and attribution of atmospheric extreme events is, therefore, a key research field in which many groups are currently working by applying different methodologies and computational tools. Machine learning and deep learning methods have arisen in the last years as powerful techniques to tackle many of the problems related to atmospheric extreme events. This paper reviews machine learning and deep learning approaches applied to the analysis, characterization, prediction, and attribution of the most important atmospheric extremes. A summary of the most used machine learning and deep learning techniques in this area, and a comprehensive critical review of literature related to ML in EEs, are provided. The critical literature review has been extended to extreme events related to rainfall and floods, heatwaves and extreme temperatures, droughts, severe weather events and fog, and low-visibility episodes. A case study focused on the analysis of extreme atmospheric temperature prediction with ML and DL techniques is also presented in the paper. Conclusions, perspectives, and outlooks on the field are finally drawn.
Sušan Fister, Jorge Pérez-Aracil, César Peláez-Rodríguez, Javier Del Ser, Sancho Salcedo-Sanz (2022) Journal of Advances in Modelling Earth Systems
Abstract
In this paper three customised Artificial Intelligence (AI) frameworks, considering Deep Learning (convolutional neural networks), Machine Learning algorithms and data reduction techniques are proposed, for a problem of long-term summer air temperature prediction. Specifically, the prediction of average air temperature in the first and second August fortnights, using input data from previous months, at two different locations, Paris (France) and Córdoba (Spain), is considered. The target variable, mainly in the first August fortnight, can contain signals of extreme events such as heatwaves, like the mega-heatwave of 2003, which affected France and the Iberian Peninsula. Thus, an accurate prediction of long-term air temperature may be valuable also for different problems related to climate change, such as attribution of extreme events, and in other problems related to renewable energy. The analysis carried out this work is based on Reanalysis data, which are first processed by a correlation analysis among different prediction variables and the target (average air temperature in August first and second fortnights). An area with the largest correlation is located, and the variables within, after a feature selection process, are the input of different deep learning and ML algorithms. The experiments carried out show a very good prediction skill in the three proposed AI frameworks, both in Paris and Córdoba regions.
C. Peláez-Rodríguez, J. Pérez-Aracil, D. Fister, L. Prieto-Godino, R.C. Deo, S. Salcedo-Sanz (2022) Renewable Energy
Abstract
A novel method for prediction of the extreme wind speed events based on a Hierarchical Classification/Regression (HCR) approach is proposed. The idea is to improve the prediction skills of different Machine Learning approaches on extreme wind speed events, while preserving the prediction performance for steady events. The proposed HCR architecture rests on three distinctive levels: first, a data preprocessing level, where training data are divided into clusters and accordingly associated labels. At this point, balancing techniques are applied to increase the significance of clusters with poorly represented wind gusts data. At a second level of the architecture, the classification of each sample into the corresponding cluster is carried out. Finally, once we have determined the cluster a sample belongs to, the third level carries out the prediction of the wind speed value, by using the regression model associated with that particular cluster. The performance of the proposed HCR approach has been tested in a real database of hourly wind speed values in Spain, considering Reanalysis data as predictive variables. The results obtained have shown excellent prediction skill in the forecasting of extreme events, achieving a 96% extremes detection, while maintaining a reasonable performance in the non-extreme samples. The performance of the methods has also been assessed using forecast data (GFS) as predictors.
Antara Dasgupta, Louise Arnal, Rebecca Emerton, Shaun Harrigan, Gwyneth Matthews, Ameer Muhammad, Karen O’Regan, Teresa Pérez-Ciria, Emixi Valdez, Bart van Osnabrugge, Micha Werner, Carlo Buontempo, Hannah Cloke, Florian Pappenberger, Ilias G. Pechlivanidis, Christel Prudhomme, Maria-Helena Ramos, Peter Salamon, (2022) Journal of flood risk management
Abstract
The unprecedented progress in ensemble hydro-meteorological modelling and forecasting on a range of temporal and spatial scales, raises a variety of new challenges which formed the theme of the Joint Virtual Workshop, ‘Connecting global to local hydrological modelling and forecasting: challenges and scientific advances’. Held from 29 June to 1 July 2021, this workshop was co-organised by the European Centre for Medium-Range Weather Forecasts (ECMWF), the Copernicus Emergency Management (CEMS) and Climate Change (C3S) Services, the Hydrological Ensemble Prediction EXperiment (HEPEX), and the Global Flood Partnership (GFP). This article aims to summarise the state-of-the-art presented at the workshop and provide an early career perspective. Recent advances in hydrological modelling and forecasting, reflections on the use of forecasts for decision-making across scales, and means to minimise new barriers to communication in the virtual format are also discussed. Thematic foci of the workshop included hydrological model development and skill assessment, uncertainty communication, forecasts for early action, co-production of services and incorporation of local knowledge, Earth observation, and data assimilation. Connecting hydrological services to societal needs and local decision-making through effective communication, capacity-building and co-production was identified as critical. Multidisciplinary collaborations emerged as crucial to effectively bring newly developed tools to practice.
Antonio Manuel Gómez-Orellana, David Guijo-Rubio, Jorge Pérez-Aracil, Pedro Antonio Gutiérrez, Sancho Salcedo-Sanz, César Hervás-Martínez, (2022) Atmospheric research
Abstract
In this paper we have tackled the problem of long-term air temperature prediction with eXplainable Artificial Intelligence (XAI) models. Specifically, we have evaluated the performance of an Artificial Neural Network (ANN) architecture with sigmoidal neurons in the hidden layer, trained by means of an evolutionary algorithm (Evolutionary ANNs, EANNs). This XAI model architecture (XAI-EANN) has been applied to the long-term air temperature prediction at different sub-regions of the South of the Iberian Peninsula. In this case, the average August air temperature has been predicted from ERA5 Reanalysis data variables, obtaining good predictions skills and explainable models in terms of the input climatological variables considered. A cluster analysis has been first carried out in terms of the average air temperature in the zone, in such a way that a number of sub-regions with different air temperature behaviour have been defined. The proposed XAI-EANN model architecture has been applied to each of the defined sub-regions, in order to find significant differences among them, which can be explained with the XAI-EANN models obtained. Finally, a comprehensive comparison against some state-of-the-art techniques has also been carried out, concluding that there are statistically significant differences in terms of accuracy in favour of the proposed XAI-EANN model, which also benefits from being an XAI model.
N. Hempelmann, C. Ehbrecht, E. Plesiat, G. Hobona, J. Simoes, D. Huard, T. J. Smith, U. S. McKnight, I. G. Pechlivanidis, and C. Alvarez-Castro (2022) Journal of ISPRS Journal of Photogrammetry and Remote Sensing
Abstract
Recent advances in modelling capabilities and data processing combined with vastly improved observation tools and networks have resulted in the expansion of available weather and climate information, from historical observations to seasonal climate forecasts, as well as decadal climate predictions and multi-decadal climate change projections. However, it remains a key challenge to ensure this information reaches the intended climate-sensitive sectors (e.g. water, energy, agriculture, health), and is fit-for-purpose to guarantee the usability of climate information for these downstream users. Climate information can be produced on demand via climate resilience information systems which are existing in various forms. To optimise the efficiency and establish better information exchange between these systems, standardisation is necessary. Here, standards and deployment options are described for how scientific methods can be be deployed in climate resilience information systems, respecting the principles of being findable, accessible, interoperable and reusable. Besides the general description of OGC-API Standards and OGC-API Processes based on existing building blocks, ongoing developments in AI-enhanced services for climate services are described.
D1.1 Project Management Plan
Extreme Events in Climate Resilience Information Systems architecture with AI
This is the first draft of the architecture of the prototype for Climate Resilience Information Systems, providing technical services to detect extreme events. The deployed application packages are running with AI components. The workflow on setting up a climate application package and deploying it into the architecture.
The entire architecture follows the FAIR principles to ensure interoperability with, for example, the COPERNICUS Climate Data Store (CDS) providing technical services for the Copernicus Climate Change Services (C3S). The sections are instructions, guidelines and backgrounds around OGC API based software which can be used to set up Climate Resilience Information Systems (CRIS).
The areas use the climate application package with the ‘inpainting process’, named Duck, as an example to guide through the different steps necessary to set up application packages for technical climate services.
The backend is deployed at the Climate Computing Center (DKRZ) running as a DEMO portal.