Fires impact a suite of radiative forcing agents but fire is one of the most challenging sources of emissions to model due to a large degree of stochasticism and a wide range of climatic and human influences that can both increase and decrease the occurrence of fires. Although many Earth system models now account for fires, there is still a need for a coherent and consistent community dataset to intercompare and constrain models. We developed a historic dataset combining satellite data over the past two decades with proxy data and fire models for use in CMIP6. Since then, new satellite data has indicated that global burned area may be much higher than previously thought and several regional datasets have shed light on the question whether fire emissions are now higher or not than in the pre-industrial era. We show how the latest insight and developments will be used to construct an updated fire emissions dataset for CMIP7, and show which fire categories carry the largest uncertainty, both for the past and into the future.

The integration of climate change and resilience considerations into the decision-making processes of National Road Administrations (NRAs) represents a delicate balancing act between ambition and pragmatism. A critical question is how to establish and execute a decision case for resilience through adaptation, finding equilibrium between service level requirements for the road network and the costs and benefits associated with enhancing resilience. The ICARUS project, funded by the Conference of European Directors of Roads (CEDR) emphasizes the importance of striking the right balance between service levels and costs, analogous to the mythologic figure Icarus flying neither too high nor too low. While European NRAs acknowledge the impact of climate change on their assets and operations, the full integration of adaptation strategies remains a formidable challenge. The ICARUS project aims to bridge this gap by advancing the state of the art in climate change resilience assessments, impact evaluation, cost-benefit assessments, and the implementation of nature-based solutions, while providing practical guidance on how to use these methods for building the decision case and use in the daily processes of road authorities.

Creating resilient infrastructure is of crucial importance to society, particularly in the face of escalating climate risks. As such, investments to increase resilience should be spent wisely to yield maximum benefits to transport users and society. However, decision-making regarding infrastructure dependencies is often based solely on qualitative information, which is inadequate for understanding the effectiveness of adaptation measures and the interconnectedness of infrastructure systems. This paper describes the Resilience Assessment and Adaptation for Critical infrastructurE (RA2CE) modeling framework. RA2CE aids infrastructure stakeholders in performing resilience assessments and to identify suitable adaptation options by generating comprehensive resilience maps for climate-induced events and ultimately pinpointing viable adaptation options. The need for systems-scale modeling, including interdependency analysis, demands fast, versatile, and generic models. We illustrate RA2CE's ability to quantify damages, indirect losses, and cascading impacts, supporting long-term adaptation planning on regional and national scales. This approach ensures that investments in resilient transport planning are allocated wisely, maximizing benefits for transport users and society.

There are already numerous examples of ML models predicting wildfire occurrence or susceptibility (e.g., Forkel et al., 2019, Cilli et al., 2022). The majority only makes a prediction without post-hoc interpretation of the model and/or quantification of the reliability of individual predictions. We want to take the ML model beyond its predicted values and learn about wildfire drivers from the model. Our main goal is to discover meaningful patterns in wildfire data that can be interpreted and understood to extract knowledge from the data. Our approach combines state-of-the-art methods for feature attribution, dimensionality reduction and clustering to identify the most representative decisions of the ML model leading to its predictions. We introduce a novel, multi-stage clustering methodology for subgroup discovery based on SHAP (SHapley Additive exPlanation) values, UMAP (Uniform Manifold Approximation and Projection) dimensionality reduction and hierarchical density-based clustering (HDBSCAN). With this approach, it is possible to identify a group of parameters that can be used to predict whether a wildfire is expected to occur or not. For this we build upon existing datasets of fire occurrences in both the Netherlands and Italy. Central to our methodology is the use of SHAP values to define subgroups (i.e., combinations of parameter values that can describe whether a fire occurs or not). As such, it eliminates noisy information from the dataset, however, preserving the aspects crucial to clustering and mitigating the effect of fluctuations in feature values that only make a small contribution to the model outcome. We enhance the clustering performance and interpretability of results by reducing multidimensional SHAP values to two dimensions before clustering through UMAP. We constructed decision-rules for each cluster that identify and differentiate the clusters, which results in highly discriminative and easily interpretable subgroup descriptions. This approach prevents large and overlapping rule sets, which often occurs when clustering is based on the raw feature space and requires manual filtering by experts. We validated our results and approach with a more conventional procedure that directly clusters in the feature space, skipping the ML model and SHAP values calculation. As a supplement to the decision rules, the model's effectiveness is assessed for each prediction in all subgroups. This differs from the conventional approach, which relies on performance metrics for an entire test set. Based on the two case studies, we conclude that supervised clustering effectively characterizes wildfire occurrence, attributing it to a set of influencing factors, both in the feature space and the spatial domain. Our approach also provides valuable insights into the performance of the ML model under diverse conditions, highlighting situations where predictions demand careful consideration. Forkel, M., Andela, N., Harrison, S. P., et al. (2019). Emergent relationships with respect to burned area in global satellite observations and fire-enabled vegetation models, Biogeosciences, 16, 57–76, https://doi.org/10.5194/bg-16-57-2019. Cilli, R., Elia, M., D’Este, M., et al. (2022). Explainable artificial intelligence (XAI) detects wildfire occurrence in the Mediterranean countries of Southern Europe. Sci Rep 12, 16349. https://doi.org/10.1038/s41598-022-20347-9.

Landscape fires are usually not associated with temperate Europe, yet not all temperate countries record statistics indicating that actual risks remain unknown. Here we introduce new wildfire statistics for The Netherlands, and summarize significant events and fatalities. The period 2017–2022 saw 611 wildfires and 405 ha burned per year, which Copernicus’ European Forest Fire Information System satellite data vastly underestimate. Fires burned more heathland than forest, were small (mean fire size 1.5 ha), were caused by people, and often burned simultaneously, in Spring and in Summer drought. Suppression, restoration and traffic delays cost 3 M€ year −1 . Dozens of significant events illustrate fire has never been away and has major societal impact amidst grave concerns for firefighter safety. Since 1833, 31 fatalities were reported. A legal framework is needed to ensure continuity of recordkeeping, as the core foundation of integrated fire management, to create a baseline for climate change, and to fulfill international reporting requirements.

Intergovernmental Panel on Climate Change (IPCC) assessments are the trusted source of scientific evidence for climate negotiations taking place under the United Nations Framework Convention on Climate Change (UNFCCC). Evidence-based decision-making needs to be informed by up-to-date and timely information on key indicators of the state of the climate system and of the human influence on the global climate system. However, successive IPCC reports are published at intervals of 5–10 years, creating potential for an information gap between report cycles. We follow methods as close as possible to those used in the IPCC Sixth Assessment Report (AR6) Working Group One (WGI) report. We compile monitoring datasets to produce estimates for key climate indicators related to forcing of the climate system: emissions of greenhouse gases and short-lived climate forcers, greenhouse gas concentrations, radiative forcing, the Earth's energy imbalance, surface temperature changes, warming attributed to human activities, the remaining carbon budget, and estimates of global temperature extremes. The purpose of this effort, grounded in an open-data, open-science approach, is to make annually updated reliable global climate indicators available in the public domain (https://doi.org/10.5281/zenodo.11388387, Smith et al., 2024a). As they are traceable to IPCC report methods, they can be trusted by all parties involved in UNFCCC negotiations and help convey wider understanding of the latest knowledge of the climate system and its direction of travel. The indicators show that, for the 2014–2023 decade average, observed warming was 1.19 [1.06 to 1.30] °C, of which 1.19 [1.0 to 1.4] °C was human-induced. For the single-year average, human-induced warming reached 1.31 [1.1 to 1.7] °C in 2023 relative to 1850–1900. The best estimate is below the 2023-observed warming record of 1.43 [1.32 to 1.53] °C, indicating a substantial contribution of internal variability in the 2023 record. Human-induced warming has been increasing at a rate that is unprecedented in the instrumental record, reaching 0.26 [0.2–0.4] °C per decade over 2014–2023. This high rate of warming is caused by a combination of net greenhouse gas emissions being at a persistent high of 53±5.4 Gt CO2e yr−1 over the last decade, as well as reductions in the strength of aerosol cooling. Despite this, there is evidence that the rate of increase in CO2 emissions over the last decade has slowed compared to the 2000s, and depending on societal choices, a continued series of these annual updates over the critical 2020s decade could track a change of direction for some of the indicators presented here.

Transportation plays a pivotal role in society in the accessibility of socio-economic functions, such as education and health services. At the same time these transport networks are put under pressure due to increasing demands and the often-increasing occurrence of climate-induced events. To increase resilience of the transportation network to disruptions, network criticality has been used to prioritise segments of the network for interventions. Here we present how equity principles can be applied in the context of decision making for resilient infrastructure. This is done for both a data-rich (The Hague, The Netherlands) and data-poor (Pontianak, Indonesia) environment. The results show that depending on the underlying equity principle different intervention locations are prioritized and changes the impact for different socio-economic groups and the general population.

The effects of natural hazards on road networks are evident and climate change will lead to more intense and more frequent impacts. This paper describes the approach and results of the resilience assessment and adaptation measure evaluation that has been conducted for the Dutch main highway network, by making use of the novel RA2CE modeling framework. The forthcoming results have been used to evaluate the desired level of resilience and to write an implementation agenda for adaptation to ensure a climate resilient highway network in 2050. The paper shows the necessity and challenges to link these resilience analyses to existing policy goal assessment frameworks as well as to asset management processes.

About half of the anthropogenic CO2 emissions remain in the atmosphere and half are taken up by the land and ocean1. If the carbon uptake by land and ocean sinks becomes less efficient, for example, owing to warming oceans2 or thawing permafrost3, a larger fraction of anthropogenic emissions will remain in the atmosphere, accelerating climate change. Changes in the efficiency of the carbon sinks can be estimated indirectly by analysing trends in the airborne fraction, that is, the ratio between the atmospheric growth rate and anthropogenic emissions of CO2 (refs. 4–10). However, current studies yield conflicting results about trends in the airborne fraction, with emissions related to land use and land cover change (LULCC) contributing the largest source of uncertainty7,11,12. Here we construct a LULCC emissions dataset using visibility data in key deforestation zones. These visibility observations are a proxy for fire emissions13,14, which are — in turn — related to LULCC15,16. Although indirect, this provides a long-term consistent dataset of LULCC emissions, showing that tropical deforestation emissions increased substantially (0.16 Pg C decade−1) since the start of CO2 concentration measurements in 1958. So far, these emissions were thought to be relatively stable, leading to an increasing airborne fraction4,5. Our results, however, indicate that the CO2 airborne fraction has decreased by 0.014 ± 0.010 decade−1 since 1959. This suggests that the combined land–ocean sink has been able to grow at least as fast as anthropogenic emissions. By generating a land use and land cover change emissions dataset using visibility data from two key deforestation regions, analysis of the data suggests a decrease in the CO2 airborne fraction since 1959.

Quantifying natural hazard impacts on critical infrastructure networks inherently involves uncertainties which makes decision-making complex. Here, we present an approach on how to account for uncertainties in the resilience assessment and in adaptation planning. These uncertainties stem from the hazard, exposure, vulnerability, and end-user data, as well as economic valuation. The consequences of natural hazards on critical infrastructure networks such as road transport networks has been proven to be evident, illustrated by recent flooding events in Western Europe. Due to climate change, many of these hazards may intensify and occur more frequently. Over the past years this has invoked progress in research that has led to an increased understanding of the effects of natural hazards on infrastructure networks. Currently, most analyses focus on the estimation of exposure, vulnerability, and the estimation of (annual expected) damages to the infrastructure assets and socio-economic losses for the users. This is subsequently used to identify hotspots for potential measures. The next step is to include adaptation in maintenance and construction planning. However, this step is often not linked to the assessment preceding the hotspot selection and because uncertainties in the assessment are not quantified, this results in decision making under (very deep) uncertainty. Here, we show the results for the Dutch highway network where we used the RA2CE - Resilience Assessment and Adaptation for Critical infrastructurE - platform, which makes use of hazard maps, user defined vulnerability curves and traffic information to produce resilience and risk maps for the infrastructure networks (resulting annual expected damages for the road operator and socio-economic losses for the road user), but also offers the possibility to perform cost-benefit analyses for proposed adaptation measures. Based on the cost-effectiveness analysis of potential measures, economically viable intervention strategies can be defined, including spatially explicit cost-benefit ratios to demonstrate economic performance of the different strategies. However, cost-benefit assessments should acknowledge the uncertain future related to climate change and socio-economic developments. Therefore, we progress the current state of the art by adding an uncertainty analysis, which takes into account all identified uncertainties in the model chain. This is based on Monte Carlo analyses providing insight in the sensitivity to all uncertainties in the process stemming from hazard, exposure, vulnerability and traffic data, as well as from the changes to the future related to climate change and socio-economic developments. The results provide an increased insight in the robustness of the strategies, instead of only one (best guess) prediction. It further allows the user and decision-maker not only to look at the expected change, but also at the high-impact, low-likelihood events. Based on validation with decision-makers future research has been identified to include black swans (unknown-unknown events) in decision-making, but also progressing on the user level, by for example including equity.

Germany, Belgium and the Netherlands were hit by extremeprecipitation and flooding in July 2021. This brief communication providesan overview of the impacts to large-scale critical infrastructure systemsand how recovery has progressed. The results show that Germany and Belgiumwere particularly affected, with many infrastructure assets severely damagedor completely destroyed. Impacts range from completely destroyed bridges andsewage systems, to severely damaged schools and hospitals. We find that(large-scale) risk assessments, often focused on larger (river) floodevents, do not find these local, but severe, impacts due to criticalinfrastructure failures. This may be the result of limited availability ofvalidation material. As such, this brief communication not only will help tobetter understand how critical infrastructure can be affected by flooding,but also can be used as validation material for future flood riskassessments.

Climate change is already being felt in Europe, unequivocally affecting the regions’ geo-structures. Concern over this is rising, as reflected in the increasing number of studies on the subject. However, the majority of these studies focused only on slopes and on a limited geographical scope. In this paper, we attempted to provide a broader picture of potential climate change impacts on the geo-structures in Europe by gathering the collective view of geo-engineers and geo-scientists in several countries, and by considering different geo-structure types. We also investigated how geo-structural concerns are being addressed in national adaptation plans. We found that specific provisions for geo-structural adaptation are generally lacking and mainly come in the form of strategies for specific problems. In this regard, two common strategies are hazard/risk assessment and monitoring, which are mainly implemented in relation to slope stability. We recommend that in future steps, other geo-structures are likewise given attention, particularly those assessed as also potentially significantly affected by climate change. Countries considered in this study are mainly the member countries of the European Large Geotechnical Institutes Platform (ELGIP).

Traditional flood risk studies often focus on direct economic impact, such as property damage or agricultural loss. However, the impact of floods is not limited to these direct damages. In fact societal costs and/or cascading effects are often much higher than the direct impact of floods. Cascading effects, such as access to healthcare and infrastructure accessibility are vital components for efficient emergency response management. This requires methodologies to quickly analyze the impact of large-scale floods on infrastructure networks. In this case study, the use of satellite-based flood maps are examined in combination with network criticality in the Mandalay region in central Myanmar. This region was severely affected by flooding after heavy monsoon rains in 2019. Many regions in the world are affected by this type of floods every year, resulting in large scale evacuations and limited access to health care. During these type of events, the transportation network is a crucial part for emergency response, as it is used for the delivery of goods, evacuation and deployment of emergency hospitals. The core of this study is a methodology to assess near real-time flood extents based on Sentinel-1 satellite imagery and the impact on network criticality. These tools were used to analyze the redundancy of the infrastructure network and quantify cascading impacts of flood hazards such as road accessibility and access to medical services. The methodology shows potential for operational use by linking with flood early warning systems (e.g. Delft-FEWS) enabling impact-based forecasting.

Spatially distributed anthropogenic and open burning emissions are fundamental data needed by Earth system models. We describe the methods used for generating gridded datasets produced for use by the modeling community, particularly for the Coupled Model Intercomparison Project Phase 6. The development of three sets of gridded data for historical open burning, historical anthropogenic, and future scenarios was coordinated to produce consistent data over 1750-2100. Historical data up to 2014 were provided with annual resolution and future scenario data in 10-year intervals. Emissions are provided on a sectoral basis, along with additional files for speciated non-methane volatile organic compounds (NMVOCs). An automated framework was developed to produce these datasets to ensure that they are reproducible and facilitate future improvements. We discuss the methodologies used to produce these data along with limitations and potential for future work.

Abstract. Spatially distributed anthropogenic and open burning emissions are fundamental data needed by Earth system models. We describe the methods used for generating gridded data sets produced for use by the modelling community, particularly for the Coupled Model Inter-comparison Project Phase 6. The development of three sets of gridded data for historical open burning, historical anthropogenic, and future scenarios were coordinated to produce consistent data over 1750–2100. Historical data up to 2014 were provided with annual resolution and future scenario data in 10-year intervals. Emissions are provided on a sectoral basis, along with additional files for speciated non-Methane Volatile Organic Compounds (NMVOCs). An automated framework was developed to produce these datasets to ensure that they are reproducible and facilitate future improvements. We discuss the methodologies used to produce these data along with limitations and potential for future work.

Recent climate changes have increased fire-prone weather conditions in many regions and have likely affected fire occurrence, which might impact ecosystem functioning, biogeochemical cycles, and society. Prediction of how fire impacts may change in the future is difficult because of the complexity of the controls on fire occurrence and burned area. Here we aim to assess how process-based firee-nabled dynamic global vegetation models (DGVMs) represent relationships between controlling factors and burned area. We developed a pattern-oriented model evaluation approach using the random forest (RF) algorithm to identify emergent relationships between climate, vegetation, and socio-economic predictor variables and burned area. We applied this approach to monthly burned area time series for the period from 2005 to 2011 from satellite observations and from DGVMs from the "Fire Modeling Intercomparison Project" (FireMIP) that were run using a common protocol and forcing data sets. The satellite-derived relationships indicate strong sensitivity to climate variables (e.g. maximum temperature, number of wet days), vegetation properties (e.g. vegetation type, previous-season plant productivity and leaf area, woody litter), and to socio-economic variables (e.g. human population density). DGVMs broadly reproduce the relationships with climate variables and, for some models, with population density. Interestingly, satellite-derived responses show a strong increase in burned area with an increase in previous-season leaf area index and plant productivity in most fire-prone ecosystems, which was largely underestimated by most DGVMs. Hence, our pattern-oriented model evaluation approach allowed us to diagnose that veg-etation effects on fire are a main deficiency regarding fireenabled dynamic global vegetation models' ability to accurately simulate the role of fire under global environmental change.

Spatially distributed anthropogenic and open burning emissions are fundamental data needed by Earth system models. We describe the methods used for generating gridded data sets produced for use by the modelling community, particularly for the Coupled Model Inter-comparison Project Phase 6. The 15 development of three sets of gridded data for historical open burning, historical anthropogenic, and future scenarios were coordinated to produce consistent data over 1750-2100. Historical data up to 2014 were provided with annual resolution and future scenario data in 10-year intervals. Emissions are provided on a sectoral basis, along with additional files for speciated non-Methane Volatile Organic Compounds (NMVOCs). An automated framework was developed to produce these datasets to ensure that they are 20 reproducible and facilitate future improvements. We discuss the methodologies used to produce these data along with limitations and potential for future work.