Explore the vast range of titles available.

Despite continuous improvements, precipitation forecasts are still not as accurate and reliable as those of other meteorological variables. A major contributing factor to this is that several key processes affecting precipitation distribution and intensity occur below the resolved scale of global weather models. Generative adversarial networks (GANs) have been demonstrated by the computer vision community to be successful at super‐resolution problems, that is, learning to add fine‐scale structure to coarse images. Leinonen et al. (2020, https://doi.org/10.1109/TGRS.2020.3032790) previously applied a GAN to produce ensembles of reconstructed high‐resolution atmospheric fields, given coarsened input data. In this paper, we demonstrate this approach can be extended to the more challenging problem of increasing the accuracy and resolution of comparatively low‐resolution input from a weather forecasting model, using high‐resolution radar measurements as a “ground truth.” The neural network must learn to add resolution and structure whilst accounting for non‐negligible forecast error. We show that GANs and VAE‐GANs can match the statistical properties of state‐of‐the‐art pointwise post‐processing methods whilst creating high‐resolution, spatially coherent precipitation maps. Our model compares favorably to the best existing downscaling methods in both pixel‐wise and pooled CRPS scores, power spectrum information and rank histograms (used to assess calibration). We test our models and show that they perform in a range of scenarios, including heavy rainfall. Plain Language Summary The processes that lead to precipitation (rainfall) happen on a very small scale. Weather forecast computer models work on much larger scales, so rainfall is often poorly predicted. In this paper, we develop a method that enhances the resolution of rainfall forecasts by a factor of 10, and makes the forecasts more accurate. We generate many samples of what the rainfall pattern could be, which gives an idea of the uncertainty in the forecast. Our method is based on machine learning and neural networks, which means that we use many past examples of weather forecasts, together with the rainfall that actually happened, and our method “automatically” learns how the forecasts can be improved. We use an existing idea called “Generative Adversarial Networks,” which has been used very successfully in image‐related tasks, such as producing realistic higher‐resolution images from low‐resolution ones. Our task is similar to producing a high‐resolution image from a low‐resolution one, hence this approach is promising. Our method outperforms a variety of existing approaches, and even produces good predictions for the most extreme rainfall situations in our data set. These are the scenarios that cause the most real‐world disruption, the most useful events to produce good forecasts for. Key Points We use generative adversarial neural networks to post‐process global weather forecast model output over the UK We produce more realistic precipitation forecasts than the input forecast data, at 10X resolution, with excellent statistical properties We match or outperform a state‐of‐the‐art pointwise downscaling scheme, while also producing spatially coherent images

In the atmosphere, microphysics refers to the microscale processes that affect cloud and precipitation particles and is a key linkage among the various components of Earth's atmospheric water and energy cycles. The representation of microphysical processes in models continues to pose a major challenge leading to uncertainty in numerical weather forecasts and climate simulations. In this paper, the problem of treating microphysics in models is divided into two parts: (i) how to represent the population of cloud and precipitation particles, given the impossibility of simulating all particles individually within a cloud, and (ii) uncertainties in the microphysical process rates owing to fundamental gaps in knowledge of cloud physics. The recently developed Lagrangian particle‐based method is advocated as a way to address several conceptual and practical challenges of representing particle populations using traditional bulk and bin microphysics parameterization schemes. For addressing critical gaps in cloud physics knowledge, sustained investment for observational advances from laboratory experiments, new probe development, and next‐generation instruments in space is needed. Greater emphasis on laboratory work, which has apparently declined over the past several decades relative to other areas of cloud physics research, is argued to be an essential ingredient for improving process‐level understanding. More systematic use of natural cloud and precipitation observations to constrain microphysics schemes is also advocated. Because it is generally difficult to quantify individual microphysical process rates from these observations directly, this presents an inverse problem that can be viewed from the standpoint of Bayesian statistics. Following this idea, a probabilistic framework is proposed that combines elements from statistical and physical modeling. Besides providing rigorous constraint of schemes, there is an added benefit of quantifying uncertainty systematically. Finally, a broader hierarchical approach is proposed to accelerate improvements in microphysics schemes, leveraging the advances described in this paper related to process modeling (using Lagrangian particle‐based schemes), laboratory experimentation, cloud and precipitation observations, and statistical methods. Plain Language Summary In the atmosphere, microphysics—the small‐scale processes affecting cloud and precipitation particles such as their growth by condensation, evaporation, and melting—is a critical part of Earth's weather and climate. Because it is impossible to simulate every cloud particle individually owing to their sheer number within even a small cloud, atmospheric models have to represent the evolution of particle populations statistically. There are critical gaps in knowledge of the microphysical processes that act on particles, especially for atmospheric ice particles because of their wide variety and intricacy of their shapes. The difficulty of representing cloud and precipitation particle populations and knowledge gaps in cloud processes both introduce important uncertainties into models that translate into uncertainty in weather forecasts and climate simulations, including climate change assessments. We discuss several specific challenges related to these problems. To improve how cloud and precipitation particle populations are represented, we advocate a “particle‐based” approach that addresses several limitations of traditional approaches and has recently gained traction as a tool for cloud modeling. Advances in observations, including laboratory studies, are argued to be essential for addressing gaps in knowledge of microphysical processes. We also advocate using statistical modeling tools to improve how these observations are used to constrain model microphysics. Finally, we discuss a hierarchical approach that combines the various pieces discussed in this article, providing a possible blueprint for accelerating progress in how microphysics is represented in cloud, weather, and climate models. Key Points Microphysics is an important component of weather and climate models, but its representation in current models is highly uncertain Two critical challenges are identified: representing cloud and precipitation particle populations and knowledge gaps in cloud physics A possible blueprint for addressing these challenges is proposed to accelerate progress in improving microphysics schemes

Adverse health outcomes caused by extreme heat represent the most direct human health threat associated with the warming of the Earth's climate. Socioeconomic, demographic, health, land cover, and temperature determinants contribute to heat vulnerability; however, nationwide patterns of residential and race/ethnicity disparities in heat vulnerability in the United States are poorly understood. This study aimed to develop a Heat Vulnerability Index (HVI) for the United States; to assess differences in heat vulnerability across geographies that have experienced historical and/or contemporary forms of marginalization; and to quantify HVI by race/ethnicity. Principal component analysis was used to calculate census tract level HVI scores based on the 2019 population characteristics of the United States. Differences in HVI scores were analyzed across the Home Owners' Loan Corporation (HOLC) “redlining” grades, the Climate and Economic Justice Screening Tool (CEJST) disadvantaged versus non‐disadvantaged communities, and race/ethnicity groups. HVI scores were calculated for 55,267 U.S. census tracts. Mean HVI scores were 17.56, 18.61, 19.45, and 19.93 for HOLC grades “A”–“D,” respectively. CEJST‐defined disadvantaged census tracts had a significantly higher mean HVI score (19.13) than non‐disadvantaged tracts (16.68). The non‐Hispanic African American or Black race/ethnicity group had the highest HVI score (18.51), followed by Hispanic or Latino (18.19). Historically redlined and contemporary CEJST disadvantaged census tracts and communities of color were found to be associated with increased vulnerability to heat. These findings can help promote equitable climate change adaptation policies by informing policymakers about the national distribution of place‐ and race/ethnicity‐based disparities in heat vulnerability. Plain Language Summary As the Earth's climate warms, extreme heat is the most direct threat to human health. Due to various socioeconomic, demographic, health, and environmental factors, some individuals and populations are more vulnerable to adverse health events caused by extreme heat. Publicly available data were obtained for each of these factors, and statistical analysis yielded a quantitative measure of heat vulnerability for 55,267 U.S. census tracts. Of these census tracts, those that have experienced historical and/or contemporary forms of marginalization were associated with increased vulnerability to heat. Additionally, non‐White race/ethnicity groups were associated with increased vulnerability to heat and were overrepresented in the census tracts with the highest vulnerability. These results can inform policymakers of the places and race/ethnicity groups most vulnerable to heat, and can therefore be used to develop equity‐promoting climate change adaptation policies. Key Points Historically “redlined” and contemporary Climate and Economic Justice Screening Tool disadvantaged communities were found to be associated with increased vulnerability to heat Communities of color were associated with increased vulnerability to heat and were overrepresented in the most vulnerable census tracts Identifying place and race/ethnicity‐based disparities in heat vulnerability can help promote equitable climate change adaptation policies

Saltwater intrusion is a critical concern for coastal communities due to its impacts on fresh ecosystems and civil infrastructure. Declining recharge and rising sea level are the two dominant drivers of saltwater intrusion along the land‐ocean continuum, but there are currently no global estimates of future saltwater intrusion that synthesize these two spatially variable processes. Here, for the first time, we provide a novel assessment of global saltwater intrusion risk by integrating future recharge and sea level rise while considering the unique geology and topography of coastal regions. We show that nearly 77% of global coastal areas below 60° north will undergo saltwater intrusion by 2100, with different dominant drivers. Climate‐driven changes in subsurface water replenishment (recharge) is responsible for the high‐magnitude cases of saltwater intrusion, whereas sea level rise and coastline migration are responsible for the global pervasiveness of saltwater intrusion and have a greater effect on low‐lying areas. Plain Language Summary Coastal watersheds around the globe are facing perilous changes to their freshwater systems. Driven by climatic changes in recharge and sea level working in tandem, sea water encroaches into coastal groundwater aquifers and consequently salinizes fresh groundwater, in a process called saltwater intrusion. To assess the vulnerability of coastal watersheds to future saltwater intrusion, we applied projections of sea level and groundwater recharge to a global analytical modeling framework. Nearly 77% of the global coast is expected to undergo measurable salinization by the year 2100. Changes in recharge have a greater effect on the magnitude of salinization, whereas sea level rise drives the widespread extensiveness of salinization around the global coast. Our results highlight the variable pressures of climate change on coastal regions and have implications for prioritizing management solutions. Key Points First global analysis of future saltwater intrusion vulnerability responding to spatially variable recharge and sea level rise is provided Recharge drives the extreme cases of saltwater intrusion, while sea level rise is responsible for its global pervasiveness Nearly 77% of global coastal areas below 60° north will undergo saltwater intrusion by 2100

Losses from natural hazards are escalating dramatically, with more properties and critical infrastructure affected each year. Although the magnitude, intensity, and/or frequency of certain hazards has increased, development contributes to this unsustainable trend, as disasters emerge when natural disturbances meet vulnerable assets and populations. To diagnose development patterns leading to increased exposure in the conterminous United States (CONUS), we identified earthquake, flood, hurricane, tornado, and wildfire hazard hotspots, and overlaid them with land use information from the Historical Settlement Data Compilation data set. Our results show that 57% of structures (homes, schools, hospitals, office buildings, etc.) are located in hazard hotspots, which represent only a third of CONUS area, and ∼1.5 million buildings lie in hotspots for two or more hazards. These critical levels of exposure are the legacy of decades of sustained growth and point to our inability, lack of knowledge, or unwillingness to limit development in hazardous zones. Development in these areas is still growing more rapidly than the baseline rates for the nation, portending larger future losses even if the effects of climate change are not considered. Key Points More than half of the structures in the conterminous United States are exposed to potentially devastating natural hazards Growth rates in hazard hotspots exceed the national trend Risk assessments can be improved by considering multiple hazards, mitigation history and fine‐scale data on the built environment

Subseasonal rainfall forecast skill is critical to support preparedness for hydrometeorological extremes. We assess how a process‐informed evaluation, which subsamples forecasting model members based on their ability to represent potential predictors of rainfall, can improve monthly rainfall forecasts within Central America in the following month, using Costa Rica and Guatemala as test cases. We generate a constrained ensemble mean by subsampling 130 members from five dynamic forecasting models in the C3S multimodel ensemble based on their representation of both (a) zonal wind direction and (b) Pacific and Atlantic sea surface temperatures (SSTs), at the time of initialization. Our results show in multiple months and locations increased mean squared error skill by 0.4 and improved detection rates of rainfall extremes. This method is transferrable to other regions driven by slowly‐changing processes. Process‐informed subsampling is successful because it identifies members that fail to represent the entire rainfall distribution when wind/SST error increases. Plain Language Summary Subseasonal rainfall forecasts provide alerts multiple weeks ahead. These forecasts present an opportunity to facilitate anticipatory actions yet are often unreliable to use when preparing for extreme weather. We develop a method to optimize rainfall forecasts by selecting individual members from a large ensemble of dynamic forecasting model outputs based on their ability to represent potential predictors of rainfall. We test our method on monthly rainfall forecasts within Central America in the following month, using Costa Rica and Guatemala as key test cases. We select members from five contributing models of the C3S multimodel ensemble using regional predictors, including wind direction and sea surface temperatures (SSTs). Our results show improvements in the detection of low and high rainfall extremes. This method is transferrable to other regions driven by slowly‐changing processes like SSTs and is beneficial for operational forecasters who can leverage regional expertise of relevant rainfall‐generating processes to subsample better performing ensemble members for their regions. Key Points Subsampling members using sea surface temperatures and zonal wind improves subseasonal ensemble rainfall forecasts in Central America In multiple months and locations mean squared error skill increases by 0.4 and extreme rainfall skill improves by 0.5 (Heidke skill) Process‐informed subsampling is useful because the models' representation of rainfall degrades as process error increases

Numerical weather prediction models require ever‐growing computing time and resources but, still, have sometimes difficulties with predicting weather extremes. We introduce a data‐driven framework that is based on analog forecasting (prediction using past similar patterns) and employs a novel deep learning pattern‐recognition technique (capsule neural networks, CapsNets) and an impact‐based autolabeling strategy. Using data from a large‐ensemble fully coupled Earth system model, CapsNets are trained on midtropospheric large‐scale circulation patterns (Z500) labeled 0–4 depending on the existence and geographical region of surface temperature extremes over North America several days ahead. The trained networks predict the occurrence/region of cold or heat waves, only using Z500, with accuracies (recalls) of 69–45% (77–48%) or 62–41% (73–47%) 1–5 days ahead. Using both surface temperature and Z500, accuracies (recalls) with CapsNets increase to ∼80% (88%). In both cases, CapsNets outperform simpler techniques such as convolutional neural networks and logistic regression, and their accuracy is least affected as the size of the training set is reduced. The results show the promises of multivariate data‐driven frameworks for accurate and fast extreme weather predictions, which can potentially augment numerical weather prediction efforts in providing early warnings. Key Points A data‐driven extreme weather prediction framework based on analog forecasting and deep learning pattern‐recognition methods is proposed Extreme surface temperature events over North America are skillfully predicted using only midtropospheric large‐scale circulation patterns More advanced deep learning methods are found to yield better forecasts, encouraging novel methods tailored for climate/weather data

Global warming is shifting the thermal dynamics of lakes, with resulting climatic variability heavily affecting their mixing dynamics. We present a dual ensemble workflow coupling climate models with lake models. We used a large set of simulations across multiple domains, multi‐scenario, and multi GCM‐ RCM combinations from CORDEX data. We forced a set of multiple hydrodynamic lake models by these multiple climate simulations to explore climate change impacts on lakes. We also quantified the contributions from the different models to the overall uncertainty. We employed this workflow to investigate the effects of climate change on Lake Sevan (Armenia). We predicted for the end of the 21st century, under RCP 8.5, a sharp increase in surface temperature (4.3±0.7K)$(4.3\\pm 0.7\\,\\mathrm{K})$and substantial bottom warming (1.7±0.7K)$(1.7\\pm 0.7\\,\\mathrm{K})$ , longer stratification periods (+55 days) and disappearance of ice cover leading to a shift in mixing regime. Increased insufficient cooling during warmer winters points to the vulnerability of Lake Sevan to climate change. Our workflow leverages the strengths of multiple models at several levels of the model chain to provide a more robust projection and at the same time a better uncertainty estimate that accounts for the contributions of the different model levels to overall uncertainty. Although for specific variables, for example, summer bottom temperature, single lake models may perform better, the full ensemble provides a robust estimate of thermal dynamics that has a high transferability so that our workflow can be a blueprint for climate impact studies in other systems. Plain Language Summary Lakes are threatened by climate change because of effects related to the increasing temperature, long stratification, and ice disappearance. One of the best tools to predict these effects on lakes is numerical modeling of lakes that benefit from climate modeling. Climate modeling is normally done globally or in the so‐called general circulation model (GCM) or more detailed simulations on regional levels (RCM) like the CORDEX data set. In this study, we used the CORDEX data, which employed several climate models from several regions (domains) for several climatic scenarios (emissions scenarios) to force multiple lake models. This approach gave us an extensive prediction about various possible outputs. We applied this approach to Lake Sevan (Armenia), a large mountain lake. Our study predicted for the worst‐case scenario, an increase of the surface temperature by almost 4.3 K by the end of the 21st century, 1.75 K for bottom temperature, a total disappearance of ice cover, and about 55 extra days of stratification, showing its vulnerability for climate change. This optimized workflow uses the strength of a wide variety of models on the climate and lake levels to better understand the impact of climate change and quantify the sources of uncertainty in the workflow. Key Points Dual multi‐model ensemble of climate data and lake models is used for robust projections of climate change impacts Variance decomposition effectively identified the sources of uncertainty and contributions of different models to the overall uncertainty Significant warming, longer stratification periods, and loss of ice cover are predicted for Lake Sevan by the end of the 21st century

Western North America experienced an unprecedented extreme heat event (EHE) in 2021, characterized by high temperatures and reduced air quality. There were approximately 740 excess deaths during the EHE in the province of British Columbia, making it one of the deadliest weather events in Canadian history. It is important to understand who is at risk of death during EHEs so that appropriate public health interventions can be developed. This study compares 1,614 deaths from 25 June to 02 July 2021 with 6,524 deaths on the same dates from 2012 to 2020 to examine differences in the prevalence of 26 chronic diseases between the two groups. Conditional logistic regression was used to estimate the odds ratio (OR) for each chronic disease, adjusted for age, sex, and all other diseases, and conditioned on geographic area. The OR [95% confidence interval] for schizophrenia among all EHE deaths was 3.07 [2.39, 3.94], and was larger than the ORs for other conditions. Chronic kidney disease and ischemic heart disease were also significantly increased among all EHE deaths, with ORs of 1.36 [1.18, 1.56] and 1.18 [1.00, 1.38], respectively. Chronic diseases associated with EHE mortality were somewhat different for deaths attributed to extreme heat, deaths with an unknown/pending cause, and non‐heat‐related deaths. Schizophrenia was the only condition associated with significantly increased odds of EHE mortality in all three subgroups. These results confirm the role of mental illness in EHE risk and provide further impetus for interventions that target specific groups of high‐risk individuals based on underlying chronic conditions. Plain Language Summary Western North America experienced the most severe extreme heat event (EHE) ever recorded in the region during the summer of 2021. There were approximately 740 more deaths than usual in British Columbia, Canada during the EHE, which made it one of the deadliest weather events in Canadian history. This study compares people who died during the EHE with people who died at the same time of year in other years to identify differences between the two groups with respect to 26 chronic diseases. We found that people with schizophrenia were at much higher risk than others during the EHE. People with chronic kidney disease and ischemic heart disease were also at increased risk. This information will be used to help develop programs that support people at higher risk during future EHEs. Key Points British Columbia experienced an unprecedented extreme heat event (EHE) in summer 2021 associated with a 95% increase in population mortality Deaths during the EHE and previous years were compared with respect to chronic diseases present at time of death Schizophrenia was most strongly associated with higher risk of death during the EHE

A new climate model, HadGEM3 N96ORCA1, is presented that is part of the GC3.1 configuration of HadGEM3. N96ORCA1 has a horizontal resolution of ~135 km in the atmosphere and 1° in the ocean and requires an order of magnitude less computing power than its medium‐resolution counterpart, N216ORCA025, while retaining a high degree of performance traceability. Scientific performance is compared to both observations and the N216ORCA025 model. N96ORCA1 reproduces observed climate mean and variability almost as well as N216ORCA025. Patterns of biases are similar across the two models. In the northwest Atlantic, N96ORCA1 shows a cold surface bias of up to 6 K, typical of ocean models of this resolution. The strength of the Atlantic meridional overturning circulation (16 to 17 Sv) matches observations. In the Southern Ocean, a warm surface bias (up to 2 K) is smaller than in N216ORCA025 and linked to improved ocean circulation. Model El Niño/Southern Oscillation and Atlantic Multidecadal Variability are close to observations. Both the cold bias in the Northern Hemisphere (N96ORCA1) and the warm bias in the Southern Hemisphere (N216ORCA025) develop in the first few decades of the simulations. As in many comparable climate models, simulated interhemispheric gradients of top‐of‐atmosphere radiation are larger than observations suggest, with contributions from both hemispheres. HadGEM3 GC3.1 N96ORCA1 constitutes the physical core of the UK Earth System Model (UKESM1) and will be used extensively in the Coupled Model Intercomparison Project 6 (CMIP6), both as part of the UK Earth System Model and as a stand‐alone coupled climate model. Plain Language Summary In this article, a new version of the climate model currently used in the United Kingdom (HadGEM3) is presented and analyzed. The circulation of the atmosphere and the oceans is simulated on a relatively coarse spatial grid with a grid cell size of about 120 km. The advantage of using a coarse spatial grid is that less computing power (on a supercomputer) is needed compared to using a finer grid. This gives an opportunity to do many more simulations of the ways in which Earth's climate may evolve in the decades and centuries ahead. We have carefully compared a simulation of the climate around the year 2000 with climate observations from that time and with a simulation from the same model with a finer spatial grid. Our results show that our new, coarse‐grid version is representing the current climate reasonably well, for instance, with regards to climate variability in the tropics and major ocean currents. However, there are clear differences between the two models. In the coarse‐grid model, the ocean surface is too cold in the northwest Atlantic, while in the fine‐grid version it is too warm in the Southern Ocean around Antarctica. We look into explanations for these inaccuracies. Key Points A low‐resolution, traceable version of the current Met Office Hadley Centre climate model HadGEM3 GC3.1 is presented The scientific performance is comparable to the medium‐resolution version, while requiring much less computational resources In the low‐resolution version the Southern Ocean warm bias is reduced, linked with a more realistic ocean circulation