Search

A comprehensive understanding of human-induced changes to rainfall is essential for water resource management and infrastructure design. However, at regional scales, existing detection and attribution studies are rarely able to conclusively identify human influence on precipitation. Here we show that anthropogenic aerosol and greenhouse gas (GHG) emissions are the primary drivers of precipitation change over the United States. GHG emissions increase mean and extreme precipitation from rain gauge measurements across all seasons, while the decadal-scale effect of global aerosol emissions decreases precipitation. Local aerosol emissions further offset GHG increases in the winter and spring but enhance rainfall during the summer and fall. Our results show that the conflicting literature on historical precipitation trends can be explained by offsetting aerosol and greenhouse gas signals. At the scale of the United States, individual climate models reproduce observed changes but cannot confidently determine whether a given anthropogenic agent has increased or decreased rainfall.

The catastrophic December 2022-January 2023 nine atmospheric rivers in California underscore the urgent need to better understand such high-risk weather extremes. Here we applied a machine learning clustering tool to understand the activity of atmospheric river clusters. Reanalysis results show that clusters with high density, that is the time fraction under atmospheric river conditions within a cluster, exhibit more frequent high-category atmospheric rivers, alongside an increased likelihood for extreme precipitation and severe land surface responses. The key circulation patterns of atmospheric river clusters are primarily attributed to subseasonal variability. Furthermore, the occurrence and density of atmospheric river clusters are modulated by the daily variability of the geopotential height field. Climate model projections suggest that atmospheric river clusters with higher density and higher categories will be more frequent as warming level increases. Our findings emphasize the important role of atmospheric river clusters in the development of climate adaptation and resilience strategies.

Atmospheric rivers (ARs) are extreme weather events that can alleviate drought or cause billions of US dollars in flood damage. By transporting significant amounts of latent energy towards the poles, they are crucial to maintaining the climate system's energy balance. Since there is no first-principle definition of an AR grounded in geophysical fluid mechanics, AR identification is currently performed by a multitude of expert-defined, threshold-based algorithms. The variety of AR detection algorithms has introduced uncertainty into the study of ARs, and the thresholds of the algorithms may not generalize to new climate datasets and resolutions. We train convolutional neural networks (CNNs) to detect ARs while representing this uncertainty; we name these models ARCNNs. To detect ARs without requiring new labeled data and labor-intensive AR detection campaigns, we present a semi-supervised learning framework based on image style transfer. This framework generalizes ARCNNs across climate datasets and input fields. Using idealized and realistic numerical models, together with observations, we assess the performance of the ARCNNs. We test the ARCNNs in an idealized simulation of a shallow-water fluid in which nearly all the tracer transport can be attributed to AR-like filamentary structures. In reanalysis and a high-resolution climate model, we use ARCNNs to calculate the contribution of ARs to meridional latent heat transport, and we demonstrate that this quantity varies considerably due to AR detection uncertainty.

BackgroundThe eHealth Literacy Scale (eHEALS) has been widely adopted by researchers to understand how eHealth literacy can be put into context. eHealth researchers need to know how to promote positive health behavior changes across college students, given the importance of the Internet to acquire and use health information. The American College Health Association identified a set of key health issues that affect college students today. By understanding how eHEALS might be related to college students’ maintenance of their health and their use of online health resources, researchers will be provided with a better understanding of eHealth literacy and its pragmatic implications for health campaigns and future interventions. ObjectiveThe goal of the study was to examine what eHEALS reveals about college student health behaviors identified by the American College Health Association. To understand college student current health maintenance and their intentions to maintain their health and use online resources, the theory of planned behavior was used as the theoretical framework for the study. MethodsData were collected via a survey of 422 college students that included the eHEALS measure and questions about health issues based on the recommendations of the American College Health Association. These questions asked about college student current health, subsequent use of online health resources, and their intention to maintain their health and make use of such resources in the future. ResultseHEALS was positively and significantly associated with all 8 areas of health issues identified by the American College Health Association for college student current maintenance of health and use of online health resources and for future intention of health maintenance and use of online resources. Key issues that emerged with eHealth literacy were maintaining safe sex practices and seeking out related information, seeking out information on an exercise regime, information on vaccinations, and maintaining a balanced diet. ConclusionsThese results suggest several areas that may be targeted for future health campaigns toward college students. In addition, eHEALS was found to be a useful instrument for college students in the United States. Lastly, these results point to a need to deliver targeted information to college students, particularly since eHEALS captures literacy based on positively phrased items.

The deadly and destructive nature of tropical cyclones (TCs) makes understanding their response to future climate change of the utmost importance. TC genesis hinges on multiple factors, including an initial disturbance. African easterly waves (AEWs) have been shown to serve as such disturbances for TCs developing in the North Atlantic. It is therefore crucial to understand the relationship between AEWs and TCs and how this relationship may be affected by climate change. In this study, we examine the AEW-TC relationship in historical and future climates using three models from the HighResMIP PRIMAVERA simulations. The AEWs and TCs were tracked in the model data using objective tracking algorithms, and AEW and TC tracks were then matched together if they were close to each other in space and time. The strength of the AEWs was measured using the eddy kinetic energy and the curvature vorticity of the waves. TC strength and intensity were measured using potential intensity and lifetime maximum 10 m windspeed. We found that future changes in the frequency of AEWs are not a good indicator of future TC activity. However, AEW strength, as well as environmental conditions conducive to strong TCs, are good indicators of AEWs that develop into TCs in both historical and future climates.

Despite the emerging influence of anthropogenic climate change on the global water cycle, at regional scales the combination of observational uncertainty, large internal variability, and modeling uncertainty undermine robust statements regarding the human influence on precipitation. Here, we use output from global climate models in a perfect-data sense to develop a framework for conducting regional detection and attribution (D&A) for precipitation, starting with the contiguous United States (CONUS) where observational uncertainty is lower than in other regions. Our unified approach can simultaneously detect systematic trends in mean and extreme precipitation, attribute trends to anthropogenic forcings, compute the effects of forcings as a function of time, and map the effects of individual forcings. Model output is used to conduct a set of tests that yield a parsimonious representation for characterizing seasonal precipitation over the CONUS for the historical record (1900 to present day), which ensures our D&A is insensitive to structural uncertainty. Our framework is developed using synthetic data in a Pearl-causal perspective wherein causality can be identified using intervention-based simulations. While the hypothesis-based framework and accompanying generalized D&A formula we develop should be widely applicable, we include a strong caution that the hypothesis-guided simplification of the formula for the historical climatic record of CONUS as described in this paper will likely fail to hold in other geographic regions and under future warming.

Historical trends in monsoon low pressure systems (LPS), the dominant rain-bearing weather system of South Asia, have been difficult to assess due to changes in the observing network. Future projections have also remained uncertain because prior studies concluded that many coarse-resolution climate models do not accurately simulate LPS. Here, we examine changes in South Asian monsoon LPS simulated by an ensemble of global models, including some with high spatial resolution, that we show skillfully represent LPS. In the ensemble mean, the number of strong LPS (monsoon depressions) decreased over the last 65 years (1950–2014) by about 15% while no trend was detected for weaker LPS (monsoon lows). The reduction in depression counts then moderated, yielding no trend in the periods 1980–2050 or 2015–2050. The ensemble mean projects a shift in genesis from ocean to land and an increase in LPS precipitation of at least 7% K−1, which together contribute to a projected increase in seasonal mean and extreme precipitation over central India.

Atmospheric rivers (ARs) are intensive poleward moisture transport events that are essential to the global hydrological cycle and are often linked to extreme weather events. We categorize the winter North Pacific ARs into two “flavors”: wind-dominated (windy ARs) and moisture-dominated (wet ARs) using 40 years of hourly data from fifth generation of the European Centre for Medium-Range Weather Forecasts Interim Reanalysis. We compare the differences between windy ARs and wet ARs including the lifecycle characteristics (such as genesis locations and changes of meteorological elements through the lifecycle), overall AR frequency, landfall impacts, and variability. The windy ARs are more likely to occur in the midlatitudes, while wet ARs are more active in the subtropics. Windy ARs are associated with intensive surface pressure lows, where the strong pressure gradient can support the strong wind within ARs. Due to larger size and longer lifetime, wet ARs are more likely to produce more precipitation over a lifecycle. By scaling the landfalling ARs, we show that wet ARs dominate the high-category ARs (Category 4 and 5) with higher spatial frequency and more precipitation, and windy ARs have higher contributions in the lower AR categories especially over British Columbia. Windy ARs are modulated by El Niño Southern Oscillation (ENSO) teleconnections via the anomalous geopotential height and extended subtropical jet. Wet ARs are affected by the anomalous sea surface temperature over the midlatitudes related to ENSO. Sensitivity analysis with an alternate AR detection algorithm shows consistent results on AR flavors but with disagreement on the amplitude.

Societies and ecosystems within and downstream of mountains rely on seasonal snowmelt to satisfy their water demands. Anthropogenic climate change has reduced mountain snowpacks worldwide, altering snowmelt magnitude and timing. Here the global warming level leading to widespread and persistent mountain snowpack decline, termed low-to-no snow, is estimated for the world’s most latitudinally contiguous mountain range, the American Cordillera. We show that a combination of dynamical, thermodynamical and hypsometric factors results in an asymmetric emergence of low-to-no-snow conditions within the midlatitudes of the American Cordillera. Low-to-no-snow emergence occurs approximately 20 years earlier in the southern hemisphere, at a third of the local warming level, and coincides with runoff efficiency declines (8% average) in both dry and wet years. The prevention of a low-to-no-snow future in either hemisphere requires the level of global warming to be held to, at most, +2.5 °C.

BackgroundThe mechanism for possible association between obesity and poor clinical outcomes from Coronavirus Disease 2019 (COVID-19) remains unclear.MethodsWe analyzed 22,915 adult COVID-19 patients hospitalized from March 2020 to April 2021 to non-intensive care using the American Heart Association National COVID Registry. A multivariable Poisson model adjusted for age, sex, medical history, admission respiratory status, hospitalization characteristics, and laboratory findings was used to calculate length of stay (LOS) as a function of body mass index (BMI). We similarly analyzed 5,327 patients admitted to intensive care for comparison.ResultsRelative to normal BMI subjects, overweight, class I obese, and class II obese patients had approximately half-day reductions in LOS (-0.469 days, P

BACKGROUNDProlonged symptoms after SARS-CoV-2 infection are well documented. However, which factors influence development of long-term symptoms, how symptoms vary across ethnic groups, and whether long-term symptoms correlate with biomarkers are points that remain elusive.METHODSAdult SARS-CoV-2 reverse transcription PCR-positive (RT-PCR-positive) patients were recruited at Stanford from March 2020 to February 2021. Study participants were seen for in-person visits at diagnosis and every 1-3 months for up to 1 year after diagnosis; they completed symptom surveys and underwent blood draws and nasal swab collections at each visit.RESULTSOur cohort (n = 617) ranged from asymptomatic to critical COVID-19 infections. In total, 40% of participants reported at least 1 symptom associated with COVID-19 six months after diagnosis. Median time from diagnosis to first resolution of all symptoms was 44 days; median time from diagnosis to sustained symptom resolution with no recurring symptoms for 1 month or longer was 214 days. Anti-nucleocapsid IgG level in the first week after positive RT-PCR test and history of lung disease were associated with time to sustained symptom resolution. COVID-19 disease severity, ethnicity, age, sex, and remdesivir use did not affect time to sustained symptom resolution.CONCLUSIONWe found that all disease severities had a similar risk of developing post-COVID-19 syndrome in an ethnically diverse population. Comorbid lung disease and lower levels of initial IgG response to SARS-CoV-2 nucleocapsid antigen were associated with longer symptom duration.TRIAL REGISTRATIONClinicalTrials.gov, NCT04373148.FUNDINGNIH UL1TR003142 CTSA grant, NIH U54CA260517 grant, NIEHS R21 ES03304901, Sean N Parker Center for Allergy and Asthma Research at Stanford University, Chan Zuckerberg Biohub, Chan Zuckerberg Initiative, Sunshine Foundation, Crown Foundation, and Parker Foundation.

A central question surrounding possible human exploration of Mars is whether crewed missions can be supported by available technologies using in situ resources. Here, we show that photovoltaics-based power systems would be adequate and practical to sustain a crewed outpost for an extended period over a large fraction of the planet’s surface. Climate data were integrated into a radiative transfer model to predict spectrally-resolved solar flux across the Martian surface. This informed detailed balance calculations for solar cell devices that identified optimal bandgap combinations for maximizing production capacity over a Martian year. We then quantified power systems, manufacturing, and agricultural demands for a six-person mission, which revealed that photovoltaics-based power generation would require (Formula presented.) 10 t of carry-along mass, outperforming alternatives over ∼50% of Mars’ surface.

Introduction: A common refrain in public health and medicine concerns the rapidly rising incidence of obesity and metabolic-like syndromes over the last two decades. A plethora of data suggests there are many roots to the issue and what to do about it. The rise of childhood obesity and the growing incidence of adult endocrine diseases like type 2 diabetes mellitus in paediatric populations are the most concerning. One of the solutions that aims to tackle health and food justice issues is the Farm-to-School programme (FTSP), which helps bring local fruit and vegetables (FVs) to schools and associated nutrition-related programmes. In the USA, a significant expansion of FTSPs has happened since the early 2000s. Numerous economic, social and nutritional benefits of these programmes have been documented, and various methodologies have been used to elucidate the benefits and barriers. Method: This paper reviews the multifaceted strengths of FTSPs including health, institutional and farm-related benefits. The paper also discusses federal- and state-level environment and the obstacles to FTSP implementation. Conclusion: FTSPs offer a comprehensive way to strengthen childhood nutrition education and stimulate local communities economically and socially. FTSPs increase the purchase of locally produced food items, resulting in higher incomes for small farmers. FTSPs have also increased community engagement and new opportunities for farming businesses. In sum, FTSPs can be considered a boon for local schools, producers, communities and better health outcomes for school-age children. Teaching children about food literacy could prove beneficial for supporting their self-efficacy regarding foods and encouraging lifetime healthy dietary behaviours.

Anthropogenic climate change is decreasing seasonal snowpacks globally, with potentially catastrophic consequences on water resources, given the long-held reliance on snowpack in water management. In this Review, we examine the changes and trickle-down impacts of snow loss in the western United States (WUS). Across the WUS, snow water equivalent declines of ~25% are expected by 2050, with losses comparable with contemporary historical trends. There is less consensus on the time horizon of snow disappearance, but model projections combined with a new low-to-no snow definition suggest ~35–60 years before low-to-no snow becomes persistent if greenhouse gas emissions continue unabated. Diminished and more ephemeral snowpacks that melt earlier will alter groundwater and streamflow dynamics. The direction of these changes are difficult to constrain given competing factors such as higher evapotranspiration, altered vegetation composition and changes in wildfire behaviour in a warmer world. These changes undermine conventional WUS water management practices, but through proactive implementation of soft and hard adaptation strategies, there is potential to build resilience to extreme, episodic and, eventually, persistent low-to-no snow conditions. Federal investments offer a timely opportunity to address these vulnerabilities, but they require a concerted portfolio of activities that cross historically siloed physical and disciplinary boundaries.

This study aims to quantify the impacts of atmosphere–ocean coupling on simulated and projected tropical cyclone (TC) precipitation globally. We used global climate model (GCM) simulations over 1950–2050 from the High Resolution Model Intercomparison Project (HighResMIP) and compared its fully coupled atmosphere–ocean GCMs (AOGCMs) with atmosphere-only GCMs (AGCMs). We find that ocean coupling generally leads to decreased TC precipitation over ocean and land. Large-scale sea surface temperature (SST) biases are critical drivers of the precipitation difference, with secondary contributions from local TC–ocean feedbacks via SST cold wakes. The two driving factors, attributed to ocean coupling in the AOGCMs, influence TC precipitation in association with decreased TC intensity and specific humidity. The AOGCMs and AGCMs consistently project TC precipitation increases in 2015–2050 relative to 1950–2014 over ocean for all basins, and for landfalling TCs in the North Atlantic and western North Pacific.

Constructing an accurate, continental, in-situ-based, kilometer-scale, long-term record of the precipitation field and its spatiotemporal changes remains a significant challenge. Here, we determine the extreme-value behavior of the NEXRAD Stage IV radar-based quantitative precipitation estimate. We find that the climatology of 5-year daily return values in the contiguous United States East of the Rocky Mountains shows only slight variability on spatial scales smaller than (Formula presented.) 100 km. In light of this finding, we test whether rain-gauge-only daily precipitation data sets can produce accurate extreme-value behavior at spatial scales finer than the spacing between gauges. We find that the 5-year daily return values are accurate at locations far from rain gauges only if the interpolation between gauges is carried out appropriately for extremes. Precipitation statistics derived from in-situ rain gauge data are therefore of sufficient spatial resolution to faithfully capture daily extremes over much of the eastern United States.

A recent study found statistically significant differences in extreme precipitation distributions over the contiguous United States (CONUS) when changing the microphysics scheme in a superparameterized global climate model. Here, we repeat the analysis globally and similarly find that differences are widespread when varying the number of predicted moments in the microphysics parameterization, but not when comparing variants of the double-moment scheme. However, contrary to the previous study in which differences largely disappeared over CONUS when 5-day simulations were conducted, we found that the signal in these shorter integrations remains within the tropics, implying a direct local effect of microphysics on precipitation extremes in these regions. The effect on precipitation is traced back to changes in vertical velocity profiles changes that are then amplified in the climatological simulations compared to the 5-day ones. Finally, the superparameterized extremes, regardless of the microphysics scheme, are shown to be larger than those from the Global Precipitation Climatology Project One-Degree Daily data set and generally smaller than those from the Tropical Rainfall Measuring Mission 3B42 data set.

While various studies explore the relationship between individual sources of climate variability and extreme precipitation, there is a need for improved understanding of how these physical phenomena simultaneously influence precipitation in the observational record across the contiguous United States. In this work, we introduce a single framework for characterizing the historical signal (anthropogenic forcing) and noise (natural variability) in seasonal mean and extreme precipitation. An important aspect of our analysis is that we simultaneously isolate the individual effects of seven modes of variability while explicitly controlling for joint inter-mode relationships. Our method utilizes a spatial statistical component that uses in situ measurements to resolve relationships to their native scales; furthermore, we use a data-driven procedure to robustly determine statistical significance. In Part I of this work we focus on natural climate variability: detection is mostly limited to DJF and SON for the modes of variability considered, with the El Niño/Southern Oscillation, the Pacific–North American pattern, and the North Atlantic Oscillation exhibiting the largest influence. Across all climate indices considered, the signals are larger and can be detected more clearly for seasonal total versus extreme precipitation. We are able to detect at least some significant relationships in all seasons in spite of extremely large (> 95%) background variability in both mean and extreme precipitation. Furthermore, we specifically quantify how the spatial aspect of our analysis reduces uncertainty and increases detection of statistical significance while also discovering results that quantify the complex interconnected relationships between climate drivers and seasonal precipitation.

Atmospheric rivers (ARs) are long and narrow filaments of vapor transport that are responsible for most poleward moisture transport outside of the tropics. Many AR detection algorithms have been developed to automatically identify ARs in climate data. The diversity of these algorithms has introduced appreciable uncertainties in quantitative measures of AR properties and thereby impedes the construction of a unified and internally consistent climatology of ARs. This paper compares nine global AR detection algorithms from the perspective of AR lifecycles following the propagation of ARs from origin to termination in the MERRA2 reanalysis over the period 1980–2017. Uncertainties in AR lifecycle characteristics, including event number, lifetime, intensity, and frequency distribution are discussed. Notably, the number of AR events per year in the Northern Hemisphere can vary by a factor of 5 with different algorithms. Although all algorithms show that the maximum AR origin (termination) frequency is located over the western (eastern) portion of ocean basins, significant disagreements appear in regional distribution. Spreads are large in AR lifetime and intensity. The number of landfalling AR events produced by the algorithms can vary from 16 to 80 events per year, although the agreement improves for stronger ARs. By examining the ARs' connections with the Madden-Julian Oscillation and El Niño Southern Oscillation, we find that the overall responses of ARs (such as changes in AR frequency, origin, and landfall activity) to climate variability are consistent among algorithms.

We examine the resolution dependence of errors in extreme sub-daily precipitation in available high-resolution climate models. We find that simulated extreme precipitation increases as horizontal resolution increases but that appropriately constructed model skill metrics do not significantly change. We find little evidence that simulated extreme winter or summer storm processes significantly improve with the resolution because the model performance changes identified are consistent with expectations from scale dependence arguments alone. We also discuss the implications of these scale-dependent limitations on the interpretation of simulated extreme precipitation. This article is part of a discussion meeting issue 'Intensification of short-duration rainfall extremes and implications for flash flood risks'.

The social cost of methane (SC-CH4) measures the economic loss of welfare caused by emitting one tonne of methane into the atmosphere. This valuation may in turn be used in cost-benefit analyses or to inform climate policies1-3. However, current SC-CH4 estimates have not included key scientific findings and observational constraints. Here we estimate the SC-CH4 by incorporating the recent upward revision of 25 per cent to calculations of the radiative forcing of methane4, combined with calibrated reduced-form global climate models and an ensemble of integrated assessment models (IAMs). Our multi-model mean estimate for the SC-CH4 is US$933 per tonne of CH4 (5-95 per cent range, US$471-1,570 per tonne of CH4) under a high-emissions scenario (Representative Concentration Pathway (RCP) 8.5), a 22 per cent decrease compared to estimates based on the climate uncertainty framework used by the US federal government5. Our ninety-fifth percentile estimate is 51 per cent lower than the corresponding figure from the US framework. Under a low-emissions scenario (RCP 2.6), our multi-model mean decreases to US$710 per tonne of CH4. Tightened equilibrium climate sensitivity estimates paired with the effect of previously neglected relationships between uncertain parameters of the climate model lower these estimates. We also show that our SC-CH4 estimates are sensitive to model combinations; for example, within one IAM, different methane cycle sub-models can induce variations of approximately 20 per cent in the estimated SC-CH4. But switching IAMs can more than double the estimated SC-CH4. Extending our results to account for societal concerns about equity produces SC-CH4 estimates that differ by more than an order of magnitude between low- and high-income regions. Our central equity-weighted estimate for the USA increases to US$8,290 per tonne of CH4 whereas our estimate for sub-Saharan Africa decreases to US$134 per tonne of CH4.

Atmospheric rivers (ARs) account for a large portion of winter precipitation in the western US. To evaluate the sources of AR and precipitation predictability at subseasonal-to-seasonal timescales, we examine the relationships between two climate modes, the El Niño/Southern Oscillation (ENSO) and Madden-Julian oscillation (MJO), and winter hydroclimate in the western US. Our analysis uses a large ensemble of the Weather Research and Forecast (WRF) model simulations from 1981 to 2017, facilitating the assessment of uncertainty in climate mode-AR relationship due to climate variability and short length of observations. Over the North Pacific basin, we find ENSO-related latitudinal shifts of ARs, but ENSO has little effect on basin-wide averaged AR frequency. Over the western US, there is some uncertainty in the ENSO-AR connection due to the impacts of ENSO indices and datasets. However, extreme El Niño events defined by the ENSO longitude index are consistently linked to increased landfalling AR activity. The MJO can lead to significantly enhanced or suppressed landfalling AR activity in the large-ensembles, depending on the phase of MJO and time lag, while observations are too short in length to robustly show this signal. ENSO substantially modulates the MJO-AR relationship, triggering variegated responses of landfalling ARs and AR precipitation in La Niña and El Niño years. Our findings highlight the need to evaluate concurrent effects of different climate modes on ARs and precipitation, and may shed light on a path toward more accurate subseasonal-to-seasonal prediction of ARs and precipitation over the western US.

We investigate how mesoscale circulations associated with convective aggregation can modulate the sensitivity of the hydrologic cycle to warming. We quantify changes in the full distribution of rain across radiative-convective equilibrium states in a cloud-resolving model. For a given Sea Surface Temperature (SST), the shift in mean rainfall between disorganized and organized states is associated with a shift in atmospheric radiative cooling, and is roughly analogous to the effect of a 4K SST increase. With rising temperatures, the increase in mean rain rate is insensitive to the presence of organization, while extremes can intensify faster in the aggregated state, leading to a faster amplification in the sporadic nature of rain. When convection aggregates, heavy rain is enhanced by 20%-30% and nonlinear behaviors are observed as a function of SST and strength of aggregation feedbacks. First, radiative- and surface-flux aggregation feedbacks have multiplicative effects on extremes, illustrating a non-trivial sensitivity to the degree of organization. Second, alternating Clausius-Clapeyron and super-Clausius-Clapeyron regimes in extreme rainfall are found as a function of SST, corresponding to varying thermodynamic and dynamic contributions, and a large sensitivity to precipitation efficiency variations in some SST ranges. The potential for mesoscale circulations in amplifying the hydrologic cycle is established. However, these nonlinear distortions question the quantitative relevance of idealized self-aggregation. This calls for a deeper investigation of relationships which capture the coupling between global energetics, aggregation feedbacks and local convection, and for systematic tests of their sensitivity to domain configurations, surface boundary conditions, microphysics, and turbulence schemes.

Identifying, detecting, and localizing extreme weather events is a crucial first step in understanding how they may vary under different climate change scenarios. Pattern recognition tasks such as classification, object detection, and segmentation (i.e., pixel-level classification) have remained challenging problems in the weather and climate sciences. While there exist many empirical heuristics for detecting extreme events, the disparities between the output of these different methods even for a single event are large and often difficult to reconcile. Given the success of deep learning (DL) in tackling similar problems in computer vision, we advocate a DL-based approach. DL, however, works best in the context of supervised learning - when labeled datasets are readily available. Reliable labeled training data for extreme weather and climate events is scarce. We create "ClimateNet" - an open, community-sourced human-expert-labeled curated dataset that captures tropical cyclones (TCs) and atmospheric rivers (ARs) in highresolution climate model output from a simulation of a recent historical period.We use the curated ClimateNet dataset to train a state-of-the-art DL model for pixel-level identification - i.e., segmentation - of TCs and ARs. We then apply the trained DL model to historical and climate change scenarios simulated by the Community Atmospheric Model (CAM5.1) and show that the DL model accurately segments the data into TCs, ARs, or "the background" at a pixel level. Further, we show how the segmentation results can be used to conduct spatially and temporally precise analytics by quantifying distributions of extreme precipitation conditioned on event types (TC or AR) at regional scales. The key contribution of this work is that it paves the way for DL-based automated, high-fidelity, and highly precise analytics of climate data using a curated expert-labeled dataset - ClimateNet. ClimateNet and the DL-based segmentation method provide several unique capabilities: (i) they can be used to calculate a variety of TC and AR statistics at a fine-grained level; (ii) they can be applied to different climate scenarios and different datasets without tuning as they do not rely on threshold conditions; and (iii) the proposed DL method is suitable for rapidly analyzing large amounts of climate model output. While our study has been conducted for two important extreme weather patterns (TCs and ARs) in simulation datasets, we believe that this methodology can be applied to a much broader class of patterns and applied to observational and reanalysis data products via transfer learning.