Search

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.

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 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 high-resolution 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, hi-fidelity and highly precise analytics of climate data using a curated expert-labelled 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.

The Arctic hydrological system is an interconnected system that is experiencing rapid change. It is comprised of permafrost, snow, glacier, frozen soils, and inland river systems. In this study, we aim to lower the barrier of using complex land models in regional applications by developing a generalizable optimization methodology and workflow for the Community Terrestrial Systems Model (CTSM), to move them toward a more Actionable Science paradigm. Further end-user engagement is required to make science such as this “fully actionable.” We applied CTSM across Alaska and the Yukon River Basin at 4-km spatial resolution. We highlighted several potentially useful high-resolution CTSM configuration changes. Additionally, we performed a multi-objective optimization using snow and river flow metrics within an adaptive surrogate-based model optimization scheme. Four representative river basins across our study domain were selected for optimization based on observed streamflow and snow water equivalent observations at 10 SNOTEL sites. Fourteen sensitive parameters were identified for optimization with half of them not directly related to hydrology or snow processes. Across fifteen out-of-sample river basins, 13 had improved flow simulations after optimization and the mean Kling-Gupta Efficiency of daily flow increased from 0.43 to 0.63 in a 30-year evaluation. In addition, we adapted the Shapley Decomposition to disentangle each parameter's contribution to streamflow performance changes, with the seven non-hydrological parameters providing a non-negligible contribution to performance gains. The snow simulation had limited improvement, likely because snow simulation is influenced more by meteorological forcing than model parameter choices. [ABSTRACT FROM AUTHOR]

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 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 high-resolution 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, hi-fidelity and highly precise analytics of climate data using a curated expert-labelled 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. [ABSTRACT FROM AUTHOR]

To mitigate the projected global warming in the 21st century, it is well recognized that society needs to cut CO2 emission and other short-lived warming agents aggressively. However, to stabilize the climate at a warming level closer to the present day, such as the "well below 2 °C" aspiration in the Paris agreement, a net-zero carbon emission by 2050 is still insufficient. The recent IPCC special report calls for a massive scheme to extract CO2 directly from the atmosphere, in addition to the decarbonization, to reach negative net emission at the mid-century mark. Another ambitious proposal is the solar radiation-based geoengineering schemes, including injecting sulfur gas into the stratosphere. Despite being in the public debate for years, these two leading geoengineering schemes have not been carefully examined under a consistent numerical modeling framework. Here we present a comprehensive analysis of climate impacts of these two geoengineering approaches using two recently available large-ensemble (> 10 members) model experiments conducted by a family of state-of-art Earth system models. The CO2 -based mitigation simulation is designed to include both emissions cut and carbon capture. The solar radiation-based mitigation simulation is designed to inject the sulfur gas strategically at specified altitudes and latitudes and run a feedback control algorithm, to avoid common problems previously identified such as the over-cooling of the Tropics and large-scale precipitation shifts. Our analysis focuses on the projected aridity conditions over the Americas in the 21st century, in detailed terms of the mitigation potential, the temporal evolution, the spatial distribution (within North and South America), the relative efficiency, and the physical mechanisms. We show that sulfur injection, in contrast to previous notions of leading to excessive terrestrial drying (in terms of precipitation reduction) while offsetting the global mean greenhouse gas (GHG) warming, will instead mitigate the projected drying tendency under RCP8.5. The surface energy balance change induced by Sulfur injection, in addition to the well-known response in temperature and precipitation, plays a crucial role in determining the overall terrestrial hydroclimate response. However, when normalized by the same amount of avoided global warming, in these simulations, sulfur injection is less effective in limiting the worsening trend of regional land aridity in the Americas, when compared with carbon capture. Temporally, the climate benefit of Sulfur injection will emerge more quickly, even when both schemes are hypothetically started in the same year of 2020. Spatially, both schemes are effective in curbing the drying trend over North America. However, for South America, the Sulfur Injection scheme is particularly more effective for the sub-Amazon region (South Brazil), while the Carbon Capture scheme is more effective for the Amazon region. We conclude that despite the apparent limitations (such as inability to address ocean acidification) and potential side effects (such as changes to the ozone layer), innovative means of Sulfur Injection should continue to be explored as a potential low-cost option in the climate solution toolbox, complementing other mitigation approaches such as emissions cut and carbon capture (Cao et al., 2017). Our results demonstrate the urgent need for multi-model comparison studies and detailed regional assessment in other parts of the world. [ABSTRACT FROM AUTHOR]

Geoengineering with stratospheric sulfate aerosols can, to some extent, be designed to achieve different climate objectives. Here we use the state‐of‐the‐art Community Earth System Model, version 1, with the Whole Atmosphere Community Climate Model as its atmospheric component (CESM1(WACCM)), to compare surface climate and stratospheric effects of two geoengineering strategies. In one, SO2 is injected into the tropical lower stratosphere at the equator to keep global mean temperature nearly constant under an RCP8.5 scenario, as has been commonly simulated in previous studies. In another, the Geoengineering Large Ensemble (GLENS), SO2 is injected into the lower stratosphere at four different locations (30°N/S and 15°N/S) to keep global mean temperature, the interhemispheric temperature gradient, and the equator‐to‐pole temperature gradient nearly unchanged. Both simulations are effective at offsetting changes in global mean temperature and the interhemispheric temperature gradient that result from increased greenhouse gases, but only GLENS fully offsets changes in the equator‐to‐pole temperature gradient. GLENS results in a more even aerosol distribution, whereas equatorial injection tends to result in an aerosol peak in the tropics. Moreover, GLENS requires less total injection than in the equatorial case due to different spatial distributions of the aerosols. Many other aspects of surface climate changes, including precipitation and sea ice coverage, also show reduced changes in GLENS as compared to equatorial injection. Stratospheric changes, including heating, circulation, and effects on the quasi‐biennial oscillation are greatly reduced in GLENS as compared to equatorial injection. Key Points: Tropospheric and stratospheric side effects of equatorial solar geoengineering are reduced if nonequatorial injections are appliedSome residual temperature effects from equatorial injection can be offset with multiple injection locationsThere are still residuals in surface climate and the stratosphere that cannot be offset with four injection locations [ABSTRACT FROM AUTHOR]