Satellite altimetry has been widely utilized to monitor global sea surface dynamics, enabling investigation of upper ocean variability from basin-scale to localized eddy ranges. However, the sparse spatial resolution of observational altimetry limits our understanding of oceanic submesoscale variability, prevalent at horizontal scales below 0.25o resolution. Here, we introduce a state-of-the-art generative diffusion model to train high-resolution sea surface height (SSH) reanalysis data and demonstrate its advantage in observational SSH downscaling over the eddy-rich Kuroshio Extension region. The diffusion-based model effectively downscales raw satellite-interpolated data from 0.25o resolution to 1/16o, corresponding to approximately 12-km wavelength. This model outperforms other high-resolution reanalysis datasets and neural network-based methods. Also, it successfully reproduces the spatial patterns and power spectra of satellite along-track observations. Our diffusion-based results indicate that eddy kinetic energy at horizontal scales less than 250 km has intensified significantly since 2004 in the Kuroshio Extension region. These findings underscore the great potential of deep learning in reconstructing satellite altimetry and enhancing our understanding of ocean dynamics at eddy scales.
{"title":"Generative Diffusion Model-based Downscaling of Observed Sea Surface Height over Kuroshio Extension since 2000","authors":"Qiuchang Han, Xingliang Jiang, Yang Zhao, Xudong Wang, Zhijin Li, Renhe Zhang","doi":"arxiv-2408.12632","DOIUrl":"https://doi.org/arxiv-2408.12632","url":null,"abstract":"Satellite altimetry has been widely utilized to monitor global sea surface\u0000dynamics, enabling investigation of upper ocean variability from basin-scale to\u0000localized eddy ranges. However, the sparse spatial resolution of observational\u0000altimetry limits our understanding of oceanic submesoscale variability,\u0000prevalent at horizontal scales below 0.25o resolution. Here, we introduce a\u0000state-of-the-art generative diffusion model to train high-resolution sea\u0000surface height (SSH) reanalysis data and demonstrate its advantage in\u0000observational SSH downscaling over the eddy-rich Kuroshio Extension region. The\u0000diffusion-based model effectively downscales raw satellite-interpolated data\u0000from 0.25o resolution to 1/16o, corresponding to approximately 12-km\u0000wavelength. This model outperforms other high-resolution reanalysis datasets\u0000and neural network-based methods. Also, it successfully reproduces the spatial\u0000patterns and power spectra of satellite along-track observations. Our\u0000diffusion-based results indicate that eddy kinetic energy at horizontal scales\u0000less than 250 km has intensified significantly since 2004 in the Kuroshio\u0000Extension region. These findings underscore the great potential of deep\u0000learning in reconstructing satellite altimetry and enhancing our understanding\u0000of ocean dynamics at eddy scales.","PeriodicalId":501166,"journal":{"name":"arXiv - PHYS - Atmospheric and Oceanic Physics","volume":"159 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142215443","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
J. Xavier ProchaskaUniversity of California, Santa CruzScripps Institution of OceanographSimons Pivot Fellow, Daniel RudnickScripps Institution of Oceanograph
Dissolved oxygen (DO) is a non-conservative tracer of interactions at the air-sea interface, respiration and photosynthesis, and advection. In this manuscript, we study extremes in the degree of oxygen saturation (SO), the ratio of DO to the maximum concentration given the water's temperature, salinity, and depth with SO=1 critically saturated. We perform the analysis with the California Underwater Glider Network (CUGN), which operates gliders on four lines that extend from the California coast to several hundred kilometers offshore, profiling to 500m depth every 3km. Since ~2017, the gliders have been equipped with a Sea-Bird 63 optode sensor to measure the DO content. We find that parcels with SO>1.1, hyperoxic extrema, occur primarily near-shore in the upper 50m of the water column and during non-winter months. Along Line 90 which originates in San Diego, these hyperoxic events occur primarily in stratified waters with shallow mixed layers. We hypothesize that photosynthesis elevates DO in sub-surface water that can not rapidly ventilate with the surface. Along the three other lines, hyperoxic extrema occur almost exclusively at the surface and are correlated with elevated Chl-a fluorescence suggesting they are primarily driven by blooms of photosynthesis. We also examine hypoxic extrema, finding that parcels with SO<0.9 and z<50m occur most frequently along the northernmost line where upwelling has greatest impact.
{"title":"Extremes of Dissolved Oxygen in the California Current System","authors":"J. Xavier ProchaskaUniversity of California, Santa CruzScripps Institution of OceanographSimons Pivot Fellow, Daniel RudnickScripps Institution of Oceanograph","doi":"arxiv-2408.12287","DOIUrl":"https://doi.org/arxiv-2408.12287","url":null,"abstract":"Dissolved oxygen (DO) is a non-conservative tracer of interactions at the\u0000air-sea interface, respiration and photosynthesis, and advection. In this\u0000manuscript, we study extremes in the degree of oxygen saturation (SO), the\u0000ratio of DO to the maximum concentration given the water's temperature,\u0000salinity, and depth with SO=1 critically saturated. We perform the analysis\u0000with the California Underwater Glider Network (CUGN), which operates gliders on\u0000four lines that extend from the California coast to several hundred kilometers\u0000offshore, profiling to 500m depth every 3km. Since ~2017, the gliders have been\u0000equipped with a Sea-Bird 63 optode sensor to measure the DO content. We find\u0000that parcels with SO>1.1, hyperoxic extrema, occur primarily near-shore in the\u0000upper 50m of the water column and during non-winter months. Along Line 90 which\u0000originates in San Diego, these hyperoxic events occur primarily in stratified\u0000waters with shallow mixed layers. We hypothesize that photosynthesis elevates\u0000DO in sub-surface water that can not rapidly ventilate with the surface. Along\u0000the three other lines, hyperoxic extrema occur almost exclusively at the\u0000surface and are correlated with elevated Chl-a fluorescence suggesting they are\u0000primarily driven by blooms of photosynthesis. We also examine hypoxic extrema,\u0000finding that parcels with SO<0.9 and z<50m occur most frequently along the\u0000northernmost line where upwelling has greatest impact.","PeriodicalId":501166,"journal":{"name":"arXiv - PHYS - Atmospheric and Oceanic Physics","volume":"9 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142215445","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Zeyi Niu, Wei Huang, Lei Zhang, Lin Deng, Haibo Wang, Yuhua Yang, Dongliang Wang, Hong Li
With the rapid development of data-driven machine learning (ML) models in meteorology, typhoon track forecasts have become increasingly accurate. However, current ML models still face challenges, such as underestimating typhoon intensity and lacking interpretability. To address these issues, this study establishes an ML-driven hybrid typhoon model, where forecast fields from the Pangu-Weather model are used to constrain the large-scale forecasts of the Weather Research and Forecasting model based on the spectral nudging method (Pangu_SP). The results show that forecasts from the Pangu_SP experiment obviously outperform those by using the Global Forecast System as the initial field (GFS_INIT) and from the Integrated Forecasting System of the European Centre for Medium-Range Weather Forecasts (ECMWF IFS) for the track forecast of Typhoon Doksuri (2023). The predicted typhoon cloud patterns from Pangu_SP are also more consistent with satellite observations. Additionally, the typhoon intensity forecasts from Pangu_SP are notably more accurate than those from the ECMWF IFS, demonstrating that the hybrid model effectively leverages the strengths of both ML and physical models. Furthermore, this study is the first to explore the significance of data assimilation in ML-driven hybrid dynamical systems. The findings reveal that after assimilating water vapor channels from the Advanced Geostationary Radiation Imager onboard Fengyun-4B, the errors in typhoon intensity forecasts are reduced.
随着气象学中数据驱动的机器学习（ML）模式的快速发展，台风路径预报变得越来越准确。然而，当前的 ML 模式仍然面临挑战，如低估台风强度和缺乏可解释性。为了解决这些问题，本研究建立了一个 ML 驱动的混合台风模式，利用盘古-天气模式的预报场来约束基于频谱推移方法（Pangu_SP）的天气研究和预报模式的大尺度预报。结果表明，在台风 "杜苏芮"（2023 年）的路径预报中，盘古_SP 试验的预报结果明显优于以全球预报系统为初始场（GFS_INIT）的预报结果，也优于欧洲中期天气预报中心综合预报系统（ECMWF IFS）的预报结果。盘古_SP 预测的台风云模式与卫星观测结果也更加一致。此外，盘古_SP 预测的台风强度也明显比 ECMWF IFS 预测的台风强度更准确，这表明混合模式有效地利用了 ML 和物理模式的优势。此外，本研究还首次探讨了数据同化在 ML 驱动的混合动力系统中的意义。研究结果表明，在同化了风云四号 B 星上高级地球静止辐射成像仪的水汽通道后，台风强度预报的误差减小了。
{"title":"Improving Typhoon Predictions by Integrating Data-Driven Machine Learning Models with Physics Models Based on the Spectral Nudging and Data Assimilation","authors":"Zeyi Niu, Wei Huang, Lei Zhang, Lin Deng, Haibo Wang, Yuhua Yang, Dongliang Wang, Hong Li","doi":"arxiv-2408.12630","DOIUrl":"https://doi.org/arxiv-2408.12630","url":null,"abstract":"With the rapid development of data-driven machine learning (ML) models in\u0000meteorology, typhoon track forecasts have become increasingly accurate.\u0000However, current ML models still face challenges, such as underestimating\u0000typhoon intensity and lacking interpretability. To address these issues, this\u0000study establishes an ML-driven hybrid typhoon model, where forecast fields from\u0000the Pangu-Weather model are used to constrain the large-scale forecasts of the\u0000Weather Research and Forecasting model based on the spectral nudging method\u0000(Pangu_SP). The results show that forecasts from the Pangu_SP experiment\u0000obviously outperform those by using the Global Forecast System as the initial\u0000field (GFS_INIT) and from the Integrated Forecasting System of the European\u0000Centre for Medium-Range Weather Forecasts (ECMWF IFS) for the track forecast of\u0000Typhoon Doksuri (2023). The predicted typhoon cloud patterns from Pangu_SP are\u0000also more consistent with satellite observations. Additionally, the typhoon\u0000intensity forecasts from Pangu_SP are notably more accurate than those from the\u0000ECMWF IFS, demonstrating that the hybrid model effectively leverages the\u0000strengths of both ML and physical models. Furthermore, this study is the first\u0000to explore the significance of data assimilation in ML-driven hybrid dynamical\u0000systems. The findings reveal that after assimilating water vapor channels from\u0000the Advanced Geostationary Radiation Imager onboard Fengyun-4B, the errors in\u0000typhoon intensity forecasts are reduced.","PeriodicalId":501166,"journal":{"name":"arXiv - PHYS - Atmospheric and Oceanic Physics","volume":"2 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142215450","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Understanding the climate dynamics at the inner edge of the habitable zone (HZ) is crucial for predicting the habitability of rocky exoplanets. Previous studies using Global Climate Models (GCMs) have indicated that planets receiving high stellar flux can exhibit climate bifurcations, leading to bistability between a cold (temperate) and a hot (runaway) climate. However, the mechanism causing this bistability has not been fully explained, in part due to the difficulty associated with inferring mechanisms from small numbers of expensive numerical simulations in GCMs. In this study, we employ a two-column (dayside and nightside), two-layer climate model to investigate the physical mechanisms driving this bistability. Through mechanism-denial experiments, we demonstrate that the runaway greenhouse effect, coupled with a cloud feedback on either the dayside or nightside, leads to climate bistability. We also map out the parameters that control the location of the bifurcations and size of the bistability. This work identifies which mechanisms and GCM parameters control the stellar flux at which rocky planets are likely to retain a hot, thick atmosphere if they experience a hot start. This is critical for the prioritization of targets and interpretation of observations by the James Webb Space Telescope (JWST). Furthermore, our modeling framework can be extended to planets with different condensable species and cloud types.
{"title":"Climate Bistability at the Inner Edge of the Habitable Zone due to Runaway Greenhouse and Cloud Feedbacks","authors":"Bowen Fan, Da Yang, Dorian S. Abbot","doi":"arxiv-2408.12563","DOIUrl":"https://doi.org/arxiv-2408.12563","url":null,"abstract":"Understanding the climate dynamics at the inner edge of the habitable zone\u0000(HZ) is crucial for predicting the habitability of rocky exoplanets. Previous\u0000studies using Global Climate Models (GCMs) have indicated that planets\u0000receiving high stellar flux can exhibit climate bifurcations, leading to\u0000bistability between a cold (temperate) and a hot (runaway) climate. However,\u0000the mechanism causing this bistability has not been fully explained, in part\u0000due to the difficulty associated with inferring mechanisms from small numbers\u0000of expensive numerical simulations in GCMs. In this study, we employ a\u0000two-column (dayside and nightside), two-layer climate model to investigate the\u0000physical mechanisms driving this bistability. Through mechanism-denial\u0000experiments, we demonstrate that the runaway greenhouse effect, coupled with a\u0000cloud feedback on either the dayside or nightside, leads to climate\u0000bistability. We also map out the parameters that control the location of the\u0000bifurcations and size of the bistability. This work identifies which mechanisms\u0000and GCM parameters control the stellar flux at which rocky planets are likely\u0000to retain a hot, thick atmosphere if they experience a hot start. This is\u0000critical for the prioritization of targets and interpretation of observations\u0000by the James Webb Space Telescope (JWST). Furthermore, our modeling framework\u0000can be extended to planets with different condensable species and cloud types.","PeriodicalId":501166,"journal":{"name":"arXiv - PHYS - Atmospheric and Oceanic Physics","volume":"11 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142215447","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
A quasi-consensus has steadily formed in the scientific literature on the fact that the prevention measures implemented by most countries to curb the 2020 COVID-19 pandemic have led to significant reductions in pollution levels around the world, especially in urban environments. Fewer studies have looked at the how these reductions at ground level translate into variations in the whole atmosphere. In this study, we examine the columnar values of aerosols at 51 mainland European stations of the Aerosol Robotic Network (AERONET). We show that when considered in the context of the long-term trend over the last decade, the columnar aerosol levels for 2020, at the regional level, do not appear exceptional. Both the yearly means and the number of episodes with extreme values for this period are within the one standard deviation of the long-term trends. We conclude that the spatially and temporally very localized reductions do not add up to statistically significant reductions in the global levels of aerosols. Furthermore, considering that pandemic lockdowns can be thought of as a simulation of a climate change mitigation scenario, we conclude that such lifestyle-based changes present a very low potential as a global climate change mitigation strategy.
{"title":"Changes in anthropogenic aerosols during the first wave of COVID-19 lockdowns in the context of long-term historical trends at 51 AERONET stations","authors":"Robert Blaga, Delia Calinoiu, Gavrila Trif-Tordai","doi":"arxiv-2408.11757","DOIUrl":"https://doi.org/arxiv-2408.11757","url":null,"abstract":"A quasi-consensus has steadily formed in the scientific literature on the\u0000fact that the prevention measures implemented by most countries to curb the\u00002020 COVID-19 pandemic have led to significant reductions in pollution levels\u0000around the world, especially in urban environments. Fewer studies have looked\u0000at the how these reductions at ground level translate into variations in the\u0000whole atmosphere. In this study, we examine the columnar values of aerosols at\u000051 mainland European stations of the Aerosol Robotic Network (AERONET). We show\u0000that when considered in the context of the long-term trend over the last\u0000decade, the columnar aerosol levels for 2020, at the regional level, do not\u0000appear exceptional. Both the yearly means and the number of episodes with\u0000extreme values for this period are within the one standard deviation of the\u0000long-term trends. We conclude that the spatially and temporally very localized\u0000reductions do not add up to statistically significant reductions in the global\u0000levels of aerosols. Furthermore, considering that pandemic lockdowns can be\u0000thought of as a simulation of a climate change mitigation scenario, we conclude\u0000that such lifestyle-based changes present a very low potential as a global\u0000climate change mitigation strategy.","PeriodicalId":501166,"journal":{"name":"arXiv - PHYS - Atmospheric and Oceanic Physics","volume":"159 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142215451","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Wuxin Wang, Weicheng Ni, Tao Han, Lei Bai, Boheng Duan, Kaijun Ren
Recent advancements in deep learning (DL) have led to the development of several Large Weather Models (LWMs) that rival state-of-the-art (SOTA) numerical weather prediction (NWP) systems. Up to now, these models still rely on traditional NWP-generated analysis fields as input and are far from being an autonomous system. While researchers are exploring data-driven data assimilation (DA) models to generate accurate initial fields for LWMs, the lack of a standard benchmark impedes the fair evaluation among different data-driven DA algorithms. Here, we introduce DABench, a benchmark dataset utilizing ERA5 data as ground truth to guide the development of end-to-end data-driven weather prediction systems. DABench contributes four standard features: (1) sparse and noisy simulated observations under the guidance of the observing system simulation experiment method; (2) a skillful pre-trained weather prediction model to generate background fields while fairly evaluating the impact of assimilation outcomes on predictions; (3) standardized evaluation metrics for model comparison; (4) a strong baseline called the DA Transformer (DaT). DaT integrates the four-dimensional variational DA prior knowledge into the Transformer model and outperforms the SOTA in physical state reconstruction, named 4DVarNet. Furthermore, we exemplify the development of an end-to-end data-driven weather prediction system by integrating DaT with the prediction model. Researchers can leverage DABench to develop their models and compare performance against established baselines, which will benefit the future advancements of data-driven weather prediction systems. The code is available on this Github repository and the dataset is available at the Baidu Drive.
深度学习（DL）领域的最新进展已导致开发出多个大型天气模型（LWM），可与最先进的（SOTA）数值天气预报（NWP）系统相媲美。迄今为止，这些模型仍依赖传统的 NWP 生成的分析场作为输入，远非自主系统。虽然研究人员正在探索数据驱动的数据同化（DA）模式，以便为 LWMs 生成精确的初始场，但标准基准的缺乏妨碍了对不同数据驱动的 DA 算法进行公平评估。在此，我们介绍 DABench，这是一个利用ERA5数据作为地面实况的基准数据集，用于指导端到端数据驱动天气预报系统的开发。DABench 有四个标准特征：(1)在观测系统模拟实验方法指导下的稀疏和噪声模拟观测；(2)熟练的预训练天气预报模型，用于生成背景场，同时公平地评估同化结果对预测的影响；(3)标准化的评估指标，用于模型比较；(4)称为 DA Transformer (DaT)的强大基线。DaT 将四维变分 DA 先验知识整合到变换器模型中，在物理状态重建方面优于 SOTA，被命名为 4DVarNet。此外，我们还举例说明了通过将 DaT 与预测模型集成，开发端到端数据驱动天气预报系统的过程。研究人员可以利用 DABench 开发自己的模型，并与既定基线比较性能，这将有利于数据驱动天气预报系统的未来发展。代码可在此 Github 代码库中获取，数据集可在百度硬盘中获取。
{"title":"DABench: A Benchmark Dataset for Data-Driven Weather Data Assimilation","authors":"Wuxin Wang, Weicheng Ni, Tao Han, Lei Bai, Boheng Duan, Kaijun Ren","doi":"arxiv-2408.11438","DOIUrl":"https://doi.org/arxiv-2408.11438","url":null,"abstract":"Recent advancements in deep learning (DL) have led to the development of\u0000several Large Weather Models (LWMs) that rival state-of-the-art (SOTA)\u0000numerical weather prediction (NWP) systems. Up to now, these models still rely\u0000on traditional NWP-generated analysis fields as input and are far from being an\u0000autonomous system. While researchers are exploring data-driven data\u0000assimilation (DA) models to generate accurate initial fields for LWMs, the lack\u0000of a standard benchmark impedes the fair evaluation among different data-driven\u0000DA algorithms. Here, we introduce DABench, a benchmark dataset utilizing ERA5\u0000data as ground truth to guide the development of end-to-end data-driven weather\u0000prediction systems. DABench contributes four standard features: (1) sparse and\u0000noisy simulated observations under the guidance of the observing system\u0000simulation experiment method; (2) a skillful pre-trained weather prediction\u0000model to generate background fields while fairly evaluating the impact of\u0000assimilation outcomes on predictions; (3) standardized evaluation metrics for\u0000model comparison; (4) a strong baseline called the DA Transformer (DaT). DaT\u0000integrates the four-dimensional variational DA prior knowledge into the\u0000Transformer model and outperforms the SOTA in physical state reconstruction,\u0000named 4DVarNet. Furthermore, we exemplify the development of an end-to-end\u0000data-driven weather prediction system by integrating DaT with the prediction\u0000model. Researchers can leverage DABench to develop their models and compare\u0000performance against established baselines, which will benefit the future\u0000advancements of data-driven weather prediction systems. The code is available\u0000on this Github repository and the dataset is available at the Baidu Drive.","PeriodicalId":501166,"journal":{"name":"arXiv - PHYS - Atmospheric and Oceanic Physics","volume":"180 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142215449","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In Bangladesh, a nation vulnerable to climate change, accurately quantifying the risk of extreme weather events is crucial for planning effective adaptation and mitigation strategies. Downscaling coarse climate model projections to finer resolutions is key in improving risk and uncertainty assessments. This work develops a new approach to rainfall downscaling by integrating statistics, physics, and machine learning and applies it to assess Bangladesh's extreme rainfall risk. Our method successfully captures the observed spatial pattern and risks associated with extreme rainfall in the present climate. It also produces uncertainty estimates by rapidly downscaling multiple models in a future climate scenario(s). Our analysis reveals that the risk of extreme rainfall is projected to increase throughout Bangladesh mid-century, with the highest risk in the northeast. The daily maximum rainfall at a 100-year return period is expected to rise by approximately 50 mm per day. However, using multiple climate models also indicates considerable uncertainty in the projected risk.
{"title":"Rapid Statistical-Physical Adversarial Downscaling Reveals Bangladesh's Rising Rainfall Risk in a Warming Climate","authors":"Anamitra Saha, Sai Ravela","doi":"arxiv-2408.11790","DOIUrl":"https://doi.org/arxiv-2408.11790","url":null,"abstract":"In Bangladesh, a nation vulnerable to climate change, accurately quantifying\u0000the risk of extreme weather events is crucial for planning effective adaptation\u0000and mitigation strategies. Downscaling coarse climate model projections to\u0000finer resolutions is key in improving risk and uncertainty assessments. This\u0000work develops a new approach to rainfall downscaling by integrating statistics,\u0000physics, and machine learning and applies it to assess Bangladesh's extreme\u0000rainfall risk. Our method successfully captures the observed spatial pattern\u0000and risks associated with extreme rainfall in the present climate. It also\u0000produces uncertainty estimates by rapidly downscaling multiple models in a\u0000future climate scenario(s). Our analysis reveals that the risk of extreme\u0000rainfall is projected to increase throughout Bangladesh mid-century, with the\u0000highest risk in the northeast. The daily maximum rainfall at a 100-year return\u0000period is expected to rise by approximately 50 mm per day. However, using\u0000multiple climate models also indicates considerable uncertainty in the\u0000projected risk.","PeriodicalId":501166,"journal":{"name":"arXiv - PHYS - Atmospheric and Oceanic Physics","volume":"7 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142215448","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Nicholas Balasus, Daniel J. Jacob, Gabriel Maxemin, Carrie Jenks, Hannah Nesser, Joannes D. Maasakkers, Daniel H. Cusworth, Tia R. Scarpelli, Daniel J. Varon, Xiaolin Wang
We use satellite observations of atmospheric methane from the TROPOMI instrument to estimate total annual methane emissions for 2019-2023 from four large Southeast US landfills with gas collection and control systems. The emissions are on average 6$times$ higher than the values reported by the landfills to the US Greenhouse Gas Reporting Program (GHGRP) which are used by the US Environmental Protection Agency (EPA) for its national Greenhouse Gas Inventory (GHGI). We find increasing emissions over the 2019-2023 period whereas the GHGRP reports a decrease. The GHGRP requires gas-collecting landfills to estimate their annual emissions either with a recovery-first model (estimating emissions as a function of methane recovered) or a generation-first model (estimating emissions from a first-order-decay applied to waste-in-place). All four landfills choose to use the recovery-first model, which yields emissions that are one-quarter of those from the generation-first model and decreasing over 2019-2023, in contrast with the TROPOMI observations. Our TROPOMI estimates for two of the landfills agree with the generation-first model, with increasing emissions over 2019-2023 due to increasing waste-in-place or decreasing methane recovery, and are still higher than the generation-first model for the other two landfills. Further examination of the GHGRP emissions from all reporting landfills in the US shows that the 19% decrease in landfill emissions reported by the GHGI over 2005-2022 reflects an increasing preference for the recovery-first model by the reporting landfills, rather than an actual emission decrease. The generation-first model would imply an increase in landfill emissions over 2013-2022, and this is more consistent with atmospheric observations.
{"title":"Satellite monitoring of annual US landfill methane emissions and trends","authors":"Nicholas Balasus, Daniel J. Jacob, Gabriel Maxemin, Carrie Jenks, Hannah Nesser, Joannes D. Maasakkers, Daniel H. Cusworth, Tia R. Scarpelli, Daniel J. Varon, Xiaolin Wang","doi":"arxiv-2408.10957","DOIUrl":"https://doi.org/arxiv-2408.10957","url":null,"abstract":"We use satellite observations of atmospheric methane from the TROPOMI\u0000instrument to estimate total annual methane emissions for 2019-2023 from four\u0000large Southeast US landfills with gas collection and control systems. The\u0000emissions are on average 6$times$ higher than the values reported by the\u0000landfills to the US Greenhouse Gas Reporting Program (GHGRP) which are used by\u0000the US Environmental Protection Agency (EPA) for its national Greenhouse Gas\u0000Inventory (GHGI). We find increasing emissions over the 2019-2023 period\u0000whereas the GHGRP reports a decrease. The GHGRP requires gas-collecting\u0000landfills to estimate their annual emissions either with a recovery-first model\u0000(estimating emissions as a function of methane recovered) or a generation-first\u0000model (estimating emissions from a first-order-decay applied to\u0000waste-in-place). All four landfills choose to use the recovery-first model,\u0000which yields emissions that are one-quarter of those from the generation-first\u0000model and decreasing over 2019-2023, in contrast with the TROPOMI observations.\u0000Our TROPOMI estimates for two of the landfills agree with the generation-first\u0000model, with increasing emissions over 2019-2023 due to increasing\u0000waste-in-place or decreasing methane recovery, and are still higher than the\u0000generation-first model for the other two landfills. Further examination of the\u0000GHGRP emissions from all reporting landfills in the US shows that the 19%\u0000decrease in landfill emissions reported by the GHGI over 2005-2022 reflects an\u0000increasing preference for the recovery-first model by the reporting landfills,\u0000rather than an actual emission decrease. The generation-first model would imply\u0000an increase in landfill emissions over 2013-2022, and this is more consistent\u0000with atmospheric observations.","PeriodicalId":501166,"journal":{"name":"arXiv - PHYS - Atmospheric and Oceanic Physics","volume":"109 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142215453","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Quantitative forecasting of average rainfall into the next season remains highly challenging, but in some favourable isolated cases may be possible with a series of relatively simple steps. We chose to explore predictions of austral springtime rainfall in SE Australia regions based on the surrounding ocean surface temperatures during the winter. In the first stage, we search for correlations between the target rainfall and both the standard ocean climate indicators as well as the time series of surface temperature data expanded in terms of Empirical Orthogonal Functions (EOFs). In the case of the Indian Ocean, during the winter the dominant EOF shows stronger correlation with the future rainfall than the commonly used Indian Ocean Dipole. Information sources with the strongest correlation to the historical rainfall data are then used as inputs into deep learning artificial neural networks. The resulting hindcasts appear accurate for September and October and less reliable for November. We also attempt to forecast the rainfall in several regions for the coming austral spring.
{"title":"Forecasting seasonal rainfall in SE Australia using Empirical Orthogonal Functions and Neural Networks","authors":"Stjepan Marcelja","doi":"arxiv-2408.10550","DOIUrl":"https://doi.org/arxiv-2408.10550","url":null,"abstract":"Quantitative forecasting of average rainfall into the next season remains\u0000highly challenging, but in some favourable isolated cases may be possible with\u0000a series of relatively simple steps. We chose to explore predictions of austral\u0000springtime rainfall in SE Australia regions based on the surrounding ocean\u0000surface temperatures during the winter. In the first stage, we search for\u0000correlations between the target rainfall and both the standard ocean climate\u0000indicators as well as the time series of surface temperature data expanded in\u0000terms of Empirical Orthogonal Functions (EOFs). In the case of the Indian\u0000Ocean, during the winter the dominant EOF shows stronger correlation with the\u0000future rainfall than the commonly used Indian Ocean Dipole. Information sources\u0000with the strongest correlation to the historical rainfall data are then used as\u0000inputs into deep learning artificial neural networks. The resulting hindcasts\u0000appear accurate for September and October and less reliable for November. We\u0000also attempt to forecast the rainfall in several regions for the coming austral\u0000spring.","PeriodicalId":501166,"journal":{"name":"arXiv - PHYS - Atmospheric and Oceanic Physics","volume":"98 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142215463","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Zili Liu, Hao Chen, Lei Bai, Wenyuan Li, Wanli Ouyang, Zhengxia Zou, Zhenwei Shi
In an era of frequent extreme weather and global warming, obtaining precise, fine-grained near-surface weather forecasts is increasingly essential for human activities. Downscaling (DS), a crucial task in meteorological forecasting, enables the reconstruction of high-resolution meteorological states for target regions from global-scale forecast results. Previous downscaling methods, inspired by CNN and Transformer-based super-resolution models, lacked tailored designs for meteorology and encountered structural limitations. Notably, they failed to efficiently integrate topography, a crucial prior in the downscaling process. In this paper, we address these limitations by pioneering the selective state space model into the meteorological field downscaling and propose a novel model called MambaDS. This model enhances the utilization of multivariable correlations and topography information, unique challenges in the downscaling process while retaining the advantages of Mamba in long-range dependency modeling and linear computational complexity. Through extensive experiments in both China mainland and the continental United States (CONUS), we validated that our proposed MambaDS achieves state-of-the-art results in three different types of meteorological field downscaling settings. We will release the code subsequently.
{"title":"MambaDS: Near-Surface Meteorological Field Downscaling with Topography Constrained Selective State Space Modeling","authors":"Zili Liu, Hao Chen, Lei Bai, Wenyuan Li, Wanli Ouyang, Zhengxia Zou, Zhenwei Shi","doi":"arxiv-2408.10854","DOIUrl":"https://doi.org/arxiv-2408.10854","url":null,"abstract":"In an era of frequent extreme weather and global warming, obtaining precise,\u0000fine-grained near-surface weather forecasts is increasingly essential for human\u0000activities. Downscaling (DS), a crucial task in meteorological forecasting,\u0000enables the reconstruction of high-resolution meteorological states for target\u0000regions from global-scale forecast results. Previous downscaling methods,\u0000inspired by CNN and Transformer-based super-resolution models, lacked tailored\u0000designs for meteorology and encountered structural limitations. Notably, they\u0000failed to efficiently integrate topography, a crucial prior in the downscaling\u0000process. In this paper, we address these limitations by pioneering the\u0000selective state space model into the meteorological field downscaling and\u0000propose a novel model called MambaDS. This model enhances the utilization of\u0000multivariable correlations and topography information, unique challenges in the\u0000downscaling process while retaining the advantages of Mamba in long-range\u0000dependency modeling and linear computational complexity. Through extensive\u0000experiments in both China mainland and the continental United States (CONUS),\u0000we validated that our proposed MambaDS achieves state-of-the-art results in\u0000three different types of meteorological field downscaling settings. We will\u0000release the code subsequently.","PeriodicalId":501166,"journal":{"name":"arXiv - PHYS - Atmospheric and Oceanic Physics","volume":"59 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142215452","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：