Vitus Benson, Ana Bastos, Christian Reimers, Alexander J. Winkler, Fanny Yang, Markus Reichstein
Accurately describing the distribution of CO$_2$ in the atmosphere with atmospheric tracer transport models is essential for greenhouse gas monitoring and verification support systems to aid implementation of international climate agreements. Large deep neural networks are poised to revolutionize weather prediction, which requires 3D modeling of the atmosphere. While similar in this regard, atmospheric transport modeling is subject to new challenges. Both, stable predictions for longer time horizons and mass conservation throughout need to be achieved, while IO plays a larger role compared to computational costs. In this study we explore four different deep neural networks (UNet, GraphCast, Spherical Fourier Neural Operator and SwinTransformer) which have proven as state-of-the-art in weather prediction to assess their usefulness for atmospheric tracer transport modeling. For this, we assemble the CarbonBench dataset, a systematic benchmark tailored for machine learning emulators of Eulerian atmospheric transport. Through architectural adjustments, we decouple the performance of our emulators from the distribution shift caused by a steady rise in atmospheric CO$_2$. More specifically, we center CO$_2$ input fields to zero mean and then use an explicit flux scheme and a mass fixer to assure mass balance. This design enables stable and mass conserving transport for over 6 months with all four neural network architectures. In our study, the SwinTransformer displays particularly strong emulation skill (90-day $R^2 > 0.99$), with physically plausible emulation even for forward runs of multiple years. This work paves the way forward towards high resolution forward and inverse modeling of inert trace gases with neural networks.
{"title":"Atmospheric Transport Modeling of CO$_2$ with Neural Networks","authors":"Vitus Benson, Ana Bastos, Christian Reimers, Alexander J. Winkler, Fanny Yang, Markus Reichstein","doi":"arxiv-2408.11032","DOIUrl":"https://doi.org/arxiv-2408.11032","url":null,"abstract":"Accurately describing the distribution of CO$_2$ in the atmosphere with\u0000atmospheric tracer transport models is essential for greenhouse gas monitoring\u0000and verification support systems to aid implementation of international climate\u0000agreements. Large deep neural networks are poised to revolutionize weather\u0000prediction, which requires 3D modeling of the atmosphere. While similar in this\u0000regard, atmospheric transport modeling is subject to new challenges. Both,\u0000stable predictions for longer time horizons and mass conservation throughout\u0000need to be achieved, while IO plays a larger role compared to computational\u0000costs. In this study we explore four different deep neural networks (UNet,\u0000GraphCast, Spherical Fourier Neural Operator and SwinTransformer) which have\u0000proven as state-of-the-art in weather prediction to assess their usefulness for\u0000atmospheric tracer transport modeling. For this, we assemble the CarbonBench\u0000dataset, a systematic benchmark tailored for machine learning emulators of\u0000Eulerian atmospheric transport. Through architectural adjustments, we decouple\u0000the performance of our emulators from the distribution shift caused by a steady\u0000rise in atmospheric CO$_2$. More specifically, we center CO$_2$ input fields to\u0000zero mean and then use an explicit flux scheme and a mass fixer to assure mass\u0000balance. This design enables stable and mass conserving transport for over 6\u0000months with all four neural network architectures. In our study, the\u0000SwinTransformer displays particularly strong emulation skill (90-day $R^2 >\u00000.99$), with physically plausible emulation even for forward runs of multiple\u0000years. This work paves the way forward towards high resolution forward and\u0000inverse modeling of inert trace gases with neural networks.","PeriodicalId":501166,"journal":{"name":"arXiv - PHYS - Atmospheric and Oceanic Physics","volume":"20 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142215462","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}
Storm-scale convection-allowing models (CAMs) are an important tool for predicting the evolution of thunderstorms and mesoscale convective systems that result in damaging extreme weather. By explicitly resolving convective dynamics within the atmosphere they afford meteorologists the nuance needed to provide outlook on hazard. Deep learning models have thus far not proven skilful at km-scale atmospheric simulation, despite being competitive at coarser resolution with state-of-the-art global, medium-range weather forecasting. We present a generative diffusion model called StormCast, which emulates the high-resolution rapid refresh (HRRR) model-NOAA's state-of-the-art 3km operational CAM. StormCast autoregressively predicts 99 state variables at km scale using a 1-hour time step, with dense vertical resolution in the atmospheric boundary layer, conditioned on 26 synoptic variables. We present evidence of successfully learnt km-scale dynamics including competitive 1-6 hour forecast skill for composite radar reflectivity alongside physically realistic convective cluster evolution, moist updrafts, and cold pool morphology. StormCast predictions maintain realistic power spectra for multiple predicted variables across multi-hour forecasts. Together, these results establish the potential for autoregressive ML to emulate CAMs -- opening up new km-scale frontiers for regional ML weather prediction and future climate hazard dynamical downscaling.
{"title":"Kilometer-Scale Convection Allowing Model Emulation using Generative Diffusion Modeling","authors":"Jaideep Pathak, Yair Cohen, Piyush Garg, Peter Harrington, Noah Brenowitz, Dale Durran, Morteza Mardani, Arash Vahdat, Shaoming Xu, Karthik Kashinath, Michael Pritchard","doi":"arxiv-2408.10958","DOIUrl":"https://doi.org/arxiv-2408.10958","url":null,"abstract":"Storm-scale convection-allowing models (CAMs) are an important tool for\u0000predicting the evolution of thunderstorms and mesoscale convective systems that\u0000result in damaging extreme weather. By explicitly resolving convective dynamics\u0000within the atmosphere they afford meteorologists the nuance needed to provide\u0000outlook on hazard. Deep learning models have thus far not proven skilful at\u0000km-scale atmospheric simulation, despite being competitive at coarser\u0000resolution with state-of-the-art global, medium-range weather forecasting. We\u0000present a generative diffusion model called StormCast, which emulates the\u0000high-resolution rapid refresh (HRRR) model-NOAA's state-of-the-art 3km\u0000operational CAM. StormCast autoregressively predicts 99 state variables at km\u0000scale using a 1-hour time step, with dense vertical resolution in the\u0000atmospheric boundary layer, conditioned on 26 synoptic variables. We present\u0000evidence of successfully learnt km-scale dynamics including competitive 1-6\u0000hour forecast skill for composite radar reflectivity alongside physically\u0000realistic convective cluster evolution, moist updrafts, and cold pool\u0000morphology. StormCast predictions maintain realistic power spectra for multiple\u0000predicted variables across multi-hour forecasts. Together, these results\u0000establish the potential for autoregressive ML to emulate CAMs -- opening up new\u0000km-scale frontiers for regional ML weather prediction and future climate hazard\u0000dynamical downscaling.","PeriodicalId":501166,"journal":{"name":"arXiv - PHYS - Atmospheric and Oceanic Physics","volume":"9 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142215454","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}
Recently, Transformers have gained traction in weather forecasting for their capability to capture long-term spatial-temporal correlations. However, their complex architectures result in large parameter counts and extended training times, limiting their practical application and scalability to global-scale forecasting. This paper aims to explore the key factor for accurate weather forecasting and design more efficient solutions. Interestingly, our empirical findings reveal that absolute positional encoding is what really works in Transformer-based weather forecasting models, which can explicitly model the spatial-temporal correlations even without attention mechanisms. We theoretically prove that its effectiveness stems from the integration of geographical coordinates and real-world time features, which are intrinsically related to the dynamics of weather. Based on this, we propose LightWeather, a lightweight and effective model for station-based global weather forecasting. We employ absolute positional encoding and a simple MLP in place of other components of Transformer. With under 30k parameters and less than one hour of training time, LightWeather achieves state-of-the-art performance on global weather datasets compared to other advanced DL methods. The results underscore the superiority of integrating spatial-temporal knowledge over complex architectures, providing novel insights for DL in weather forecasting.
{"title":"LightWeather: Harnessing Absolute Positional Encoding to Efficient and Scalable Global Weather Forecasting","authors":"Yisong Fu, Fei Wang, Zezhi Shao, Chengqing Yu, Yujie Li, Zhao Chen, Zhulin An, Yongjun Xu","doi":"arxiv-2408.09695","DOIUrl":"https://doi.org/arxiv-2408.09695","url":null,"abstract":"Recently, Transformers have gained traction in weather forecasting for their\u0000capability to capture long-term spatial-temporal correlations. However, their\u0000complex architectures result in large parameter counts and extended training\u0000times, limiting their practical application and scalability to global-scale\u0000forecasting. This paper aims to explore the key factor for accurate weather\u0000forecasting and design more efficient solutions. Interestingly, our empirical\u0000findings reveal that absolute positional encoding is what really works in\u0000Transformer-based weather forecasting models, which can explicitly model the\u0000spatial-temporal correlations even without attention mechanisms. We\u0000theoretically prove that its effectiveness stems from the integration of\u0000geographical coordinates and real-world time features, which are intrinsically\u0000related to the dynamics of weather. Based on this, we propose LightWeather, a\u0000lightweight and effective model for station-based global weather forecasting.\u0000We employ absolute positional encoding and a simple MLP in place of other\u0000components of Transformer. With under 30k parameters and less than one hour of\u0000training time, LightWeather achieves state-of-the-art performance on global\u0000weather datasets compared to other advanced DL methods. The results underscore\u0000the superiority of integrating spatial-temporal knowledge over complex\u0000architectures, providing novel insights for DL in weather forecasting.","PeriodicalId":501166,"journal":{"name":"arXiv - PHYS - Atmospheric and Oceanic Physics","volume":"42 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142215467","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}
Growing evidence is found in observations and numerical modelling of the importance of steep seafloor topography for turbulent diapycnal mixing leading to redistribution of suspended matter and nutrients, especially in waters with abundant internal tides. One of the remaining questions is the extent of turbulent mixing away from and above nearly flat topography, which is addressed in this paper. Evaluated are observations from an opportunistic, week-long mooring of high-resolution temperature sensors above a small seafloor slope in about 1200 m water depth of the Eastern Mediterranean. The environment has weak tides, so that near-inertial motions and -shear dominate internal waves. Vertical displacement shapes suggest instabilities to represent locally generated turbulent overturns, rather than partial salinity-compensated intrusions dispersed isopycnally from turbulence near the slope. This conclusion is supported by the duration of instabilities, as all individual overturns last shorter than the mean buoyancy period and sequences of overturns last shorter than the local inertial period. The displacement shapes are more erratic than observed in stronger stratified waters in which shear drives turbulence, and better correspond with predominantly buoyancy-driven convection-turbulence. This convection-turbulence is confirmed from spectral information, generally occurring dominant close to the seafloor and only in weakly stratified layers well above it. Mean turbulence values are 10-100 times smaller than found above steep ocean topography, but 10 times larger than found in the open-ocean interior.
{"title":"Intrusions and turbulent mixing above a small Eastern Mediterranean seafloor-slope","authors":"Hans van Haren","doi":"arxiv-2408.07992","DOIUrl":"https://doi.org/arxiv-2408.07992","url":null,"abstract":"Growing evidence is found in observations and numerical modelling of the\u0000importance of steep seafloor topography for turbulent diapycnal mixing leading\u0000to redistribution of suspended matter and nutrients, especially in waters with\u0000abundant internal tides. One of the remaining questions is the extent of\u0000turbulent mixing away from and above nearly flat topography, which is addressed\u0000in this paper. Evaluated are observations from an opportunistic, week-long\u0000mooring of high-resolution temperature sensors above a small seafloor slope in\u0000about 1200 m water depth of the Eastern Mediterranean. The environment has weak\u0000tides, so that near-inertial motions and -shear dominate internal waves.\u0000Vertical displacement shapes suggest instabilities to represent locally\u0000generated turbulent overturns, rather than partial salinity-compensated\u0000intrusions dispersed isopycnally from turbulence near the slope. This\u0000conclusion is supported by the duration of instabilities, as all individual\u0000overturns last shorter than the mean buoyancy period and sequences of overturns\u0000last shorter than the local inertial period. The displacement shapes are more\u0000erratic than observed in stronger stratified waters in which shear drives\u0000turbulence, and better correspond with predominantly buoyancy-driven\u0000convection-turbulence. This convection-turbulence is confirmed from spectral\u0000information, generally occurring dominant close to the seafloor and only in\u0000weakly stratified layers well above it. Mean turbulence values are 10-100 times\u0000smaller than found above steep ocean topography, but 10 times larger than found\u0000in the open-ocean interior.","PeriodicalId":501166,"journal":{"name":"arXiv - PHYS - Atmospheric and Oceanic Physics","volume":"24 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142215468","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}
Capitalizing on the recent availability of ERA5 monthly averaged long-term data records of mean atmospheric and climate fields based on high-resolution reanalysis, deep-learning architectures offer an alternative to physics-based daily numerical weather predictions for subseasonal to seasonal (S2S) and annual means. A novel Deep UNet++-based Ensemble (DUNE) neural architecture is introduced, employing multi-encoder-decoder structures with residual blocks. When initialized from a prior month or year, this architecture produced the first AI-based global monthly, seasonal, or annual mean forecast of 2-meter temperatures (T2m) and sea surface temperatures (SST). ERA5 monthly mean data is used as input for T2m over land, SST over oceans, and solar radiation at the top of the atmosphere for each month of 40 years to train the model. Validation forecasts are performed for an additional two years, followed by five years of forecast evaluations to account for natural annual variability. AI-trained inference forecast weights generate forecasts in seconds, enabling ensemble seasonal forecasts. Root Mean Squared Error (RMSE), Anomaly Correlation Coefficient (ACC), and Heidke Skill Score (HSS) statistics are presented globally and over specific regions. These forecasts outperform persistence, climatology, and multiple linear regression for all domains. DUNE forecasts demonstrate comparable statistical accuracy to NOAA's operational monthly and seasonal probabilistic outlook forecasts over the US but at significantly higher resolutions. RMSE and ACC error statistics for other recent AI-based daily forecasts also show superior performance for DUNE-based forecasts. The DUNE model's application to an ensemble data assimilation cycle shows comparable forecast accuracy with a single high-resolution model, potentially eliminating the need for retraining on extrapolated datasets.
{"title":"DUNE: A Machine Learning Deep UNet++ based Ensemble Approach to Monthly, Seasonal and Annual Climate Forecasting","authors":"Pratik Shukla, Milton Halem","doi":"arxiv-2408.06262","DOIUrl":"https://doi.org/arxiv-2408.06262","url":null,"abstract":"Capitalizing on the recent availability of ERA5 monthly averaged long-term\u0000data records of mean atmospheric and climate fields based on high-resolution\u0000reanalysis, deep-learning architectures offer an alternative to physics-based\u0000daily numerical weather predictions for subseasonal to seasonal (S2S) and\u0000annual means. A novel Deep UNet++-based Ensemble (DUNE) neural architecture is\u0000introduced, employing multi-encoder-decoder structures with residual blocks.\u0000When initialized from a prior month or year, this architecture produced the\u0000first AI-based global monthly, seasonal, or annual mean forecast of 2-meter\u0000temperatures (T2m) and sea surface temperatures (SST). ERA5 monthly mean data\u0000is used as input for T2m over land, SST over oceans, and solar radiation at the\u0000top of the atmosphere for each month of 40 years to train the model. Validation\u0000forecasts are performed for an additional two years, followed by five years of\u0000forecast evaluations to account for natural annual variability. AI-trained\u0000inference forecast weights generate forecasts in seconds, enabling ensemble\u0000seasonal forecasts. Root Mean Squared Error (RMSE), Anomaly Correlation\u0000Coefficient (ACC), and Heidke Skill Score (HSS) statistics are presented\u0000globally and over specific regions. These forecasts outperform persistence,\u0000climatology, and multiple linear regression for all domains. DUNE forecasts\u0000demonstrate comparable statistical accuracy to NOAA's operational monthly and\u0000seasonal probabilistic outlook forecasts over the US but at significantly\u0000higher resolutions. RMSE and ACC error statistics for other recent AI-based\u0000daily forecasts also show superior performance for DUNE-based forecasts. The\u0000DUNE model's application to an ensemble data assimilation cycle shows\u0000comparable forecast accuracy with a single high-resolution model, potentially\u0000eliminating the need for retraining on extrapolated datasets.","PeriodicalId":501166,"journal":{"name":"arXiv - PHYS - Atmospheric and Oceanic Physics","volume":"98 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142215469","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}
Haoyu Qin, Yungang Chen, Qianchuan Jiang, Pengchao Sun, Xiancai Ye, Chao Lin
Deep Learning based Weather Prediction (DLWP) models have been improving rapidly over the last few years, surpassing state of the art numerical weather forecasts by significant margins. While much of the optimization effort is focused on training curriculum to extend forecast range in the global context, two aspects remains less explored: limited area modeling and better backbones for weather forecasting. We show in this paper that MetMamba, a DLWP model built on a state-of-the-art state-space model, Mamba, offers notable performance gains and unique advantages over other popular backbones using traditional attention mechanisms and neural operators. We also demonstrate the feasibility of deep learning based limited area modeling via coupled training with a global host model.
{"title":"MetMamba: Regional Weather Forecasting with Spatial-Temporal Mamba Model","authors":"Haoyu Qin, Yungang Chen, Qianchuan Jiang, Pengchao Sun, Xiancai Ye, Chao Lin","doi":"arxiv-2408.06400","DOIUrl":"https://doi.org/arxiv-2408.06400","url":null,"abstract":"Deep Learning based Weather Prediction (DLWP) models have been improving\u0000rapidly over the last few years, surpassing state of the art numerical weather\u0000forecasts by significant margins. While much of the optimization effort is\u0000focused on training curriculum to extend forecast range in the global context,\u0000two aspects remains less explored: limited area modeling and better backbones\u0000for weather forecasting. We show in this paper that MetMamba, a DLWP model\u0000built on a state-of-the-art state-space model, Mamba, offers notable\u0000performance gains and unique advantages over other popular backbones using\u0000traditional attention mechanisms and neural operators. We also demonstrate the\u0000feasibility of deep learning based limited area modeling via coupled training\u0000with a global host model.","PeriodicalId":501166,"journal":{"name":"arXiv - PHYS - Atmospheric and Oceanic Physics","volume":"21 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142215464","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 ProchaskaAffiliate of the Ocean Sciences Department, University of California, Santa CruzDepartment of Astronomy & Astrophysics, UCSCKavli IPMUScripps Institution of Oceanography, University of California, San Diego, Robert J. FrouinScripps Institution of Oceanography, University of California, San Diego
Since 1978, sensors on remote-sensing satellites have provided global, multi-band images at optical wavelengths to assess ocean color. In parallel, sophisticated radiative transfer models account for attenuation and emission by the Earth's atmosphere and ocean, thereby estimating the water-leaving radiance or and remote-sensing reflectance Rrs. From these Rrs measurements, estimates of the absorption and scattering by seawater are inferred. We emphasize an inherent, physical degeneracy in the radiative transfer equation that relates Rrs to the absorption and backscattering coefficients a and b_b, aka inherent optical properties (IOPs). Because Rrs depends solely on the ratio of b_b to a, meaning one cannot retrieve independent functions for the non-water IOPs, a_nw and b_bnw, without a priori knowledge. Moreover, water generally dominates scattering at blue wavelengths and absorption at red wavelengths, further limiting retrievals of IOPs in the presence of noise. We demonstrate that all previous and current multi-spectral satellite observations lack the statistical power to measure more than 3 parameters total to describe a_nw and b_bnw. Due to the ubiquitous exponential-like absorption by color dissolved organic matter at short wavelengths (l<500nm), multi-spectral Rrs do not permit the detection of phytoplankton absorption a_ph without very strict priors. Furthermore, such priors lead to biased and uncertain retrievals of a_ph. Hyperspectral observations may recover a 4th and possibly 5th parameter describing only one or two aspects of the complexity of a_ph. These results cast doubt on decades of literature on IOP retrievals, including estimates of phytoplankton growth and biomass. We further conclude that NASA/PACE will greatly enhance our ability to measure the phytoplankton biomass of Earth, but challenges remain in resolving the IOPs.
{"title":"On the Peril of Inferring Phytoplankton Properties from Remote-Sensing Observations","authors":"J. Xavier ProchaskaAffiliate of the Ocean Sciences Department, University of California, Santa CruzDepartment of Astronomy & Astrophysics, UCSCKavli IPMUScripps Institution of Oceanography, University of California, San Diego, Robert J. FrouinScripps Institution of Oceanography, University of California, San Diego","doi":"arxiv-2408.06149","DOIUrl":"https://doi.org/arxiv-2408.06149","url":null,"abstract":"Since 1978, sensors on remote-sensing satellites have provided global,\u0000multi-band images at optical wavelengths to assess ocean color. In parallel,\u0000sophisticated radiative transfer models account for attenuation and emission by\u0000the Earth's atmosphere and ocean, thereby estimating the water-leaving radiance\u0000or and remote-sensing reflectance Rrs. From these Rrs measurements, estimates\u0000of the absorption and scattering by seawater are inferred. We emphasize an\u0000inherent, physical degeneracy in the radiative transfer equation that relates\u0000Rrs to the absorption and backscattering coefficients a and b_b, aka inherent\u0000optical properties (IOPs). Because Rrs depends solely on the ratio of b_b to a,\u0000meaning one cannot retrieve independent functions for the non-water IOPs, a_nw\u0000and b_bnw, without a priori knowledge. Moreover, water generally dominates\u0000scattering at blue wavelengths and absorption at red wavelengths, further\u0000limiting retrievals of IOPs in the presence of noise. We demonstrate that all\u0000previous and current multi-spectral satellite observations lack the statistical\u0000power to measure more than 3 parameters total to describe a_nw and b_bnw. Due\u0000to the ubiquitous exponential-like absorption by color dissolved organic matter\u0000at short wavelengths (l<500nm), multi-spectral Rrs do not permit the detection\u0000of phytoplankton absorption a_ph without very strict priors. Furthermore, such\u0000priors lead to biased and uncertain retrievals of a_ph. Hyperspectral\u0000observations may recover a 4th and possibly 5th parameter describing only one\u0000or two aspects of the complexity of a_ph. These results cast doubt on decades\u0000of literature on IOP retrievals, including estimates of phytoplankton growth\u0000and biomass. We further conclude that NASA/PACE will greatly enhance our\u0000ability to measure the phytoplankton biomass of Earth, but challenges remain in\u0000resolving the IOPs.","PeriodicalId":501166,"journal":{"name":"arXiv - PHYS - Atmospheric and Oceanic Physics","volume":"18 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142215465","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 recent years, several studies have been made in which atmospheric and oceanic data were used to decompose horizontal velocity statistics into a rotational component, associated with vertical vorticity, and a divergent component, associated with horizontal divergence. Making the assumption of statistical homogeneity in a horizontal plane, this can be accomplished by relating the rotational and divergent components of the difference between the velocities at two points to the corresponding longitudinal and transverse components, where the longitudinal and transverse directions are parallel respectively perpendicular to the line between the points. In previous studies, the decomposition has most often been made under the assumption of statistical isotropy. Some attempts have also been made to analyse the anisotropic problem. We derive the full anisotropic equations relating the rotational, divergent and the rotational-divergent components of the second order structure functions to the longitudinal, transverse and longitudinal-transverse components and solve the equations analytically. We also derive some results for third order structure functions, with special focus on the components associated with cyclone-anticyclone asymmetry. Based on the analysis of these components and results from previous analyses of aircraft data, it is concluded that there is an exclusively rotational flow component that is giving rise to strong dominance of cyclonic motions in the upper troposphere and a strong dominance of anticyclonic motions in the lower stratosphere in the range of scales from ten to one thousand km
{"title":"Helmholtz decompositions of horizontal structure functions including components associated with cyclone-anticyclone symmetry breaking","authors":"Erik Lindborg","doi":"arxiv-2408.05734","DOIUrl":"https://doi.org/arxiv-2408.05734","url":null,"abstract":"In recent years, several studies have been made in which atmospheric and\u0000oceanic data were used to decompose horizontal velocity statistics into a\u0000rotational component, associated with vertical vorticity, and a divergent\u0000component, associated with horizontal divergence. Making the assumption of\u0000statistical homogeneity in a horizontal plane, this can be accomplished by\u0000relating the rotational and divergent components of the difference between the\u0000velocities at two points to the corresponding longitudinal and transverse\u0000components, where the longitudinal and transverse directions are parallel\u0000respectively perpendicular to the line between the points. In previous studies,\u0000the decomposition has most often been made under the assumption of statistical\u0000isotropy. Some attempts have also been made to analyse the anisotropic problem.\u0000We derive the full anisotropic equations relating the rotational, divergent and\u0000the rotational-divergent components of the second order structure functions to\u0000the longitudinal, transverse and longitudinal-transverse components and solve\u0000the equations analytically. We also derive some results for third order\u0000structure functions, with special focus on the components associated with\u0000cyclone-anticyclone asymmetry. Based on the analysis of these components and\u0000results from previous analyses of aircraft data, it is concluded that there is\u0000an exclusively rotational flow component that is giving rise to strong\u0000dominance of cyclonic motions in the upper troposphere and a strong dominance\u0000of anticyclonic motions in the lower stratosphere in the range of scales from\u0000ten to one thousand km","PeriodicalId":501166,"journal":{"name":"arXiv - PHYS - Atmospheric and Oceanic Physics","volume":"2 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142215466","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}
Firstly, by establishing a prediction model for global sea-level rise and calculating with Maple, it is shown that the global sea-level rise rate in 2009 is 2.68 mm/a. The height and rate of global sea-level rise will be about 9.11 cm and 3.22 mm/a in 2020. Based on the study and the actual land subsidence in Shanghai Lingang New City, the rate of relative sea-level rise near Lingang New City is calculated to be 12.68 mm/a in 2009. Then, through setting up the extrapolation prediction model with a linear trend term and a significant tidal cycle, the rise rate of average sea-level near Lingang New City was predicted. The result showed it will be 0.33 mm/a in 2020.
{"title":"Prediction of Sea Level Rise near Shanghai","authors":"Yi Zheng","doi":"arxiv-2408.06387","DOIUrl":"https://doi.org/arxiv-2408.06387","url":null,"abstract":"Firstly, by establishing a prediction model for global sea-level rise and\u0000calculating with Maple, it is shown that the global sea-level rise rate in 2009\u0000is 2.68 mm/a. The height and rate of global sea-level rise will be about 9.11\u0000cm and 3.22 mm/a in 2020. Based on the study and the actual land subsidence in\u0000Shanghai Lingang New City, the rate of relative sea-level rise near Lingang New\u0000City is calculated to be 12.68 mm/a in 2009. Then, through setting up the\u0000extrapolation prediction model with a linear trend term and a significant tidal\u0000cycle, the rise rate of average sea-level near Lingang New City was predicted.\u0000The result showed it will be 0.33 mm/a in 2020.","PeriodicalId":501166,"journal":{"name":"arXiv - PHYS - Atmospheric and Oceanic Physics","volume":"21 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142215470","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}
Laura H. Yang, Daniel J. Jacob, Haipeng Lin, Ruijun Dang, Kelvin H. Bates, James D. East, Katherine R. Travis, Drew C. Pendergrass, Lee T. Murray
Deploying hydrogen technologies is one option to reduce energy carbon dioxide emissions, but recent studies have called attention to the indirect climate implications of fugitive hydrogen emissions. We find that biases in hydroxyl (OH) radical concentrations and reactivity in current atmospheric chemistry models may cause a 20% overestimate of the hydrogen Global Warming Potential (GWP). A better understanding of OH chemistry is critical for reliable estimates of the hydrogen GWP.
{"title":"Model underestimates of OH reactivity cause overestimate of hydrogen's climate impact","authors":"Laura H. Yang, Daniel J. Jacob, Haipeng Lin, Ruijun Dang, Kelvin H. Bates, James D. East, Katherine R. Travis, Drew C. Pendergrass, Lee T. Murray","doi":"arxiv-2408.05127","DOIUrl":"https://doi.org/arxiv-2408.05127","url":null,"abstract":"Deploying hydrogen technologies is one option to reduce energy carbon dioxide\u0000emissions, but recent studies have called attention to the indirect climate\u0000implications of fugitive hydrogen emissions. We find that biases in hydroxyl\u0000(OH) radical concentrations and reactivity in current atmospheric chemistry\u0000models may cause a 20% overestimate of the hydrogen Global Warming Potential\u0000(GWP). A better understanding of OH chemistry is critical for reliable\u0000estimates of the hydrogen GWP.","PeriodicalId":501166,"journal":{"name":"arXiv - PHYS - Atmospheric and Oceanic Physics","volume":"10 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141940511","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：