Atmospheric Fields Research Articles

Downscaling of Historical Wind Fields over Switzerland Using Generative Adversarial Networks

Abstract Near-surface wind is difficult to estimate using global numerical weather and climate models, because airflow is strongly modified by underlying topography, especially that of a country such as Switzerland. In this article, we use a statistical approach based on deep learning and a high-resolution digital elevation model to spatially downscale hourly near-surface wind fields at coarse resolution from ERA5 reanalysis from their original 25-km grid to a 1.1-km grid. A 1.1-km-resolution wind dataset for 2016–20 from the operational numerical weather prediction model COSMO-1 of the national weather service MeteoSwiss is used to train and validate our model, a generative adversarial network (GAN) with gradient penalized Wasserstein loss aided by transfer learning. The results are realistic-looking high-resolution historical maps of gridded hourly wind fields over Switzerland and very good and robust predictions of the aggregated wind speed distribution. Regionally averaged image-specific metrics show a clear improvement in prediction relative to ERA5, with skill measures generally better for locations over the flatter Swiss Plateau than for Alpine regions. The downscaled wind fields demonstrate higher-resolution, physically plausible orographic effects, such as ridge acceleration and sheltering, that are not resolved in the original ERA5 fields. Significance Statement Statistical downscaling, which increases the resolution of atmospheric fields, is widely used to refine the outputs of global reanalysis and climate models, most commonly for temperature and precipitation. Near-surface winds are strongly modified by the underlying topography, generating local flow conditions that can be very difficult to estimate. This study develops a deep learning model that uses local topographic information to spatially downscale hourly near-surface winds from their original 25-km resolution to a 1.1-km grid over Switzerland. Our model produces realistic high-resolution gridded wind fields with expected orographic effects but performs better in flatter regions than in mountains. These downscaled fields are useful for impact assessment and decision-making in regions where global reanalysis data at coarse resolution may be the only products available.

Rolling vs. seasonal PMF: real-world multi-site and synthetic dataset comparison

Abstract. Particulate matter (PM) has become a major concern in terms of human health and climate impact. In particular, the source apportionment (SA) of organic aerosols (OA) present in submicron particles (PM1) has gained relevance as an atmospheric research field due to the diversity and complexity of its primary sources and secondary formation processes. Moreover, relatively simple but robust instruments such as the Aerosol Chemical Speciation Monitor (ACSM) are now widely available for the near-real-time online determination of the composition of the non-refractory PM1. One of the most used tools for SA purposes is the source-receptor positive matrix factorisation (PMF) model. Even though the recently developed rolling PMF technique has already been used for OA SA on ACSM datasets, no study has assessed its added value compared to the more common seasonal PMF method using a practical approach yet. In this paper, both techniques were applied to a synthetic dataset and to nine European ACSM datasets in order to spot the main output discrepancies between methods. The main advantage of the synthetic dataset approach was that the methods' outputs could be compared to the expected “true” values, i.e. the original synthetic dataset values. This approach revealed similar apportionment results amongst methods, although the rolling PMF profile's adaptability feature proved to be advantageous, as it generated output profiles that moved nearer to the truth points. Nevertheless, these results highlighted the impact of the profile anchor on the solution, as the use of a different anchor with respect to the truth led to significantly different results in both methods. In the multi-site study, while differences were generally not significant when considering year-long periods, their importance grew towards shorter time spans, as in intra-month or intra-day cycles. As far as correlation with external measurements is concerned, rolling PMF performed better than seasonal PMF globally for the ambient datasets investigated here, especially in periods between seasons. The results of this multi-site comparison coincide with the synthetic dataset in terms of rolling–seasonal similarity and rolling PMF reporting moderate improvements. Altogether, the results of this study provide solid evidence of the robustness of both methods and of the overall efficiency of the recently proposed rolling PMF approach.

Spatial and seasonal variations of near-inertial kinetic energy in the upper South China Sea: Role of synoptic atmospheric systems

The spatial distribution and seasonal variation of near-inertial kinetic energy (NIKE) in the upper ocean of the South China Sea (SCS) are examined using the Global 1/12° Analysis Hybrid Coordinate Ocean Model and the Navy Coupled Ocean Data Assimilation reanalysis product. The annual mean NIKE in the upper SCS is characterized by a rapid decay with depth and a southwestward decrease from the west of Luzon Island to the southern SCS, reflecting the pattern of near-inertial energy input by the atmospheric wind field. Owing to the changes of near-inertial wind forcing as well as the mixed layer depth, NIKE in the upper SCS exhibits a pronounced seasonal cycle. The magnitude of mixed layer NIKE averaged in the SCS in November-January is found to be approximately twice of that in April-June. Further analysis shows that the variation of near-inertial wind forcing in the SCS can, to a large extent, be explained by northerly cold surges during the winter SCS monsoon and strong tropical cyclones from the Tropical Pacific. In addition, the wind reversal during the monsoon transitional period and the low-level convergence formed by the elevated terrain of islands also contribute to oceanic near-inertial energy in the SCS.

Downscaling multi-model climate projection ensembles with deep learning (DeepESD): contribution to CORDEX EUR-44

Abstract. Deep learning (DL) has recently emerged as an innovative tool to downscale climate variables from large-scale atmospheric fields under the perfect-prognosis (PP) approach. Different convolutional neural networks (CNNs) have been applied under present-day conditions with promising results, but little is known about their suitability for extrapolating future climate change conditions. Here, we analyze this problem from a multi-model perspective, developing and evaluating an ensemble of CNN-based downscaled projections (hereafter DeepESD) for temperature and precipitation over the European EUR-44i (0.5∘) domain, based on eight global circulation models (GCMs) from the Coupled Model Intercomparison Project Phase 5 (CMIP5). To our knowledge, this is the first time that CNNs have been used to produce downscaled multi-model ensembles based on the perfect-prognosis approach, allowing us to quantify inter-model uncertainty in climate change signals. The results are compared with those corresponding to an EUR-44 ensemble of regional climate models (RCMs) showing that DeepESD reduces distributional biases in the historical period. Moreover, the resulting climate change signals are broadly comparable to those obtained with the RCMs, with similar spatial structures. As for the uncertainty of the climate change signal (measured on the basis of inter-model spread), DeepESD preserves the uncertainty for temperature and results in a reduced uncertainty for precipitation. To facilitate further studies of this downscaling approach, we follow FAIR principles and make publicly available the code (a Jupyter notebook) and the DeepESD dataset. In particular, DeepESD is published at the Earth System Grid Federation (ESGF), as the first continental-wide PP dataset contributing to CORDEX (EUR-44).

Intraseasonal Melting of Northern Barents Sea Ice Forced by Circumpolar Clockwise-Propagating Atmospheric Waves during Early Summer

Abstract Arctic sea ice intraseasonal variation (ISV) is crucial for understanding and predicting atmospheric subseasonal variations over the middle and high latitudes but unclear. Sea ice concentration (SIC) over the northern Barents Sea (NBS) features large ISV during the melting season (April–July). Based on the observed SIC, this study finds that the NBS SIC ISV in the melting season is dominated by 30–60-day periodicity. The composite analysis, using 34 significant 30–60-day sea ice melting events during 1989–2017, demonstrates that 30–60-day circumpolar clockwise-propagating atmospheric waves (CCPW) are concurrent with the NBS SIC ISV, which features zonal wavenumber 1 along 65°N and a typical quasi-barotropic structure. Further analysis finds that the 30–60-day surface air temperature (SAT) evidently leads the SIC variations by nearly 6 days over the NBS, which is primarily caused by low-level meridional thermal advection linked with the 30–60-day CCPW. The positive anomalies of the downward sensible heat and longwave radiative fluxes, caused by the increased SAT and atmospheric moisture, play the dominant roles in melting the sea ice on the 30–60-day time scale over the NBS. The increased atmospheric moisture is mainly ascribed to the increased horizontal moisture advection influence by the 30–60-day CCPW. This study strongly suggests that the atmospheric ISV is a crucial precursor for NBS sea ice intraseasonal changes in boreal summer, and more accurate subseasonal predictions of atmospheric circulation, temperature, and moisture are indispensable for improving sea ice subseasonal prediction over the Arctic region. Significance Statement Northern Barents Sea (NBS) sea ice intraseasonal variation (ISV) is crucial for understanding mid- to high-latitude climate variations as well as new trans-Arctic shipping predictions but lacks solid knowledge. This study found that the 30–60-day variation is the dominant ISV periodicity of NBS sea ice change during summer, which is essentially modulated by circumpolar clockwise-propagating atmospheric waves. The atmospheric wave-induced meridional thermal advection modulates the surface temperature and atmospheric moisture, causes the changes of downward sensible heat and longwave radiative fluxes, and eventually dominantly regulates the 30–60-day sea ice variations. The mechanism of sea ice ISV strongly suggests that accurately predicting the atmospheric fields is indispensable for obtaining more accurate sea ice subseasonal prediction.

Quiet Sun Center to Limb Variation of the Linear Polarization Observed by CLASP2 Across the Mg ii h and k Lines

The CLASP2 (Chromospheric LAyer Spectro-Polarimeter 2) sounding rocket mission was launched on 2019 April 11. CLASP2 measured the four Stokes parameters of the Mg ii h and k spectral region around 2800 Å along a 200″ slit at three locations on the solar disk, achieving the first spatially and spectrally resolved observations of the solar polarization in this near-ultraviolet region. The focus of the work presented here is the center-to-limb variation of the linear polarization across these resonance lines, which is produced by the scattering of anisotropic radiation in the solar atmosphere. The linear polarization signals of the Mg ii h and k lines are sensitive to the magnetic field from the low to the upper chromosphere through the Hanle and magneto-optical effects. We compare the observations to theoretical predictions from radiative transfer calculations in unmagnetized semiempirical models, arguing that magnetic fields and horizontal inhomogeneities are needed to explain the observed polarization signals and spatial variations. This comparison is an important step in both validating and refining our understanding of the physical origin of these polarization signatures, and also in paving the way toward future space telescopes for probing the magnetic fields of the solar upper atmosphere via ultraviolet spectropolarimetry.

Exploring a Directional Measurement Method of Three-Dimensional Electric Field Intensity in the Atmosphere

The traditional process of measuring a three-dimensional (3D) electric field in the atmosphere may suffer from limited accuracy of the electric field component intensity due to multiple factors, such as the 3D electric field sensor’s pose change due to the sensor-carrier motion and the non-parallel positions between the 3D electric compass and field sensor’s measurement axes. Thus, this paper proposes a modified 3D electric field directional decomposition algorithm to overcome this problem. Specifically, the developed solution utilizes a parameter identification and fine-tuning scheme that solves the inaccurate measurement problem caused by the non-coaxial 3D positions between the electric field sensor and the compass. The measurement results highlight that the proposed method eliminates the instability of the electric field intensity measurement errors caused by the 3D electric field sensor changes because the measurement accuracy of the corrected electric field component intensity does not change with the sensor’s rotation angle. Our technique reduces the measurement error of the electric field intensity originating from the non-parallel position 3D electronic compass and field sensor measurement axes, further improving the measurement accuracy of the atmospheric field intensity.

Probing Magnetic Fields in Protoplanetary Disk Atmospheres through Polarized Near-IR Light Scattered by Aligned Grains

Magnetic fields play essential roles in protoplanetary disks. Magnetic fields in the disk atmosphere are of particular interest, as they are connected to the wind-launching mechanism. In this work, we study the polarization of the light scattered off of magnetically aligned grains in the disk atmosphere, focusing on the deviation of the polarization orientation from the canonical azimuthal direction, which may be detectable in near-IR polarimetry with instruments such as VLT/SPHERE. We show with a simple disk model that the polarization can even be oriented along the radial (rather than azimuthal) direction, especially in highly inclined disks with toroidally dominated magnetic fields. This polarization reversal is caused by the anisotropy in the polarizability of aligned grains and is thus a telltale sign of such grains. We show that the near-IR light is scattered mostly by μm-sized grains or smaller at the τ = 1 surface and such grains can be magnetically aligned if they contain superparamagnetic inclusions. For comparison with observations, we generate synthetic maps of the ratios of U ϕ /I and Q ϕ /I, which can be used to infer the existence of (magnetically) aligned grains through a negative Q ϕ (polarization reversal) and/or a significant level of U ϕ /I. We show that two features observed in the existing data, an asymmetric distribution of U ϕ with respect to the disk minor axis and a spatial distribution of U ϕ that is predominantly positive or negative, are incompatible with scattering by spherical grains in an axisymmetric disk. They provide indirect evidence for scattering by aligned nonspherical grains.

Tropical Cyclone Wind Field Reconstruction and Validation Using Measurements from SFMR and SMAP Radiometer

Accurate information on tropical cyclone position, intensity, and structure is critical for storm surge prediction. Atmospheric reanalysis datasets can provide gridded, full coverage, long-term and multi-parameter atmospheric fields for the research on the impact of tropical cyclones on the upper ocean, which effectively makes up for the uneven temporal and spatial distribution of satellite remote sensing and in situ data. However, the reanalysis data cannot accurately describe characteristic parameters of tropical cyclones, especially in high wind conditions. In this paper, the performance of the tropical cyclone representation in ERA5 (European Centre for Medium-Range Weather Forecasts Reanalysis 5th Generation) is investigated and analyzed with respect to IBTrACS (International Best Track Archive for Climate Stewardship) during the period 2018–2020. Comparisons demonstrate that ERA5 winds significantly underestimate the maximum wind speed during the tropical cyclones (>30 m/s) compared to those provided by IBTrACS. An effective wind reconstruction method is examined to enhance tropical cyclone intensity representation in reanalysis data in 94 cases of 31 tropical cyclones 2018–2020. The reconstructed wind speeds are in good agreement with the SFMR (Stepped Frequency Microwave Radiometer) measured data and SMAP (Soil Moisture Active Passive) L-band radiometer remotely sensed measurements. The proposed wind reconstruction method can effectively improve the accuracy of the tropical cyclone representation in ERA5, and will benefit from the establishment of remote sensing satellite retrieval model and the forcing fields of the ocean model.

Relocation of GATE from the Pacific to the Atlantic

Abstract This article documents historically the planning of the Global Atmospheric Research Program’s (GARP) Atlantic Tropical Experiment (GATE), the largest atmospheric field program of all time. In its earliest planning, GATE was called the Tropical Meteorological Experiment (TROMEX) and was designed to be in the tropical western Pacific. For reasons including concerns of the U.S. Department of Defense, the international project was relocated to the tropical Atlantic and renamed GATE.

Wind tunnel experimental calibration of hemispherical 7-hole probe pressure–velocity parametric equation

The multi-hole probe can measure the velocity and three-dimensional direction of the flow field at the same time, so it is often used to measure the three-dimensional flow field. Compared with other flow field measuring instruments, the multi-hole probe has stronger environmental adaptability and stability, and can better measure the three-dimensional flow field of the middle atmosphere. Therefore, a hemispherical 7-hole probe was designed, a pressure–velocity parameterized equation was established based on the theory of flow around a sphere, and a new calibration method was developed based on this. The calibration is carried out in a subsonic low speed wind tunnel, multiple combinations of flow parameters (inflow velocity and flow angles) are adjusted during the calibration. The results are compared with the numerical simulation results, both are quite close, with a speed measurement deviation of less than 5% and an angle measurement deviation of less than 1°. Our results establish the practicality of the hemispherical 7-hole probe and the simplified calibration procedure, both of which improve calibration efficiency and lower probe calibration costs.

The Climate Characteristics of the Northeast China Cold Vortex against the Background of Global Warming

In this study, by using the ERA5 data of the atmospheric circulation field that was re-analyzed by the ECMWF (European Centre for Medium-Range Weather Forecasts), we revealed the features of the Northeast China Cold Vortex (NCCV) from 1950 to 2020 (including active days, occurrence time of NCCV processes, and process durations). This study focused on a comparative analysis of the differences in the NCCV’s climate characteristics in the cold and warm periods to help future predictions. The results revealed the following: From 1950 to 2020, the NCCV occurred 2961 times on 9782 days. The average annual occurrence time of NCCV processes, annual average of cold vortex days, and average process duration of the NCCVs were 41.7 times, 137.8 days, and 3.6 days, respectively. These indicators of the NCCVs showed an increasing trend, but the trend was not significant. The NCCVs occurred most frequently in May, followed by June, and were located at the southernmost point in June. Therefore, it had the most active days and a relatively long process duration in May and June, significantly impacting Northeast China. During the cold period (1950–1980), the annual occurrence time of NCCV processes, number of cold vortex days, and the process duration of the NCCVs all showed an increasing trend, while in the warm period, these showed a decreasing trend. In addition, the durations of the NCCVs decreased significantly in the warm period, which indicated that the NCCV processes continued to weaken after climate warming. During the warm period (1981–2020), the frequency and active days of the NCCVs throughout the year and most months increased, and its general location was more southerly than in the cold period. Moreover, the annual average occurrence time of NCCV processes, number of active days, and average duration of the NCCV in the warm period were more than those in the cold period. Finally, the NCCVs continued for longer in autumn and winter than in spring and summer, and the durations of the NCCVs increased in warm periods.

Development of Moist Singular Vectors in GRAPES-GEPS and a Preliminary Evaluation

ABSTRACT In this study, moist singular vector (MSV) was developed based on GRAPES-GEPS (Global/Regional Assimilation and Prediction System – Global Ensemble Prediction System), the adjoint model of large-scale condensation and cumulus deep convection in GRAPES-4DVar (Four-dimensional variational assimilation). Five consecutive days of numerical experiments were performed for a preliminary evaluation of MSV. The singular values, horizontal distribution structure, spread of MSVs perturbation and its influence on the ensemble prediction were compared for each group of tests. The results showed that in the middle and high latitudes of the northern and southern hemispheres, the addition of both linearized moist physical processes increased the spread of the mid- and low-level SVs, but the linearized large-scale condensation (LC) process plays a leading role in the structure of MSV. The analysis of ensemble forecast shows the inclusion of moist linearized physical processes led to a greater effect of MSV on the rainfall levels of 10 and 25 mm and a slight improvement in anomaly correlation coefficient (ACC) of the atmospheric circulation field, and more obvious improvement due to linearized large-scale condensation. In the future, continuous multi-year testing and tropical-specific analyses are required for operation.

Q-SAT for atmosphere and gravity field detection: Design, mission and preliminary results

Q-SAT, a small spherical satellite designed and developed by Tsinghua university, was successfully launched on Aug 6, 2020 at Jiuquan satellite launch center in China. The mission of Q-Sat is the joint measurement of long-wavelength Earth gravity field and upper atmosphere density, which is of great significance for improving spacecraft orbit prediction and promoting the development of satellite gravity measurement technology. The satellite is designed in spherical shape innovatively and deploys high dynamic dual-frequency GPS as the main payload. On-orbit data shows that the atmospheric density detection precision is in the order of 10−14kg/m3 and the gravity recovery precision achieves the level of 30 orders. All mission success criteria and objectives were achieved. This paper presents the design, space mission, key technologies, science capabilities and preliminary results of the satellite.

Over-the-Horizon Channel of Radio Communication at VHF Band via Artificial Plasma Clouds

Metal cesium (Cs) vapor, released at an altitude in the upper atmosphere, generates an artificial space plasma cloud with high electron density. It can act as a scatter to realize radio communication across an over-the-horizon distance. A full-life model of radio propagation via an artificial plasma cloud is established based on its space–time evolution in this article. In addition, this model is verified by the sounding data of the rocket release experiment carried out in the United States. The multipath effect of plasma cloud scattering is studied by using ray tracing, and then, the coherent bandwidth is estimated for this spatial channel. Also, the theoretical channel model of artificial plasma cloud is established by using its space–time evolution and atmospheric wind field parameters. Considering common modulation modes and receiver noise levels, the potential communication performance is evaluated by using the proposed channel model, such as the transmission rate, the bit error rate (BER), and the available time for building up and maintaining communication links. It is shown that the maximal communication rate is up to 200 kb/s, BER overall maintains below 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">−7</sup> , and the available time for a single release is about 30–60 min; for the case of this article, they are release mass of 2–10 kg, communication distances of 400–1200 km, and radio frequencies of 30–70 MHz at the VHF band. This demonstrates that the artificial channel can propagate radio waves at a much higher frequency, wider bandwidth, and higher rate than those of HF communication. Hopefully, it could be a solution to overcome the intolerable shortcoming and adverse effects on HF communication due to ionospheric disturbances especially in cases of maritime, emergency, and minimum communications

On the Efficiency of the Integration of Satellite Microwave Radiometric Means for Monitoring Atmospheric Humidity Fields in the Zones of Hurricane Origin

On the Efficiency of the Integration of Satellite Microwave Radiometric Means for Monitoring Atmospheric Humidity Fields in the Zones of Hurricane Origin

This Looks Like That There: Interpretable Neural Networks for Image Tasks When Location Matters

Abstract We develop and demonstrate a new interpretable deep learning model specifically designed for image analysis in Earth system science applications. The neural network is designed to be inherently interpretable, rather than explained via post hoc methods. This is achieved by training the network to identify parts of training images that act as prototypes for correctly classifying unseen images. The new network architecture extends the interpretable prototype architecture of a previous study in computer science to incorporate absolute location. This is useful for Earth system science where images are typically the result of physics-based processes, and the information is often geolocated. Although the network is constrained to only learn via similarities to a small number of learned prototypes, it can be trained to exhibit only a minimal reduction in accuracy relative to noninterpretable architectures. We apply the new model to two Earth science use cases: a synthetic dataset that loosely represents atmospheric high and low pressure systems, and atmospheric reanalysis fields to identify the state of tropical convective activity associated with the Madden–Julian oscillation. In both cases, we demonstrate that considering absolute location greatly improves testing accuracies when compared with a location-agnostic method. Furthermore, the network architecture identifies specific historical dates that capture multivariate, prototypical behavior of tropical climate variability. Significance Statement Machine learning models are incredibly powerful predictors but are often opaque “black boxes.” The how-and-why the model makes its predictions is inscrutable—the model is not interpretable. We introduce a new machine learning model specifically designed for image analysis in Earth system science applications. The model is designed to be inherently interpretable and extends previous work in computer science to incorporate location information. This is important because images in Earth system science are typically the result of physics-based processes, and the information is often map based. We demonstrate its use for two Earth science use cases and show that the interpretable network exhibits only a small reduction in accuracy relative to black-box models.

Performance Evaluation of the RANS Models in Predicting the Pollutant Concentration Field within a Compact Urban Setting: Effects of the Source Location and Turbulent Schmidt Number

Computational Fluid Dynamics (CFD) is used to accurately model and predict the dispersion of a passive scalar in the atmospheric wind flow field within an urban setting. The Mock Urban Setting Tests (MUST) experiment was recreated in this work to test and evaluate various modeling settings and to form a framework for reliable representation of dispersion flow in compact urban geometries. Four case studies with distinct source locations and configurations are modeled using Reynolds-Averaged Navier–Stokes (RANS) equations with ANSYS CFX. The performance of three widely suggested closure models of standard k−ε, RNG k−ε, and SST k−ω is assessed by calculating and interpreting the statistical performance metrics with a specific emphasis on the effects of the source locations. This work demonstrates that the overprediction of the turbulent kinetic energy by the standard k−ε counteracts the general underpredictions by RANS in geometries with building complexes. As a result, the superiority of the standard k−ε in predicting the scalar concentration field over the two other closures in all four cases is observed, with SST k−ω showing the most discrepancies with the field measurements. Additionally, a sensitivity study is also conducted to find the optimum turbulent Schmidt number (Sct) using two approaches of the constant and locally variable values.

Calculation theory and experiment verification of sedimentation potential of the complex particle system

In this paper, the equations for calculating the sedimentation potential of complex particle systems composed of solid-liquid mixtures are established, and the multifactorial coupling effects are also taken into consideration, especially different types and shapes of particles in the system, suspended particle concentration, ion concentration, solid-fluid interaction, as well as the coupling effect of the electric field induced by the sedimentation current, etc. In this context, the theoretical calculation results are proved to be consistent with the test results. The equations established in the paper are of substantial value in exploring the principles of particle behavior control in the fields of the chemical industry, metallurgy, pharmacy, air pollution, water environment treatment, etc. Moreover, this paper further discovers the sedimentation potential characteristics of the "time-switch" function of the uniform particle system, and based on this finding, the logical calculation unit is constructed, which can be expected to have a wide range of potential applications in the fields of hydrology, atmosphere, water quality, and other intelligent monitoring.

The Influence of Sub‐Footprint Cloudiness on Three‐Dimensional Horizontal Wind From Geostationary Hyperspectral Infrared Sounder Observations

AbstractRadiance measurements from a geostationary hyperspectral infrared sounder (GeoHIS) with high temporal resolution not only provide a continuous weather cube of atmospheric temperature and moisture information at different pressure levels, but also enable derivation of three‐dimensional (3D) horizontal winds by tracking atmospheric water vapor features. However, GeoHIS radiances are influenced by sub‐footprint cloudiness, which needs to be considered in tracking the moisture features for deriving the atmospheric wind fields. By combining the collocated high spatial resolution cloud information from an imager onboard the same platform, the 3D horizontal wind retrievals can be improved, and the influence of sub‐footprint cloudiness on winds can be quantified for better applications. Using data from the Advanced Geostationary Radiation Imager (AGRI) and Geostationary Interferometric Infrared Sounder onboard the same experimental geostationary satellite Fengyun‐4A, it is found that 3D horizontal wind retrievals can be derived under both clear and partially clear skies with reasonable accuracy. Sub‐footprint cloud information provides noticeable improvement in wind retrievals; higher/lower clouds have more/less influence while thicker/thinner clouds have more/less influence, respectively, on the wind product. The sub‐footprint cloudiness (cloud‐top pressure and cloud coverage) provides a good indication of the quality flag for quantitative applications of the 3D horizontal wind product.