Xingming Liang

Radiative Transfer Models (RTMs) are pivotal in remote sensing and data assimilation, providing simulations of satellite-measured radiances and corresponding Jacobians—matrices that quantify how radiative quantities change in response to variations in atmospheric or surface parameters. These Jacobians lie at the heart of variational data assimilation and drive a wide range of inverse modeling applications. Although traditional RTMs yield high-fidelity results, they are computationally intensive, which is increasingly problematic given the high spatial, temporal, and spectral resolutions measured by modern sensors. Recent advances in artificial intelligence (AI) have led to AI-derived “fast” RTMs that substantially accelerate forward predictions, (e.g., brightness temperatures). However, accurately estimating and stabilizing AI-derived Jacobians remains a key challenge. In this paper, we present a new AI-CRTM framework based on a Residual Neural Network (ResNet), augmented by physical constraints using the reference Jacobians from the traditional Community Radiative Transfer Model (CRTM). By leveraging residual connections, the proposed approach mitigates vanishing gradients and captures complex, multi-scale dependencies in radiative transfer physics. Physical constraints are applied during training to align the learned partial derivatives with trusted RTM references, thus ensuring physically consistent Jacobians. The AI_CRTM was initially trained and tested for simulated measurements from the Advanced Technology Microwave Sounder (ATMS) in global clear-sky conditions. Forward-model accuracy is accessed through comparisons with both CRTM-simulated radiances and high-spectral-resolution line-by-line simulations from the MonoRTM model. The results demonstrate that incorporating these constraints, alongside ResNet’s residual structure, significantly improves both forward prediction accuracy and Jacobian fidelity. This hybrid strategy addresses ill-posed issues in AI Jacobian inference, broadens applicability to different atmospheric scenarios, and paves the way for rapid and robust data assimilation.

Space-borne microwave sounding instruments have become vital data sources for weather prediction and climate change studies. Among the various radiometer configurations, the total power microwave radiometer is particularly appealing for current and future operational satellites due to its superior sensitivity and simple design. However, its performance is vulnerable to degradation caused by receiver gain fluctuations, electronic 1/f noise, and other time-varying receiver characteristics. For numerical weather prediction (NWP) users, 1/f noise introduces interchannel correlations, complicating the assimilation of affected observations and reducing their accuracy. Addressing this noise issue in the ground data-processing system is essential to enhance the utility of microwave sounding data. This article focuses on the characterization and mitigation of noise in current and future microwave sounding instruments, with particular emphasis on the impact of 1/f noise. Various methods are applied to quantitatively characterize noise features in both the frequency and time domains. Additionally, the influence of calibration parameters on 1/f noise is analyzed. Based on these findings, we propose a mitigation algorithm for reducing noise during the on-orbit calibration of microwave sounding instruments, aiming to improve the quality of retrieved data for operational use.

The NOAA-21 ATMS, launched in November 2022, collected two-dimensional lunar scan data on March 10, 2023, during its commissioning phase at a Moon phase angle of 34°. The raw data were calibrated using MiCalPS, a microwave calibration and geolocation tool developed at the University of Maryland. Analysis showed that NOAA-21’s lunar antenna gain is generally lower than NOAA-20’s due to sampling rate differences. After correcting for beam pointing errors, disk-averaged lunar brightness temperatures (T^DISK_B,Moon) were derived for 23–183 GHz. These were compared to NOAA-20 data at 0° phase angle and predictions from the microwave lunar radiative transfer model (ML-RTM). The observed model differences were: −3.3 K (K band), 0.01 K (V), 2.0 K (W), −2.3 K (low-G), and 1.6 K (high-G), consistent with predicted phase delay trends. Further observations at varying Moon phases are recommended to enhance MLRTM validation.

The Advanced Technology Microwave Sounder (ATMS) is an important satellite instrument that provides vital data on atmosphere temperature and moisture for weather forecasting, climate research, and help us plan for extreme weather. However, its coarse resolution and angular dependence have long been a challenge for improving image visualization. This study proposes a method to enhance the imagery visualization for ATMS, combining limb correction with artificial intelligence (AI) resolution enhancement. Measurement data from the ATMS onboard NOAA-20 was utilized to train the limb-correction method, which was then validated using newly acquired NOAA-21 ATMS data. The AI resolution enhancement was performed using Enhanced Super-Resolution Generative Adversarial Networks (ESRGAN), which increased the pixel resolution by a factor of four. The high-resolution Advanced Microwave Scanning Radiometer 2 (AMSR2) data served as a reference to initially and quantitatively evaluate the resolution enhancement method. The combined method of limb correction and AI resolution enhancement produced an angular-dependence-free and high-resolution ATMS image, resulting in a significant improvement in image visualization, including surface and atmosphere information, and allows for clear identification of severe weather events. For the swift identification and analysis of tropical cyclones in the upcoming season, as of this writing, this proposed method has been routinely employed to produce high-quality global ATMS images, and these images are showcased and tested in the NOAA internal high-resolution imagery visualization system – JSTAR Mapper. Moreover, concentrated efforts are being made to further enhance these images in preparation for an official release .

Frequent false signals of precipitation in satellite passive microwave retrievals over the Aral Sea have been identified as being caused by an outdated surface database. The database includes the surface type, elevation, and the percentage of the primary surface type in each grid. It was also found that the grid resolution of 1/6 degree (∼18 km at the equator) of the outdated surface data was too coarse to process global precipitation measurement mission microwave imager data, which have a field of view resolution of 31.68 km ² for channels at 89, 166.5, and 183.31 GHz). In this article, we generated a new surface database at a resolution of $0.05^\circ \times 0.05^\circ $ (∼5.6 km at the equator). Using the new surface data, the false precipitation problem above is addressed. At the same time, the retrieval accuracy for other parameters such as total precipitable water over the Aral sea is significantly improved as well. The resolution of the database is suitable for most microwave and infrared sounding measurements of atmospheric temperature and moisture profiles as well as precipitation. By comparing this new surface data against the outdated surface data, we also see the loss of permanent ice in Antarctica and a dramatic reduction of water surface over the Aral sea. Using a 40-year record of remote sensing data, we can observe the steady decrease in size of the Aral sea, as a result of regional water use policies and natural climate change.

In this article, we introduce a fully connected deep neural network algorithm to emulate the Community Cadiative Transfer Model (FCDN_CRTM) simulation of brightness temperatures (BTs) from the Advanced Technology Microwave Sounder (ATMS) channels for clear-sky cases over ocean surfaces. The FCDN_CRTM fine-tuned through three sensitivity experiments with respect to sample-size determination, model separation, and introduction of novel features toward improving the accuracy of the model. In addition to the BT simulation, we evaluated the Jacobians with respect to surface and atmospheric parameters. Atmosphere profiles from the European Centre for Medium-Range Weather Forecasts, sea surface temperature from the Canadian Meteorology Centre, and ATMS sensor data records were used as FCDN_CRTM inputs. In comparison to CRTM, the FCDN_CRTM minus CRTM mean biases were within several hundredths of a Kelvin (K), and the corresponding standard deviations (SDs) were between 0.05 and 0.15 K for all ATMS bands. The accuracies for both mean bias and SD were consistent throughout the evaluation period, which spanned approximately 1 year beyond the period of the FCDN_CRTM training dataset. Furthermore, the model Jacobians generally compared well with CRTM Jacobians in terms of surface temperature, wind speed, air temperature, and (log) water vapor. The performance of the FCDN_CRTM forward and Jacobian model indicate potential for use in data assimilation and physical retrieval systems, such as the NOAA operational Microwave Integrated Retrieval System.

Earth's surface reflectance is an important parameter affecting ultraviolet (UV) and visible (VIS) radiance calculations at the top of the atmosphere because many UV and VIS channels can acquire information about the surface and atmosphere. This article provides the theoretical basis for deriving the surface reflectance from satellite-measured UV and VIS observations at window and lower sounding channels with the help of the community radiative transfer model (CRTM) and collocated atmospheric profiles such as ozone, water vapor, and aerosols. Cirrus cloud may be included in the calculation as long as the observations contain enough reflected radiation from the surface. An explicit equation with three scalar parameters $\alpha $ , $\beta $ , and $\delta $ is obtained for users to calculate Lambertian surface reflectance from the observation. The expressions for the three parameters are somewhat complicated and computationally expansive. We found a simple and smart way that can exactly calculate the three parameters with quasi-linear functions. Numerical experiments using the CRTM simulations have demonstrated the algorithm accuracy for the surface reflectance retrieval better than 2.0E-14. As a case study, measured surface reflectance and the derived surface reflectance over desert from satellite UV measurements are compared. The derived surface reflectance from Suomi National Polar-orbiting Partnership. Visible Infrared Imaging Radiometer Suite (VIIRS) observations and the VIIRS reflectance product are compared as well. In addition, this methodology can also be used to calculate microwave and infrared surface emissivity with scatterings and solar radiation by adding the surface Planck radiance at the surface temperature.

This experiment is the first ultraviolet radiance assimilation for atmospheric ozone in the troposphere and stratosphere. The experiment has provided better understanding of which observations need to be assimilated, what bias correction scheme may be optimal, and how to obtain surface reflectance. A key element is the extension of the Community Radiative Transfer Model (CRTM) to handle fully polarized radiances, which presents challenges in terms of computational resource requirements. In this study, a scalar (unpolarized) treatment of radiances was used. The surface reflectance plays an important role in assimilating the nadir mapper (NM) radiance of the Ozone Mapping and Profiler Suite (OMPS). Most OMPS NM measurements are affected by the surface reflection of solar radiation. We propose a linear spectral reflectance model that can be determined inline by fitting two OMPS NM channel radiances at 347.6 and 371.8 nm because the two channels have near zero sensitivity on atmospheric ozone. Assimilating a transformed reflectance measurement variable, the N value can overcome the difficulty in handling the large dynamic range of radiance and normalized radiance across the spectrum of the OMPS NM. It was found that the error in bias correction, surface reflectance, and neglecting polarization in radiative transfer calculations can be largely mitigated by using the two estimated surface reflectance. This study serves as a preliminary demonstration of direct ultraviolet radiance assimilation for total column ozone in the atmosphere.

The authors would like to make the following correction of