A global network of ground-based instruments has been using moonlight to measure atmospheric aerosols at night, producing the world's largest such dataset. Employing an empirical correction method that accounts for the Moon's complex and variable brightness results in nighttime measurements that agree closely with daytime measurements during day/night transition intervals under stable aerosol conditions. 

This paper describes the Multi-angle Imaging SpectroRadiometer (MISR) Research Product Algorithm (MRPA) and analyzes its performance. The MRPA an aerosol retrieval algorithm based on the MISR aerosol Research Algorithm (RA), which explored and introduced innovations into MISR aerosol retrievals. Previously, the RA ran on a limited, case-by-case basis. The MRPA is an automated version optimized so users can create 1.1 km-resolution, quality aerosol-type results for most of the MISR data record.

Air pollution can change rapidly above cities, while ground monitors show conditions at only one height. We used a laser-based instrument to distinguish urban pollution particles from dust and estimate how fine-particle pollution changed with height over Beijing. Testing against independent monitoring data showed that accounting for particle type reduced errors by 38 %. This approach can improve understanding of pollution layers and their movement.

MOD07_L2 lacks full validation across China. Using 18-year sounding data (74 stations), we evaluated its temperature accuracy, lower tropospheric stability (LTS), and inversion (TIL) detection. Nighttime accuracy exceeds daytime; upper layers outperform lower. Negative biases exist at 850/920 hPa, especially daytime. LTS is better by day (r=0.74) than night (r=0.57). TIL detection is low (6.6% day, 13.1% night). This guides MOD07_L2 use in related fields.

China’s Fengyun-4M, planned for launch in late 2026, will be the first geostationary satellite to carry a passive microwave instrument, opening a new era of weather observation. To better use its more than one hundred channels, we developed a retrieval method and an optimized channel setup for all-sky temperature profile retrieval over ocean. The results indicate that this algorithm shows low errors while reducing computing time, which is a preferred choice for future operational use.

This study evaluates the location accuracy of a European-wide lightning detection network using lightning events near tall structures across Europe (2012–2025). A clustering approach identifies strikes likely hitting structures, enabling large-scale validation. Results show a median accuracy of ~125 m, consistent with previous studies, and reveal spatial patterns and temporal improvement due to network upgrades. The method provides a scalable proxy for performance assessment.

To mitigate N₂O emissions, the processes behind the emissions need to be well understood. Because of this, we developed a new measurement setup for precise and simultaneously fast airborne in-situ N₂O measurements. We developed an empirical water vapor correction enabling measurements in regions with high humidity and found no significant influence of different flight parameters on the measurements during two campaigns. This setup can help to measure N₂O in humid and undersampled regions.

Fog threatens transportation but is hard to monitor. Ground sensors see only single points and satellites miss vertical detail. We developed a radar-based method to estimate visibility inside fog based on how fog droplets block light. The method finds water content and droplet sizes from radar signals, then calculates how far one can see. It provides three-dimensional fog monitoring without extra instruments.

Lightning is a vital meteorological indicator, an Essential Climate Variable, and a key factor in atmospheric chemistry. EUMETSAT’s Meteosat Third Generation Lightning Imager (MTG-LI), operational since July 2024, is Europe’s first space-based geostationary lightning location system devoted to lightning diagnosis and characterization. This manuscript assesses its performance over 18 months, serving as a foundational reference for users.

We validated the GHGSat fleet against a large set of precisely measured emissions from 2021 to 2026 and built a model linking detection probability to leak rate, wind, retrieval noise, and resolution. In typical conditions the satellites detect a leak of about 100 kilograms per hour half the time and quantify its rate accurately. This sets a clear benchmark for how reliably satellites detect and measure methane leaks, which is essential for trustworthy emissions reporting and mitigation.

This paper presents an ice hydrometeor retrieval algorithm for (sub)millimeter-wave radiometers. The algorithm combines Bayesian Monte Carlo Integration with the Optimal Estimation Method, extending retrieval capability beyond the limitations of the a priori database. Experiments using simulations and airborne observations show that the algorithm effectively reduces discrepancies between observations and simulations and improves the accuracy of retrieved ice cloud microphysical properties.

The Meteosat Third Generation Lightning Imager (MTG-LI) is the first European space mission for monitoring lightning activity over Europe, Africa, and their surrounding areas. During its first year of observations from July 2024 to June 2025, MTG-LI detected a total of 595 × 10⁶ lightning flashes. The highest lightning frequency was found in central Africa. MTG-LI is an excellent instrument for monitoring total lightning activity on operational and climatological scales.

Droplet-freezing experiments help us study substances that trigger ice formation in the atmosphere, but existing ways to show uncertainty or compare results are lacking. We adapt confidence intervals used in medical time-to-even studies, then test them using realistic simulations. They work well across most of the freezing range, though less well at its limits. We also compare tests for differences between curves and show that the best choice depends on how the curves differ.

Detecting and quantifying airborne pollutants near the ground surface and over inaccessible terrain, such as wetlands or unpaved roads, poses a challenge for ground-based measurements, such as trucks. Small drones equipped with precise and rapid-response laser-based sensors can conveniently fly over such areas and collect data. Here we report a sensor for natural gas detection and show that it can accurately quantify leak rates that may be missed by other mobile detection methods.

This paper proposes a new method for detecting aerosols. The method combines atmospheric detection LiDAR technology and intelligent driving technology. Obstacle avoidance constraints and physical constraints during cruise detection motion are fully considered. Experiments are also conducted for three different application scenarios: continuous vertical detection, scanning detection, and unmanned intelligent cruise detection.

Consistent cloud detection across lidar instruments and sites remains challenging. We develop a transferable artificial intelligence method using multi-wavelength lidar signals to improve cloud boundary detection with limited expert labels. Tests at independent sites and comparisons with weather balloon observations demonstrate its reliability. It will be deployed in the China Aerosol Raman Lidar Network (CARLNET) to support reliable automated cloud and aerosol processing.

We introduce an AI-based approach to identify unwanted signals in weather radar images using a neural network. To address the challenge of acquiring sufficient amounts of labeled radar images, the network is trained on synthetically generated radar images. By evaluating the segmentation performance of the trained network on experimentally measured radar images with expert-labeled ground truth, we demonstrate that it outperforms a state-of-the-art method currently used at Deutscher Wetterdienst.

Satellite estimates of airborne particles can be inaccurate when light reflected by land is poorly estimated. We developed a method for China's DaQi-1 satellite that uses machine learning to correct this land signal while retaining a physics-based calculation. Comparisons with ground observations showed reduced errors in spring and autumn–winter. The method also reduced errors the following spring without retraining, suggesting more reliable monitoring over complex land surfaces.

The Radon Tracer Method (RTM) is a top-down approach to estimate greenhouse gas emissions. While simple in principle, incorrect use can complicate interpretation of results. Based on observations from a range of contrasting sites, this article reviews the underlying assumptions and key considerations for applying the RTM. It also introduces the concept of coupling RTM analyses with nocturnal stability classification, to reduce uncertainty of fetch estimates and improve interpretation of results.

Global satellite navigation system occultation data is one of the important sources for weather forecasting. As more and more occultation missions are launched, it's becoming really important to understand how these constellations work together. This study found that combining data from 12 constellations greatly increases time and space coverage, but some missions are still essential for observing tropical regions. These findings provide guidance for future satellite planning.