Guizhou Provincial Meteorological Bureau

Abstract Autumn precipitation plays a crucial role in agricultural irrigation and hydropower generation in southwestern China (SWC) and the Yangtze River Basin (YRB), yet its long-term changes and underlying mechanisms remain unclear. Based on the ERA5 reanalysis and station observation data, this study reveals significant declines in autumn precipitation across the SWC and YRB regions from 1973 to 2023. These declines are attributed to weakened water vapor transports from the Bay of Bengal and the South China Sea to China land areas. Further investigation indicates that increasing surface pressure over the Tibetan Plateau and decreasing pressure over the northwestern Pacific in the past several decades have led to a decreased ocean-to-land pressure gradient over East Asia during the autumn. This change in pressure gradient induces easterly and northerly trends in water vapor fluxes over East Asia, reducing moisture transport from the tropical oceans and resulting in decreased precipitation in SWC and the YRB. Both statistical analyses and model sensitivity experiments demonstrate that these pressure changes are associated with the tropical Pacific and Atlantic sea surface temperature (SST) warmings. Specifically, warm SSTs in the western Pacific trigger northeastward-propagating Rossby waves, increasing (decreasing) surface pressure over the Tibetan Plateau (northwestern Pacific). Meanwhile, warm SSTs in the tropical North Atlantic can influence the upstream Rossby wave patterns, lowering surface pressure over the northwestern Pacific. These findings improve our understanding of the changes and variability in autumn precipitation across China, providing important implications for predicting and managing regional droughts under global warming.

Abstract. Cloud vertical structure plays a central role in regulating Earth’s radiation balance and hydrological cycle, yet it remains poorly represented in weather and climate models due to limited high-resolution observations. Using a newly established national network of 80 Ka-band cloud radars, we provide the first high-spatiotemporal-resolution characterization of cloud vertical structure across China for 2024 and quantify its thermodynamic, dynamical, and land-surface controls. An improved retrieval algorithm accounting for height-dependent radar sensitivity and clutter suppression is applied to derive cloud boundaries. The national annual mean cloud occurrence frequency is 56.7 %, dominated by single-layer clouds (34.7 %), followed by two-layer (14.7 %) and multi-layer clouds (7.1 %). Single-layer clouds prevail over arid northwestern China, whereas multi-layer clouds are more frequent in humid southeastern regions. Cloud base height exhibits strong seasonality, with higher values in summer and lower values in winter, and distinctly lower bases over the Tibetan Plateau. Diurnally, summer clouds show a pronounced afternoon peak between 3 and 9 km, while winter clouds are mainly confined below 3 km with a near-sunrise maximum. Thermodynamic conditions exert primary control on cloud vertical development. Higher low-level humidity favors deeper clouds and higher tops, whereas stronger lower-tropospheric stability suppresses vertical growth. Wind shear generally limits cloud depth, though moderate shear may enhance organization under unstable conditions. Land-surface characteristics further modulate cloud base height, with higher bases over barren land and lower bases over forests. These results provide national-scale observational benchmarks for improving cloud parameterizations in numerical models.

This study conducts heavy rainfall identification tasks based on satellite data, focusing on the issue that classical Stacking ensemble methods rely on the implicit weight learning mechanism of meta-learner, which results in a lack of specificity in base model weight allocation. To address this, this work proposes the Enhance Stacking (EN-Stacking) framework, with improvements implemented from two key technical dimensions: establishing a sample weight generation mechanism and introducing an attention network as a weight constraint module. Through the collaborative optimization of base model training and sample weight generation, as well as attention network training and dynamic weight allocation for base models, this framework effectively improves the interpretability and recognition accuracy of the heavy rainfall identification model, and provides a more reliable computer technology solution for satellite data-driven extreme weather identification.

Cardiovascular and cerebrovascular diseases are critical public health challenges influenced by environmental and meteorological factors. Understanding the association between these factors and disease incidence can provide valuable insights for disease prevention and control.This study analyzed data from Anshun City, western Guizhou, collected between January 2018 and December 2022. A Distributed Lag Non-linear Model (DLNM) was employed to evaluate the lagged and non-linear effects of meteorological variables (e.g., temperature, precipitation, wind speed) and air pollutants (e.g., PM2.5, SO2) on the incidence of cardiovascular and cerebrovascular diseases. Covariates such as seasonality and time trends were included to adjust for confounding effects.The results revealed significant associations between meteorological factors, air pollution, and disease incidence. Increased precipitation and SO2 concentrations significantly elevated the risk of cardiovascular and cerebrovascular diseases, particularly at a lag of 25–30 days (e.g., RR for SO2 = 1.19, 95% CI: 1.10–1.28). Conversely, higher average and maximum wind speeds demonstrated a protective effect (e.g., RR for maximum wind speed = 0.70, 95% CI: 0.62–0.78). Seasonal patterns and temperature variations further influenced disease incidence.These findings highlight the complex interactions between meteorological factors and air pollution in influencing cardiovascular and cerebrovascular disease risk. The study provides evidence for targeted public health interventions and emphasizes the importance of incorporating meteorological and environmental data into disease prevention strategies.

Abstract Heat waves (HWs), intensifying under climate change, critically modulate planetary boundary‐layer (PBL) turbulence through poorly constrained mechanisms. Leveraging unique radar wind profiler network measurements across three Beijing during the record‐breaking 2023 summer (16 HW days), we quantify the turbulence dissipation rate (ε) variations under anticyclone driven HWs (hereafter called Type AC, 40% dominance). The mean ε in the PBL during HWs is elevated by ∼55%, demonstrating heat‐amplified turbulence. Divergent forcing regimes emerges–surface‐air temperature difference (Ts−Ta) governs PBL turbulence generation while vertical wind shear (VWS) dominates mechanical mixing aloft. Intriguingly, clouds play a dual role: they enhance VWS‐induced turbulence under normal conditions but reduce heat driven turbulence. These findings establish the first observational evidence of synoptic‐scale thermal‐dynamic decoupling in urban PBLs during extreme heat, providing mechanistic insights for improving megacity air quality forecasting and heat‐stress resilience strategies.

<strong class="journal-contentHeaderColor">Abstract.</strong> Low-level atmospheric turbulence plays a critical role in cloud dynamics and aviation safety. Nevertheless, height-resolved turbulence profiles remain scarce, largely owing to observational challenges. By leveraging collocated radar wind profiler (RWP) and radiosonde observations from 29 stations across China during 2023, a high-resolution dataset of low-level turbulence-related parameters are generated based on spectral width method. This dataset includes squared Brunt&ndash;Vaisala frequency (<em>N²</em>), turbulent dissipation rate (&epsilon;), and vertical eddy diffusivity (&kappa;), inner scale (<em>l₀</em>), and buoyancy length scale (<em>L_B</em>), which are provided twice daily at 00 and 12 UTC with a vertical resolution of 120 m, covering altitudes from 0.12 km to 3.0 km above ground level. Spatial analysis reveals significant regional disparities in turbulence-related parameters across China, where &epsilon;, &kappa; and <em>L_B</em> are higher in northwest and north China compared to south China, while <em>N²</em> and <em>l₀</em> display an inverse spatial pattern. This contrasting geographical distributions suggest distinct atmospheric instability across China. In terms of seasonality, turbulence-related variables showed maxima during spring and summer. Vertical profiles characteristics show distinct altitudinal dependencies, &epsilon;, <em>L_B</em> and &kappa; exhibit progressive attenuation with altitude, while <em>N²</em> and <em>l₀</em> increase with height. Statistical analysis indicates that &epsilon; and &kappa; follow log-normal distributions, whereas <em>l₀</em> and <em>L_B</em> align with Gamma distributions. This dataset is publicly accessible <a href="https://doi.org/10.5281/zenodo.14959025">https://doi.org/10.5281/zenodo.14959025</a> (Meng and Guo, 2025), which provides crucial insights into the fine-scale structural evolution of low-level turbulence. The preliminary findings based on the dataset have great implications for improving our understanding of pre-storm environment and conducting scientific planning and guiding of low-level flight routes in the emerging low-altitude economy in China.

<strong class="journal-contentHeaderColor">Abstract.</strong> Low-level convergence and cyclonic circulation are one of the most important dynamic variables in governing the initiation and development of convective storms. Our ability to obtain high-resolution horizontal divergence and vertical vorticity profiles, nevertheless, remains limited largely due to the lack of vertical wind observations. To fill this data gap, a high-density mesonet consisting of six radar wind profilers (RWP) sites has been operated in Beijing, which allowed for continuous observations of the three-dimensional winds with high vertical resolution. This paper aims to produce a temporally continuous horizontal divergence and vertical vorticity dataset at the vertical resolution of 120 m, which are derived from horizontal winds measured by the RWP mesonet in Beijing by using the triangle method. This dataset is generated at intervals of 6-minute for the whole year of 2023, covering the altitude range of 0&ndash;5 km. The dynamic variables from RWP mesonet are found to scatter sharply, as opposed to those from ERA5 that are concentrated around zero, especially at the high altitudes. Particularly, the negative divergence and positive vorticity are detected in the low-level troposphere up to 1 h in advance of the occurrence of rainfall events, and their magnitudes are increasingly becoming greater when the time comes closer to the rainfall onset, exhibiting the key role that the dataset plays in rainfall nowcasting. This is indicative of, to some extent, the effectiveness of high-resolution divergence and vorticity dataset in Beijing. The dataset is publicly available at <a href="https://doi.org/10.5281/zenodo.14176969" target="_blank" rel="noopener">https://doi.org/10.5281/zenodo.14176969</a> (Guo et al., 2024a), which is of significance for a multitude of scientific research and applications, including convection initiation, air quality forecasting, among others. Therefore, the findings highlight the urgent need of exploiting the dynamic variables from the RWP mesonet measurements to better characterize the pre-storm environment.

Accurate photovoltaic (PV) power forecasting is crucial for stable grid integration, particularly under rapidly changing weather conditions. This study presents an ultra-short-term forecasting model that integrates sky imager data and meteorological radar data, achieving significant improvements in forecasting accuracy. By dynamically tracking cloud movement and estimating cloud coverage, the model enhances performance under both clear and cloudy conditions. Over an 8-day evaluation period, the average forecasting accuracy improved from 67.26% to 77.47% (+15%), with MSE reduced by 39.2% (from 481.5 to 292.6), R2 increased from 0.724 to 0.855, NSE improved from 0.725 to 0.851, and Theil’s U reduced from 0.42 to 0.32. Notable improvements were observed during abrupt weather transitions, particularly on 1 July and 3 July, where the combination of MCR and sky imager data demonstrated superior adaptability. This integrated approach provides a robust foundation for advancing ultra-short-term PV power forecasting.