Global warming drives increased rainfall variability, fundamentally because of enlarged moisture holding capacity of the atmosphere. But what does this increase mean in practical terms for rainfall spectrum? Here, we explore the link between changes in rainfall variability and rainfall spectrum and project hydroclimate volatility changes over eastern China, using a combination of long-term observed daily precipitation records and multiple decadal convection-permitting climate simulations. The increase in rainfall variability is reflected in a shift in the rainfall spectrum—from light to more intense precipitation—indicating heightened hazards of both droughts and flash floods. These wider swings between wet and dry extremes are driven by more frequent occurrences of high daily convective available potential energy and convective inhibition values, and intensified by stronger East Asian summer monsoon circulations. Using a combined measure of precipitation intensity and dry spell duration, we project a 29.4% increase in the precipitation volatility index by the end of the 21st century under a high emissions scenario. Our findings highlight the growing importance of rainfall variability in shaping regional hydroclimate volatility and underscore the need to accelerate a shift in water management strategies toward integrated drought and flood risk management.

BACKGROUND:Fungal keratitis (FK) is a vision-threatening corneal infection whose seasonal and long-term dynamics may be shaped by climate change and air pollution, yet exact environmental impact still unclear. This study the effects of meteorological and air pollution variables on FK incidence in Beijing, China, from 2000 to 2019 and developed an environmental risk index for prospective guidance. METHODS:A retrospective hospital-based time-series analysis of microbiologically confirmed FK cases at a tertiary eye center was conducted. Monthly meteorological variables (air temperature, precipitation, relative humidity, wind speed, ultraviolet irradiance) and air pollutants (SO₂, NO, NO₂, NOx, CO, O₃, PM₂.₅; 2006-2019) were compiled. Annual FK counts were forecast using a generalized additive model (GAM). Spearman correlation and Distributed Lag Non-linear Model (DLNM) were employed to analyze associations and lagged effects of environmental variables on FK. Fungal Active Index (FAI) was developed using the Random Forest model to predict infection risks. RESULTS:FK incidence increased over time, with GAM forecasting continued growth to ∼301 cases annually by 2029. DLNM identified significant non-linear, delayed effects of multiple meteorological factors and pollutants. High temperature increased FK risk (cumulative RR 1.79, 75th vs. median over lag 0-6 months), while precipitation and wind speed showed complex risk patterns across lags. Men, young adults, and middle-aged individuals were identified as the most vulnerable groups. The FAI demonstrated a moderate correlation with monthly fungal infection cases (ρ=0.867, P < 0.001). CONCLUSIONS:Climate and air pollution have non-linear, delayed effects on FK, varying by pathogen type and patient demographics. The FAI serves as a predictive tool for real-time guidance on fungal infection risks.

High-resolution simulation of urban carbon emissions is critical for achieving national carbon-peaking and carbon-neutrality targets. Previous studies that rely on sectoral fossil-fuel emission inventories often fail to capture the spatial heterogeneity and dynamic evolution of CO2 fluxes within cities. In this study, we develop the Natural–Anthropogenic Carbon Estimation Scheme (NACES) by fusing multi-source data, and map urban CO2 fluxes to 100 m grid cells to perform hourly CO2 flux simulations for four megacities: Beijing, Shanghai, Guangzhou, and Shenzhen. NACES reproduces the temporal variability and absolute magnitudes of CO2 fluxes well at eight eddy-covariance (EC) sites, with R2 reaching 0.81 at the Beijing Shuangqing station. While exhibiting broadly consistent large-scale trends with ODIAC (Open-source Data Inventory for Anthropogenic CO2) and EDGAR (Emissions Database for Global Atmospheric Research) inventory products, NACES can effectively capture 74.66%–82.93% of diurnal carbon-flux variability in megacities and distinctly uncover critical emission hotspots that conventional kilometer-scale inventories have masked. All four cities display pronounced diurnal variability in carbon fluxes, yet differ markedly in peak fluxes and spatial heterogeneity: Beijing’s high fluxes are concentrated in the urban core and along major traffic arteries; Shanghai’s largest emission zones are centered on the central business district and large thermal power plants; Guangzhou shows a multi-centric pattern; and Shenzhen exhibits a relatively even spatial distribution. Time-series analysis from 2019 to 2023 indicates that Shanghai was most affected by COVID-19 mobility restrictions and lockdowns, with mean annual CO2 flux declining by 12.70%; Beijing, Guangzhou, and Shenzhen experienced declines of 7.54%, 3.44%, and 3.50%, respectively. NACES effectively reconstructs the locations, magnitudes, and temporal dynamics of high-flux corridors and point sources, enabling a refined depiction of the coupled natural–anthropogenic carbon cycle and addressing limitations of existing inventories in characterizing intra-urban variability.

heat storage (Q(s)) is a critical contributor to the urban heat island (UHI) effect, yet its spatiotemporal patterns and quantitative contribution remain poorly understood. Urban surface thermal properties directly influence Q(s), but accurately quantifying these properties at a regional scale remains a challenge. This study retrieves urban thermal inertia (TI) and hourly Q(s) based on heat conduction theory, using 24-h cycles of Himawari-8 land surface temperature (LST), in three urban agglomerations of China. The quantitative impact of Q(s) on near-surface air temperature (T-a) is also investigated at both city and local climate zones (LCZs) scales. With validation of Q(s) against flux tower data, the correlation coefficient can be up to 0.92 and root-mean-square error (RMSE) is 39 W/m(2). Results show that there is a significantly stronger correlation of Q(s) with Ta during nighttime (R-2 > 0.95) than daytime, confirming its dominant control in nocturnal UHI. To be specific, across LCZs, high-rise and open-built areas exhibited greater nocturnal Q(s) (and correspondingly higher Ta) compared to compact mid-lowrise neighborhoods. Notably, water bodies remain higher Qs (higher Ta) than vegetated surfaces. Entropy-based thermodynamic assessment suggests high-rise buildings exhibit higher but faster UHI change than low-rise. For future urban planning, we recommend prioritizing the heat mitigation of high-rise buildings and strategically planning blue space to balance its daytime cooling against nocturnal warming effects.

This study employs the WRF-based RMAPS-ST numerical weather prediction model to conduct 1-km resolution simulations of the Beijing Winter Olympic Games on February 2022. An inter-comparison of the SH, the SMS and the IUM scale-aware PBL schemes is conducted. The performance of YSU as a traditional PBL scheme is also evaluated as a baseline. The investigation focuses on their ability to forecast over three distinct competition areas, Beijing (BJ), Yanqing (YQ), and Zhangjiakou (ZJK), each characterized by different terrain. The results reveal that the IUM and SMS schemes generally outperform the YSU and the SH schemes, particularly in the BJ area, where IUM exhibits the lowest root-mean-square error (RMSE) for 2-m temperature and 10-m wind speed. In YQ, with its complex topography, there are no significant discrepancies among the schemes, while SMS improves wind speed forecasts in ZJK. An abrupt nocturnal warming and a gusty wind case are chosen to further demonstrate the performance in all schemes. Results show that IUM scheme can reproduce these two cases, attributed to its stronger turbulence mixing length. These findings underscore the importance of scale-aware PBL schemes in improving high-resolution weather forecasts and diverse meteorological conditions. The IUM scheme, in particular, emerges as a promising tool for future numerical weather prediction applications.

Urban morphology exerts a strong control on flow behavior within the roughness sublayer (RSL), yet the interactions between building configurations and momentum transport are still not well understood. In this study, we use large-eddy simulations (LES) and machine learning to examine how key morphological features affect turbulence and drag in idealized building clusters. Our LES results show that dispersive momentum flux (DMF) contributes significantly to total momentum transport, with the ratio of DMF to turbulent momentum flux (TMF) ranging from to across most configurations and dropping below in the densest arrays. Likewise, dispersive kinetic energy (DKE) often surpasses of turbulent kinetic energy (TKE) within the urban canopy layer. These results highlight the need to incorporate both turbulent and dispersive contributions in urban turbulence modeling. The bulk drag coefficient () also exhibits strong sensitivity to urban form, driven primarily by pressure differences between windward and leeward surfaces. Among the morphological parameters considered, the street-canyon aspect ratio emerges as the dominant factor, showing an correlation with . We also show that drag estimates are dependent on the diagnostic formulation used, particularly whether dispersive contributions are included. To enable efficient parameterization of the complex vertical turbulence structure, we develop a U-Net-based machine-learning model capable of reconstructing vertical profiles of kinetic energy (including both TKE and DKE components) from morphological inputs alone. Together, these findings refine our understanding of momentum and energy exchange in urban environments and provide a foundation for improved parameterizations of urban canopy processes in larger-scale meteorological models.

Recent precipitation extremes have shattered historical records over the Beijing-Tianjin-Hebei (BTH) region, causing devastating losses of life and property. While climate change is expected to increase the probability of precipitation extremes, how the frequency and intensity of rarest events will change remains unclear. Here, we develop an integrated framework combining long-term observations, large-ensemble Earth system simulations, and convection-permitting regional modeling to project the likelihood and magnitude of such “rareness” precipitation events. Using the Community Earth System Model Large Ensemble (CESM2-LE), we find that the likelihood of regional events comparable to the July 2023 BTH extreme precipitation event (“23.7 BTH” event) increase by 159% under the SSP3-7.0 scenario, primarily driven by thermodynamic intensification linked to more frequent moisture-abundant conditions. We further find that the local intensity of the most extreme future storms may increase by approximately 30%, with hourly precipitation rate nearly doubling. Our framework provides a robust pathway to quantify the frequency and magnitude of unprecedented regional extremes, offering critical implications for flood management, hazard mitigation, and climate adaptation planning.

Vegetated patches, such as tree clusters in urban parks or agricultural fields, play a critical role in modulating local microclimates through shade and evapotranspiration. This study investigates the sensitivity of the micrometeorological environment within and around a patch of trees to background solar radiation (250, 500, 750, and 1000 Wm−2), air temperature (25, 30, 35, and 40 °C) and relative humidity (20%, 40%, 60%, and 80%) using a large-eddy simulation method. The finite extent of the forest patch generates pronounced edge effects, including airflow deceleration at the leading edge, internal boundary layer development, and sharp horizontal gradients in temperature and humidity across the patch. These edge-driven dynamics are strongly modulated by the background climate. Our results indicate that near-surface cooling and increased humidity are most effective at low radiation levels. As incoming radiation is beyond 500 W/m2, cooling near the forest floor saturates, while warming intensifies in the upper canopy due to increased radiation absorption. High radiative forcing reduces the relative contribution of evaporative cooling, limiting the magnitude and spatial footprint of the canopy-induced thermal anomaly. Similarly, rising background temperatures weaken near-surface cooling and exacerbate upper-canopy warming, consistent with increased stomatal resistance under warmer conditions. Background relative humidity further modulates these responses. Cooling is strongest under dry conditions (e.g., 20%) and diminishes with increasing humidity. Under highly humid conditions (e.g. 80%), evaporative cooling is suppressed, thus reducing the canopy-induced cooling effect. Nevertheless, a net cooling effect persists near the ground across all simulated scenarios, suggesting that urban forests remain a useful thermal mitigation strategy in a warming climate, although their efficacy depends strongly on background meteorological conditions and vegetation physiological characteristics.

Urbanization can significantly modify precipitation patterns. This review systematically analyzes 262 studies (1980-2024) across China, focusing on how urbanization alters precipitation. A new quality assessment framework including nine objective criteria was developed to evaluate scientific rigor and data quality of studies. Papers meeting the required standard were selected for in-depth analysis. Unified variables were calculated using standardized formulas, enabling direct comparison between regions and cities. Results showed that most of the analyzed literature (77%) concentrates on the three major urban agglomerations: the Beijing-Tianjin-Hebei (BTH), the Yangtze River Delta (YRD), and the Pearl River Delta (PRD), with a consistent pattern characterized by a reduction in light precipitation and a rise in moderate-to-extreme events. A consistent north-south gradient was observed in the urban contribution to precipitation (Cu), with values increasing from the BTH (30.8%) to the YRD (32.4%) and peaking in the PRD (56.1%) at the regional level. At the city level, the contributions of Beijing (18.6%), Shanghai (30.0%), and Guangzhou (34.2%) followed the same pattern. These spatial differences are likely influenced by regional climatic conditions and anthropogenic activities. The urban-rural difference in precipitation rate (RPTP u-r) at city level are significantly smaller than those at the broader urbanagglomeration level. The RPTPu-r of Beijing, Shanghai and Guangzhou are 0.4%, 2.1%, 1.1% per decade, respectively, much lower than those of the BTH, YRD and PRD region (7.1%, 9%, 13.4% per decade, respectively). This study provides a systematic analysis of recent advances, knowledges the gaps, offers concrete recommendations for advancing the field.

This study leverages a comprehensive set of CMIP6 GCM outputs, downscaled via a convolutional neural network (CNN), to examine Beijing's temperature data spanning from 1961 to 2100. The ensemble of downscaled CMIP6 results indicates a consistent rise in Beijing's annual mean maximum (Tmax) and minimum (Tmin) temperatures from 1961 to 2100. Additionally, the potential impacts on local temperature-related extreme events are assessed using 8 indices of extreme climate events. The findings show that hot days (SU30) are projected to see a significant increase in both intensity (CSU30) and frequency (CSU30 frequency) over time and across scenarios, surpassing historical levels through a gradual step-up increase, a pattern similarly observed in warm nights (TR). The frequency and intensity of extreme temperatures associated with cold events (FD, ID) in Beijing are expected to decrease markedly, particularly in the central southern part of the study area. Regarding the temporal variation of urban-rural contrast, Tmin is on a downward trend under the SSP585 scenario, while that of Tmax is likely to remain stable under both the SSP245 and SSP585 scenarios in the future. The difference in the intensity of extreme weather events (SU30, CSU30, ID, FD, CFD, TR) between urban and suburban areas is growing, yet the frequency of extreme events (CSU30 frequency, CFD frequency) is on the decline.

Urbanization and cropland irrigation modify land surface water and energy budgets in different ways; however, few observational studies have explicitly quantified their contrasts. Using high-resolution observations from over 2000 surface weather stations and urban and irrigation fraction data, this study investigated the individual and combined effects of urbanization and cropland irrigation on surface air temperature over the Beijing–Tianjin–Hebei (BTH) region in China, where highly urbanized areas and heavily irrigated croplands exist together. The results indicate that (1) the daytime irrigation cooling (with surface air temperature decreasing by ~0.1–0.5 °C at irrigated stations) was non-negligible in late autumn, early winter, and later spring months, when winter wheat irrigation mainly occurred over the BTH region, while a slight warming was observed at many irrigated stations during the nighttime. By contrast, urban warming was most pronounced in the nighttime, especially in winter, and the daytime warming at urban sites was much weaker and comparable to the magnitude of cooling induced by concurrent irrigation for winter wheat. (2) Collectively, the vast stretches of irrigated croplands helped mitigate urban warming, and their combined effect on the daytime surface air temperature over the whole region resulted in a slight cooling of ~0.2 °C in some of the winter wheat-growing months. (3) The contrasting temperature changes due to urbanization and irrigation were spatially variable. Beijing was predominantly characterized by urban warming, while Shijiazhuang, with extensive irrigation, exhibited irrigation cooling (or slight warming) during the daytime (or nighttime) in most of the winter wheat-growing months, which could be a possible contributor to the daytime cooling (or stronger nighttime warming) at urban sites. This work highlights the temperature contrasts between urban areas and surrounding irrigated croplands, as well as the potential role of extensive irrigation in mitigating (or enhancing) daytime (or nighttime) urban warming.

Understanding how complex terrain influences urban wind patterns and cooling effects is crucial for addressing urban heat island (UHI) impacts and promoting climate-resilient cities. This study explores how terrain and urbanization jointly shape local wind systems and cooling in the Hangzhou Bay area, China, utilizing observational data from up to 574 meteorological stations (July-September, 2017-2021). Mixed local circulations occur 29-57 % of the time. In this region, strong UHI intensity can advance or delay local wind onset by 1-2 h and modify wind direction transitions, with clear west-east contrasts linked to the area's specific terrain-mountains to the west, sea to the east. These findings reflect localized interactions rather than universal patterns. Nocturnal mountain winds contribute over 2.5 degrees Ch of cooling, especially from 17:00 to 19:00, with effects lasting similar to 9 h - significantly stronger than those of daytime sea breezes. However, these winds may intensify UHI effects by disproportionately cooling rural over urban areas. The results highlight the importance of region-specific station selection when assessing UHI, to avoid misinterpretations driven by local wind influences. This work provides actionable insights for improving urban ventilation and thermal comfort in complex terrain settings.

The urban neighborhood serves as the fundamental unit for fine-scale management of urban carbon emissions, playing a critical role in achieving urban carbon neutrality and sustainable development. Carbon source-sink data at the urban neighborhood scale exhibit significant spatial heterogeneity, and accurate estimation of CO2 fluxes helps to better understand the relationship between urban carbon processes and both anthropogenic and natural factors. This study employed multi-source data and the SUEWS urban land surface model to simulate CO2 fluxes at the neighborhood scale. High temporal (hourly) and spatial (20 m) resolution CO2 fluxes were obtained in the typical mixed-use areas surrounding the IAP and RCEES stations in Beijing. Carbon fluxes from traffic, buildings, human metabolism, soil, and vegetation were quantified for the years 2016, 2019, and 2020. The results show that the SUEWS model effectively captures the temporal and spatial dynamics of CO2 flux. The multi-year average CO2 flux at IAP and RCEES stations was 28.17 and 12.87 kg CO2/m2 per year, respectively. Traffic accounted for the largest share of CO2 emissions, contributing more than 50%, followed by emissions from buildings and human metabolism. The study also evaluated the potential for carbon reduction in urban neighborhoods under future low-carbon policies. Under the moderate emission scenario SSP2-4.5, along with the implementation of strong policy measures, including 80% rooftop greening and electric vehicle adoption, carbon emissions in urban neighborhoods could be reduced by approximately 60%. This study provides essential data and technical support for urban CO2 reduction through fine-scale CO2 flux calculations.

Integration of weather and climate forecasting is currently the frontier of numerical modeling development in China, and dynamic downscaling allows for improving the performance and resolution of global climate models to the weather scale. Focusing on the “23.7” extreme rainstorm (July 29, 00:00 - August 2, 00:00 UTC) in the Beijing-Tianjin-Hebei region (BTH), this study assesses predictions from the China Meteorological Administration Climate Prediction System version 3 (CMA-CPSv3, 45 km resolution) and 9-km dynamic downscaling hindcasts from the Weather Research and Forecasting model (WRF-9km). Unlike traditional climate anomalies approaches, direct outputs are used for evaluation, similar to weather forecasting tests. By examining, both the CMA-CPSv3 forecasts and the WRF-9km hindcasts offer a 5-day prediction window for this rainstorm. They successfully predict the rainstorms and related atmospheric circulations from July 24th onward, aligning with observed and reanalyzed data. WRF-9km, with the higher resolution and optimized physical processes, outperforms CMA-CPSv3, particularly in precipitation spatial distribution and center intensity. The WRF-9km 7/24 hindcast exhibits the most significant enhancement compared to the corresponding CMA-CPSv3 forecast. This improvement is notably reflected in the substantial increase in spatial correlation, rising from 0.68 to 0.79, as well as a reduction in the difference of center values, decreasing from -51% to -20%. Furthermore, the WRF-9km 7/24 hindcast also improves the Critical Success Index by 0.08, the Success Rate by 0.08, and the Probability of Detection by 0.29 for heavy rainfall (over 25.0 mm/d). However, improvements in large-scale circulations with WRF-9km are limited, which may restrict advancements in predictability. In conclusion, the WRF-9km can enhance the performance and resolution of CMA-CPSv3 predictions, which can serve as one route for CMA-CPSv3 to achieve weather-climate integration.

Increasing urban tree cover is a widely recommended strategy for mitigating urban heat, as trees are expected to cool cities through evaportranspiration and shade provision. In real-world scenarios, both shading and transpirational cooling processes depend on prevailing climate conditions (e.g. radiation, wind, humidity) and tree attributes (e.g. shape and size of greenery space, tree height, leaf-area density, stomatal conductance). Quantifying these factors is crucial for assessing the roles of trees in mitigating the urban heat island effect and shaping local climates. Here, we seek to explore the cooling effect of a finite-size forest in response to changes in local climate conditions using a large-eddy simulation method. Airflow over the finite-size forest experiences abrupt changes in wind, temperature and humidity due to the transition from the sunlit openings to the relatively sheltered forest interior. The vertical velocity changes sign at both the leading and trailing edges, indicating the presence of local recirculations therewith due to the blocking effect imposed by the dense canopy. The dense canopy cover intercepts the incoming solar radiation, resulting in cooler ground temperatures beneath the forest compared to open areas. Additionally, the forest canopy release water vapor to the atmosphere through evapotranspiration, leading to increased humidity levels around the forest. The cooler and more humid air from the forest mixes with the warmer air from the open areas, yielding slightly cooler temperatures in the near-field of forest wake. Simulation results under various scenarios are compared to identify trends and relationships between urban trees and local climate conditions. These findings can be used to inform future field campaigns over forests of finite size with distinct edges and planning strategies aimed at improving microclimate via urban greenery.

Traditional wildfire spread prediction models often struggle to simulate fire propagation accurately in complex terrain or under strong wind conditions due to their semi-empirical nature and simplified treatment of fire-atmosphere interactions. This study presents a novel rapid fire spread model that integrates Briggs' buoyant plume theory, the Rothermel fire spread model, and Huygens' principle of wave propagation. The model is designed to simulate fire behavior in grasslands with complex terrain,enabling detailed representation of fire front dynamics. By analytically solving flame plume equations, the model quantifies the combustion heat production and its impact on surrounding wind fields through buoyant lifting effects. This innovative two-way coupling of atmosphere-fire interactions enhances the model's ability to simulate fire behavior under various environmental conditions. The model incorporates multiple factors, such as the wind speed, terrain, flame plume entrainment, fuel moisture content, packing ratio, and effective flame width, overcoming the limitations of traditional models in accurately capturing the shape of fire fronts. Validation through numerical simulations demonstrates that the model effectively reproduces both the temporal and spatial evolution of fire fronts across different wind and terrain conditions. In benchmark tests with a 100-meter ignition line, the model shows remarkable agreement with fully-coupled dynamic simulations, with a mere-10.38% deviation in fire spread rate while accurately replicating fire front patterns. The proposed model offers high computational efficiency and can serve as a valuable tool for wildfire risk assessment and emergency response.

Accurate urban-resolving climate data are essential for urban climate research and applications. However, General Circulation Models (GCMs) often lack the resolution and urban representation needed to provide reliable fine-scale climate information over urban areas. Convection-permitting modeling (CPM) has emerged as a promising solution to this challenge, despite its computational demands. Evaluating the added value of CPM for specific regions is crucial. In this study, we utilized the Weather Research and Forecasting (WRF) model coupled with a single-layer urban canopy model, as a regional climate model, to assess the performance and added value of CPM at both regional (urban clusters) and local (megacity) scales. With an optimized dynamic downscaling scheme, we conducted 3-km-resolution CPM and 9-km-resolution dynamic downscaling modeling (DDM) during the summer of 2020 in eastern China, where most cities and urban clusters are located. At the local scale, CPM well reproduced observed precipitation rates at daily and sub-daily time scales, greatly improved the overestimation of drizzle-to-light rainfall events and underestimation of heavy-to-torrential rain events in ERA5 reanalysis data. Additionally, CPM effectively captured diurnal variations in precipitation across six sub-regions of eastern China, a capability lacking in DDM and ERA5. Moreover, CPM successfully reproduced the observed urban heat island intensity in Beijing by capturing the heterogeneous air temperature distribution, outperforming ERA5 and DDM. Our findings highlight the considerable added value of CPM in simulating sub-daily precipitation variations and urban heat island intensity over urban areas of China. These insights will greatly enhance future high-resolution regional climate simulations and climate change projections over urban areas in China.

Top‐down methods commonly use atmospheric CO 2 concentration observations to constrain carbon source and sinks. Despite the increase in spaceborne and ground‐based concentration measurements, atmospheric inversions are usually limited by uncertainties in chemical transport models (CTMs) when relating fluxes to observed CO 2 mole fractions. CO 2 eddy covariance (EC) flux measurements have been widely used to directly measure CO 2 fluxes over various ecosystems, but they have rarely been used as constraints in top‐down estimations. In this study, we focused on the development of a novel fluxes assimilation scheme through direct flux observations within an Ensemble Square Root Filter assimilation framework. The assimilation scheme avoided some of complexities of concentration observation assimilations. The methodology was primarily applied to typical regions in west China, taking advantage of eight long‐term ecosystem EC sites. Moreover, four sets of assimilation experiments were designed to quantify the impacts of observational constraints by flux and concentration measurements. Generally, results indicate that the monthly and hourly statistics of the a posteriori fluxes constrained by flux observations agreed well with flux measurements, demonstrating reasonable performance in seasonal and diurnal variations. Specifically, assimilation results demonstrated the advantage of a posteriori estimates inferred from flux measurements during growing season, as compared to results inferred from concentrations, while some limitation still exists in monthly budget estimates. Nevertheless, it is important to note that current results are only a mathematical optimum. CO 2 biospheric fluxes can be estimated more reliably and robustly at the regional scale given considerably more flux observations for efficient constraint.

The integration of weather and climate prediction represents the current frontier in the development of numerical modeling in China. Dynamic downscaling serves as a pivotal approach, improving the performance and resolution of global climate models to the weather scale. Focusing on the '23.7' extreme rainstorm (July 29, 00:00-August 2, 00:00 UTC) in the Beijing-Tianjin-Hebei region, this study assesses predictions from the China Meteorological Administration Climate Prediction System version 3 (CMA-CPSv3, 45 km resolution) and 9-km dynamic downscaling hindcasts from the Weather Research and Forecasting model (WRF-9 km). In contrast to the conventional climate anomaly approaches, direct outputs are used for evaluation, similar to weather forecasting tests. By examining, both the CMA-CPSv3 predictions and the WRF-9 km hindcasts provide a 5-day prediction window for this rainstorm. They successfully predict the rainstorms and related atmospheric circulations from July 24th onward, aligning with observed and reanalyzed data. WRF-9 km, with the higher resolution and optimised physical processes, outperforms CMA-CPSv3, especially in precipitation spatial distribution and center intensity. The WRF-9 km 7/24 hindcast demonstrates the most significant enhancement compared to the corresponding CMA-CPSv3 prediction. This improvement is notably reflected in the substantial increase in spatial correlation, from 0.68 to 0.79, as well as a reduction in the difference of center values, decreasing from -51% to -20%. Furthermore, the WRF-9 km 7/24 hindcast improves the Critical Success Index by 0.08, the Success Rate by 0.08, and the Probability of Detection by 0.29 for heavy rainfall (over 25.0 mm/d). However, improvements in large-scale circulations with WRF-9 km are limited, which may restrict advancements in predictability. In conclusion, the WRF-9 km enhances the performance and resolution of CMA-CPSv3 predictions, which can be regarded as a viable pathway for CMA-CPSv3 to achieve weather-climate integration.