The dust weather in the Qinghai-Xizang Plateau essentially results from the synergistic effect of three key factors： sand supply， persistent strong winds and unstable atmospheric stratification. Owing to the unique topography， sparse vegetation， and complex thermodynamic and dynamic environment of the Qinghai-Xizang Plateau， dust weather occurs under different conditions from those in typical desert regions. As a major threat to aviation safety， dust weather superimposed on the harsh high-altitude environment poses multidimensional challenges to flight safety， airport operations and maintenance management. It is crucial to understand the mechanisms of dust weather and address its impacts so as to ensure flight safety and improve aviation service quality. Through extensive literature research， the study explores the progress and trends in the research on the impact of dust weather on aviation in the Qinghai-Xizang Plateau. In terms of flight safety， dust weather sharply reduces visibility， leading to severe weakening of the pilots’ visual ability and increased risk of navigation deviation. Aircraft landings under strong crosswinds are prone to deviate from the runway； the high-velocity sand grain impacts on the fuselage and engine inlets can alter aerodynamic structural characteristics， accelerate surface wear， and reduce aircraft performance； In terms of aviation operation and maintenance， dust weather forces airports to reduce operational capacity， extend aircraft turnaround time， and substantially increase mechanical maintenance and scheduling costs. The systematic review and in-depth analysis of the unique formation mechanisms of dust weather on the Qinghai-Xizang Plateau and its impacts on the aviation industry provide theoretical support and orientation for improving flight safety， operational efficiency， and service quality at plateau airports， and outlines prospects for future policy-making and research.

Wind field characteristics in the lower and middle tropospheric wind fields are of considerable significance for the development of weather systems as well as aviation safety， and have long been a key focus of meteorological research. Due to the influence of surface uplift， the boundary layer processes over the Qinghai-Xizang Plateau exhibit pronounced interactions with the upper and middle tropospheric processes of the surrounding atmosphere. Based oncontinuous observation data from the wind profile radar deployed at Hongyuan Station （with altitude of 3494 m） located in the eastern part of the Qinghai-Xizang Plateau， this study systematically analyzes the spatial distribution patterns and temporal variation characteristics of wind fields within the range of 6 km above the ground. The results reveal distinctive variations in the wind field characteristics within the low levels of the Hongyuan region. A clear wind transition layer is identified in the lower and middle troposphere. Above this transition layer， westerly winds prevail， while below the layer， southwesterly and northerly winds dominate. Notably， northerly winds are observed in the late afternoon and nighttime， with southwesterly winds prevailing at other times. The vertical mixing of wind speed is relatively uniform below the transition layer. This layer is thickest in summer and thinnest in winter， which is associated with the seasonal variations of the large-scale westerly circulation and local convective activities. Wind speed within 6 km above Hongyuan shows significant diurnal variations， with the largest diurnal range in winter and the smallest in summer. A stable maximum horizontal wind speed layer persists at approximately 1800 m above ground level （AGL） from 05:00 to 09:00 （Beijing Time）. Multiple low-level jet structures are observed in the lower atmosphere， with 74. 33% occurring below 1 km AGL. The occurrence frequency peaks in June， and the average intensity is the highest in February. Southwesterly low-level jets are the dominant type， followed by northerly low-level jets， which are more likely to trigger precipitation. The findings of this study advance our understanding of lower tropospheric atmospheric variations over the eastern Qinghai-Xizang Plateau.

Using lightning location system data collected in the Liangshan region from 2010 to 2023， this study applied K-means clustering to delineate lightning-activity zones and employed a two-component Gaussian mixture model and the Institute of Electrical and Electronics Engineers （IEEE） cumulative probability model for statistical analysis. The spatiotemporal distributions， statistical characteristics， and probability models of the frequency density and current amplitude of cloud-to-ground lightning return strokes were systematically investigated. The results show that： （1） lightning activity in the Liangshan region exhibits a significant negative spatial correlation between frequency and current amplitude. High-frequency lightning is concentrated in the central part of the region， whereas strong lightning events occur mainly in the western high-altitude strong-lightning zone； （2） the temporal variations in lightning frequency and current amplitude are asynchronous. Lightning frequency peaks on summer afternoons， consistent with thermally driven convection， whereas the highest-amplitude positive return strokes occur mainly between 09:00 （Beijing time， the same as after） and 11:00 in February and March； （3） the logarithmic probability density of lightning current amplitude generally exhibits a bimodal structure， and the two-component Gaussian mixture model provides a markedly better fit than the single Gaussian model， indicating that lightning activity in the region may be influenced by different physical processes； （4） the IEEE cumulative probability model describes the local distribution of lightning current amplitude more accurately than the Chinese electric power industry standard DL/T 620-1997. This study provides a refined set of IEEE model parameters for each physical zone， offering a direct scientific basis for regionalized and differentiated lightning-protection design.

The Qinghai-Xizang Plateau hosts numerous lakes， most of which undergo seasonal freezing. Observations from multiple lakes have revealed a water warming phenomenon during the ice-covered period. However， the widely applied CLM4-LISSS （Community Land Model version 4 - Lake， Ice， Snow， Sediment Simulator） in regional climate modeling fails to simulate this under-ice warming effect， which hinders studies on lake-atmosphere interactions and their regional climatic and ecological impacts. Therefore， this paper improved the CLM4-LISSS model in four aspects： radiation transfer， convective mixing， lake ice surface albedo， and water/ice extinction coefficients. Using in-situ observed meteorological data and lake thermal condition data from Ngoring Lake， we evaluated the performance of the improved model and revealed the impact of the revised parameterization schemes on simulated lake water temperatures. The result indicates substantial biases in the CLM4-LISSS model's simulation of Ngoring Lake's water temperature during ice-covered periods， stemming from four core structural limitations： complete absorption of shortwave radiation by the top ice layer with no downward transmission to underlying water， erroneous inclusion of lake ice in vertical convective mixing during density instability， overestimation of lake ice surface albedo， and overparameterization of water extinction coefficients alongside complete omission of ice extinction effects， to address these deficiencies， we implemented targeted revisions to the model's parameterization schemes by modifying shortwave radiation partitioning to allow partial penetration into sub-ice water layers， introducing the nswater parameter to exclude ice-containing layers from convective mixing processes， calibrating ice surface albedo from 0. 6 to 0. 3 which aligns with in-situ observations， adjusting water extinction coefficients from 0. 29 m^-1 to 0. 15 m^-1， and incorporating ice extinction coefficients of 2. 0 m^-1， these improvements yielded significant improvements in simulating both under-ice warming during frozen periods and deep-water temperature dynamics during summer thermal stratification. During the ice-covered period， the 2 m water temperature RMSE reduced from 3. 20 ℃ to 1. 30 ℃ with BIAS improving from -2. 82 ℃ to -0. 36 ℃， while the 13 m water temperature RMSE decreased from 2. 52 ℃ to 0. 60 ℃ with BIAS shifting from -2. 52 ℃ to -0. 12 ℃. During summer stratification， the 13 m water temperature RMSE dropped from 4. 28 ℃ to 0. 95 ℃ with BIAS improving from -3. 80 ℃ to -0. 31 ℃.

To address the scarcity of precipitation observations and poor universality of precipitation products over the complex plateau terrain in the Yellow River Source Region， a precipitation retrieval model based on Fengyun-4A （FY-4A） satellite data and eXtreme Gradient Boosting （XGBoost） machine learning algorithm.Multiple cloud-physical characteristic variables were constructed using multi-channel brightness temperature data， while Bayesian optimization was demployed for hyperparameter tuning.Dual models （classification and regression） were trained using gridded and station-based precipitation data respectively， with case studies validating model applicability in the study region.The results shows： （1） The classification model achieved 100% precision on training samples and over 80% accuracy on test sets； （2） the regression model exhibited a determination coefficient （R²） of 0.985 for gridded training data， whereas only 0.452 for station-based data， with smaller errors for precipitation intensities <10.0 mm·h^-1； （3） the validation of cases demonstrated effective spatial pattern capture， yet systematic underestimation occurred for intense precipitation cores （>10.0 mm·h^-1）.Future research should integrate this model with numerical weather prediction frameworks to incorporate cloud microphysical processes， and develop multi-scale retrieval systems by assimilating microwave band data from complementary satellite platforms to support regional precipitation estimates.

As part of the surface energy balance， anthropogenic heat has a profound impact on the climate and ecological environment of urban areas. Developing a broadly applicable method for estimating anthropogenic heat is essential for its effective coupling with numerical models and for improving the accuracy of urban environmental simulations. Based on the Remote Sensing Surface Energy Balance （RS-SEB） model with MODIS （MODerate-resolution Imaging Spectroradiometer） remote sensing products and meteorological station observation data， combined with refined Local Climate Zone （LCZ） classification map， this study quantitatively estimates anthropogenic heat of Beijing under clear sky conditions in October 2018 and couples with the urban canopy model of the WRF model. The results show that both the control and sensitivity experiment successfully represent the diurnal evolution characteristics of near-surface temperature with a single-peaked pattern. However， temperatures are generally overestimated during nighttime and underestimated during daytime. Comparing three different sensitivity experiments： NO_AH （no anthropogenic heat）， DEF_AH （default anthropogenic heat parameters）， and CAL_AH （anthropogenic heat parameters updated based on remote sensing estimates）， during the daily maximum temperature period and the daytime 12:00 （local time， the same below） -14:00， the CAL_AH experiment shows improved simulation performance for the average temperature at each station compared with the other two groups， alleviating the underestimation of temperature. The root mean square error （RMSE） of daily maximum temperature decreases from 1. 72 ℃ to 1. 37 ℃， the mean bias reduces from -1. 28 ℃ to -0. 69 ℃， and the simulation errors at stations in urban built-up areas is significantly reduced， with the improvement exceeding that of suburban stations， though overall nighttime accuracy deteriorates. For different urban building types， the daytime temperature distribution is relatively concentrated with small differences. Among them， compact and open mid-to-high-rise buildings （LCZ1~2， LCZ4~5） are important types causing high urban temperatures after introducing anthropogenic sensible heat. From the perspective of spatial distribution differences， the CAL_AH experiment simulates a stronger urban heat island effect during 12:00 -14:00， with an average urban temperature increase of 1. 08 ℃ and a suburban increase of 0. 28 ℃. This study provides a basis for the refined characterization of urban anthropogenic heat and its application in numerical models， which is of great significance for improving the simulation capabilities of urban thermal environment and high temperature risk assessment.

Focusing on the Yellow River Source Region， this study employed the Land Surface Energy Balance System （SEBS） model combined with FY-4A geostationary meteorological satellite data and meteorological observations to estimate daily clear-sky net radiation flux， soil heat flux， sensible heat flux， and latent heat flux （derived via the energy balance equation） during the growing season （April-October 2023）.Daily evapotranspiration （ET） was subsequently calculated.The results indicate that hourly variations in energy fluxes exhibit a distinct unimodal pattern， peaking around noon before declining.At midday， sensible heat flux， latent heat flux， and net radiation exceeded 200.0 W·m^-2， 300.0 W·m^-2， and 1000.0 W·m^-2， respectively.Spatially， the energy fluxes display longitudinal zonation： net radiation and sensible heat flux were significantly higher in the western and central regions than in the east， whereas latent heat flux showed the opposite trend， with higher values in the eastern part.For ET， the magnitude followed Hongyuan > Dari > Qingshuihe Station， with a clear northwest-to-southeast increasing gradient.Seasonal variations were pronounced， with summer ET > spring > autumn.The regional average daily ET during the growing season was approximately 3.0 mm.Spatially， daily ET in April， May， and September mostly ranged between 1.0~3.0 mm， while June-August saw widespread values exceeding 4.0 mm.By October， mean daily ET dropped below 1.0 mm.Validation confirmed the reliability of SEBS-derived ET estimates for the Yellow River Source Region.The high-temporal-resolution FY-4A satellite data enabled spatially explicit ET mapping， revealing regional water consumption patterns and providing valuable insights for sustainable water resource management.

Soil evaporation is a key process in regional water cycling and land surface energy exchange， and its accurate simulation is essential for understanding hydrothermal processes in arid regions. However， model applicability in drylands remains insufficiently validated using multi-source field observations. In this study， lysimeter measurements， eddy-covariance observations， soil moisture and temperature profiles， and meteorological data were integrated in the Alxa Desert region of Inner Mongolia， China， to evaluate the applicability of HYDRUS-1D for simulating soil evaporation under arid conditions. The results showed that， under the bare-soil configuration adopted in this study， the model reproduced the total amount of bare-soil evaporation and its daily dynamics well. The simulated evaporation agreed closely with lysimeter measurements， with a high coefficient of determination （R²=0. 99） and a relatively small percent bias （PBIAS=-8. 45%） in cumulative evaporation. The model also reproduced the hydrothermal dynamics of the shallow soil layer （0~20 cm） reasonably well. However， marked deviations occurred in the middle and deep soil layers （20~100 cm）， characterized by premature soil moisture responses after rainfall， overestimated peak water contents followed by faster recession， and systematic overestimation of soil temperature. Compared with eddy-covariance observations， the model underestimated latent heat flux， whereas the deviations in sensible heat flux and soil heat flux were relatively small. These discrepancies mainly reflect the fact that the current bare-soil configuration did not explicitly account for root water uptake and vegetation transpiration. The results indicate that HYDRUS-1D， under the current configuration， is applicable to simulating bare-soil evaporation and near-surface hydrothermal exchange but cannot fully represent middle- and deep-layer hydrothermal processes or latent heat dissipation under actual vegetated conditions. Future studies should further incorporate root water uptake and vegetation transpiration into the HYDRUS-1D framework to improve its ability to simulate long-term soil water movement in arid regions.

To investigate the specific effects and extent of the influence of soil temperature and soil moisture content-particularly soil temperature variations-on soil thermal conductivity， so as to provide a reference for land surface models to consider the influence of soil temperature changes on thermal conductivity.This study utilizes observational data from grassland and forest underlying surfaces collected at the Pingliang Land Surface Process and Severe Weather Observation Station， Chinese Academy of Sciences， from December 2017 to July 2021， to comparatively analyze the relationships among soil thermal conductivity， soil water content， and soil temperature.Partial correlation analysis was employed to quantify the independent effects of water content and temperature on thermal conductivity.The results show that： （1） The grassland typically enters the freezing phase 1~7 days earlier than the forest and initiates thawing 1~6 days later， the grassland has 1~4 additional days of freeze-thaw events compared to the forest areas.（2） Both surfaces exhibit pronounced seasonal cycles， with soil thermal conductivity， water content， and temperature peaking in summer and reaching minima in winter.Thermal conductivity closely tracks variations in water content.At 5 cm depth， the grassland shows higher annual mean soil water content （~0.12 m³·m⁻³） and temperature （~1.3 ℃） than the forest， along with greater seasonal and diurnal variability.（3） The peak soil temperature at 5 cm depth in the forest lags behind that of the grassland by approximately 2 hours， and the diurnal amplitudes of thermal conductivity， water content， and temperature are all larger in the grassland than in the forest.（4） Under dry conditions （soil water content ≤ 0.12 m³·m⁻³）， soil temperature remains significantly， albeit weakly， correlated with thermal conductivity （partial correlation coefficient r=0.169， p<0.001）， indicating that both water content and temperature should be jointly considered in land surface models for arid and semi-arid regions.

Water scarcity issues triggered by climatic aridity and humidityrepresent the core challenge affecting human survival， ecological security and sustainable development. In recent years， significant changes in temperature and precipitation have occurred in Shaanxi. Accurately identifying variations in hydroclimatic conditions and understanding the driving mechanisms are of great importance for water resource management. Based on the water cycle process， this study comprehensively analyzes the characteristics of climatic aridity and humidity variations in Shaanxi using precipitation minus evapotranspiration （P-ET）， and summarizes the physical mechanisms responsible for the interdecadal dry-wet variations. The results show a drying trend in summer over Shaanxi during the period 1981 -2022， with a more pronounced trend in the central and southern regions. Further analysis detected a distinct transition from wet to dry conditions in summer around 1996. Analysis of the contributions of climatic factors to summer evapotranspiration in Shaanxi before and after the shift reveals that the increase in vapor pressure deficit accounts for the largest contribution （53. 0%） to the enhancement of evapotranspiration. The increases in net radiation and near-surface temperature contribute 28. 3% and 18. 5%， respectively. Changes in all climatic factors lead to weakened summer evapotranspiration in northern Shaanxi and strengthened evapotranspiration in the central and southern parts. However， because the decrease in precipitation is greater than the decrease in evapotranspiration， precipitation-evapotranspiration shows a declining trend across Shaanxi， resulting in a drier summer climate. Moisture transport is closely related to summer dry-wet variations in Shaanxi. During 1996 -2022， summer circulation over Shaanxi and its surrounding areas is influenced by an anticyclonic anomalous circulation， leading to predominant moisture divergence and a significant reduction in net moisture income. Interdecadal sea surface temperature （SST） anomalies in the northwestern North Atlantic play an important role in influencing summer dry-wet conditions in Shaanxi. When the interdecadal SST in the northwestern North Atlantic is anomalously warm， it excites an atmospheric wave train that propagates downstream across Eurasia. This results in positive geopotential height anomalies and an anticyclonic anomalous circulation over Shaanxi and adjacent regions， consequently leading to moisture divergence and a tendency towards drier summer in Shaanxi.

This study conducts a multi-scale diagnostic analysis of a heavy rainfall event on the northern foot of the Qinling Mountains under the absence of a low-level jet. It reveals the coupling mechanism of terrain-front-mesoscale convective vortex （MCV） interactions. The results show that latent heating released by precipitation condensation generates a warm and moist air cap， which covers the near-surface cold air cushion formed by the Qinling terrain blocking behind the cold front. Deep convection is triggered by the combined effects of terrain uplift， frontal forced ascent， and mid-level MCV rotational uplift. The 700 hPa MCV establishes a positive feedback loop through dynamic pumping and moist potential vorticity （MPV）， leading to strong upward motion that forms the ascending branch of a vertical secondary circulation over the heavy rainfall area. The strong MPV gradient from low to high levels， along with an eastward tilt of the negative MPV center relative to the positive center， serves as a key predictor for rainfall initiation and eastward movement of the rainfall center. The terrain convergence lines below 925 hPa within 80 km in front of the northern piedmont of the Qinling Mountains， as well as the formation， movement direction and moving speed of mid-level vortices at 700~500 hPa， are key predictors that require special attention in routine forecasting of this type of weather system.

By setuping platinum resistance temperature sensors （PT100） and optical fibre temperature sensors （OS）， a distributed temperature observation system was constructed in an expansion-type cloud chamber， enabling optimized spatial monitoring of the chamber temperature field. The spatio-temporal evolution of the gas temperature distribution and the influence of wall temperature on air temperature in the enclosed space of the cloud chamber was systematically investigated by conducting a series of associated experiments with two typical conditions， thermally stable and dynamic expansion conditions. The results show that： （1） during the condition of thermally stability， the temperature inside the chamber mainly changed through the heat transfer mode of the inner wall of the chamber. Under static equilibrium conditions， the vertical gas temperature difference within 120 min was ≤0.1 ℃， and the horizontal difference was ≤0.05 ℃； during the uniformly temperature-controlling process from 0 to -40 ℃， the initial deviation between the upper and the lower temperatures in the chamber was more than 4 ℃ due to the influence of the wall temperature， and then gradually reduced to <2 ℃；the rapid temperature-control process from 27 °C to -20 °C took approximately 78 min， with a cooling rate of 0.6 °C·min⁻¹， exhibiting a fast-to-slow trend. （2） Under the condition of environmental change， i.e. dynamic gas expansion in the cloud chamber， the maximum value of the horizontal gas temperature difference in the chamber was 2 ℃， and the vertical gas distribution of the temperature presented a significant gradient characteristic， as temperature systematically increasing with height within the chamber： the temperature was slightly higher in the upper part and decreased towards the lower part. This conclusion reveals the dynamic response characteristics of gas temperature distribution under different thermodynamic conditions and its relationship with the gas expansion， providing multidimensional empirical evidence for the precise characterisation of the thermodynamic parameters of the experimental subject. （3） During the expansion process， the gas was heated by the chamber wall， which was no longer controlled and maintained at a constant temperature， resulting in a real cooling rate significantly lower than the theoretically calculated adiabatic expansion cooling rate. The time constant for the heating of the gas within the chamber—that is， the relaxation time required for the gas to return to a steady state following dynamic changes， is approximately 500 s. This parameter characterizes the reheating efficiency of the cloud chamber wall to the internal gas after the expansion process， and directly affects the effect of cloud expansion and the maintenance time of cloud.

To better understand the microphysical characteristics of extreme hourly precipitation events during the Meiyu period in the eastern Jianghuai region and their differences from ordinary precipitation， as well as to improve the quantitative estimation of extreme precipitation， this study uses hourly observed precipitation and raindrop size distribution（DSD） data from the Taizhou national meteorological station from 2017 to 2024. The 98th percentile of climatic precipitation data（1991 -2020） is used as the threshold during the Meiyu period. The DSD characteristics of extreme hourly precipitation in the eastern Jianghuai region are analyzed and compared with those of ordinary precipitation during the Meiyu period. The results show that： （1） extreme hourly precipitation is primarily generated by convective or mixed clouds. The precipitation amount is not determined directly by the frequency of minute rain intensity but is the result of the combined effect of their intensity and frequency. The average values of minite rainfall intensity （R）， rainwater content （W）， and total raindrop number concentration（N_T ） for such events are significantly higher than those of ordinary precipitation. The contribution rate of large raindrops （D>2. 5 mm） to these parameters is relatively higher， while that of small raindrops （D<1 mm） is conversely lower. （2） The raindrop number concentrations in the composite and concentrated DSD of extreme hourly precipitation are higher， while the maximum DSD overlaps with that of ordinary precipitation. The Γ distribution provides good fits for the composite， concentrated， and maximum DSDs of extreme hourly precipitation. The average values of the three fitted parameters （the intercept parameter： lgN ₀； the shape parameter： μ； the slope parameter： Λ）in the distribution function are significantly smaller， and the DSD show a stable spectral shape. The maximum DSD approximates the M-P distribution， which shows strong general applicability in describing the extreme states of the DSD. （3） The average values of D_m and 1gN_W for extreme hourly precipitation are larger， closely resembling those of "oceanic" convective precipitation and dominated by the warm rain process. The fitting relationships of D_m， lgN_T and lgN_W with R show better positive correlation. The DSD adjusts more rapidly toward an equilibrium state， with the response of raindrop size to changes in R being more pronounced than that of raindrop number concentration. （4） The Z-R relationship for ordinary precipitation is close to the operational radar standard relationship. The differences mainly occur during extreme hourly precipitation periods. The fitting relationship of extreme hourly precipitation is Z=178. 8R ^{1. 53}. For the same Z， radar tends to underestimate R， which requires high vigilance.

Using airborne detection data from February 28， 2021， combined with MODIS satellite data， ground-based radar data， and NCEP FNL reanalysis data， this study analyzes the cloud microphysical characteristics of aircraft icing in central and southern Hebei Province. This icing event was the result of the combined influence of a low trough cold front， a low-level jet stream， and a surface inversion trough. Vertical detection results reveal that the cloud system exhibited a three-layer structure overall： the lower layer consisted of mixed-phase clouds with relatively high supercooled water content， while the middle and upper layers were cold clouds. A temperature inversion layer observed between 5900 m and 6100 m promoted the accumulation of supercooled water， with a maximum value of 0. 18 g·m⁻³ observed at its base. Concurrently， the inversion layer significantly suppressed vertical convective mixing， thereby forming a typical microphysical environment characterized by "abundant liquid water but scarce ice crystals". The aircraft experienced the most severe icing in this region. The probability distribution of icing， derived using an icing probability algorithm based on MODIS-retrieved cloud top temperature， cloud optical thickness， cloud effective particle radius， and liquid water path， shows high consistency with the actual aircraft observations. The findings of this study reveal the critical role of the thermal structure of the inversion layer in regulating the distribution of supercooled water， providing a reference for aircraft icing warnings in North China.

The central urban region of Chongqing， situated within a canyon encircled by two mountainous barriers， exhibits a complex terrain that significantly hinders the natural dispersion of atmospheric pollutants. The scavenging effect of rainfall plays a pivotal role in reducing atmospheric pollutants. An analysis of rainfall process and raindrop size distribution （DSD）， and their effects on air pollutants was conducted using observational data from the central urban area（2016 -2021）. The results indicate that rainfall exerts a significant scavenging effect on atmospheric pollutants. However， O₃ exhibits a distinct response regulated by solar radiation， resulting in a scavenging mechanism that differs markedly from other pollutants. The scavenging efficiencys for the six pollutants rank in descending order as follows： PM₁₀， PM_2.5， SO₂， NO _x， O₃， and CO， with CO showing the lowest efficiency due to its poor solubility in water. For pollutants excluding O₃， an enhancement in the scavenging efficiency is observed in tandem with escalated maximum rainfall intensities， extended rainfall durations， and increased cumulative rainfall totals. The study also underscores the more pronounced influence of the mass-weighted mean diameter of raindrops （D_m ） and the mean raindrop concentration （N_t ） on the removal of aerosol particles as opposed to gaseous pollutants. Optimal removal of aerosol particles is observed to occur during the flood season， particularly when D_m >0. 5 mm and N_t >1250 m^-3. Moreover， high scavenging efficiencies for aerosols during the flood season are primarily associated with medium-to-low values of the raindrop spectrum shape parameter （μ ≤ 8. 0） and low values of the slope parameter （λ ≤ 7. 6 mm⁻¹）.

Under the process of climate warming and humidification， extreme precipitation events have occurred frequently in the arid and semi-arid regions of northwest China.The fragility of the ecological environment in this area makes it highly vulnerable to secondary disasters induced by extreme precipitation， resulting in strong disaster-causing effects.Using daily precipitation data and atmospheric reanalysis data， this study focuses on extreme precipitation events in the arid and semi-arid regions of northwest China during the summer （June-August） from 1980 to 2022.Based on the K-means clustering method， these extreme precipitation events are classified into three types， which occur in the mid-eastern Loess Plateau， the Ili River Valley Basin， and the Qilian Mountains respectively.Additionally， an analysis is conducted on their frequencies， large-scale circulation patterns， and water vapor transport of each type of extreme precipitation.The main results show that： （1） Over the past 40 years， the frequency of summer extreme precipitation in the arid and semi-arid regions of northwest China has shown an increasing trend.While a slight decreasing trend and amplified interannual oscillation occurred since the 21st century.Among these events， the extreme precipitation occurring in the mid-eastern Loess Plateau has the strongest intensity but the lowest cumulative days， whereas that in the Ili River Valley Basin has the weakest intensity but the highest cumulative days.（2） The large-scale circulation patterns corresponding to the three types of extreme precipitation exhibit a "low pressure in west and high pressure in east" pattern over the East Asian continent， however， significant differences exist in their intensity and location.Furthermore， anomalous anticyclone of the Western Pacific Subtropical High， along with water vapor transport via the Southwest Monsoon， the southerly winds on the eastern side of the Qinghai-Tibet Plateau， and the mid-latitudes， serves as important sources of water vapor affecting the three types of extreme precipitation.（3） Since the 21st century， extreme precipitation frequencies have increased in the eastern part of the arid and semi-arid regions of northwest China， while it has decreased rapidly in the western part.The frequency variations of mid-eastern Loess Plateau， the Ili River Valley Basin extreme precipitation are significantly influenced by the mid-latitude quasi steady Rossby wave activities， whereas， Type of the Qilian Mountains is notably affected by anomalous low pressure over the Tibetan Plateau.The above results provide a certain theoretical basis and application value for deeply understanding of the spatiotemporal Characteristics of extreme precipitation in northwest China， improving the accuracy of summer extreme precipitation prediction， and enhancing disaster prevention and mitigation efforts.

Under the background of global warming， drought has emerged as a critical disaster severely impacting the security of global terrestrial ecosystems， with pronounced regional heterogeneity. Sun-induced chlorophyll fluorescence （SIF）， which captures the physiological status of vegetation， has demonstrated the potential for early drought detection. However， the coarse spatial resolution of existing SIF products limits their application in regional-scale precision drought monitoring. This study integrates multi-source remote sensing data， including GOSIF， MODIS， and meteorological station observations， selects vegetation indices （NDVI and EVI）， reflectance bands （RED and NIR）， and topographic factors as independent variables， and utilizes an adaptive genetic algorithm to optimize the model parameters of Support Vector Regression （SVR）， Random Forest （RF）， and Convolutional Neural Network （CNN）. A downscaling model was subsequently constructed by applying Bayesian Model Averaging （BMA） to integrate the results from these individual models through weighted averaging， thereby generating a high-resolution SIF product for drought monitoring. The results indicate that the BMA-based downscaling model outperformed any single model （SVR， RF， or CNN） optimized with the adaptive genetic algorithm， achieving higher accuracy on both training and testing datasets. The BMA model yielded an R² of 0. 785， an RMSE of 0. 0453， and a Bias of 0. 009， compared to the SVR model， which obtained an R² of 0. 62， an RMSE of 0. 0603， and a Bias of 0. 0095. The downscaled SIF product derived from the BMA model accurately captured drought severity characteristics and exhibited the highest correlation and consistency with the Palmer Drought Severity Index （PDSI）， achieving Pearson correlation coefficient of 0. 7408 and consistency index of 0. 7639. During a drought event in June 2022， the BMA-downscaled SIF data effectively identified localized fine-scale drought patches and core drought areas at the county level， thereby overcoming the limitations of coarse-resolution drought monitoring at sub-provincial scales. This approach provides valuable technical support for precision drought monitoring and facilitates improved disaster prevention and mitigation at regional scales. This investigation into the advantages and disadvantages of BMA as a downscaling technique offers critical technical insights for the evaluation and application of statistical downscaling approaches in producing high-resolution， regional-scale remote sensing data.

Based on daily precipitation data from 44 meteorological stations in the Sichuan-Chongqing region during 1970 -2024， together with ERA5 reanalysis data， this study evaluated and validated the applicability of the daily-scale Dry-Wet Abrupt Alternation Index （DWAAI） by selecting historical typical dry-wet abrupt alternation events. The spatiotemporal characteristics of dry-wet abrupt alternation events during the flood season （May-September） in the Sichuan-Chongqing region from 1970 to 2024 were systematically analyzed. In addition， the atmospheric circulation characteristics associated with a representative severe dry-wet abrupt alternation event that occurred in 2020 were investigated. The results indicated that the daily-scale DWAAI exhibited good accuracy and rationality in monitoring dry-wet abrupt alternation events in the Sichuan-Chongqing region. The intra-annual monthly distribution of these events showed significant differences， characterized by a unimodal pattern with a pronounced peak in July. From an interannual perspective， no significant trend was observed in the overall station-based occurrence frequency， whereas event intensity displayed a weak increasing tendency. Most dry-wet abrupt alternation events were dominated by light intensity， and the cumulative frequency of drought-to-flood transitions exceeded that of flood-to-drought transitions. Spatially， dry-wet abrupt alternation events during the flood season occurred most frequently in the Sichuan Basin and the Chongqing region， while the western Sichuan Plateau experienced the fewest events， with pronounced regional heterogeneity. High-incidence areas of severe drought-to-flood events were mainly distributed along a north-south belt from Bazhong to Xuyong in the Sichuan Basin， as well as in Wanzhou （Chongqing） and the Wenjiang area of Sichuan. In contrast， severe flood-to-drought events rarely occurred at most stations on the western Sichuan Plateau， with high-frequency centers concentrated around Wanyuan and Wenjiang. These findings provided important data support and scientific evidence for monitoring and predicting extreme weather and climate variability in the Sichuan-Chongqing region， reducing disaster losses， and formulating effective disaster prevention and mitigation strategies.

This study focuses on an extreme precipitation event that occurred in North China from July 28 to August 1， 2023， referred to as the "23.7" extreme precipitation event.The purpose of this research is to assess the performance of various parameterization schemes within the WRF model in simulating this event， identify the optimal scheme combinations， and diagnose the physical causes of simulation biases.Using the Weather Research and Forecasting （WRF-ARW） model version 4.5.1， a total of 24 sensitivity experiments were conducted with six different microphysics schemes and four cumulus parameterization schemes.The precipitation simulation results were systematically evaluated using various spatial-temporal verification metrics， including the Threat Score （TS）， the Fractions Skill Score （FSS） and so on.The results indicate that： （1） The dynamical downscaling simulations effectively reproduced the orientation and area of rainfall associated with this extreme precipitation event， with accuracy improving as model resolution increased.（2） Among all scheme combinations， the combination of the WSM6 microphysics scheme and the Kain-Fritsch cumulus scheme （WSM6-KF） demonstrated the highest overall skill in simulating the precipitation area， intensity， and evolution trend， particularly excelling in the simulation of heavy rainfall events and above.In-depth diagnosis revealed that this combination's strength lies in its ability to accurately reproduce the distribution of deep graupel and intense updrafts corresponding to the heavy precipitation center， thereby capturing the key microphysical and dynamical features of this extreme precipitation event.（3） Despite the strong performance of the optimal scheme （WSM6-KF）， systematic biases were still present.Comparisons with ERA5 reanalysis data indicated that this scheme underestimated the water vapor transport pathways and moist layer thickness in the terrain-forced uplift regions of the Taihang Mountains during the later stages of precipitation， which directly resulted in an underestimation of peak precipitation and a westward shift of some rainfall areas.This study provides references for future operational forecasting of extreme precipitation in this region and insights for model improvement.

Under the background of frequent extreme weather and increasing serious secondary disasters such as geological disasters and floods in small and medium-sized rivers， it is of great significance to enhance the ability to forecast and early warning of flood meteorological risk in small and medium-sized basins. The Yiluo River Basin in the middle reaches of the Yellow River was selected as the study area. Based on multi-source data such as geographical information， meteorological and hydrological observations， and combined with digitized remote sensing information of the actual river network at fourth- to fifth-order streams， the Yiluo River Basin was precisely divided into 19 subbasins. Utilizing the localized SWAT distributed hydrological model， the key hydrological sections of Yiluo River and the daily runoff series of Heishiguan， Baimasi and Longmenzhen Hydrological stations in past 44 years were reconstructed to classify the flood grades. Taking the warning flood as an example， the statistical methods such as correlation coefficient analysis， P-III frequency curve and quadratic function fitting were used to establish the fitting curves of the maximum cumulative areal rainfall and the maximum runoff increase of the above three key hydrological stations and their sections， thus the critical thresholds of areal rainfall for different grades were obtained. On this basis， the areal rainfall threshold transplantation scheme was adopted to clarify the different grades flood-inducing critical surface rainfall thresholds for the 19 sub-basins of the Yiluo River， and ultimately obtained a set of graded flood-inducing critical area rainfall threshold indicators for small and medium-sized basins in the study area. The threshold indicators were applied and verified in the Yiluo River Basin during the flood season in 2024， showing that the overall hit rate of flood based on the research threshold was 55. 6%， among them， the early warning hit rate of single-day heavy rainfall was 100%， for continuous precipitation processes， using 48-hour cumulative area-rainfall early warning， the hit rate reached 88. 9%. This study provides research ideas and references for flood prevention， forecasting and early warning services in areas with insufficient or no hydrological data in the small and medium-sized river basins， and provides a valuable reference and technical support for flood disaster prevention and mitigation.

Hailstorms are a major severe weather hazard in Hebei Province， which can cause severe economic losses in a short period of time.This study computed the Maximum Expected Size of Hail （MESH）， Probability of Hail （POH）， Probability of Severe Hail （POSH）， Vertical Integrated Liquid （VIL）， and Vertical Integrated Liquid Density （VILD） using 6-minute-interval three-dimensional reflectivity data from ground-based radar observation data from Hebei Province between 2019 and 2023.A total of 245 genuine hail events and 1373 non-hail events were chosen as analysis samples based on radar data and hail observations from national stations in Hebei Province.Five widely used hail identification factors were assessed for performance.Significant distributional variations between hail and non-hail events were found for MESH， VIL， VILD and POSH using kernel density estimation methodology.POSH has the best hail detection performance among single parameters， according to the skill score evaluation based on the confusion matrix.This is mostly because it efficiently filters out interference from low reflectivity elements by taking into account both the 0 ℃ layer and -20 ℃ layer height information.Multi-parameter combinations， on the other hand， show more stability since they more accurately capture the variety of hail formation and complex meteorological conditions， which effectively lowers the likelihood of false alarms.POSH>10%， VIL>14 kg·m⁻²， and VILD>1.2 g·m⁻³ were the three parameters that showed the best results.When compared to single-parameter methods， its Critical Success Index （CSI） increased by about 3%， with higher accuracy being associated with the peak CSI.However， the accuracy of composite parameter-based hail size prediction is constrained by the influence of mixed hydrometeor types， complicated scattering and attenuation events， variability in hail size distribution， and horizontal advection effects during hailstone fall.This study suggests MESHra （range average）， an enhanced hail size algorithm.This approach successfully improves the connection between hail size observations and forecasts by averaging the echo intensities surpassing 40 dBZ across all height layers within a 9 km grid distance and recalculating MESH.The multi-parameter combination identification method was applied to two hail events in Hebei Province in 2025 （June 9 and June 18）， successfully identifying all eight hail cases and providing advance warning of hail 6 to 60 minutes in advance.While the conventional MESH model exhibited lower errors in forecasting large hail with diameters over 18 mm， the enhanced MESHra model showed lower errors in predicting small hail with diameters under 10 mm.In Hebei Province， 85% of hailstones had a diameter of less than 10 mm between 2019 and 2023.For operational hail forecasting in Hebei Province， the MESHra hail prediction is therefore reasonably appropriate.However， the MESH forecast is better suited for anticipating hail size when the atmosphere is favourable to huge hail.

Based on statistical analysis of elevated convection cases during the 2010 -2024 cold seasons in Henan Province using conventional observations and ERA5 reanalysis data， this study examines spatiotemporal distributions， synoptic configurations， and formation mechanisms. Results indicate that： （1） Characteristic configurations feature prominent 500 hPa troughs， intense 700 hPa southwesterly warm-moist jets， and a deep cold dome with northeasterlies below 850 hPa. Key sounding signatures include pronounced cold domes exceeding 150 hPa thickness and frontal inversions with tops at 850~700 hPa where over half of cases exhibit inversion-layer temperature differences greater than 4 °C. Thermodynamic instability is best characterized by elevated most unstable energy and thermal contrast between 700 hPa and 500 hPa. （2） Moisture transport at 700 hPa and its associated high specific humidity provide essential vapor conditions for elevated convection during the cold season. Strong upward motion is observed above 800 hPa， with its core centered at 500~400 hPa. Intense deep-layer vertical wind shear （>20 m·s^-1） and pronounced baroclinicity develop between the northerly flow beneath the inversion layer and the southwesterly jet atop the inversion layer. Such vertical wind shear configuration proves conducive to the sustenance and development of cold-season convection. （3） Cold season elevated convective instability mechanisms encompass four distinct types. The dominant conditional instability type exhibits deep vertical convection characterized by the strongest updraft intensity， with convective penetration extending significantly above the -20 ℃ level， frequently triggering thunderstorms and hailstorms. The hybrid mechanism type integrates conditional instability with conditional symmetric instability， displaying the most extensive updraft coverage at moderate intensity and capable of producing multiphase hazards including thunderstorms， hail and freezing precipitation mixtures such as freezing rain and sleet. Both the conditional symmetric instability type manifested through narrow slantwise convection and the frontogenesis circulation type featuring broad multicellular updraft structures demonstrate lower convective development heights with weaker intensities， predominantly generating thunderstorms and convective heavy winter precipitation events.