Wildfires enhance post-fire debris flow (PFDF) hazards by altering catchment hydrology and sediment dynamics, yet the mechanisms governing their spatiotemporal evolution during the first rainy season remain poorly constrained. Following the March 2024 Yajiang wildfire (Sichuan, China, 278.81 km2 burned), this study analyzes the first high-resolution PFDF inventory (506 events across 211 catchments) derived from field surveys and drone mapping. PFDF magnitude scales with catchment area, with narrow-steep catchments dominating activity due to amplified sediment connectivity and transport capacity. Temporal clustering of PFDFs occurred in May–July (79.6

Previous studies on precipitation caused by tropical cyclones (TCs) have largely focused on strong TCs while systematic research on weak systems remains limited. This study utilized the high-resolution (3 km) weather research and forecasting (WRF) model to investigate the sensitivity of heavy precipitation generated by weak TC Mulan (2022) over the South China Sea to cumulus and microphysics parameterization schemes. Six cumulus parameterizations and five microphysics schemes were configured with simulations validated against gauge observations. Results indicated that cumulus parameterizations had a significant influence on precipitation simulation while the microphysics schemes exhibited a relatively minor impact in this case. The combination of the New Tiedtke cumulus scheme and the WRF Single-Moment 6-class (WSM6) microphysics scheme yielded the best simulation of precipitation compared with the observations. Further investigation revealed that cumulus parameterizations modulated simulated large-scale circulation, moisture transportation, and vertical velocities. The New Tiedtke scheme reproduced a more northward TC track and an intensified southeast jet along the coast of Southeast China, aligning with the observed heavy precipitation zones and providing favorable dynamic and thermodynamic conditions for precipitation. In contrast, the Betts-Miller-Janji & cacute; (BMJ) scheme resolved cloud-environment interactions inadequately, resulting in excessive convection and latent heating, which amplified cumulus precipitation compared to the New Tiedtke scheme. Among the cumulus parameterization schemes, the Kain-Fritsch (KF) and BMJ schemes underperformed due to their overestimation of deep convection while the New Tiedtke and Multi-scale KF (MSKF) schemes showed better performance. This study provides a valuable reference for further precipitation prediction research in the study region and adjacent areas.

To explore the vertical structure of precipitation in eastern China, multi-source remote sensing has been deployed to Jinan, a city in eastern China. Based on the data of K-band micro-rain radar and Ka-band cloud radar, a typical stratiform cloud precipitation experienced by Jinan city in eastern China from 12:00 on April 22, 2023 to 00:00 on April 23, 2023 is analyzed in detail. The results show that: 1) Both the Micro Rain Radar and the cloud radar can observe the bright band at the altitude of about 2.4 km, and the evolution of the bright band with time and height can be clearly observed. 2) The particles are mainly mixed phase in the bright band, and the particles above the bright band is mainly solid, such as snowflakes, while they are mainly liquid below the bright band, such as raindrops. 3) The characteristics of raindrop size distribution below the bright band show that the evaporation process in the early precipitation period is greater than that in the mature period, and the collisional coalescence in the mature period is more active. 4) Cloud radar can infer near-ground collisional coalescence and other microphysical processes from the radar reflectivity factor, which is consistent with the evolution of raindrop size distribution of Micro Rain Radar at different heights, so the synchronous observation of radar with different wavelengths can better show the microphysical processes of precipitation.

China is highly susceptible to landslides and debris flow disasters as it is a mountainous country with unique topography and monsoon climate. In this study, an efficient statistical model is used to predict the landslide risk in China under the Representative Concentration Pathway 8.5 by 2050, with the precipitation data from global climate models (GCMs) as the driving field. Additionally, for the first time, the impact of future changes in land use types on landslide risk is explored. By distinguishing between landslide susceptibility and landslide risk, the results indicate that the landslide susceptibility in China will change in the near future. The occurrence of high-frequency landslide risks is concentrated in southwestern and southeastern China, with an overall increase in landslide frequency. Although different GCMs differ in projecting the future spatio-temporal distribution of precipitation, there is a consensus that the increased landslide risk in China’s future is largely attributed to the increase in extremely heavy precipitation. Moreover, alterations in land use have an impact on landslide risk. In the Huang-Huai-Hai Plain, Qinghai Tibet Plateau, and Loess Plateau, changes in land types can mitigate landslide risks. Conversely, in other areas, such changes may increase the risk of landslides. This study aims to facilitate informed decision-making and preparedness measures to protect lives and assets in response to the changing climate conditions.

Hydrological models play an important role in water resources management and extreme events forecasting, and they are sensitive to the underlying conditions. This study aims to evaluate the impact of different soil-type maps and land-use maps on hydrological simulations and watershed responses by applying the WRF-Hydro (Weather Research and Forecasting Model Hydrological modeling system) distributed hydrological model to the Xijiang River basin. WRF-Hydro runs for four different scenarios for the period 1992-2013. FAO (Food and Agriculture Organization) and GSDE (Global soil-type maps and land-use maps are freely combined to form four scenarios. It is found that soil moisture and surface runoff are sensitive to soil-type maps, and absorbed shortwave radiation is found to be the least sensitive to soil-type maps. Absorbed shortwave radiation and heat flux are sensitive to land-use maps. The model performance of simulating soil moisture has increased when the soil-type map changes from FAO to GSDE and the land-use map changes from MODIS to CNLUCC for most stations. When the soil-type map changes from FAO to GSDE and the land-use map changes from MODIS to CNLUCC, the biases of simulating streamflow decrease. This study shows that the performance of the offline WRF-Hydro is significantly influenced by soil-type and land-use maps, and better simulation results can be obtained with more realistic underlying surface maps.

High-resolution gridded meteorological datasets offer valuable data sources for studying climate extremes. This paper evaluates and compares six high-resolution gridded datasets and two global climate extremes indices datasets. The study finds that, in terms of accuracy, most indices of gridded datasets are more consistent than those of observatory data. Among the gridded datasets, CN05 and CMFD_CDAT exhibit the best performance. The study examines the spatial patterns and temporal evolution of daily values for each dataset. Although there are numerical deviations between the datasets, they demonstrate similar long-term trends and spatial distribution patterns of cool day, cool night, warm day, warm night, and extreme wet days over the past three decades. The study also explores the accuracy of extreme indices of gridded datasets under different station density scenarios. The results suggest that the datasets are more accurate in regions with dense observations. However, the method used to generate the dataset can also affect the final climate indices calculation. In western China, where there is a complex topography and limited observation stations, the background fields of gridded datasets, such as reanalysis data and satellite data, can compensate for the defects of sparse observations. Therefore, when selecting a suitable gridded data product for the study of climate extremes, researchers should consider whether there are sufficient observational data sources and reliable background fields.

On 20 July 2021, a sudden rainstorm happened in central and northern Henan Province, China, killing at least 302 people. This extreme precipitation event incurred substantial socioeconomic impacts and resulted in serious losses. Accurate monitoring of such rainstorm events is crucial. In this study, qualitative and quantitative methods are used to comprehensively evaluate the abilities of 10 high-resolution satellite precipitation products [CMORPH-Raw (Climate Prediction Center morphing technique), CMORPH-RT, PERSIANN-CCS (Precipitation Estimation from Remotely Sensed Information Using Artificial Neural Networks), GPM IMERG-Early (Integrated Multisatellite Retrievals for Global Precipitation Measurement), GPM IMERG-Late, GSMaP-Now (Global Satellite Mapping of Precipitation), GSMaP-NRT, FY-2F, FY-2G, and FY-2H] in capturing this extreme rainstorm event, as well as their performances in monitoring different precipitation intensities. The results show that these satellite precipitation products are able to capture the spatial distributions of the rainstorm (e.g., its location in central and northern Henan), but all products have underestimated the amount of precipitation in the rainstorm center. With the increase in precipitation intensity, the hit rate decreases, the threat score decreases, and the false alarm rate increases. CMORPH-RT is better at capturing the rainstorm than CMORPH-Raw, and it depictes the rainstorm process well; GPM IMERG-Late is more accurate than GPM IMERG-Early; GSMaP-NRT has performed better than GSMaP-Now; and PERSIANN-CCS and FY-2F perform poorly. Among the products, CMORPH-RT performs the best, which has accurately captured the center of the rainstorm, and is also the closest to the station-based observations. In general, the satellite precipitation products that integrate infrared and passive microwave data are found to be better than those that only make use of infrared data. The satellite precipitation retrieval algorithm and the amount of passive microwave data have a relatively greater impact on the accuracy of satellite precipitation products.

On 20 July 2021, an abrupt rainstorm occurred in central and northern Henan province, China and killed at least 302 people. Such kind of extreme precipitation events have great impacts on social and economic development and human lives. Therefore, accurate monitoring of such rainstorm events is crucial. Qualitative and quantitative methods are used to comprehensively evaluate 10 high-resolution satellite precipitation products (CMORPH-Raw, CMORPH-RT, PERSIANN-CCS, GPM IMERG-Early, GPM IMERG-Late, GSMaP-Now, GSMaP-NRT, FY-2F, FY-2G and FY-2H) of this extreme rainstorm event, and the ability of these 10 products to represent different levels of precipitation are also analyzed. The results show that these 10 satellite precipitation products can capture spatial distribution characteristics of the rainstorm that is located in central and northern Henan, but all satellite precipitation products underestimate the precipitation value at the rainstorm center. As the precipitation level increases, the hit rate and the TS score decrease, and the false alarm rate increases. CMORPH-RT has a better ability to capture the rainstorm than CMORPH-Raw, and shows the ability to capture rainstorm process. GPM IMERG-Late is more accurate than GPM IMERG-Early. Compared to GSMAP-Now, the performance of GSMAP-NRT enhances. PERSIANN-CCS and FY-2F show a poor performance and cannot capture this extreme rainstorm event. Among them, CMORPH-RT has the best performance, since it accurately captures the rainstorm center and magnitude that are the closest to observations. In general, satellite precipitation products that integrate infrared and passive microwave information are better than the products that only use infrared information. The satellite precipitation retrieval algorithm and the amount of passive microwave data have a great impact on the accuracy of satellite precipitation products.

Predicting the change trend of water quality is an important evidence to maintain and govern the current water quality. The Grey Markov SC GM（1,1）model was used to forecast the index of permanganate in water source. The result shows the Grey Markov SC GM (1,1) model can not only reveal the overall change trend of water quality of the water source, but also overcome the error arising from the random fluctuation. As a result it is a feasible and practical method.

Based on the observational data of CN05. 1, this article evaluated and compared the performance of 39 CMIP5 (Coupled Model Intercomparison Project phase 5) models and 34 CMIP6 (CMIP phase 6) models in simulating surface air temperature of the three provinces in Northeast China (the Heilongjiang, Jilin and Liaoning Province) by Taylor diagram, skill scores (S value) and the composite rating indicators (M-r). Results showed that: 1) The CMIP6 models possess a relatively higher capability in simulating the temperature from the interannual variations of regionally averaged surface air temperature, spatial distributions of annual mean surface air temperature and its trend than CMIP5 models but shows a negative bias; 2) CMIP5 and CMIP6 preferred ensemble mean (MMES and MME6) generally performs better than the individual models. Compared with the MME5, MME6 shows a significant improvement in simulating the climatological spatial distribution of temperature and its trend, whilst is slightly inferior in simulating the interannual variations of regionally averaged temperature. Generally speaking, CMIP6 models exhibit a significant improvement compared to the CMIP5 models and MME6 has been proven effective at simulating the temperature in Northeast China.

Post-processing methods can be used to reduce the biases of hydrological models. In this research, six post-processing methods are compared: quantile mapping (QM) methods, which include four kinds of transformations, and two newly established machine learning frameworks [support vector regression (SVR) and convolutional neural network (CNN)] based on meteorological data and variation mode decomposition (VMD)-decomposed streamflow. These post-processing methods are applied to a distributed model (WRF-Hydro), and the evaluation is carried out over five watersheds with different areas in South China. The post-processing methods are separately applied to calibrated and uncalibrated models. The results show that the SVR- and CNN-based post-processing methods perform better than the QM methods in terms of daily streamflow simulations in different areas with different topographies in the Xijiang River basin. There are large uncertainties in the QM post-processing methods. The CNN-based post-processing performs slightly better than the SVR-based post-processing, but both methods can markedly improve the simulated streamflow. The CNN- and SVR-based post-processing frameworks are suitable for both calibration and test periods. The differences between post-processing with uncalibrated and calibrated models are quite small for SVR- and CNN-based post-processing, but large for QM post-processing. For WRF-Hydro, the CNN- and SVR-based post-processing methods consume much less time and computational resources than model calibration.

Study region: Xijiang River, South China. Study focus: This paper discusses the application of WRF-Hydro, a distributed hydrological model, to a complicated watershed. The model performance on simulating streamflow, soil moisture, soil temperature and evapotranspiration is evaluated. Changes and characteristics of streamflow and related variables simulated by the model are analyzed. New hydrological insights for the region: In this study, thirteen sensitive parameters used in this model are tested in large and small watersheds of the basin. It is found that large basins are more sensitive to base flow parameters than small basins. The WRF default soil type dataset is replaced by the Beijing Normal University (BNU) soil type dataset that is more accurate. The model can simulate temporal changes of streamflow as well as temporary variabilities of hydrological variables. The model can be applied in small and large watersheds. The trend of streamflow in the basin is spatially uneven with an increasing trend in the upper part and a decreasing trend in the lower part. The changes in streamflow are partly related to precipitation and evapotranspiration trends. The trends of 30-day maximum/minimum streamflow are not significant, but the reversals demonstrate significant changes during 1980-2018. This study is expected to serve as a reference for the application of the model in this basin and model calibration over different sizes of study area.

2017年6月18日14时,北京门头沟区石羊沟流域爆发山洪泥石流,造成人员伤亡.使用站点降水、融合降水、雷达降水和卫星降水驱动WRF-Hydro水文模型运行,评估不同降水资料在本次山洪模拟中的效能.结果表明,各种降水的空间分布和时间演变较为相似,但除雷达降水外,其他降水均存在较大程度的低估.利用雷达降水模拟出的山洪与实际山洪在出现时间和洪峰流量上最为相近,利用其他降水模拟的山洪出现时间相近但低估了洪峰.目前常规台站观测降水暂不满足对小尺度突发山洪的研究和预警需求,亟需融合更多的自动站和雷达、卫星等实时高分辨率降水产品.

天气预报是指一周内至两周时间尺度的气象预报,而月季及以上时间尺度的预报则属于气候预测范畴.中国的气候预测起步很早,无论在研究工作中还是在业务应用上都取得了显著成就.文中扼要回顾了这些研究和业务发展成就,重点包括:对于季风和梅雨、寒潮的早期认知和后期研究发现、早期气候预测业务发展概况、动力气候预测的早期探索、动力-统计气候预测方法的研制和应用、气候预测模式的发展以及初始化和多模式集合预测、东亚气候系统变异的全方位探索、气候预测范畴的不断拓展和气候预测研究的不断创新.也对未来气候预测研究和业务发展提出了几个重大挑战性课题,涉及不同时间尺度气候变异过程之间的相互作用、季节内至年代际气候预测、气候系统模式及初始化、动力-统计相结合的气候预测方法等方面.

A rainfall threshold for landslide occurrence at a national scale in China has rarely been developed in the early warning system for landslides. Based on 771 landslide events that occurred in China during 1998–2017, four groups of rainfall thresholds at different quantile levels of the quantile regression for landslide occurrences in China are defined, which include the original rainfall event–duration (E–D) thresholds and normalized (the accumulated rainfall is normalized by mean annual precipitation) (EMAP–D) rainfall thresholds based on the merged rainfall and the Climate Prediction Center Morphing technique (CMORPH) rainfall products, respectively. Each group consists of four sub-thresholds in rainy season and non-rainy season, and both are divided into short duration (<48 h) and long duration (≥48 h). The results show that the slope of the regression line for the thresholds in the events with long durations is larger than that with short durations. In addition, the rainfall thresholds in the non-rainy season are generally lower than those in the rainy season. The E–D thresholds defined in this paper are generally lower than other thresholds in previous studies on a global scale, and a regional or national scale in China. This might be due to there being more landslide events used in this paper, as well as the combined effects of special geological environment, climate condition and human activities in China. Compared with the previous landslide model, the positive rates of the rainfall thresholds for landslides have increased by 16%–20%, 10%–17% and 20%–38% in the whole year, rainy season and non-rainy season, respectively.

滑坡演变是一个长期且复杂的过程,地质灾害体在各类数据中有不同的表现特征,同时各类数据源在滑坡不同阶段有一定的适用性.采用多源数据融合方法,利用各类数据源在滑坡体中不同的特征与适用性,研究黄泥坝子滑坡在滑前-滑中-滑后动态演变过程中的变形破坏特征和时空演化规律.结果 表明,黄泥坝子滑坡变形破坏过程可分为4个阶段:启动阶段、加速变形(加速滑移)阶段、前缘扩展(减速滑移)阶段、渐进稳定阶段;黄泥坝子滑坡是在自重、降雨渗透、地震及人类工程活动造成的震动等多效应影响作用下形成的蠕滑-拉裂式滑坡.总结各类数据在滑坡不同阶段的应用,滑前阶段可应用合成孔径雷达技术、光学影像与地形数据确定潜在滑坡体;滑中阶段可利用光学遥感影像分析滑坡堆积体整体变形与演变趋势,利用全球定位系统(global positioning system,GPS)持续观测局部变形;滑后阶段可通过现场调查确定地质灾害体工程地质特征.

根据DELWARE温度和降水数据、GLDAS蒸散发数据和湄公河干流9个水文站的实测径流,采用回归分析、均值T检验和低通滤波,分析了该流域气候和径流在1950-2017年间的变化情况,经分析表明流域内气候和径流在研究时段内有较大变化,而且在不同的月份呈现不同的变化特征.流域年平均温度整体呈增加趋势,2008年后的平均温度相对2008年前平均温度有显著增加;流域年平均降水的变化幅度不大;流域平均蒸散发在12月一次年2月呈下降趋势,其他月份呈增加趋势,2008-2017年月平均蒸散发与1950-2007年月平均蒸散发相比大幅提升,尤其是在6-10月;湄公河流域年径流没有显著变化,但径流在12月一次年4月呈上升趋势,7-10月呈下降趋势,其中,上升趋势比下降趋势显著,1-4月径流上升趋势在2008年之后更为显著;最小径流在2008年后有显著增加趋势,最大径流在2008年后呈下降趋势;年流量逆转次数自20世纪90年代起有明显升高趋势.通过比较温度、降水、蒸散发和径流在不同时间段的变化情况,可以看出径流在2008年后变化趋势和气候自然变化关系不显著,但可能跟大坝蓄水能力显著提高等人为活动有较大关系.

基于动力降尺度预测系统,中国科学院大气物理研究所竺可桢-南森国际研究中心对2018年夏季我国极端降水日数及滑坡泥石流灾害的发生风险进行了超前4个月的实时预测试验.与实测结果相比,该系统对2018年夏季我国极端降水日数空间分布的预测与实况基本相符,但大部分地区存在明显低估;滑坡泥石流的预测结果与目前统计的由于降水引发的滑坡泥石流灾害事件的分布基本吻合.此次预测试验表明,中国科学院大气物理研究所竺可桢-南森国际研究中心发展的动力降尺度预测系统对我国夏季极端降水和滑坡泥石流灾害具有一定的预测能力,具有实时预测价值.

对跨南海西南次海盆及两侧陆缘的一条1050km长的、包括海底地震(OBS)、长排列多道地震和重磁在内的综合地球物理探测剖面(CFT)进行了构造成像和研究.在多道地震成像基础上建立了CFT剖面初始速度模型,进而通过初至波层析成像方法反演了CFT剖面的速度结构模型,在重力异常资料的约束下建立了CFT剖面的综合地壳结构模型.讨论了沿CFT剖面出现的下地壳高速体、龙门海山的低密度物质等地质问题.结果 表明,下地壳高速层在北部陆坡、西南海盆和南部南沙地块均有分布,厚度在0～4km之间,可能与陆缘下地壳物质和地幔物质熔融混合,以及深海盆海底扩张期间构造拉伸导致地幔蛇纹岩化有关.

利用滑坡敏感性分布和降雨阈值公式建立了一个滑坡泥石流统计模型,该模型可以用于中国大尺度范围内的滑坡泥石流预警.使用CMORPH卫星降水驱动该统计模型,对2016-2017年的106起滑坡泥石流事件进行了验证分析.结果表明,该模型能较好地预警大多数滑坡泥石流事件,其中对72.1％的雨季滑坡泥石流事件能较好预警,但对非雨季的事件只有35％能较好预警,对雨季的预警效果明显优于非雨季.由于滑坡泥石流主要发生在雨季,因此该模型总体上具有较好的效能.该模型对于强降雨引发的快速滑坡事件具有较好的预警能力,但对于由强度较小、持续时间较长的降雨引发的慢过程滑坡事件的预警效果有待提升.利用该统计模型以及CMORPH实时卫星降水产品,可以建立滑坡泥石流大尺度实时预警系统,对滑坡泥石流减灾防灾具有一定意义.