Accurate precipitation forecast verification (PFV) is essential for improving forecasting models and supporting disaster management. However, current PFV methods remain limited, point-to-point matching is overly sensitive to minor errors, whereas spatial verification typically necessitates parameter tuning or heuristic rules derived from expert knowledge, which constrains their availability. To tackle these issues, we are inspired by the success of deep learning in image verification through extracting high-level features, and thus propose a self-supervised contrastive learning-based PFV method (CLPFV). First, CLPFV uses precipitations augmentation (displacement, intensity, area) to simulate actual forecast errors and construct positive and negative training sample pairs. Subsequently, with a novel loss function proportionally penalizing forecast errors, a backbone network is trained in CLPFV to extract high-level precipitation features. Finally, the cosine similarity of features is calculated as CLPFV's verification score. Experiments demonstrate that CLPFV outperforms traditional and spatial verifications in different degrees of forecast errors and aligns better with expert assessments. In general, CLPFV offers an efficient deep learning solution for PFV tasks.

In numerical weather prediction, dynamic initialization (DI) of tropical cyclones (TCs) and bogus data assimilation (BDA) are two widely used vortex initialization methods, both of which substantially affect the track and intensity of simulated TCs. This study investigated the impact mechanisms of DI and BDA using the Typhoon Regional Assimilation and Prediction System, which is a moving nested configuration of the Weather Research and Forecasting model designed specifically for TC prediction. Sensitivity experiments demonstrated that the DI approach tends to systematically overestimate TC intensity, primarily owing to moisture updates. Adopting BDA can alleviate this overestimation but often introduces underestimation of TC intensity, thereby limiting its applicability. To address these limitations, this study proposed an optimized strategy that leverages the strengths of both methods: BDA is activated when the bogus vortex intensity is close to the observations, whereas DI is used when the bogus vortex intensity is substantially underestimated. To avoid DI-induced overestimation, full-variable updates are applied only under strong predicted TC intensification, and DI is deactivated during weakening phases. The optimized strategy, tested on six TC cases, reduced errors in TC track, minimum sea level pressure, and maximum wind speed forecasts by 21.6%, 38.6%, and 33.8%, respectively, relative to a non-optimized DI and BDA experiment. Improved TC track and intensity predictions also enhanced quantitative precipitation forecast scores across thresholds by 9.6%-54.8%. These results highlight the critical role of accurate vortex initialization for reliable TC forecasting.

During the summer monsoon season, with prevailing southwesterly winds, rainfall usually peaks in the afternoon on northern Hainan Island (HNI), China. On days of stronger prevailing winds, convection initiation (CI) often occurs in the north central HNI. Through simulations of a typical case of 20 June 2017 using a 2 km grid, we identify a primary CI mechanism over northern HNI that is driven by daytime boundary-layer (BL) vertical mixing and the resulting low-level convergence. Overnight and into the early morning, a wedge-like wind structure forms in the BL along an east-west vertical cross-section over northern HNI, which is characterized by a low-level southeasterly wind component, deeper in the east and shallower in the west, undercutting the northwesterly wind component aloft. By noon, surface-heating-induced BL vertical mixing transports the upper-level northwesterly component to the surface on the west side, creating a zone of convergence with the low-level southeasterly component from the east and triggering convection around noon. The wedge structure arises from the orographic effects of the mountains over southern HNI, when southerly flows bypassing the mountains from the west side gain more westerly component, while the flows climbing over the mountains descend on the leeside and cause isentropic surfaces to dip, lowering the layer possessing westerly component. Sensitivity experiments show that sea breezes are not responsible for initiating the convection in north central HNI, contrary to earlier suggestions, while BL horizontal convective rolls play only a secondary role in modulating the timing and locations of CI.

Previous studies have shown that there are significant errors in temperature forecasts in convection-permitting numerical weather prediction models, causing biases in wind and precipitation forecasts. In this study, the main sources of temperature forecast errors that stem from the uncertainty of physical parameterization schemes were investigated. The temperature forecasts were evaluated based on daily predictions over a month-long period. Overall, for all tested schemes, the largest biases originated from errors in cloud forecasting. Key factors affecting both temperature and cloud fraction predictions within each physical scheme were examined. These included key radiation processes, shallow cumulus convection, land-surface types, and the utilization of an urban model. For the radiation scheme, the number of quadrature points affected the amount of incoming and outgoing radiation and was a crucial setting that influenced the temperature forecasts. The subgrid cloud parameterization and the order of streams in the radiation scheme further enhanced the forecast differences. For the shallow convection scheme, different parameter settings mainly altered the cloud fraction simulation in cloudy areas rather than in clear skies. The subsequent temperature forecast was sensitive to the cloud fraction simulation. Parameter tuning was needed prior to applying the shallow convection scheme. For different land-surface datasets and urban model applications, accurate identification of land-surface types is crucial as it reflects the real surface albedo. This accuracy significantly influences the reliability of temperature forecasts. In addition, the more realistic heat capacity settings of buildings in the urban model improved the temperature forecast at the city scale. Temperature is one of the critical factors that can significantly influence wind and precipitation forecasts. In numerical weather prediction models, various physical schemes can affect temperature forecasts. Some schemes predict consistently higher or lower temperature than others, leading to systematic forecast biases. However, the reasons behind these discrepancies are not fully understood. This study aims to identify the primary factors that drive differences in forecasts between different physical schemes. Our findings indicate that the choice of the number of quadrature points in radiation scheme, the tuning of key parameters in shallow convection scheme, and the adoption of more realistic underlying surface types and urban canopy models are crucial for accurate temperature forecasting.

This study uses three double-moment bulk microphysics schemes to simulate four rainstorm events in China, aiming to evaluate the simulated raindrop size distribution (RSD) in heavy rainfall. Both the Thompson and Milbrandt-Yau schemes exhibit nearly constant near-surface raindrop mean diameter Dmwhen the rainwater content exceeds 2 g m23, noticeably diverging from Dm values retrieved by polarimetric radar. The variability in Dm between deep and shallow convection also steadily decreases from the top of warm-cloud layer to the surface. The National Severe Storms Laboratory (NSSL) scheme shows an unrealistic increase in the near-surface Dmvalues with increasing rainwater content. By investigating the relative contribution to the total raindrop number transfer rate, the raindrop self-collection and breakup processes are identified to control all the simulated RSDs in heavy rainfall. In these schemes, the raindrop selfcollection/breakup rate is anomalously positively correlated with the rainwater content. The self-collection/breakup efficiency Ec is a simple function of Dm in the schemes. In Thompson and Milbrandt-Yau, Ec is positive below a cutoff Dm value, indicating net raindrop self-collection, and negative above it, indicating net raindrop breakup. In contrast, Ec is always nonnegative in NSSL, suggesting net raindrop self-collection rather than breakup. Two sensitivity experiments are designed by swapping the Ec values between Thompson and NSSL. Simulated RSD characteristics from the two schemes are effectively exchanged. Such changes in self-collection and breakup settings directly affect extreme rainfall forecasting by altering raindrop fall velocities and affecting other warm-rain processes, with an impact of up to 30% on the maximum hourly rainfall forecasts. SIGNIFICANCE STATEMENT: Accurate parameterization of cloud microphysical processes and representation of raindrop size distribution characteristics are essential for high-accuracy quantitative precipitation forecast in numerical models. This study evaluates the simulated raindrop size distribution in multiple rainstorm events and the ability of key microphysical process parameterizations in heavy rainfall. Deficiencies in cloud parameterizations and raindrop size distribution simulations are revealed and found to obviously affect the quantitative forecasting of heavy rainfall. The findings also provide valuable insights into improving heavy rainfall forecasting from the perspective of cloud microphysics parameterization.

Including the sub‐grid terrain solar radiative effect (STSRE) of East Asia in the convection‐permitting Weather Research and Forecasting model can clearly improve the forecast skill of Meiyu rainfall in the Yangtze‐Huaihe River Basin (YHRB). However, the STSRE of which region mainly leads to the improvement of Meiyu rainfall forecast remains unclear. This study systematically explores the impacts of local and remote STSRE on the Meiyu rainfall forecast, their relative importance, and underlying physical mechanisms. Results show that relative to the remote STSRE of Tibetan Plateau (TP), the local STSRE of YHRB leads to relatively larger improvement of Meiyu rainfall forecast with the root mean square error decreased by 6.24% and Taylor score increased by 2.76%. Further mechanism analysis indicates that the STSRE of the TP weakens the TP heat source, leading to weakening and westward shifting of the South Asian High, which in turn suppresses the divergence in the upper troposphere and the ascending motion over the YHRB and thereafter improves Meiyu rainfall forecast. In contrast, the STSRE of the YHRB decreases surface incident solar radiation and surface thermal conditions, producing a cold anomaly and anticyclonic circulation difference in the lower troposphere over the YHRB. This anticyclonic difference directly weakens southwesterly water vapor transport, thereby reducing the overestimation of Meiyu rainfall. This study highlights the importance of local terrain (rather than the TP) in enhancing Meiyu rainfall forecast accuracy for 3–4 days weather forecasts.

The Northeast China cold vortex (NECV) is a major weather system producing heavy rainfall in northern China, yet the influence of complex terrain, especially orographic gravity waves (OGWs), on such heavy rainfall remains poorly understood. This study investigates the impact of OGW drag (OGWD) parameterization on an NECV heavy rainfall event over the southern Yanshan Mountains on 6 July 2011, using the Weather Research and Forecasting Model at 3-km resolution. Results show that the OGWD parameterization can weaken the NECV circulation and diminish the orographic lifting and moisture transport over the southern slope of the Yanshan Mountains given the decelerated upslope flow. Therefore, the overestimation of the heavy rainfall intensity in the absence of OGWD parameterization was alleviated significantly, indicating the importance of OGWD parameterization even in high-resolution numerical models. However, the parameterization of OGWD introduced weak but widespread spurious rainfall ahead of the Taihang Mountains, as it decelerated the northwesterly downslope winds on the southeastern slope of the Taihang Mountains which enhanced the upslope moisture transport. This spurious rainfall was mitigated significantly when using a revised OGWD scheme accounting for the nonhydrostatic effect (NHE) on the surface momentum flux of vertically propagating OGWs. The NHE more notably attenuated the OGWD over the Taihang Mountains than over the Yanshan Mountains, which strengthened the NECV northwesterly flow downgliding the Taihang Mountains and inhibited the moisture transport.

Subgrid-scale (SGS) turbulent mixing is essential to convection-permitting simulations where turbulent fluxes are partially resolved and partially subgrid scale. This study investigates the characteristics of the three-dimensional (3D) SGS fluxes of an idealized squall line in a weak sheared environment. The 3D SGS fluxes on kilometer-scale grids are obtained by coarse graining a benchmark large-eddy simulation (LES) conducted with the Advanced Regional Prediction System model. Countergradient (CG) transport is found in both horizontal and vertical SGS fluxes in the updraft region, which results from nonlocal transport associated with the tilted updraft. Using moist-conserved variables for the identifica-tion of CG fluxes, the spatial distribution and the occurrence rates of the CG and the downgradient fluxes are investigated. A scale-similarity (Hgrad) and a conventional gradient-diffusion [turbulence kinetic energy (TKE)] SGS closures are then evaluated at kilometer-scale resolutions against the LES. Both the offline evaluation and the online simulations demonstrate improvements of the Hgrad closure over the TKE closure, mostly due to the former's ability to represent CG fluxes. Through sensitivity experiments, we investigate the role of horizontal flux parameterization in predicting the structure and intensity of deep convection with tilted updraft. SIGNIFICANCE STATEMENT: Current-day numerical weather prediction models operate on kilometer-scale grids that permit partially explicit resolution of deep convection. Accurate parameterization of the unresolved subgrid-scale (SGS) turbulence is key to improving kilometer-scale simulations of deep convection. Former studies on SGS turbulence for kilometer-scale grids are often based on upright deep convection and point to the essential role of vertical SGS flux parameterization. We investigate a vertically tilted deep convective system under the influence of the cold pool. Our findings show that the contribution of the horizontal SGS fluxes may be comparable to the vertical SGS fluxes in kilometer-scale simulations of convective storms with tilted structures.

This study evaluated the precipitation forecast produced by the operational China Meteorological Administration Mesoscale model (CMA-MESO) during the "super violent" Meiyu season of 2020. Generally, CMA-MESO, which runs with similar to 3-km-grid resolution, is able to reproduce the distribution and diurnal variation of precipitation. However, the precipitation amount is greatly overestimated, especially in eastern coastal areas of China. Precipitation in that region usually occurs with two peaks: one in the morning that mostly reflects organized precipitation systems, and the other in the afternoon generated mostly by local convection. Analyses showed that overestimation of low-level wind speed is the main reason for the overestimation of precipitation. CMA-MESO produces low-level winds that are overly strong, which greatly enhance the predicted convergence at night, leading to overestimation of precipitation. Additionally, the stronger wind speed increases the estimated transport of water vapor to the eastern coastal area, producing fake convection near the coastal mountains as the perturbed wind direction turns toward the mountain area in the afternoon. In comparison with ERA5, CMA-MESO tends to overestimate (underestimate) the temperature in the northwest (southeast), and the larger temperature gradient increases the pressure gradient, resulting in the stronger low-level wind speed. The resolution of the operational China Meteorological Administration Mesoscale model (CMA-MESO) was upgraded to 3 km in 2020. However, the overall performance of the precipitation forecast has not been evaluated comprehensively, and the main factors causing precipitation forecast biases are not well understood. This study analyzed the CMA-MESO precipitation forecasts during the 2020 super Meiyu season. Generally, CMA-MESO well reproduced the distribution and propagation of precipitation, but the intensity was overestimated, especially in eastern coastal areas of China. CMA-MESO tended to produce an overly strong southwesterly low-level wind that transported too much warm moist air to eastern coastal areas, resulting in excessive rainfall. Further comparative evaluation suggested that the overestimation of low-level winds might be related to the larger NW-SE temperature gradient of CMA-MESO. The findings of the current study could provide guidance for improving physical parameterization in the future. Performance of the operational CMA-MESO in China at convection-permitting resolution was evaluated CMA-MESO can well reproduce the distribution and propagation of precipitation, but overestimates precipitation in eastern coastal areas Overestimation of low-level wind speed is the main reason for the overestimation of precipitation in eastern coastal areas

The concave mountainous area of Pearl River Delta is a summer rainfall hotspot along the South China coast due to the presence of warm-moist monsoon flow and complex orography. This study evaluated the performance of a convection-permitting Weather Research and Forecasting (WRF) model in forecasting nocturnal rainfall in this area, focusing on days with low level southwesterly winds during the summers of 2013–2015. Results showed that the nocturnal rainfall exhibited two centers, one located along the large-scale northern mountains and the other along the small-scale Huadu Hill. WRF demonstrated superior performance in predicting rainfall over the northern mountainous region. In contrast, WRF significantly underestimated nocturnal rainfall both near local Huadu Hill and in the foothill area of northern mountains, which were strongly influenced by local forcings. Using high-resolution analyses from Variational Doppler Radar Analysis System (VDRAS), which assimilated both Doppler radar and Automatic Weather Stations observations, we firstly investigated the mesoscale mechanism governing the convection initiation (CI) of a typical localized nocturnal convection. Results showed that the enhanced prevailing low-level southerly winds, combined with local circulation induced by the urban heat island effect and orographic forcings, led to the formation of low-level convergence and strong updrafts before CI. Subsequently, we identified the sources of forecast errors in triggering CI. Results revealed that WRF severely underestimated thermal contrast between the Guangdong-Hong Kong-Macao Greater Bay Area urban agglomeration and the concave mountains, leading to the absence of the northeastern/northern inflows toward the cities. Consequently, low level convergence and updrafts near the CI position were too weak to lift air parcels above the severely overestimated level of free convection, thereby failing to trigger the convection.VDRAS-based sensitivity experiments, with a specific focus on assimilating surface temperature, validated the crucial role of urban-mountain thermal contrast on local winds that triggered the nocturnal convection.This study underscores the significance of urban-mountain thermal contrast and local circulations in determining nighttime precipitation formation and prediction, particularly in geographically complex regions characterized by concave mountains and urban agglomerations. The findings highlight the need for improved representation of these local forcing mechanisms in numerical weather prediction models to enhance their accuracy in forecasting nocturnal rainfall events.

With the development of refined numerical forecasts, problems such as score distortion due to the division of precipitation thresholds in both traditional and improved scoring methods for precipitation forecasts and the increasing subjective risk arising from the scale setting of the neighborhood spatial verification method have become increasingly prominent. To address these issues, a general comprehensive evaluation method (GCEM) is developed for cross-scale precipitation forecasts by directly analyzing the proximity of precipitation forecasts and observations in this study. In addition to the core indicator of the precipitation accuracy score (PAS), the GCEM system also includes score indices for insufficient precipitation forecasts, excessive precipitation forecasts, precipitation forecast biases, and clear/rainy forecasts. The PAS does not distinguish the magnitude of precipitation and does not delimit the area of influence; it constitutes a fair scoring formula with objective performance and can be suitable for evaluating rainfall events such as general and extreme precipitation. The PAS can be used to calculate the accuracy of numerical models or quantitative precipitation forecasts, enabling the quantitative evaluation of the comprehensive capability of various refined precipitation forecasting products. Based on the GCEM, comparative experiments between the PAS and threat score (TS) are conducted for two typical precipitation weather processes. The results show that relative to the TS, the PAS better aligns with subjective expectations, indicating that the PAS is more reasonable than the TS. In the case of an extreme-precipitation event in Henan, China, two high-resolution models were evaluated using the PAS, TS, and fraction skill score (FSS), verifying the evaluation ability of PAS scoring for predicting extreme-precipitation events. In addition, other indices of the GCEM are utilized to analyze the range and extent of both insufficient and excessive forecasts of precipitation, as well as the precipitation forecasting ability for different weather processes. These indices not only provide overall scores similar to those of the TS for individual cases but also support two-dimensional score distribution plots which can comprehensively reflect the performance and characteristics of precipitation forecasts. Both theoretical and practical applications demonstrate that the GCEM exhibits distinct advantages and potential promotion and application value compared to the various mainstream precipitation forecast verification methods.

An EF4-rated supercell tornado occurred on 3 July 2019 in Kaiyuan, China, causing heavy casualties. A three-level nested-grid high-resolution numerical simulation is used to investigate the initiation of the tornadic supercell. Automatic weather station (AWS) data, FY-4A visible satellite data, and Doppler radar data are used to verify the model simulation. The most important aspects of the simulated presupercell mesoscale convective system (MCS) and the initiation of the supercell agree with observations. Detailed investigation of the model results reveals that the initial cells form fi rst above a convective boundary layer (CBL) on the dry side of a surface dryline. Above the CBL is a moist layer in terms of relative humidity, and the layer is stable. Convectively generated gravity waves (GWs) emanating from the MCS and propagating southward along the stable layer above the CBL provide localized forcing for the actual triggering of initial cells at specific locations. The associated perturbation potential temperature and vertical velocity patterns confirm that the GWs trigger a series of cloud bands. The additional lifting by the updraft of a horizontal convective roll in the CBL underneath the GW updraft works together to promote faster growth of the initial cell that later becomes the supercell. Examination of the Scorer parameter profiles shows favorable conditions for vertical trapping of GWs along the waveguide in the stable layer, preventing the radiation of wave energy to the upper levels.

The parameterization of orographic gravity wave drag (OGWD) is essential for accurate numerical weather prediction in regions of complex terrain. Current OGWD schemes assume hydrostatic orographic gravity waves (OGWs) but the parameterized OGWs in fi ne-resolution models with narrow subgrid-scale orography can be significantly affected by the nonhydrostatic effects (NHE). In our recent work, the OGWD scheme in the Model for Prediction Across Scales (MPAS) was revised by accounting for the NHE on the surface wave momentum fl ux of upward-propagating OGWs. Herein, the revised OGWD scheme is implemented in the Weather Research and Forecasting (WRF) Model to evaluate its performance in short-range weather forecast. Two sets of 36-h WRF simulations are conducted for nine Northeast China cold vortices (NECVs) that occurred in the warm season of 2011 using the original and revised OGWD schemes. Results show that the WRF Model tends to underestimate the intensity of the NECVs, producing too high geopotential height. When accounting for the NHE in the OGWD scheme, the NECV intensity biases are significantly reduced. Analyses reveal that the NHE act to weaken the lower-tropospheric OGWD by decreasing the surface wave momentum fl ux, which strengthens the NECV in the lower troposphere. Consequently, the strengthened low-level cyclonic circulation increases the posttrough cold advection to the southwest of the NECV which in turn enhances the NECV in the mid-upper troposphere with reduced geopotential height. The NHE are found to increase as the model horizontal resolution increases, suggesting greater importance of NHE in the OGWD parameterization of high-resolution numerical models.

Convective initiation ahead of a surface cold front within the Northeast China cold vortex of June 1, 2021, is investigated using observations and a convection-permitting simulation. The initiation of the convection is elevated without identifiable surface mesoscale forcing. Air feeding the initial convective cells originates from at least 1 km above the ground where parcel-based convective available potential energy is sufficiently large. An elevated front with large equivalent potential temperature gradient and flow convergence across is primarily responsible for the initiation of elevated convection and later organization into a deep convective line. The elevated front is similar to the upper cold front within extratropic cyclones studied elsewhere, but its altitude is lower in this case. Backward-trajectory analysis of air parcels on the cold side of the elevated front shows that dry air intrusion reinforces the elevated front by increasing the equivalent potential temperature gradient and moment convergence across the front. The convergence forcing combined with access to convective unstable air at the leading edge of the westerly flows east of the elevated front causes the initiation of convection. The development and spreading of a surface cold pool help push the convective line eastward away from the surface cold front located to the west. Orographic gravity waves provide additional forcing so that convection cells are initiated first at the upward branches of the gravity waves, but the overall organization of initial cells into a solid convective line is controlled mainly by mesoscale structures associated with the elevated front.

The low-echo centroid (LEC) storm, characterized by the dominance of warm-rain processes and high precipitation efficiency (Vitale & Ryan, 2013), is generally associated with high-intensity rainfall events in tropical and subtropical regions (Hamada et al. 2015). While many double-moment microphysical parameterization schemes are primarily designed for deep convection or cold-rain processes, research on their performance in simulating LEC warm-cloud precipitation systems is limited. In this study, we investigate an extreme rainfall case that occurred on 20 July 2016 in northern China, resulting in over 600 mm of maximum 24-hour accumulated rainfall. According to radar observations, the case is characterized by LEC structure. Using the Weather Research and Forecasting (WRF) model with 4 km grid spacing, we simulate this event employing the Morrison, Thompson, and Milbrandt-Yau double-moment microphysics schemes. Evaluation based on simulated infrared brightness temperature (BT) using a radiative transfer model and simulated reflectivity reveals that while different microphysics schemes generally predict rainfall amount, location, and propagation accurately, they fail to replicate the three-dimensional cloud structures. The simulated convective cores (>35dBZ) are higher than -10 °C, indicating active cold-cloud processes, while the observed LEC suggests the dominance of warm-cloud processes. The model produces an excessive number of upper-level clouds and overshooting clouds, also overpredict the cloud-top height. Sensitivity experiments show that simulated brightness temperatures are primarily influenced by the concentration of cloud ice particles. Morrison and Milbrandt-Yau microphysics schemes produces an overabundance of cloud ice particles in the upper layer, leading to the overproduction of uppe-level cloud and incorrect representation of cloud top height. Warm-rain processes are not fully developed, and the cold-rain processes are not effectively restrained, resulting in unrealistic cloud structure. By adjusting microphysical processes in the schemes, such as increasing cloud water number concentration, the simulated convective cores align more closely with the observed ones. In summary, while the current microphysics schemes effectively simulate rainfall intensity and propagation, there is a clear need for improvement in simulating the cloud particle distribution and vertical structure of LEC storms. Our findings underscore the importance of refining microphysical parameterization schemes for accurate simulation of extreme rainfall events.

Readme for GCEM Data and Code 1 Data 1.1 Observed Precipitation Data 1.1.1 /DataSoftware/01Observed_precipitation_data/2019071612 Hourly precipitation from 00:00 to 12:00 UTC on July 16, 2019 for Case 1 surfr01h.nc surfr02h.nc surfr03h.nc surfr04h.nc surfr05h.nc surfr06h.nc surfr07h.nc surfr08h.nc surfr09h.nc surfr10h.nc surfr11h.nc surfr12h.nc 1.1.2 /DataSoftware/01Observed_precipitation_data/2020061312 Hourly precipitation from 00:00 to 12:00 UTC on June 13,2020 for Case 2 surfr01h.nc surfr02h.nc surfr03h.nc surfr04h.nc surfr05h.nc surfr06h.nc surfr07h.nc surfr08h.nc surfr09h.nc surfr10h.nc surfr11h.nc surfr12h.nc 1.2 Forecasted Precipitation Data 1.2.1 /DataSoftware/02Forecasted_precipitation_data/2019071612 Data for Case 1 during 00:00-12:00 UTC on July 16, 2019 WRF3.2019071600000.nc (initial field at 12:00 UTC on July 16, 2019) WRF3.2019071600012.nc (12-hour accumulated precipitation during 00:00–12:00 UTC on July 16, 2019) 1.2.2 /DataSoftware/02Forecasted_precipitation_data/2020061312 Data for Case 2 during 00:00-12:00 UTC on June 13,2020 WRF3.2020061300000.nc (initial field at 12:00 UTC on June 13,2020) WRF3.2020061300012.nc (12-hour accumulated precipitation during 00:00–12:00 UTC on June 13,2020) 2Code and Configuration Files 2.1 GCEM of grid precipitation forecast data (/DataSoftware/03Software_Configuration_Results_of_GCEM) pastonc6hd2.f90 Main program, reads observed and forecasted precipitation data, performs GCEM verification, and outputs result files. module_skinput.f90 Subprogram, module for reading one or more observed precipitation grid file module_ybinput.f90 Subprogram, module for reading the start (or end) forecasted precipitation grid file mod_uxpasid2.f90 Subprogram, module for used to perform GCEM verification on forecasted data module_outnc.f90 Subprogram, module for outputting the verification results in netCDF file format compilePAS10mmd2.sh Used to compile source files to generate executable file under Linux r12hfile.txt Configuration file, used to specify the latitude and longitude range for data source and verification pastonc6hd2.exe Executable file 2.2 TS of grid precipitation forecast data (/DataSoftware/04Software_Configuration_Results_of_TS-Score) tsmain01.f90 Main program, reads observed and forecasted precipitation data, performs TS verification, and outputs result files. module_skinput.f90 Subprogram, module for reading one or more observed precipitation grid file module_ybinput.f90 Subprogram, module for reading the start (or end) forecasted precipitation grid file module_uxtsiTure.f90 Subprogram, module for used to perform TS verification on forecasted data compileTS.sh Used to compile source files to generate executable file under Linux r12hfile.txt Configuration file, used to specify the latitude and longitude range for data source and verification tsmain01.exe Executable file 2.3 PAS mini-program (/DataSoftware/05Software_of_PAS) pas10ux.f90 Main program, used to perform PAS verification on single point precipitation forecast mod_uxpasid2.f90 Subprogram, Module for PAS of single point forecast compilePAS10ux.sh Used to compile source files to generate executable file under Linux pas10ux.exe Executable file 3 Output files 3.1 GCEM verification results (/DataSoftware/03Software_Configuration_Results_of_GCEM/Results_GCEM) rainverd2019071612012.nc Result file in netCDF format rainverd2019071612012.nc.txt Result explanation file in netCDF format rainverd2020061312012.nc Result file in netCDF format rainverd2020061312012.nc.txt Result explanation file in netCDF format 3.2 TS verification results (/DataSoftware/04Software_Configuration_Results_of_TS-Score/Results_TS) ts12h2019071612.txt TS result file in text format ts12h2020061312.txt TS result file in text format 4 Compiling Environment The verification program runs in a UNIX environment and requires the intel compiler (v2017) and the netCDF (v4.6.1) support library UNIX Environment Settings # .bashrc module load intel/intel-compiler-2017.5.239 module load intelmpi/2019.6.154 export F90=ifort export NETCDF=/public/software/mathlib/netcdf/4.6.1_intel-2017_mpi-2017_hdf5-1.8.20-intel2017 export NETCDF_LIB=$NETCDF/lib export NETCDF_INC=$NETCDF/include export PATH=$NETCDF/bin:$PATH export LD_LIBRARY_PATH=$NETCDF/lib:$LD_LIBRARY_PATH 5 Compiling and Running Steps 5.1 The steps for case 1 during 00:00–12:00 UTC on July 16, 2019 1. Creating an installation and running sub-directory mkdir p2019 2. Copying data sources, code files and configuration files to this directory 3. Running in this directory ./compilePAS10mmd2.sh Compile to generate executable file (pastonc6hd2.exe) ./compileTS.sh Compile to generate executable file (tsmain01.exe) 4. Modifying the configuration file (r12hfile.txt) Mainly modifying the data source path for lines 4, 19, and 23 5. Run the executable files ./pastonc6hd2.exe > outnc12hd22019071612.txt Creating the GCEM result file (rainverd2019071612012.nc), procedure file (outnc12hd22019071612.txt) ./tsmain01.exe >ts12h2019071612.txt Creating the TS result and procedure file (ts12h2019071612.txt) 5.2 The steps for case 2 during 00:00–12:00 UTC on June 13,2020 1. Creating an installation and running sub-directory mkdir p2020 2. Copying data sources, code files and configuration files to this directory 3. Running in this directory ./compilePAS10mmd2.sh Compile to generate executable file (pastonc6hd2.exe) ./compileTS.sh Compile to generate executable file (tsmain01.exe) 4. Modifying the configuration file (r12hfile.txt) Mainly modifying the data source path for lines 4, 19, and 23 5. Run the executable files ./pastonc6hd2.exe > outnc12hd22020061312.txt Creating the GCEM result file (rainverd2020061312012.nc), procedure file (outnc12hd22020061312.txt) ./tsmain01.exe >ts12h2020061312.txt Creating the TS result and procedure file (ts12h2019071612.txt) 5.3 The steps for PAS mini-program 1. Creating an installation and running sub-directory mkdir pas 2. Copying code files and configuration files to this directory 3. Running in this directory ./compilePAS10ux.sh Compile to generate executable file (pastonc6hd2.exe) 4. linking the executable file as pas ln -sf pas10ux.exe pas 5. Running the PAS mini-program for example: ./pas 15 20 Parameter 1: 15 represents observed precipitation Parameter 2: 20 represents forecasted precipitation Output: 0.895 1 The following instructions for specific usage: pas rainsk rainyb [level] Input parameters Parameter 1 (rainsk): observed precipitation (mm) Parameter 2 (rainyb): forecasted precipitation (mm) Parameter 3 (level):Specifing magnitude (Optional, default to ≥0.1mm) Onput parameters Parameter 1 (ipas): Pas score value (0-1) or correct value of no precipitation forecast (1); -999.000 represents default. Parameter 2 (iTure): 0 indicates that the rating is correct for a no precipitation forecast; 1 indicates a PAS score of ≥ the specified magnitude; 9 indicates that it is not in the no precipitation test, nor is it the verification the specified magnitude; -999 indicates default. 6Module code main interface description 6.1 skinput() subroutine skinput(skfile,skfilenum,rain,gridskx,gridsky,longitude,latitude) integer,intent(in) :: skfilenum character(len=200),dimension(skfilenum),intent(in) :: skfile real,dimension(:,:),allocatable,intent(out) :: rain integer,intent(out) :: gridskx,gridsky usage: Read a set of observed precipitation data files and output grid accumulated precipitation skfile, A set of filenames that are arrays of strings (input) skfilenum, Number of files (input) rain, Accumulated precipitation, rain(nx,ny) (output) gridskx, grid points, nx (output) gridsky, grid points, ny (output) gridlon, Longitude array, gridlon(nx) (output) gridlat, Latitude array, gridlat(ny) (output) 6.2 ybinput() subroutine ybinput(ybfile,apcp,gridybx,gridyby,gridyblon,gridyblat) character(len=200),intent(in) :: ybfile real,dimension(:,:),allocatable,intent(out) :: apcp,gridyblat,gridyblon integer,intent(out) :: gridybx,gridyby usage: Read a set of forecasted precipitation data files and output forecast grid precipitation ybfile, Forecast file (input) apcp, forecasted precipitation array, apcp(nx, ny) (output) gridybx, Number of grid points for forecast data, nx (output) gridyby, Number of grid points for forecast data, ny (output) gridyblon, Longitude of forecast data, gridyblon(nx, ny) (output) gridyblat, Latitude of forecast data, gridyblat(nx, ny) (output) 6.3 uxpasid2() subroutine uxpasid2(ui,xi,level,pas,iTure,iclass,ieps) real,intent(in) :: ui,xi,level real,intent(out) :: pas,ieps integer,intent(out) :: iTure,iclass usage: Read in the observed and forecasted precipitation, and output the PAS score result rainsk, Observed precipitation (input) rainyb, Forecasted precipitation (input) level, Specifing magnitude (input) ipas, Pas score value (0-1) or correct value of no precipitation forecast (1) iTure, 0 indicates that the rating is correct for a no precipitation forecast; 1 indicates a PAS score of ≥ the specified magnitude; 9 indicates that it is not in the no precipitation test, nor is it the verification the specified magnitude; -999 indicates default. iclass, 0 indicates the category (no precipitation forecast is correct) 1 indicates the category of insufficient precipitation forecast (observation u<10mm) 2 indicates the category of excessive(or equal) precipitation forecast(observation u<10mm) 3 indicates the category of insufficient precipitation forecast (observation u≥10mm) 4 indicates the category of excessive(or equal) precipitation forecast(observation u≥10mm) -999 indicates default. ieps, 0 indicates the forecasted and observed precipitation are equal <0 indicates insufficient precipitation forecast >0 indicates excessive precipitation forecast -999 indicates default. 6.4 outpasnc() subroutine outpasnc(title,vtime,vhour,gridncx,gridncy,gridnclon,gridnclat,rainncsk,rainncyb,& pasc,pas01,pas10,pas25,pas50, pas2p5,pas5,pas15, pascnc,pasnc01,pasnc10,& pasnc25,pasnc50,pasnc2p5,pasnc5,pasnc15,ipsnc,epsnc,iepsnc,ips,eps,ieps) character(len=10),intent(in) :: vtime,title integer,intent(in) :: vhour,gridncx,gridncy character(len=10) :: chour real,dimension(gridncx),intent(in) :: gridnclon real,dimension(gridncy),intent(in) :: gridnclat real,dimension(gridncx,gridncy),intent(in) :: rainncsk,rainncyb,pascnc,pasnc01 real,dimension(gridncx,gridncy),intent(in) :: pasnc10,pasnc25,pasnc50,ipsnc,epsnc,iepsnc real,dimension(gridncx,gridncy),intent(in) ::pasnc2p5,pasnc5,pasnc15 real,intent(in) :: pasc,pas01,pas10,pas25,pas50,ips,eps,ieps,pas2p5,pas5,pas15 usage: Output data to a netCDF format file title, File name tag (d) vtime, time string (yyyymmdddhh, eg. 2019071612) vhour, Accumulated precipitation duration (12) gridncx, x grid points (240) gridncy, y grid points (200) gridnclon, x grid points longitude array gridnclat, y grid points latitude array rainncsk, Observed precipitation rainncyb, Forecasted precipitation pasc, PASC pas01, ≥0.1mm PAS pas10, ≥10mm PAS pas25, ≥25mm PAS pas50, ≥50mm PAS pas2p5, ≥2.5mm PAS pas5, ≥5mm PAS pas15, ≥15mm PAS pascnc, PASC array pasnc01, ≥0.1mm PAS array pasnc10, ≥10mm PAS array pasnc25, ≥25mm PAS array pasnc50, ≥50mm PAS array pasnc2p5, ≥2.5mm PAS array pasnc5, ≥5mm PAS array pasnc15, ≥15mm PAS array ipsnc, IPS array epsnc, EPS array iepsnc, IEPS array ips, IPS eps, EPS ieps IEPS

We have successfully incorporated a 3‐dimensional sub‐grid terrain solar radiative effect (3D STSRE) parameterization scheme into a convection‐permitting Weather Research and Forecasting model (WRF_CPM) in this study. Impacts of 3D STSRE scheme on the ability of WRF_CPM in forecasting the precipitation in summer over the Tibetan Plateau (TP) and nearby regions with complex terrain have been systematically addressed by conducting experiments without and with the 3D STSRE scheme. Results show that the application of 3D STSRE scheme can obviously mitigate the overestimation of surface solar radiation (SSR) and rainfall over TP and nearby regions, especially over the areas with much more rugged terrain (i.e., southern TP) in the WRF_CPM without 3D STSRE scheme. Further mechanism analyses indicate that the decreased surface heating induced by the reduction of SSR reduces the intensity of the thermal‐low pressure over the TP, which leads to the diminished strength of southwesterly winds and thereafter the weaker convergence of moisture flux over the southern TP. Moreover, the weakened surface thermal forcing makes the local atmosphere more stable, suppressing the vertical water vapor transport and local convection. These effects greatly alleviate the overestimation of precipitation over the southern TP produced by the WRF_CPM without the 3D STSRE scheme.

Based on potential vorticity (PV) thinking, northeast China cold vortex (NCCV) corresponds to an upper‐level high PV anomaly from stratospheric PV downward intrusion. Within a vortex, latent heating from precipitation would produce a vertical dipole of PV anomalies that would affect structures and evolution of the vortex. In this paper, three‐dimensional structures and evolution of NCCV and, effects of latent heating from precipitation along bent‐back, cold and warm fronts on them are investigated based on convection‐allowing simulations for an intense NCCV case during 8–17 June 2012. Trajectory analysis shows that the negative upper‐level diabatic PV anomaly from bent‐back frontal precipitation, near the vortex center, is the dominant contributor to erosion of high PV in the vortex core region as it is advected in, leading to the weakening of the vortex. The negative PV anomalies along the cold and warm fronts, at the east‐to‐southeast side of the vortex, are mostly advected downstream away from the NCCV. In the middle troposphere, positive PV anomalies are primarily generated along fronts and the accumulated positive PV anomalies filling the vortex region help to reinforce the low‐level cyclonic circulation. The lower‐level PV is affected by surface heating and cooling through their effects on static stability, but such effects are periodic and create mainly diurnal variations. The NCCV eventually decays as the upper‐level vortex weakens due to significant PV erosion.