Displays

By using a multi-source data set consisting of high resolution satellite, radar, lightning, and model data this study presents the analysis of characteristics of deep convective systems over Germany and first results of a new model to predict the remaining lifetime of existing thunderstorms. Contrary to previous studies, the analysis was performed for the full mixture of observed convective systems regardless of their organization type, since our focus is an operational forecasting environment where no simple method is available to differentiate organization types. Basis for the study are all deep convective cell detections in satellite data (using Cb-TRAM, Thunderstorm Tracking and Monitoring) in a five month period (June 2016, May, June, and July 2017, and June 2018). The lifetimes of all cells are normalized, averaged and separated into life cycle phases to investigate the behavior of the parameters from the different data sources during the detected lifetime. Furthermore, the thunderstorm cells are sorted by their lifetime to determine differences between the characteristics of long- and short-lived convective systems. Parameters with predictive skill are then combined with fuzzy logic to determine the actual stage of a thunderstorm, and to nowcast its remaining lifetime. It will be shown that the new lifetime prediction model contributes to an improvement of the thunderstorm nowcasting.

How to cite: Zöbisch, I., Forster, C., Zinner, T., and Wapler, K.: Analysis and nowcasting of deep convective systems over Germany in multi-source data, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-3521, https://doi.org/10.5194/egusphere-egu2020-3521, 2020.

Based on the characteristics of sudden and local, short life history, serious disasters and so on, the severe convection weather system is difficult to be captured by the conventional meteorological observation network, and is still challenging for catastrophic weather forecasting. In order to improve the service ability in strong weather monitoring and prediction, the following researches have been carried out recently: (1) The new mesocyclone and tornado vortex feature recognition algorithms are developed and proved to be successfully in identifying tornado vortex characteristics in more than a dozen tornado cases. Extracted from Doppler radar volume scan data, a number of parameters (exceed thirty) have been used in the research on the automatic recognition and warning technology of classified severe convective weather (downburst, tornado, hail and short-time strong precipitation). Based on large sample data and results of a variety of analysis methods, a thunderstorm winds Bayes discriminant model has also been established. The testing results show that its Heidke skill score is 0.836, along with the accuracy rate and hit rate are greater than 95%, and the empty rate is below 5%. (2) Rapid update cycle forecast system can effectively improve the quality of model initial values that is very suitable for short time forecast application. For the sake of improving severe thunderstorm prediction, a novel pseudo-observation and assimilation approach involving water vapor mass mixing ratio is proposed to better initialize numerical weather prediction (NWP) at convection-resolving scales. In addition, a new set of simplified and parameterized dual-polarization radar simulators for horizontal reflectivity (Z_H), differential reflectivity (Z_DR), specific difference phase (K_DP), and correlation coefficient (ρ_HV) have been co-developed, and some preliminary data assimilation experiments have shown that the assimilation of dual polarization variables including differential reflectivity and specific difference phase in addition to radar radial velocity and horizontal reflectivity can help improve the accuracy of initial conditions for model hydrometer variables and ensuing model forecasts. (3) Although not yet mature enough for meteorological application, blending technology which is expected to overcome the deficiency of the quantitative precipitation (QPF) by a mesoscale NWP model for the short term at convective scales and the rapidly descending skill of rainfall forecast based on radar extrapolation method beyond the first few hours is under development and debugging, and also has potential in enhancing the ability of rainfall forecast within the nowcasting period. (4) The above methods and systems were applied and provided technical support for meteorological services during the 7^th Wuhan World Military Games in 2019, and a good service effect had been achieved.

How to cite: Cui, C., Xiao, Y., Lai, A., and Du, M.: Advances in the Study of Severe Convection Weather Nowcasting in Central China, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-1656, https://doi.org/10.5194/egusphere-egu2020-1656, 2020.

Nowcasts (short-range forecasts) of rainfall can be used for providing early warning of flash floods. Thus, they are of high societal importance especially in densely populated urban areas. Weather radars are ideally suited for this purpose due to their good spatial coverage and high spatial and temporal resolution (e.g. 1 km and 5 minutes).

A novel approach to radar-based rainfall nowcasting is proposed. The forecast model consists of two components: horizontal advection and temporal evolution of rainfall intensities. The advection velocities are estimated from radar-measured rain rate fields using a pattern matching method. A smooth advection field is obtained by interpolating the motion to areas with no precipitation. The extrapolation is done using a semi-Lagrangian scheme.

The temporal evolution of rainfall intensities is described in Lagrangian coordinates by using a spatiotemporal process model. Such models are ubiquitous in environmental and physical sciences. This study presents the first attempt to apply such a model to three-dimensional rainfall measurements to capture the vertical structure of rainfall processes. This is done by using a linear integro-differential equation with the Markovian assumption (i.e. the next time step depends conditionally on the previous one). Spatial dependencies are modeled via a convolution kernel. To reduce the dimensionality of the parameter estimation, the kernel is parametrized by a trivariate Gaussian function, and the model is formulated and implemented in the Fourier domain. Finally, the parameter estimation is done in the Bayesian framework by applying a Markov Chain Monte Carlo (MCMC) method with Gibbs sampling.

The operational feasibility of the proposed model is evaluated by using data from the NEXRAD WSR-88D radar deployed in Fort Worth, Texas. Measurements from 14 elevation angles are used by restricting the analyses to liquid precipitation below the melting layer. The data processing chain consists of 1) temporal interpolation within radar volumes, 2) clutter filtering, 3) attenuation correction, 4) melting layer detection, 5) polarimetric rain rate estimation based on reflectivity, specific differential phase and differential reflectivity and 6) interpolation to a three-dimensional grid.

The focus of the validation is on higher rain rates (> 5 mm/h) using 10 events during 2018-2019 with mixed convective and stratiform rainfall. Predicted rain rates from the nowcasting model are compared to observations from low-angle radar scans. Using standard verification scores (e.g. equitable threat score and mean absolute error), it is shown that for rainfall rates between 5-25 mm/h, the proposed method can yield up to 30% improvement compared to state of the art extrapolation nowcasting methods. This is attributed to using the spatiotemporal model and vertical profile information obtained from three-dimensional input data.

How to cite: Pulkkinen, S., Venkatachalam, C., and von Lerber, A.: Precipitation nowcasting using spatiotemporal models and volumetric radar data, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-5681, https://doi.org/10.5194/egusphere-egu2020-5681, 2020.

The Remote Sensing of Electrification, Lightning, and Mesoscale/Microscale Processes with Adaptive Ground Observations (RELAMPAGO) international field campaign occurred June 1, 2018, to April 30, 2019. This campaign was comprised of more than 150 scientists from 10 organizations. Data was collected to investigate different phases of the life cycle of thunderstorms that occur in Argentina to better understand the physical mechanisms that cause the initiation and growth of organized convective systems in some of the most intense storms on the planet. The main focus of the project was to develop new conceptual models for forecasting extreme weather events that will hopefully lead to reductions in future loss of life and property.

This presentation shows the performance of a recently developed model for estimating ice mass aloft, a key component in the atmospheric electrification process, and a method for nowcasting lightning activity using C-band weather radar and Global Lightning Dataset (GLD360) data from RELAMPAGO. This nowcasting method uses a grid-based approach to make specific forecasts of lightning in space and time. The method estimates ice mass aloft in the region where electrification occurs using a numerical optimization approach to essentially reframe a simplified bulk microphysical model into a completely data-driven model. Previous results using WSR-88D S-band radar data in the United States showed that using this model significantly improved nowcasts of first-flash lightning occurrence versus the traditional weather radar-based ice mass estimator as well as using lightning flash-rate density directly.

How to cite: Ruzanski, E., Chandrasekar, V., and Arias, I.: Nowcasting lightning during RELAMPAGO, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-13272, https://doi.org/10.5194/egusphere-egu2020-13272, 2020.

Accurate and timely hydrological forecasts highly depend on their meteorological input. Current numerical weather predictions (NWP) do not have sufficiently high spatial and temporal resolutions for adequate use for short lead times (less than six hours ahead) in fast-responding mountainous, lowland and polder catchments. Therefore, radar rainfall nowcasting, the process of statistically extrapolating the most recent radar rainfall observation, is increasingly used. However, as most studies consist of analyses based on a relatively small sample of generally 2–15 events, best practices for the use and choice of these algorithms within operational forecasting systems are not yet available. In this study, we aim to determine the predictive skill of radar rainfall nowcasting algorithms for the short-term predictability of rainfall, in which we focus on different lowland catchments in the Netherlands. We concentrate particularly on the dependency of the forecast skill on catchment and environmental characteristics, such as event type and duration, seasonality, catchment size and location with regard to the radar location and prevailing wind direction. For this purpose, we performed a large-sample analysis of 1481 events spread over four event durations and twelve lowland catchments (6.5–957 km²). Four algorithms were tested and compared with Eulerian Persistence: Rainymotion Sparse and DenseRotation, pySTEPS deterministic and pySTEPS probabilistic with 20 ensemble members. Maximum skillful lead times increased for longer event durations, due to the more persistent character of these events. In all cases, pySTEPS deterministic attained the longest maximum skillful lead times: 25 min for 1-h, 39 min for 3-h, 56 min for 6-h and 116 min for 24-h durations. During winter, when more persistent stratiform precipitation is present, we found three times lower mean absolute errors than for nowcasts during summer with more convective precipitation. For the fractions skill score, higher skill was obtained with increasing grid cell sizes. This was advantageous for larger catchments, whereas some catchments became smaller than the grid size after upscaling. Catchment location mattered as well: up to two times higher skillful lead times were found downwind of the radars than upwind, given the prevailing wind direction. The pySTEPS algorithms outperformed Rainymotion benchmark algorithms due to rainfall field evolution estimations with cascade decomposition and an autoregressive model. We found that most errors still originate from growth and dissipation processes which are not or only partially (stochastically) accounted for.

How to cite: Imhoff, R., Brauer, C., Overeem, A., Weerts, A., and Uijlenhoet, R.: Comparing four radar rainfall nowcasting algorithms for 1481 events, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-13376, https://doi.org/10.5194/egusphere-egu2020-13376, 2020.

This study uses C-band polarimetric radar (C-POL) measurements to classify the hydrometeors and retrieve rain drop size distributions (DSDs) during the IMFRE (Investigative Monsoon Frontal Rainfall Experiment) field campaign in central China.  Three types of precipitation in a Meiyu frontal heavy rainfall case are classified to further investigate the microphysical characteristics and processes based on C-POL observations and classified hydrometeors. When raindrops fall from the freezing level, collision–coalescence plays an equally important role as break-up and/or evaporation in stratiform regions, but is the dominant process for convective-related precipitation and is an attribute of intensive precipitation. There are more supercooled liquid water droplets above the freezing level in convective cores due to strong updrafts, which can bring more cloud droplets into the upper levels to help the formation of graupel and hail. To the best of our knowledge, this work is the first time that C-POL-classified hydrometeors and rain-parameter retrievals have been validated against in-situ aircraft and surface disdrometer measurements over the middle reaches of the Yangtze River Valley in central China, which will pave the way for future studies related to Meiyu frontal rainfall systems.

How to cite: Wang, B., Dong, X., Fu, Z., and Zhou, L.: Investigation of Hydrometeors Using a C-band Poloarimetric Radar and In-situ Measurements during IMFRE in Central China, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-1657, https://doi.org/10.5194/egusphere-egu2020-1657, 2020.

The majority of heavy rainfall and flooding events in the central China during the Meiyu season are caused by the multi-scale monsoon frontal systems. However, there are limited studies of the vertical distributions of monsoon frontal rainfall. This work for the first time analyzed the vertical structures of the different stages of Meiyu precipitation systems over the Yangtze‐Huai River Valley in central China using measurements and retrievals from the Global Precipitation Measurement Mission Dual‐Frequency Precipitation Radar (GPM‐DPR) and Feng Yun satellites. GPM‐DPR retrieved near‐surface rain and drop size distributions were first validated against the surface disdrometer measurements and showed good agreement. Then we analyzed three cases from the Integrative Monsoon Frontal Rainfall Experiment to demonstrate the different characteristics of convective precipitation (CP) and stratiform precipitation (SP) in the developing, mature, and dissipating stages of the Meiyu precipitation systems, respectively. For statistical analysis, all Meiyu cases during the period 2016–2018 detected by GPM‐DPR were collected and classified into different types and stages. In the stratiform regions of Meiyu precipitation systems, coalescence slightly overwhelms break‐up and/or evaporation processes, but it was dominant in the convective regions when raindrops fall. There were large numbers of large ice particles during the developing stage due to strong updrafts and abundant moisture, whereas there were both large ice and liquid particles in the mature stage. The vertical structures of the SP examined in this study were similar to those over the ocean regions due to high relative humidity but different to the mountainous west regions of the USA. The findings of the stage‐dependent SP vertical structures provide better understanding of the evolution of monsoon frontal precipitation, as well as the associated microphysical properties, and provide insights to improve microphysical parameterization in future models.

How to cite: Sun, Y., Dong, X., Cui, W., Zhou, Z., Fu, Z., Zhou, L., Deng, Y., and Cui, C.: Vertical Structures of Typical Meiyu Precipitation Events Retrieved from GPM‐DPR, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-2512, https://doi.org/10.5194/egusphere-egu2020-2512, 2020.

The high-resolution INCA system (Integrated Nowcasting through Comprehensive Analysis) provides gridded analyses and short-term forecasts of meteorological fields at a horizontal resolution of 1 km and at 5 to 60 min temporal resolution. After the nowcasting part, INCA fields are blended into AROME or a statistically optimized combination of NWP models, thus providing a seamless chain of forecasting fields at various scales. As an operational product of the national Austrian weather service (ZAMG), INCA is employed for a number of applications, such as for hydrological runoff modelling, severe weather warnings, in road safety, agriculture and in the renewable energy sector. The product development is embedded into the research activities of the last year period which, amongst others, included the development of new blending methods into the state of the art NWP models, a new approach for precipitation analysis and nowcasting as well as the evaluation of wind, temperature and humidity analyses at 100 m horizontal resolution. For precipitation, a new radar-raingauge merging algorithm has been developed, which is based on station density and radar minimum beam height. Precipitation nowcasting uses optical flow and nested subdomains for breaking down the displacement vectors to the target grid. In the sub-kilometer version, the improvements due to the topography-relevant features (i.e. altitude, slope) in Alpine areas are shown as well as the potential benefits of high resolution nowcasting in urban areas. The methods and related results will be presented along with a comprehensive verification.

How to cite: Bica, B., Hadzimustafic, J., Atencia, A., Kulmer, M., Hollosi, B., Schneider, S., Kann, A., and Wang, Y.: Recent developments in the INCA analysis and nowcasting system, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-22584, https://doi.org/10.5194/egusphere-egu2020-22584, 2020.

Numerical weather prediction models with high resolution (of order kms or less) can deliver very accurate low-level winds. The problem is that one cannot afford to run simulations at very high resolution over global or other large domains for long periods because the computational power needed is prohibitive.

Instead, we propose using neural networks to downscale low-resolution wind-field simulations (input) to high-resolution fields (targets) to try to match a high-resolution simulation. Based on short-range forecasts of wind fields (at the 100m level) from the ECMWF ERA5 reanalysis, at 31km resolution, and the HRES (deterministic) model version, at 9km resolution, we explore two complementary approaches, in an initial “proof-of-concept” study.

In a first step, we evaluate the ability of U-Net-type convolutional neural networks to learn a one-to-one mapping of low-resolution input data to high-resolution simulation results. By creating a compressed feature-space representation of the data, networks of this kind manage to encode important flow characteristics of the input fields and assimilate information from additional data sources. Next to wind vector fields, we use topographical information to inform the network, at low and high resolution, and include additional parameters that strongly influence wind-field prediction in simulations, such as vertical stability (via the simple, compact metric of boundary layer height) and the land-sea mask. We thus infer weather-situation and location-dependent wind structures that could not be retrieved otherwise.

In some situations, however, it will be inappropriate to deliver only a single estimate for the high-resolution wind field. Especially in regions where topographic complexity fosters the emergence of complex wind patterns, a variety of different high-resolution estimates may be equally compatible with the low-resolution input, and with physical reasoning. In a second step, we therefore extend the learning task from optimizing deterministic one-to-one mappings to modelling the distribution of physically reasonable high-resolution wind-vector fields, conditioned on the given low-resolution input. Using the framework of conditional variational autoencoders, we realize a generative model, based on convolutional neural networks, which is able to learn the conditional distributions from data. Sampling multiple estimates of the high-resolution wind vector fields from the model enables us to explore multimodalities in the data and to infer uncertainties in the predictand.

In a future customer-oriented extension of this proof-of-concept work, we would envisage using a target resolution higher than 9km - say in the 1-4km range - to deliver much better representivity for users. Ensembles of low resolution input data could also be used, to deliver as output an “ensemble of ensembles”, to condense into a meaningful probabilistic format for users. The many exciting applications of this work (e.g. for wind power management) will be highlighted.

How to cite: Kern, M., Höhlein, K., Hewson, T., and Westermann, R.: Towards Operational Downscaling of Low Resolution Wind Fields using Neural Networks, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-5447, https://doi.org/10.5194/egusphere-egu2020-5447, 2020.

    Weather forecasts play an important role in the Olympic game,especially the mountain snow projects, which will help to find a "window period" for the game. The winter Olympics track is located on very complex terrain, and a detailed weather forecast is needed. A Post-processing method based on machine learning is used for the future-10-days weather prediction with 1-km spatial resolution and 1-hour temporal resolution, which can greatly improve accuracy and refinement of numerical weather prediction(NWP). The ECWMF/RMAPS model data and the automatic weather station data(AWS) from 2015-2018 are prepared for the training data and test data, included 48 features and 4 labels (the observed 2m temperature, relative humidity , 10m wind speed and wind direction ). The model data are grid point, while the AWS data are station point. We take the nearest 9 model point to predict the station point, instead of making an interpolation between the grid point and station point. Then the feature number will be 48*9 in dataset. The interpolation error from grid point to station is eliminated,and the spatial distribution is considered to some extent. Machine leaning method we used are SVM, Random Forest, Gradient Boosting Decision Tree(GBDT) and XGBoost. We find that XGBoost method performs best, slightly better than GBDT and Random Forest. It is noted that we did some feature engineering work before training, and we found that it’s not that the more features, the better the model, while 10 features are enough. Also there is an interesting thing that the features that closely related the labels values becomes less important as the forecast time increases,such as the model outputed 2m temperature, 10m wind speed and wind direction. While some features that forecasters don’t pay attention to become more important in the 6-10 days prediction, such as latent heat flux, snow depth and so on. So it’s necessary to train the model based on dynamic weight parameters for different forecast time. Through the post-processing based on the machine learning method, the forecast accuracy has been greatly improved compared with EC model. The averaged forecast accuracy of 0-10 days for 2m relative humidity, 10m wind speed and direction has been increased by almost 15%, and the temperature accuracy has been increased by 20%~40% ( 40% for 0-3 days, and the accuracy decreased with the forecast time ). 

How to cite: Yanyan, K., Haochen, L., Jiangjiang, X., and Yingxin, Z.: Post-processing for NWP Outputs Based on Machine Learning for 2022 Winter Olympics Games over Complex Terrain, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-10463, https://doi.org/10.5194/egusphere-egu2020-10463, 2020.

Visualization of numerical weather prediction data and atmospheric observations has always been an important and ubiquitous tool in weather forecasting. Visualization research has made much progress in recent years, in particular with respect to techniques for ensemble data, interactivity, 3D depiction, and feature-detection. Transfer of new techniques into weather forecasting, however, is slow.

In this contribution, we will discuss the potential of recent developments in 3D and ensemble visualization research for weather forecasting. We will introduce our work on 3D feature-detection methods for jet-stream and front features, which facilitate analysis of the evolution of jet-stream core lines and frontal surfaces in an (ensemble) forecast. The techniques have been integrated into the 3D visual ensemble analysis framework Met.3D (https://met3d.wavestoweather.de), in which they can be combined with traditional 2D depictions as well as further 3D visual elements and be displayed in an interactive 3D context. We will present and discuss 3D ensemble forecast products created with Met.3D based on forecast data from ECMWF and DWD, and demonstrate their use in the exploration of example cases including an extratropical transition over the North Atlantic and a European winter storm.

In addition, we will introduce new semi-operational 3D forecast products based on our techniques that we provide experimentally on the web, in order to gather user feedback and to initiate discussion about potential benefit of such products for operations.

How to cite: Rautenhaus, M., Hewson, T., Modali, K. R., Beckert, A., and Kern, M.: Interactive 3D visual analysis in weather forecasting, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-15011, https://doi.org/10.5194/egusphere-egu2020-15011, 2020.

Lightning strikes from convective storms are a serious safety concern for public and businesses alike. Accurate assessment of local lightning risk is therefore crucial for various industries. However, it is usually not possible to obtain lightning climatologies with reasonable spatial detail, due to the scarcity of well distributed, long term observations. At this respect, meteorological models serve as a useful tool for creating lightning risk maps, provided that their output can be verified with available observations. In this study, a high resolution (3 km) lightning risk map has been constructed for Turkey, using output from Weather Research and Forecasting Model (WRF). The model was forced by the ECMWF’s ERA-5 reanalysis data, and run for the period of January 2014 – December 2018 (5 years). Simulations were conducted on high-performance computers offered by Amazon Web Services. Lightning flash rates were estimated from WRF output using the parameterization scheme proposed by McCaul et al. (2009). Model-derived lightning rates have been calibrated and validated by observed lightning data for the determined region. The spatial pattern and average rate of lightning flashes over the validation region have been found to agree reasonably well with available observations. The high resolution lightning risk map produced in this study is the first one for Turkey that is based on numerical modeling, and it will serve as an objective guidance for location-based lightning risk assessment in the country.

How to cite: Yılmaz, M. Y., Göktürk, O. M., and Fidan, G.: A High-Resolution, Model-Based Lightning Risk Map for Turkey, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-5031, https://doi.org/10.5194/egusphere-egu2020-5031, 2020.

As the grid resolution continues to increase, it is essential to utilize more data to describe the occurrence and evolution of mesoscale system more precisely. Based on Local Analysis and Prediction System (LAPS), NCEP reanalysis data (FNL) is used as the background field. By improving and developing LAPS, a high spatial-temporal resolution mesoscale data is generated and a sample database is established by fusing a variety of observation data.

Moreover, the synoptic and physical variables that may affect the short- duration heavy precipitation are fully considered when establishing the diagnosis-statistical forecast model. After the calculation, three discriminant formulas of strong and weak precipitation, heavy precipitation classification I and II are obtained.

Furthermore, the grouping accuracy of the discriminant formula and T_s score were calculated, and after the independent sample test and typical case test, it can be concluded that these discriminant formulas can be used to distinguish strong and weak precipitation and heavy precipitation classification.

How to cite: Ke, D., Wang, X., and Cui, C.: Characteristics and classifications of short-duration heavy precipitation in mesoscale system evolution, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-7810, https://doi.org/10.5194/egusphere-egu2020-7810, 2020.

1.Introduction

A key issue in developing the ensemble prediction technique is the recognition of uncertain factors in numerical forecasting and how to use appropriate perturbation techniques to reflect these uncertain processes and improve ensemble prediction levels.Plenty of corresponding perturbation techniques have been developed. Such as Initial uncertainty and model uncertainty,In addition to the influence of IC and model uncertainty ,precipitation is closely related to terrain.The influence of terrain on the heavy rain includes the following three aspects:(1) The terrain has significant effect on the climatic distribution of precipitation.(2)The windward slope and leeward slope and other dynamic effects generated by terrain impact the triggering and intensity of precipitation.(3)The thermal effect is triggered by the heating of land surface of terrain at different height and latent heat release when airflow rises ,and this thermal action makes mountain precipitation closely related to terrain distribution .What are the terrain uncertainties in the model?(1)Different vertical coordinate systems lead to significant differences in terrain treatment(2)The conversion from real terrain to model terrain is closely associated with the resolution of the model and different terrain interpolation schemes, and it affects the simulation results of precipitation .(3)Measuring error of real terrain, etc.In this report, A terrain perturbation scheme (ter) has been firstly incorporated into an ensemble prediction system (EPS) and preliminarily tested in the simulation of the extremely heavy rain event occurred on 21 July, 2012 in Beijing, along with other three perturbation schemes.

2.Case,data and schemes

(1)Case: Based on the extremely heavy rain case in Beijing on July 21,2012, maximum precipitation center more than 400mm.(2)Data: GEPS of NCEP were used as initial background fields and lateral boundary condition , surface and upper-level observation of GTS,Rain gauge etc.(3)Model: WRFv4.3, 9km horizontal resolution ,511*511 grid point, 51 vertical layers,KF Eta,WSM6,etc(4)Experiments schemes: Four different perturbation schemes were used in the experiments and six members in each experiment. Sch_1(IC) considered the IC uncertainty ,the parameterization schemes were same but IC/LBC came from different GEPS members. Sch_2(phy) considered the Phy uncertainty ,the IC/LBC were same but PHY schemes were comprised of different parameterization schemes. Sch_3-4(ter and icter) considered the terrain uncertainty ,the second aspect of terrain uncertainty was considered in this study. Two different model terrain smoothing schemes and 3 terrain interpolation schemes were used to reflect the forecast error caused by terrain height. Icter is the mixed scheme of ter and ic.

3.Preliminary test and results

(1)Precipitation is closely related to terrain, terrain uncertainties have significant effect on the intensity and falling area of precipitation.(2) Only a simple terrain perturbation can produce a significant forecast spread , and its ensemble mean forecast is also improved compared with control forecast. for this case, it has a slightly positive contribution to the spread and probability forecast of precipitation on the basis of not impacting the quality of ensemble mean forecast.(3) In this case, the magnitude of spread generated by the terrain perturbation scheme is significantly smaller than that generated by the initial perturbation and physics process perturbation schemes.

How to cite: jun, L., jun, D., yu, L., and jianyu, X.: Preliminary study on terrain uncertainty and its perturbing scheme, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-1856, https://doi.org/10.5194/egusphere-egu2020-1856, 2020.

The objective extrapolation forecast is the main method for 0-1 hour convective storm nowcasting. Radar echo extrapolation was performed by using the 6 minute interval radar mosaics obtained from the radar images of 8 multi-radars in Beijing-Tianjin-Hebei region. A comparative study of two extrapolated forecasts of eighteen typical convective precipitation events occurred in Beijing-Tianjin-Hebei region from 2016 to 2018 was conducted. Compared with the tracking radar echoes by correlation method, the variational echo tracking method utilizes variational technique to compute the motion vector fields, and uses two strict constraints to get a better motion vector field. The results indicated that the variational echo tracking method performed better in prediction of the radar echo pattern, echo location, and echo intensity at 30- and 60-min forecast lead times: 1) A comparative study of the two extrapolated forecasts of four precipitation events in Beijing-Tianjin-Hebei region was conducted. The result indicated that the radar echo location, the echo pattern and echo intensity produced by the variational echo tracking method were closer to the real observation within one hour. 2) Quantitative evaluation for the two extrapolated forecasts of the eighteen typical convective precipitation events was conducted. Compared with the tracking radar echoes by correlation method, the probability of detection and the critical success index of the 30- or 60-min extrapolated forecast produced by the variational echo tracking method were higher, meanwhile the false alarm rate was lower when the radar echo threshold was 35dBz and 45dBz. Also, a quantitative evaluation classified by the weather type indicated that the variational echo tracking method performed better than the tracking radar echoes by correlation method in most weather types.

How to cite: Wu, J., Chen, M., Qin, R., Gao, F., and Song, L.: The variational echo tracking method and its application in convective storm nowcasting, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-1293, https://doi.org/10.5194/egusphere-egu2020-1293, 2020.

The Ili Valley is an area with frequent heavy rain in Xinjiang. In this paper, a heavy rainstorm process in this area on June 26, 2015 is taken as an example. The observational data and WRF high-resolution numerical simulation results are used to analyze the synoptic background and the process of the precipitation. The results show that: (1) The Central Asian low vortex and the upper-level jet provides a favorable circulation background for this heavy rain. Northerly winds and westerly winds forms a low-level convergence line in Ili Valley. (2) In addition to the convergence of low-level airflow, the uplifting effect of the terrain on the westerly winds also intensifies the low-level ascending motion. At the same time, the uplifting effect of the terrain on the northerly winds causes the middle-level ascending motion. After the low-level ascending motion is connected with that of the middle level, precipitation begins to occur. The convection further develops, superimposed with the upwards phase of upper-level wave, and the precipitation increases strongly. (3) Through spectral analysis methods, the characteristics of the upper-level wave are obtained, and the wave is an inertial gravity wave. It is further obtained from the mesoscale three-dimensional Eliassen-Palm (EP) flux that during the period of heavy precipitation, the energy of the upper-level inertial gravity wave is transported down to the low level of the precipitation area. (4) Convective instability plays an important role in the enhancement of the precipitation in the Ili Valley. The analysis of potential divergence further indicates that the convective instability in the precipitation area is mainly caused by the vertical shear part of potential divergence, while the divergence part of the potential divergence can strengthen the convective instability in the leeward slope of the terrain. It indicates that the dynamic and thermodynamic factors are coupled with each other, which affects the precipitation location, intensity and evolution.

How to cite: Huang, X. and Zhou, Y.: Numerical Simulation and Mechanism Analysis of an Extreme Precipitation in Ili Valley, Xinjiang, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-1717, https://doi.org/10.5194/egusphere-egu2020-1717, 2020.

Central Asian Vortices (CAVs) are deep cyclonic systems that occur in the Central Asian and are identified at the 500 hPa level. CAVs are significantly associated with many convective events in the Xinjiang province. In order to strengthen the understanding of the mesoscale systems development mechanisms in torrential rain under the influence of CAVs, we analyzes the rainstorm process occurred in the Aksu region that is near the west of Tianshan Mountains, during June 17 to 18, 2013 basing on a variety of data. The results show that the precipitation process occurs under the background of the circulation of the two ridges in a trough over the middle and high latitudes, and the CAV provides favorable large-scale dynamic and water vapor conditions for this rainstorm. The convergence line is the important mesoscale system, which is formed by the superposition of the CAV circulation and the flow stream around the special topography of the west Tianshan Mountains. Due to the difference of thermal properties between the mountain and desert, the slope wind drives convergence line to move and the strong convection developed along the convergence line triggers strong precipitation in the Aksu region. The WRF is able to well simulate not only the location and intensity of the heavy rain but also the evolution of wind field. Preliminary analysis combined with observations and simulation data shows that under the blockage of west Tianshan Mountains, the south wind accumulates and convergences near the valley. As a result, a local convergence line is formed. Meanwhile, with the adjustment of the large-scale circulation situation, especially after the CAV moves to the vicinity of the Aksu area, one part of the westward flow that comes from the south of the vortex turns into northwest wind after crossing the west Tianshan Mountains, and the other part turns into the northeast wind after passing through the Yili Valley, these two flow aggravate the northerly airflow and enhance the intensity of convergence, thereby promote the formation of mesoscale convergence lines and strengthen it. The eastward airflow-induced water vapor accumulates in front of the southern foot of the Tianshan Mountains, and strengthens as the convergence line moves towards southeast with the enhancement of the valley wind at night. Accompanied with the convergence uplift, the accumulation of water vapor at the foot of the mountain promotes the release of unstable energy and brings heavy precipitation to the Aksu region.

How to cite: Guo, N. and Zhou, Y.: Analysis of a severe precipitation process in Aksu Area under the background of the Central Asian Vortex, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-1718, https://doi.org/10.5194/egusphere-egu2020-1718, 2020.

During the Meiyu period, floods are prone to occur in the middle and lower reaches of the Yangtze River due to the highly concentrated and heavy rainfall, which caused huge life and economic losses. Based on numerical simulation by assimilating Doppler radar, radiosonde, and surface meteorological observations, the evolution mechanism for the initiation, development and decaying of a Meiyu frontal rainstorm that occurred from 4th to 5th July 2014 is analyzed in this study. Results show that the numerical experiment can well reproduce the temporal variability of heavy precipitation and successfully simulate accumulative precipitation and its evolution over the key rainstorm area. The simulated “rainbelt training” is consistent with observed “echo training” on both spatial structure and temporal evolution. The convective cells in the mesoscale convective belt propagated from southwest to northeast across the key rainstorm area, leading to large accumulative precipitation and rainstorm in this area. There existed convective instability in lower levels above the key rainstorm area, while strong ascending motion developed during period of heavy rainfall. Combined with abundant water vapor supply, the above condition was favorable for the formation and development of heavy rainfall. The Low level jet (LLJ) provided sufficient energy for the rainstorm system, and the low-level convergence intensified, which was an important reason for the maintenance of precipitation system and its eventual intensification to rainstorm. At its mature stage, the rainstorm system demonstrated vertically tilted structure with strong ascending motion in the key rainstorm area, which was favorable for the occurrence of heavy rainfall. In the decaying stage, unstable energy decreased, and the rainstorm no longer had sufficient energy to sustain. The rapid weakening of LLJ resulted in smaller energy supply to the convective system, and the stratification tended to be stable in the middle and lower levels. The ascending motion weakened correspondingly, which made it hard for the convective system to maintain.

How to cite: Li, H., Hu, Y., and Zhou, Z.: Analysis of the Evolution of a Meiyu Frontal Rainstorm Based on Doppler Radar Data Assimilation, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-1975, https://doi.org/10.5194/egusphere-egu2020-1975, 2020.

Convection initiation (CI)  occurred near the oasis region surrounded by gobi desert in the Southwestern Xinjiang, Northwest China is investigated using a real-data, high-resolution Weather Research and Forecasting (WRF) simulation. Observations revealed that many CIs occurred successively near oasis region, some of which developed significantly in both size and intensity and eventually become a strong mesoscale convective system (MCS). The WRF simulation captured the general features of the CIs and MCS. Lagrangian vertical momentum budgets were conducted along the backward trajectories of air parcels within three convective cells. The total vertical acceleration was decomposed into dynamic and buoyant components. The results showed that the buoyant acceleration played a decisive role for about half of the air parcels during the CI, which was contributed by the dry air buoyancy. However, the dynamic acceleration mainly contributed during the CI for about one fourth of the air parcels. The dynamic acceleration can be further decomposed into five terms based on anelastic approximation. The positive dynamic acceleration was mainly caused by the vertical twisting term associated with the mid-level vertical shear, while the extension term contributed negatively to the dynamic acceleration. The other two terms related to horizontal curvature and height variation of density were negligibly small.

How to cite: Ming, J. and Abulikemu, A.: A numerical study of convection initiation in Southwestern Xinjiang, Northwest China, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-2285, https://doi.org/10.5194/egusphere-egu2020-2285, 2020.

Using hourly air pollutants concentration from six environmental monitor stations, meteorological data and wind profile radar data in winter during 2013-2015, the influences of shallow foehn on diffusion conditions and air pollution concentration over Urumqi were analyzed. The results showed the occurrence frequency of shallow foehn was 57.3% in Urumqi in winter. The flow depth, base height and top height of shallow foehn were about 1500 m, 600 m and 2100 m, respectively. The maximum mixing layer depth, the inversion depth, the temperature difference between the top and bottom of inversion layer on foehn days were 200 m lower, 344m thicker and 4.4℃ higher than the corresponding values on non-foehn days, respectively. However, the differences of wind speed and inversion intensity between on foehn days and on non-foehn days were slight. Also, the frequency of each pollution level on foehn days was higher than on non-foehn days with extra frequency of 18% from level Ⅲ to level Ⅵ. Moreover, there was foehn existence on days with air pollution level Ⅵ. Except for O₃, the other five air pollutant concentrations at each environmental station on foen days were all higher than on non-foehn days but with similar diurnal variation. The spatial distributions of six air pollutants on foehn days and non-foehn days were almost same. Overall, the air quality at south urban area was relative excellent than other areas.

How to cite: keming, Z.: The Influence of Shallow Foehn on Atmospheric Diffusion Conditions and Air Quality over Urumqi in Winter, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-2579, https://doi.org/10.5194/egusphere-egu2020-2579, 2020.

It is well known that climate changes sometimes may cause natural disasters，especially the disastrous weather days，as downpour，flood，landslide and mudslide，and their derivatives disasters not only have relationship with precipitation，but also，closely with rainfall intensity. In the practice of Xinjiang disaster prevention，it’s urgent to know the temporal and spatial distributions of precipitation intensity and the maximum of precipitation intensity in different recurrence periods. In this paper，based on the observed hourly precipitation data over 16 national-standard stations during May to September from 1991 to 2013 in Xinjiang，some large-scale，multisites and long-paying observed hourly precipitation data have been used firstly together with the methods of probability distributions，statistical tests，variant difference analysis and extreme value analysis，the temporal and spatial distributions and the diurnal variation of hourly rain in summer in Xinjiang have been analyzed. The results show that the hourly rain presents high frequency in northwest and low frequency in southeast of Xinjiang. The high value center of the frequency with hourly rainfall intensity over 0.1 mm·h-1 or 4 mm·h-1 both in Western Tianshan Mountains. The frequency of heavy rainfall is increasing in places such as Ruoqiang where rains less. The high frequent periods of heavy rainfall，with hourly rainfall intensity over 4 mm·h-1，are often occurred in the afternoon，and the first and second half of the night in Northern Xinjiang，while it occurs at night in Southern Xinjiang. The hourly rain frequencies share obviously different diurnal variation in all regions of Xinjiang，where the hourly rainfall is not well-distributed. The distribution characteristic of daily rain in Northern Tacheng and Altay Prefecture is bimodal and in the rest regions of Northern Xinjiang is unimodal. Nevertheless，in Southern Xinjiang，most are in bimodal distribution. The total frequency of hourly rainfall intensity larger than 0.1 mm·h-1 or 4.0 mm·h-1 in Northern and Southern Xinjiang both appears to be an evident increase trend，and it would increase more significance in Southern Xinjiang in the 2010s. The high value region of hourly rainfall intensity occurring once in 50 or 100 years，respectively 45 mm·h-1 and 50 mm·h-1，both is in the western Aksu.

How to cite: chunyan, C.: Temporal and spatial distributions of hourly rain intensity under the warm background in Xinjiang, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-2592, https://doi.org/10.5194/egusphere-egu2020-2592, 2020.

Severe downslope wind triggered by the interaction between the gap jet and the asymmetrical topography in the Tianshan canyon which caused severe disasters of trains rolloverin the Turpan Depression in Xinjiang on February 28, 2007. To understand the mechanism of downslopewindstorm between the interaction of large-scale circulation background，mesoscale system and complex topography in this extreme windstorm.We use a WRF model to simulated it.Base on the mesoscale diagnostic analysis to simulative results，We propose a mechanism for the windstorm : Under the pressure gradient between north-south sides of the Tianshan Mountain, the air parcel climbs windward slope and flowsinto the Tianshan Gorge and then forms gap jet due to effect of narrow，the jet generated gravity waves forced by the asymmetric terrain of the Tianshan Canyon, and produces a lee waves in the leeward, which transmits the energy of the gap jet to the ground，andSevere downslope windstorm formed finally. In this process, the turbulence formed by the wave breaking and the critical layer absorb the upper layer energy downward, which strengthens the energy of the gap jet,and the atmospheric stability stratification exacerbates the sinking movement, which sinks energy to the ground.

How to cite: hao, T.: Mesoscale Analysis of Severe Downslope Windstorm Caused by Gap Jet in Tianshan Canyon, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-2794, https://doi.org/10.5194/egusphere-egu2020-2794, 2020.

In order to solve the problem of excessive energy dissipation near the sub-grid scale in numerical weather model, the Stochastic Kinetic Energy Backscatter (SKEB) method is introduced into the GRAPES-REPS regional ensemble prediction system, and the first-order autoregressive stochastic process is used in the horizontal direction. Calculate the random pattern obtained by spherical harmonic expansion in the direction, calculate the local dynamic energy dissipation rate caused by the numerical diffusion scheme, construct the random flow function forcing, and convert it into horizontal wind speed disturbance, compensate the dissipated kinetic energy, and carry out A 10-day ensemble prediction test and a randomized time and space scale sensitivity test in September and October 2018 (choose 1st, 7th, 13th, 19th, and 25th), and evaluate the test results. The main conclusions of the research work are as follows: By comparing the ensemble prediction results of the test using the SKEB method and the test without the SKEB method, the use of the SKEB scheme increases the large aerodynamic energy of the GRAPES regional model in the small and medium-scale region, and improves the GRAPES model to the actual atmosphere to some extent. The simulation ability of kinetic energy spectrum; the introduction of SKEB scheme in regional ensemble prediction can significantly improve the dispersion of U and V in horizontal wind field of regional model, and the problem of insufficient dispersion of large-scale dynamic energy dissipation rate in Qinghai-Tibet Plateau region is improved. The SKEB program has improved the forecasting skills to a certain extent, such as reducing the CRPS scores of the horizontal wind fields U and V, reducing the outliers scores of the horizontal wind field, temperature, and 10 m wind speed; the introduction of the SKEB method can improve the light rain. The precipitation probability prediction skill score, but the improvement of the score did not pass the significance test, so it is considered that the SKEB method is difficult to effectively improve the probability prediction technique of precipitation.

Sensitivity tests based on the SKEB method for five time scales of random pattern (1h, 3h, 6h, 9h and 12h of the time series τ) show that the ensemble prediction is sensitive to the five time scales of the stochastic model of the SKEB method. And the 12h experiment show the best performance than the others.

How to cite: Li, H. and Chen, J.: Sensitivity analysis of SKEB method in Regional ensemble forecast system GRAPES-REPS, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-3444, https://doi.org/10.5194/egusphere-egu2020-3444, 2020.

        This study demonstrates how model bias can adversely affect the quality assessment of an ensemble prediction system (EPS) by verification metrics. A regional EPS [Global and Regional Assimilation and Prediction Enhanced System-Regional Ensemble Prediction System (GRAPES-REPS)] was verified over a period of one month over China. Three variables (500-hPa and 2-m temperatures, and 250-hPa wind) are selected to represent "strong" and "weak" bias situations. Ensemble spread and probabilistic forecasts are compared before and after a bias correction. The results show that the conclusions drawn from ensemble verification about the EPS are dramatically different with or without model bias. This is true for both ensemble spread and probabilistic forecasts. The GRAPES-REPS is severely underdispersive before the bias correction but becomes calibrated afterward, although the improvement in the spread' spatial structure is much less; the spread-skill relation is also improved. The probabilities become much sharper and almost perfectly reliable after the bias is removed. Therefore, it is necessary to remove forecast biases before an EPS can be accurately evaluated since an EPS deals only with random error but not systematic error. Only when an EPS has no or little forecast bias, can ensemble verification metrics reliably reveal the true quality of an EPS without removing forecast bias first. An implication is that EPS developers should not be expected to introduce methods to dramatically increase ensemble spread (either by perturbation method or statistical calibration) to achieve reliability. Instead, the preferred solution is to reduce model bias through prediction system developments and to focus on the quality of spread (not the quantity of spread). Forecast products should also be produced from the debiased but not the raw ensemble.

How to cite: Wang, J., Chen, J., and Du, J.: Sensitivity of ensemble forecast verification to model bias, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-4374, https://doi.org/10.5194/egusphere-egu2020-4374, 2020.

Abstract

The aim of this study is to assess the skills of a stochastic weather generator (SWG) to forecast precipitation in Europe. The SWG is based on the random sampling of circulation analogues, which is a simple form of machine learning simulation. The SWG was developed and tested by Yiou and Déandréis (2019) to forecast daily average temperature and the NAO index. Ensemble forecasts with lead times from 5 to 80 days were evaluated with CRPSS scores against climatology and persistence forecasts. Reasonable scores were obtained up to 20 days.  In this study, we adapt the parameters of the analogue SWG to optimize the simulation of European precipitations. We then analyze the performance of this SWG for lead times of 2 to 20 days, with the forecast skill scores used by Yiou and Déandréis (2019). To achieve this objective, the SWG will use ECA&D precipitation data (Haylock. 2002), and the analogues of circulation will be computed from sea-level pressure (SLP) or geopotential heights (Z500) from the NCEP reanalysis. This provides 100-member ensemble forecasts on a daily time increment. We will evaluate the seasonal dependence of the forecast skills of precipitation and the conditional dependence to weather regimes. Comparisons with “real” medium range forecasts from the ECMWF will be performed.

References

Yiou, P., and Céline D.. Stochastic ensemble climate forecast with an analogue model. Geoscientific Model Development 12, 2 (2019): 723‑34.

Haylock, M. R. et al.. A European daily high-resolution gridded data set of surface temperature and precipitation for 1950-2006. J. Geophys. Res. - Atmospheres 113, D20 (2008): doi:10.1029/2008JD010201.

Acknowledge

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 813844.

How to cite: Krouma, M., Yiou, P., Déandréis, C., and Thao, S.: Ensemble weather forecast of precipitation with a stochastic weather generator based on analogues circulation, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-4705, https://doi.org/10.5194/egusphere-egu2020-4705, 2020.

Cloud analysis module (GCAS) in GRAPES model can combine radar reflectivity data, satellite data and surface observations to provide there-dimensional cloud information. To show application effect of GCAS on 3-km resolution forecasts in arid and semi-arid areas of Northwest China, three sets of forecast experiments were conducted with GRAPES_Meso model, which includes control experiment (Con_exp), gcas experiment (Gcas_exp) and hot-start experiment (Hot_exp). The impact of cloud analysis on the prediction effect was investigated using 13 heavy rainfall cases and one month continuous experiments.

These experimental results show the use of cloud analysis can significantly improve forecasting skills of precipitation. Compared with hourly precipitation observations, Gcas_exp performed better than Con_exp and Hot_exp, which gets a higher threat scores of precipitation both for 13 cases and for one-month continuous experiments. Hot_exp presented an positive effect only in the first few hours. Oftentimes, Hot_exp got a worse forecast than Con_exp after the first several hours. In addition, gcas_exp has a positive effect on the prediction of 2m temperature, 10m wind and other variables, but forecasted composite reflectivity was stronger than its observations. Hot_exp can reduce this strength bias to some extent.

How to cite: Ren, X., Shao, A., Liu, W., and Chen, X.: Performance Evaluation of Cloud Analysis in GRAPES 3km Model over Northwest China, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-5217, https://doi.org/10.5194/egusphere-egu2020-5217, 2020.

 Based on conventional meteorological observation data and particulate matter monitoring data, combined with traceability of pollution sources, the main causes of sand-dust heavy pollution and the characteristics of dust transmission in eastern Qinghai in May 2019 were analyzed by using the principles of meteorology and trajectory analysis.The results show as following：The heavy dusty weather is mainly affected by the development of the enhanced low-slot eastward of Lake Baikal. The low-slots carry strong cold air moving eastward ,leading to heavy pollution in the eastern part of Qinghai. During the dusty weather, the cold air from the Hexi Corridor was poured into the eastern part of Qinghai from the valley . The dust from Gansu Province entered and then transported from east to west to easdtern Qinghai,causing pollution. The presence of the inversion layer stabilizes the atmospheric stratification in the eastern boundary layer of Qinghai, which is not conducive to the outward spread of pollutants caused by surface turbulence activities. The long-term maintenance of dust that cannot be diffused in time is the main cause of heavy pollution. In the early stage of sand and dust weather, the humidity conditions in the eastern part of Qinghai gradually deteriorated. Before the sand-dust occurred, the sensible heat on the ground increased significantly. The water vapor in the atmosphere weakened, the air was dry, and the water vapor condition, which is an important condition for the formation of sand and dust,was poor. The dust storm transmission route affecting the eastern part of Qinghai is transmitted from southeast to northwest. The mixed layer height and static weather index of the EC numerical forecasts have a good predictive indication during the process. The results of the trajectory analysis also indicate that the dusty weather in the eastern part of Qinghai (Xining, Haidong, etc.) was caused by backward irrigation of sandfrom the Hexi Corridor, and affected Haidong and Xining areas under the influence of the terrain.

How to cite: guo, X. and ma, Y.: Analysis of a typical heavy dust pollution weather in Semi-arid region- A Case Study in Eastern Qinghai , EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-5220, https://doi.org/10.5194/egusphere-egu2020-5220, 2020.

Using conventional and unconventional meteorological observation dataes, RMAPS-NOW( RR4DVar cloud model), the severe convective rainstorm occurred over the Beijing-Tianjin-Hebei junction area on May 17, 2019 was analyzed. Taking the observation of  Tongzhou 101 farm rainstorm (17: 30-20h precipitation 179.1mm) as an example, the results showed that this server convective rainstorm took place under a weak background, (1) Boundary conditions : The Beijing-Tianjin-Hebei area had high temperature and high humidity. The θse energy front was located in the middle of Beijing-Tianjin-Hebei. The CAPE at Beijing Observatory reached 1113 J · Kg-1 at 08h, and correction value reached 2669 J · Kg-1 at 14h. Convection cloud streets appeared around 12 o'clock in the visible image of the Himawari-8 satellite; the southeast wind jet in the boundary layer provided sufficient water  for convection development; at 20 o'clock, sounding showed that the vertical wind shear of 0-6 km increased to 17.5 m · s-1. (2) Trigger conditions: The southeast wind at the rear of the offshore High merged with sea breeze and pushed inland, formed a local convergence line with the local southerly wind in the central part of Beijing, Tianjin, and Hebei, and convection occurred at the convergence line and the θse energy front. (3) Tongzhou heavy rain was caused by two convective cells. Cell 1 was generated at the convergence line and the θse front. It developed into a server storm within 1 hour, and the composite reflectivity was > 60dBz. Subsequently, at its downstream (northwest side, the leading airflow is the southeast airflow), Cell 2 developed rapidly, and the two Celles revolved, moved over Tongzhou successively, accompanied by heavy rainfall ,hail and strong winds. (4) RMAPS-NOW data can describe the refined process of cells formation and evolution, that is, the θse field in the Beijing-Tianjin-Hebei region is extremely uneven, even in the θse front area. In the boundary layer convergence line and θse high-energy region, the convective bubble stimulated the formation of cell 1, the airflow spined up, and the development of the convective cell was strengthened. Half an hour later, a single cell was gradually separated into two cells (cell 1 and cell 2) in the upper layer, and the updraft gradually separated from the center into two rotating oblique updrafts. Seen from the echo profile, the two cells were connected by a cloud bridge and rotated clockwise. The convergence line in the boundary connected the two cells and rotated organically.

How to cite: yingxin, Z. and rui, Q.: Analysis of a server convective rainstorm in the weak background, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-5333, https://doi.org/10.5194/egusphere-egu2020-5333, 2020.

The ECMWF (European Centre for Medium-Range Weather Forecasts) precipitation type forecast products—PTYPE are verified using the weather observations of more than 2000 stations in China of the past three winter half years (October to next March). The products include the deterministic forecast from High-resolution model (HRE) and the probability forecast from ensemble prediction system (EPS). Based on the verification results, optimal probability thresholds approaches under criteria of TS maximization (TSmax), frequency match (Bias1) and HSS maximization (HSSmax) are used to improve the deterministic precipitation type forecast skill. The researched precipitation types include rain, sleet, snow and freezing rain.

The verification results show that the proportion correct of deterministic forecast of ECMWF high-resolution model is mostly larger than 90% and the TSs of rain and snow are high, next is freezing rain, and the TS of sleet is small indicating that the forecast skill of sleet is limited. The rain and snow separating line of deterministic forecasts show errors of a little south in short-range and more and more significant north following elongating lead times in medium-range. The area of sleet forecasts is smaller than observations and the freezing rain is bigger for the high-resolution deterministic forecast. The ensemble prediction system offsets these errors partly by probability forecast. The probability forecast of rain from the ensemble prediction system is smaller than the observation frequency and the probability forecast of snow is larger in short-range and smaller in medium-range than the observation frequency. However, there are some forecast skills for all of these probability forecasts. There are advantages of ensemble prediction system compared to the high-resolution deterministic model. For rain and snow, for some special cost/loss ratio events the EPS is better than the HRD. For sleet and freezing rain, the EPS is better than the HRD significantly, especially for the freezing rain.

The optimal thresholds of snow and freezing rain are largest which are about 50%~90%, decreasing with elongating lead times. The thresholds of rain are small which are about 10%~20%, increasing with elongating lead times. The thresholds of sleet are the smallest which are under 10%. The verifications show that the approach of optimal probability threshold based on EPS can improve the forecast skill of precipitation type. The proportion correct of HRD is about 92%. Bias1 and TSmax improve it and the improvement of HSSmax is the most significant which is about 94%. The HSS of HRD is about 0.77~0.65. Bias1 increases 0.02 and TSmax increases more. The improvement of HSSmax is the biggest which is about 0.81~0.68 and the increasing rate is around 4%. From the verifications of every kinds of precipitation types, it is demonstrated that the approach of optimal probability threshold improves the performance of rain and snow forecasts significantly compared to the HRD and decreases the forecast area and missing of freezing rain and sleet which are forecasted more areas and false alarms by the HRD.

Key words: ECMWF; ensemble prediction system；precipitation type forecast; approach of optimal probability threshold; verification

How to cite: Dong, Q., Zhang, F., Hu, N., and Zong, Z.: Application and Verification of the ECMWF Precipitation Type Forecast Product (PTYPE) in China, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-6348, https://doi.org/10.5194/egusphere-egu2020-6348, 2020.

In order to meet the high demand for advanced weather forecasts in the 2022 Olympic and Paralympic Winter Games, hourly wind observation data of some venues in Zhangjiakou City  is analyzed. Based on the specific gust characteristics in these venues, deviation of numerical weather prediction model is initially calculated to demonstrate the systematic bias of instantaneous wind speed forecasts derived from ECMWF. Additionally, a statistical down scaling method is further used by establishing the relationship between model forecasts and observation. Then independent samples are imported to the established equations to generate revised outputs. Tests show that the established equations have a better effect on forecasting the instantaneous wind speed than original model outputs and the corrected outputs have significantly better accuracy in predicting the instantaneous wind speed in the studied area.

How to cite: Li, J., Dong, Q., and Li, R.: Analysis of Gust Characteristics and Forecast Correction for the 2022 Olympic and Paralympic Winter Games in Zhangjiakou , EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-6627, https://doi.org/10.5194/egusphere-egu2020-6627, 2020.

Many Numerical Weather Prediction (NWP) models provide the parameter total snow depth as a Direct Model Output (DMO) surface variable. In mountain regions, however, the orographic flow modification significantly influences precipitation formation and preferential settling, leading to large model biases if DMO is directly compared to fresh snow point observations. Avalanche risk forecasts in turn require calibrated deterministic and probabilistic fresh snow forecasts, as the amount of fresh snow constitutes a crucial driver of avalanche risk.

In this study, MOSMIX-SNOW, a Model Output Statistics (MOS) product based on multiple linear regression is developed. Ground-based observations and operational forecast data of the two deterministic global NWP models ICON and ECMWF form the basis of the MOS system. MOSMIX-SNOW offers point forecasts for 20 deterministic as well as probabilistic forecast variables like the amount of fresh snow within 24h, the probability of more than 30cm of fresh snow within 24h and some basic variables like 2m temperature and dew point. The unique characteristic of MOSMIX-SNOW is the large number of observation-based, model-based and empirical predictors, which exceeds 200. Furthermore, a long historical data period of 9 years is applied for training of the MOS system. Thus, local orographic effects and large scale flow patterns are implicitly included in the MOS equations by a location and lead time specific choice of predictors. To avoid unrealistic jumps in the forecast, persistence predictors, which represent the forecast value of the previous forecast hour, are included in the MOS system. All forecasts feature a maximum lead time of +48h, have an hourly forecast resolution as well as update cycle and are available for about 15 mountain locations in the Bavarian Alps between 1100m and 2400m above sea level.

The verification analysis of the winter season 2018/19 shows that MOSMIX-SNOW forecasts offer a significantly higher forecast reliability than the raw ensemble of the regional NWP model COSMO-D2-EPS. The bias of the deterministic forecast parameters is smaller for MOSMIX-SNOW, especially for heavy snowfall events. MOSMIX-SNOW turned out to be a useful tool to support the avalanche risk forecasts on a daily basis during the snowy winter of 2018/19. Furthermore, the deterministic fresh snow forecast of MOSMIX-SNOW and other meteorological parameters like 2m-temperature serve as input for operational snowpack simulations. Measurement related noise and snow drift in the observations, however, are identified as an important source of uncertainty and the application of noise reduction techniques like a Savitzky-Golay filter are expected to have a beneficial impact on the forecast quality. MOSMIX-SNOW will become operational by end of 2020.

How to cite: Lambert, A., Trepte, S., and Ehrnsperger, F.: MOSMIX-SNOW – A Model Output Statistics Product for Fresh Snow Forecasts at Mountain Locations, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-8204, https://doi.org/10.5194/egusphere-egu2020-8204, 2020.

Impact of land surface processes on mesoscale convective initiation over Africa in ensemble model simulations: 3 Case studies using UKMO Unified Model
V S Semeena¹, C Taylor¹ and A Hartley²
1. UK Centre for Ecology and Hydrology, Wallingford, Oxfordshire, OX10 8BB, UK
2. Met Office, FitzRoy Road, Exeter, Devon, EX1 3PB, UK

The populations of the developing world have a greater need for accurate weather predictions because national economies and personal livelihoods depend very heavily on weather-sensitive factors including agriculture, water resources and public health.   Climate related risk is an obstacle in improving food security and rural livelihood in Africa.   An effective system to provide reasonable forecast can have a great positive impact on the life quality in African continent.    Thus predicting an event with accuracy is essential to provide early warning of heavy rainfall and floods that may lead to loss of life and property. Past studies have shown the importance of the land surface on the development of African convective storms. Here we are using 72 hour ensemble model simulations to evaluate the representation of land and its influence on convection in forecast models.   
Three episodes of heavy rainfall events are identified over western and eastern African region over late springtime of 2019 for this study.   A heavy rainfall event is recorded over SW Mali on 25th April 2019 followed by the development of a convective system over northern Benin on 29th April 2019.   The latter one develops into a mesoscale system on 30th April extending up to western Nigeria and this convective initiation in the afternoon and development into a larger system by late evening continues until 3rd May.    Our eastern African case examines the daytime development of convective cells which develop over southern Sudan, and grow into a mesoscale system which crosses over to Congo by midnight.  17 ensemble members simulation of the UK Met Office Unified Model (UKMO-UM) that were run for a forecasting testbed within the African SWIFT (Science for Weather Information and Forecasting Techniques) is used to understand the role of land surface temperature (LST) and soil moisture (SM) in formation of mesoscale systems.  Single-model ensemble simulations of the UM in global domain at 0.2813 X 0.1875 degree longitude, latitude resolution and the regional convection permitting (CP) model in 2 different horizontal resolutions – 8.8km and 4.4km – are performed.   Results are compared with LST from Meteosat Second Generation (MSG) satellite data and precipitation data from Global Precipitation Measurements (GPM).   Both global and regional models capture the main features though the convective initiation takes place much earlier in the models than in reality.   We notice that the representation of rivers and wetlands in the global model affects the spatial patterns of surface fluxes, in turn introducing biases into the forecast.    Further comparison of surface fluxes in the ensemble simulations of these case studies with observed LST and SM illustrate the importance of land initialisation for short term forecasts. 

How to cite: Valiyaveetil Shamsudheen, S., Taylor, C., and Hartley, A.: Impact of land surface processes on mesoscale convective initiation over Africa in ensemble model simulations: 3 Case studies using UKMO Unified Model, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-11429, https://doi.org/10.5194/egusphere-egu2020-11429, 2020.

 The accurate tropical cyclone (TC) track forecast is necessary to mitigate and prepare significant damage. TC has been predicted by the numerical models, statistical models, and machine learning methods in previous researches. However, those models are separately used for TC track forecast, and historical data with satellite images were used as input variables for machine learning without forecast data from numerical models. In this study, we corrected the TC track forecast of a numerical model by artificial neural network (ANN). TCs that occurred from 2006 to 2015 over the western North Pacific were hindcasted by the Weather Research and Forecasting (WRF) model, and all categories of TCs except for tropical depression (i.e., tropical storm, severe tropical storm, and typhoon) from June to November were included in this study. We evaluated the performance of TC track forecast in terms of duration, translation speed, and direction compared with the best track data. The simulated positions of TCs at 24-hour, 48-hour, and 72-hour forecast lead time were used as variables for training and testing ANN. To optimize the number of neurons in ANN, simulated TCs were divided into two parts; TCs in 2006-2014 for ANN optimization and those in 2015 for a blind test. Also, the output selection method based on the forecast error of the WRF was applied to exclude the outlier of ANN results. By applying the output selection, the forecast error of ANN was further reduced than that of the WRF. As a result, ANN with the output selection method could improve TC track forecast by about 15% compared to the WRF. Also, the effect of ANN tended to increase when the forecast error of the WRF was large. The output selection method was particularly effective by excluding outliers of ANN results when the forecast error of the WRF was small.

※ This research was supported by Next-Generation Information Computing Development Program through the National Research Foundation of Korea(NRF) funded by the Ministry of Science, ICT (NRF-2016M3C4A7952637).

How to cite: Kim, K., Cha, D.-H., and Im, J.: Improvement of Tropical Cyclone Track Forecast over the Western North Pacific Using a Machine Learning Method, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-13648, https://doi.org/10.5194/egusphere-egu2020-13648, 2020.

The aim of the work is to present simulation results of Weather Research and Forecasting (WRF) Model for high-resolution dynamical downscaling done over selected part of Poland. The research carried out a few unique simulations for selected days of the year 2019. For each model run different configuration of physical parameters (parametrization of boundary layer) were used. Additionally, two model runs were tested using the same configuration for physical parameterizations, but with two different spatial resolution. Additionally the sensitivity of the model in terms of spatial resolution was analyzed. Model was configured using two nested domains with 9 km and 3 km grid cell resolutions. All WRF simulations was simulated using GFS gribs with its initial time of 00 UTC. The results were compared with meteorological observations from meteorological stations. Results show high sensitivity of the obtained dynamical downscaling geophysical fields to the selected model configuration. High verifiability of air temperature forecasts was obtained using YSU and MYNN3 BL schemes. Mean Absolute Error (MAE) for temperature prediction has lower values in the summer season. Studies show the most optimal model configuration for BL for Poland area.

How to cite: Kendzierski, S.: A review of selected parameterization schemes of WRF model over Poland area in short-term weather forecast, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-21736, https://doi.org/10.5194/egusphere-egu2020-21736, 2020.