Ice clouds are mostly composed of different ice crystal habits. It is of great importance to classify ice crystal habits seeing as they could greatly impact single‐scattering properties of ice crystal particles. The single‐scattering properties play an important role in the study of cloud remote sensing and the Earth's atmospheric radiation budget. However, there are countless ice crystals with different shapes in ice clouds, and the task of empirical classification based on naked‐eye observations is unreliable, time consuming and subjective, which leads to classification results having obvious uncertainties and biases. In this paper, the images of ice crystals observed from airborne Cloud Particle Imager in China are used to establish an ice crystal data set called Ice Crystals Database in China, which consists of 10 habit categories containing over 7,000 images. We propose an automatic classification model of ice crystal habits, called TL‐ResNet152, which is a deep convolutional neural network based on the newly developed method of transfer learning. The results show that the TL‐ResNet152 model could achieve reliable performance in ice crystal habits classification with the accuracy of 96%, which is far more accurate than traditional classification methods. Achieving high‐precision automatic classification of ice crystal habits will help us better understand the radiation characteristics of ice clouds.  Reference: Xiao Haixia; Zhang Feng*; He Qianshan; Liu Pu; Yang Zhipeng; Classification of ice crystal habits observed from airborne Cloud Particle Imager by deep transfer learning, Earth and Space Science, 2019, 6: 1877-1886.

Observations have confirmed that clouds have a large impact on radiation which in turn affects cloud development. Most radiation parameterization schemes in WRF are parallel to the plane which ignores the inhomogeneous properties of clouds, while in fact almost no clouds fields are uniform. Inhomogeneity effects should be significant in Nagqu Prefecture of Tibet, which has frequent and massive local cumulus precipitation caused by intense radiation. The spherical harmonics discrete ordinate method (SHDOM) uses both spherical harmonics and discrete ordinates to represent the radiance field during different parts of the solution algorithm which can perform 3D unpolarized radiative transfer with high efficiency. In this paper we coupled the 3D SHDOM radiation scheme in WRF with microphysics scheme considering the surface albedo and underlying surface condition. We have done high resolution large-eddy simulations with the new radiation scheme of WRF to study the radiation characteristics of cumuli in Nagqu Prefecture, especially the effects of cloud top cooling on the cloud development. Comparing with the multi-observation data of ground and upper air in Nagqu Prefecture, the preliminary results show that the cooling and heating on cloud top highly impacts cumulus cloud developments in Tibet. 

As a new generation of hyperspectral sensor, the Tropospheric Monitoring Instrument (TROPOMI) on the Copernicus Sentinel-5 Precursor (S5P) was launched in October 2017 and started to globally map atmospheric trace gas constituents, clound and aerosols. In particular, aerosol parameters including UV aerosol index, optical depth, and aerosol layer height are vital to the assessment of the air quality in the lower atmosphere. The aerosol height information is derived based on absorption in the O₂ A-band in the near infrared wavelength range (759 and 770 nm). Aerosol retrieval from spaceborne measurements has been a challenging task and requires accurate and efficient forward modeling and inversion calculations. In this work, we present a conventional retrieval algorithm developed for TROPOMI/S5P O2 A-band and analyze the impact of different forward and instrument model parameters on the retrieval output. Through this work, an optimized retrieval configuration and a better understanding of the instrument characteristics can be achieved.

The Invariant Imbedding (IIM) T-matrix method is recognized as one of the most promising scattering models since it can perform the scattering simulation of the nonspherical particles in a semi-analytical way. However, because the T-matrix should be updated in each iterative process, its computational efficiency is an important issue in the actual scattering simulation. To alleviate this problem, the symmetrical properties for the nonspherical particles with symmetrical geometries are systematically investigated in this paper. Firstly, the symmetry of the U-matrix (an important matrix in the IIM T-matrix model) is derived for the particles with mirror symmetry with respect to the coordinate planes. In this case, the U-matrix is firstly decomposed into the sine and cosine components, and then its symmetrical properties are obtained by combining the spatial symmetry of the permittivity and the symmetry of the angular functions. In the second part, the symmetry of the U-matrix is derived for the particles with N-folds symmetrical geometry, and based on the symmetrical properties, the method to simplify the T-matrix iteration is further proposed. In this case, it can be found that by using the symmetry of the U-matrix, both the U-matrix and T-matrix can be rearranged into the block diagonal ones, and the calculation of the T-matrix can be decomposed into the iteration of several block sub-matrices, which can cut down the computational amount and memory consumption notably. Also, it can be seen that the derivation process also provides another point of view to understand the symmetry of T-matrix for the particles with N-folds symmetrical geometry.

Several radiative transfer schemes are compared in infrared spectra using the Rapid Radiative Transfer Model for General Circulation Models Applications (RRTMG). By calculating the root mean squared error of net flux (referred as RMSE(NF)) in various atmosphere, the sensitivities to cloud fraction are decreased in the order: non-scattering simplification (NS), the adding method of δ-two-stream discrete ordinate approximation (δ-2DDA), the adding method of δ-four-stream discrete ordinate approximation (δ-4DDA). The accuracy and efficiency of δ-2DDA and δ-4DDA schemes are studied under the contexts of using two different methods for dealing with the region where the solar and infrared spectra overlap. As one of the two overlap methods, One Band method (OB), which is used by RRTMG, has no advantage in model efficiency and is about 0.34 (0.41) W/m² higher in mean column RMSE(NF) of the δ-two-stream (δ-four-stream) schemes than Whole Bands method (WB). Moreover, a new scheme, which is simple but adequate to handle the overlap region, is derived to solve the solar energy in longwave spectra. Reference: Han Lin, Kun Wu*, Wenwen Li, 2020, Comparisons of radiative transfer schemes for infrared , spectra and the region with solar and infrared spectra overlap in RRTMG, Journal of Quantitative Spectroscopy & Radiative Transfer, 244, 106846. 

Sea salts are hygroscopic. During the deliquescence and crystallization processes, the particles could be nonspherical and inhomogeneous. Beccause of the importance of sea salt aerosol in atmospheric radiative transfer and remote sensing studies, the particle shape and inhomogeneity effects on the optical properties of sea salt aerosols should be examined. In this presentation, we report on recent progress on the polarized optical properties of sea salt aerosols. In particular, we highlight the inhomogeneity effect of sea salt aerosols on their single-scattering properties and the top of the atmosphere (TOA) polarized radiances. First, six sea salt particle models including homogeneous sphere, super-spheroids, inhomogeneous spheres with homogeneous sphere and super-spheroidal cores were defined to describe the morphologies of sea salt particles at different relative humidities. The Lorenz-Mie theory and the invariant imbedding T–matrix method were used to compute the single-scattering optical properties of the sea salt aerosols. Next, the polarized radiance at the top of the atmosphere (TOA) was simulated by using a radiative transfer model based on the adding-doubling method. Lastly, theoretial calcuations were then compared with PARASOL (Polarization and Anisotropy of Reflectances for Atmospheric Science coupled with Observations from a Lidar) observations. Based on the modeling results and the PARASOL observations, it is found that the inhomogeneity of sea salt causes salient negative polarization features at the backscattering angles (170°–175°). To explain such a phenomenon, a Debye-series formulation of light scattering by a coated-sphere is further employed to study the underlying scattering mechanism, and the geometric-optics term that causes the negative polarization is explicitly identified. These findings could be useful in remote sensing studies.

Cloudy sky spectral radiance at the top of the atmosphere has always been an important while difficult variable to simulate for fast radiative transfer models. In this paper, we focus on examining the impacts of cloud scattering properties on the spectral radiance signature of the High-spectral Resolution Infrared Atmospheric Sounder (HIRAS) onboard the Fengyun-3D (FY-3D) satellite by using the Advanced Radiative transfer Modeling System (ARMS) and the Community Radiative Transfer Model (CRTM). Cloud scattering properties used in the radiative transfer models are critical for modeling the spectral radiance under cloudy sky, which involves choices of appropriate cloud particle models and particle size distributions, etc. Multiple FY-3D HIRAS observations over Southern China and Southeast Asia with ice or liquid water cloud cover on 6 May 2018 are examined, respectively. Vertical atmospheric profiles are derived from the Modern-Era Retrospective analysis for Research and Applications, Version 2 reanalysis product. Cloud property retrievals from the Moderate Resolution Imaging Spectroradiometer are used. Cloud scattering property parameterization schemes based on spherical and nonspherical cloud particle shapes are implemented for liquid water and ice clouds in ARMS and CRTM, respectively. Results show that both ARMS and CRTM can well simulate the radiance at the HIRAS spectral ranges under liquid water cloud condition as compared with the HIRAS observation with mean absolute error (MAE) of brightness temperature of less than 1 K. However, for ice cloud conditions, ARMS model using assumed spherical ice properties exhibits large biases between simulation and observation. CRTM with nonspherical ice properties using 16-stream approximation shows MAE less than 1 K and MAE of about 1 K using 2-stream approximation.

The Directional Polarimetric Camera (DPC) carried by the GaoFen-5 satellite can continuously observe the earth in multiple bands, multiple angles, and high spatial resolution. This article uses the French multi-angle polarization load POLDER cloud recognition algorithm as a reference, and combines DPC multi-band reflectivity, polarization reflectivity, apparent pressure and other information to develop a cloud detection algorithm suitable for DPC. The algorithm is mainly divided into three parts: First, the threshold method is used to detect cloud pixels, and the apparent pressure is introduced to further restrict the conditions of clouds at different heights, and then use the 865nm band polarization Reflectance identifies the solar flare area reflected by the sea surface, and corrects the solar flare interference when the reflectance threshold is used to identify cloud pixels. In order to verify the accuracy of the algorithm, the MOD06 cloud mask product of MODIS on October 1, 2018 was compared with the results of the cloud recognition algorithm in this paper, and it was found that the cloud recognition results were in good agreement with the MOD06 products. In order to further quantitatively verify the accuracy of the cloud detection algorithm in this paper, the CALIPSO-VFM data from October 01 to 04, 2018 and the cloud detection results of this paper and the MYDO6 cloud mask product were selected to calculate the cloud/clear pixels hit rate and false alarm rate, respectively. The calculation results show that the average cloud hit rate of the algorithm is 13.501% higher than that of the MYD06 cloud mask product. The cloud error prediction rate is only 3.561% higher than that of MYD06 cloud mask products. The cloud detection algorithm proposed in this paper can provide important data support for subsequent DPC research on cloud parameters, water vapor, and aerosols. 

An early warning system for weather risk management in agricultural sector is operated by the Republic of Korea and covers major farming areas within the country. This system forecasts meteorological values and alerts individual farmers if a predicted meteorological condition may cause damage to their crops. However, meteorological forecasting can be especially difficult in complex mountainous terrain, where differences in the climate distribution, surface covering, and other conditions may cause spatial variation. Spatial downscaling using forecasts provided by the Korea Meteorological Administration has been well documented as a method to examine meteorological factors such as temperature, solar radiance, and precipitation. However, very little research has been conducted on the use of spatial downscaling forecast technology to predict relative humidity and water vapor pressure. This study examined the data collected from a network of 14 weather stations that were densely distributed in farming areas to quantify and explain the vapor pressure residuals obtained when calculating the relative humidity. Water vapor pressure was calculated using the temperature and humidity values observed at the end of each hour. Wind speeds were calculated by averaging the speeds measured over the last ten minutes of every hour.  The water vapor pressure deviation was corrected using hourly correction coefficients, whereas the remaining residuals were used to validate the vapor pressure differences due to surface covering (e.g., orchards, grasslands). Additionally, the maximum values of vapor pressure variation were calculated based on windless conditions. Our results showed that the characteristics of vapor pressure reduction at the target points depended on the wind speed and were expressed through an empirical equation. Areas with dense vegetation, such as forests and orchards, had a higher vapor pressure compared to grasslands, and there was generally very little net variation in vapor pressure during the night at wind speed of approximately 1m/s.

Traditional global climate models (GCMs) with coarse uniform resolution (UR) usually have deficiency in simulating realistic results at regional scale, while experimental global high-resolution models show benefits but also raise much computational burden. In recent years, variable resolution (VR) models with unstructured mesh are found to provide comparable results at regional scale and require less computational resources. In this study, the variable resolution CAM-MPAS model with the MPAS (Model for Prediction Across Scales) dynamical core coupled with CAM5 (Community Atmosphere Model Version 5) physics package is used to evaluate the effect of 30 km regional refinement over East Asia on the precipitation simulation. Our results show that the CAM-MPAS model can reasonably reproduce the annual and seasonal precipitation over East Asia, and the MPAS-VR simulation shows reduced mean bias and improvements in seasonal cycle, intensity distribution, and interannual variation compared with the low resolution MPAS-UR simulation. Furthermore, the major contribution to the improvements over the Tibet Plateau in the MPAS-VR experiment comes from the decrease of the grid spacing rather than the increase of the terrain resolution.

Over past years, the Weather Research and Forecasting (WRF) model has been widely used in the wind energy field, to determine the climatology of wind resource over large areas for site selection of wind farms, to assess the wind resource and annual energy production at a given site, and to forecast wind energy production of wind farms for electrical grid balancing purposes. Therefore, it is important to simulate accurately the wind speed at near-hub-height (typically ~100 m above ground) that is influenced by both mesoscale and local scale meteorological processes. This study evaluates a series of WRF simulations at resolutions from meso-scale (~5 km) to LES-scale (~60 m), i.e. 5 km, 3.3 km, 2 km, 1 km, 333 m, and 66 m, in terms of the model capability to reproduce the basic features of the wind at near-hub-height as well as the shape of wind shear. This study focuses on a few sites of Jiangsu Province, where intensive wind energy penetrations exist both onshore and offshore due to the favorable wind conditions and relatively large coastlines. The measurements of wind speeds at several altitudes are used to compare with the modeling results. The results show that the simulations with different horizontal resolutions tend to have different performance at the coastal and onshore stations. The simulations at all resolutions can roughly reproduce the wind speed at 70 m above the ground but overestimate the wind speed at the onshore site. The LES simulation captures more observed micro-scale wind features and reduces the monthly mean of modeling bias. More analysis of the difference in modeling results at various resolutions will also be discussed.

The rapid strengthening process of typhoon “Mujigae” over the offshore area was affected by various physical factors and processes including its structure, large-scale atmospheric circulation and marine environment. The improvement of typhoon intensity prediction requires the in-depth understanding of the dominant physical processes that are responsible for the rapid intensification of typhoon. Theoretical analysis of numerical simulation data was conducted to quantitatively identify such dominant physical factors and processes. The physical framework for rapid intensification of offshore typhoon with energy budget as the core was established and the high-resolution numerical simulation data of typhoon “Mujigae” were obtained by WRF. On this basis, the environmental wind field, divergent wind field and rotating wind field during the rapid intensification process of typhoon were decomposed, the kinetic energy and available potential energy budgets of typhoon rotation and divergence circulation were calculated. Analyzing the effects of the energy factors and the spatial configuration and temporal evolution of related physical quantities in the energy conversion process, the effects of barotropic, baroclinic, radiation, sensible heat and latent heat on the rapid intensification of typhoon were obtained, and the associated mechanism of the rapid intensification of typhoon over the offshore area was revealed.

We investigate this interaction using observational data and a high-resolution simulation of a hailstorm that occurred over Taizhou (Zhejiang, China) on March 19, 2014. During the hailstorm event, near-surface meso-γ vortices along a convergence line interact with hail cells. Herein the 10-m surface wind data from automatic weather stations shows that several meso-γ vortices or vortex-like disturbances existed over the convergence zone and played a vital role in the evolution of the hailstorm and the location of the hail. The model results agree with the observations and present a closer correlation between the hail and the low-level meso-γ vortices than those observed. The model simulation indicates that such low-level meso-γ vortices can be used to predict the next 10-min hail fallout zone. The low-level meso-γ vortices originated over the convergence zone and then fed back into the convergence field and provoked a stronger updraft. Vorticity budget analysis found that positive vorticity was mainly caused by the stretching and tilting term. Vorticity generation was initiated primarily by stretching and was extended by tilting. A three-dimensional flow analysis shows that the existence of low-level meso-γ vortices could help enhance a local updraft. Furthermore, the simulation reveals that the low-level meso-γ vortices existed in the bounded weak echo region at the front of the hail cell, enhancing convergence and strengthening updrafts. Graupel was broadly located between the 0°C isothermal line and the top of the clouds, roughly between the 0°C and −20°C isothermal lines. Accordingly, the hailstones grew rapidly. The suitable environment and the positive effect of the meso-γ vortices on the updrafts enabled hailstorm formation. 

Urban heat island is the most documented phenomenon of climate change. There are more than 450 cities where the phenomenon is experimentally documented. Urban overheating has a serious impact on energy, health, pollution and survivability of low income population. The present paper will present the more recent developments in the field of urban mitigation technologies. Progress on urban greenery, reflective, chromic, fluorescent and photonic materials will be presented in detail. Their potential to decrease the peak ambient temperature will be analysed. The impact on energy demand, peak electricity consumption, heat related mortality and morbidity as well as on pollution levels will presented.  

Recently urban heat situation in summer becomes problematic more and more in most Asian cites. Urban heat is the combined effect of global warming and urban warming. The latter is possibly weakened by appropriate urban planning, though global warming cannot be hindered by local municipality’s efforts. Among many heat mitigation methods, a cool roof with high reflective painting seemed to be the most effective solution and many of the previous studies showed that cool roof installation reduced urban air temperature by 1 to 2 K. In addition cool roof required no running cost unlike the other mitigation techniques. Despite many numerical studies on the cool roof applied to various cities, we found cool roof potential has not been evaluated appropriately yet because they assumed uniform and homogeneous urban forms in their simulations. We simulated the effect of the cool roof installation in Tokyo Metropolis using the WRF model with a single-layer urban canopy model, which was modified so that the actual urban parameters were incorporated in the simulation. The result showed that near surface air temperature (Ta) reduced by 0.24 K on average over Tokyo Metropolis if reflective painting (0.85 in albedo) was applied to all the buildings’ roofs and this reduction was much smaller compared to previous simulations on cool roof found in the literature. In addition, the degree of the Ta reduction linearly increased with the value of packing density of buildings. Since the urban parameters in the default WRF seem to overestimate actual urban forms for many cities as exemplified in the value of plane area index, 0.50, previous studies possibly resulted in the overestimation of temperature reduction by cool roof. 

Governments around the world have implemented measures to slow down the spread of COVID-19, resulting in a substantial decrease in the usage of motorized transportation. The ensuing decrease in the emission of traffic-related heat and pollutants is expected to impact the environment through various pathways, especially near urban areas, where there is a higher concentration of traffic. In this study, we perform high-resolution urban climate simulations to assess the direct impact of the decrease in traffic-related heat emissions due to COVID-19 on urban temperature characteristics. One simulation spans the January–May 2020 period; two additional simulations spanning the April 2019–May 2020 period, with normal and reduced traffic, are used to assess the impacts throughout the year. These simulations are performed for the city of Montreal, the second largest urban centre in Canada. The mechanisms and main findings of this study are likely to be applicable to most large urban centres around the globe. The results show that an 80% reduction in traffic results in a decrease of up to 1 °C in the near-surface temperature for regions with heavy traffic. The magnitude of the temperature decrease varies substantially with the diurnal traffic cycle and also from day to day, being greatest when the near-surface wind speeds are low and there is a temperature inversion in the surface layer. This reduction in near-surface temperature is reflected by an up to 20% reduction in hot hours (when temperature exceeds 30 °C) during the warm season, thus reducing heat stress for vulnerable populations. No substantial changes occur outside of traffic corridors, indicating that potential reductions in traffic would need to be supplemented by additional measures to reduce urban temperatures and associated heat stress, especially in a warming climate, to ensure human health and well-being.

Transit-Oriented Development (TOD) aims to optimize environmental and economic issues and the quality of life (QOL) around a station of the public transportation system in a community area. TOD planning is especially expected to enhance low carbonization, improving the urban climate with compact and efficient land-use design in an area along with public transportation such as railways. However, the comprehensive and interactive methods for the above problems have not been investigated sufficiently. The study proposes a development concept for improving the environmental value and reducing the environmental load. The improvement of the environmental value aims to enrich the greenery around buildings and communities and cause to create a thermally comfortable environment to mitigate patients by heatstroke. The reduction of the environment load will be expected to find out the comprehensive methodology of both the modal-shift from the vehicle-oriented to the public transportation-oriented society and installing low carbon energy systems for all buildings in communities. This study evaluates numerically the network of railway stations along with the Tsukuba Express (TX) line in Japan using the Weather Research and Forecasting (WRF) model with a resolution of 250-m.  The results of this study will be valuable for further studies related to the urban climate, urban planning, TOD, human health, and energy management, as this study is the first numerically and multidisciplinary attempt focuses on transport development and human influences on the local climate and lifestyle. 

High levels of urbanisation, coupled with an increasing trend in extreme weather under future climate change scenarios, combine to create significant challenges to increasing urban resilience for the future (Masson et al., 2020). Urban climate services provide tools to support decision making at a range of scales across the city (Grimmond et al., 2020) and can therefore be used to inform city resilience. This presentation looks at a prototype urban climate service which provides long-term climate change projections at the city-specific scale. The ‘City Pack’ service is a set of factsheets which provide high-level non-technical summaries of climate change projections for an individual city and the science behind them. The audience for the City Pack includes city officials, city planners and the general public. In 2019 the first ‘City Pack’ was co-developed by the Met Office and Bristol City Council. Since then, the template developed with Bristol City Council has been used to develop City Packs for a number of additional cities across the UK. Feedback from this process is ongoing, with initial findings suggesting that the expansion of the City Pack within the UK has been successful. The City Pack is now being upscaled for a number of cities in China under the CSSP (Climate Science for Service Partnership) China Project. Prototype City Packs are currently being developed for the cities of Harbin, Shanghai and Chongqing. The project is seeking to ascertain if services which are co-produced with and bespoke to one set of stakeholders, may provide an equally valuable service for other cities and what support tools need to be in place to ensure effective use and uptake. In doing so, the project is also looking to investigate the ways in which a service may be upscaled to increase the reach and impact of the climate service.

Recent extreme heat events are likely to become more frequent over the 21^st Century and exacerbated in cities due to the urban heat island effect.  Due to high population densities and a concentration of assets, urban areas are more vulnerable to climatic extremes with impacts that traverse health, infrastructure, built environment and economic activity. Recent advances in high resolution modelling enable better representation of urban processes and provide greater understanding of extreme events. By exploiting such advances in underpinning science, the Met Office is generating urban climate services for city stakeholders to plan for and manage heat stress in their city. The Met Office has been engaging with local authorities and city stakeholders in China and the UK to co-produce a prototype, two tier, urban heat climate service to enhance the resilience of urban environments to extreme heat events. The prototype is based on a strong requirement from several cities to develop an evidence base of the heat hazard and understand current and future hot spots vulnerable to extremes of heat within the city. Tier 1 uses observations and high-resolution climate data to provide city specific information of the heat hazard in a graphical factsheet format. This includes information on future changes in temperature, extreme heat indicators, frequency and duration of heatwave events, and spatial distribution of heat across the city. Tier 2 involves working closely with city stakeholders to combine the hazard information with data on health, built environment and socio-economics, to provide tailored information on heat exposure and vulnerability. This will allow users to identify highly vulnerable parts of the city network and neighbourhoods for priority action. This two-tier service can provide an evidence base to inform urban policy, design and adaptation strategies, and prepare authorities and city stakeholders for future demand on city services.

The decision model of household evacuation has been developed and discussed for long decades. The complexity of decision making under uncertainty using weather forecast and early warnings makes the model design so difficult since many levels of factors should be considered and decision path should be clarified. This research aims to construct a structural equation model (SEM) to consider factors accounting for personal cognition on risk information, including individual demographic factors, rational assessment process, psychological responses, social interactions, and cultural backgrounds. We utilize this SEM to inspect the effectiveness of current weather forecast services and flood early warnings in raising risk perceptions and the likelihood of household evacuation for Taiwan and US residents. The research results suggest both countries should use media most trusted to disseminate weather risk information, either TV, broadcast, or social media. TV, broadcast, and social media could increase individuals’ exposure to risk information across lead times. We also suggest individuals read risk information as reminders to inform themselves of risk situations so they could adjust risk perceptions across lead times and make more proper decisions pertaining to household evacuations. Correct interpretation of weather forecasts are essential capacity which should be developed pervasively in both countries, and the needs for customized weather forecasts should be met for weather forecast users in both countries. Weather forecasts are imperfect information with uncertainty and we should emphasize the uncertainty of flood hazards and risk situations at earlier lead time to the general public, and let users understand the chance of potential hazards by reading probabilistic forecasts in addition to deterministic forecasts. With proper interpretation of weather forecasts and early warning by the general public, weather forecasts and early warnings could exert effectiveness in risk management and encourage household precautionary actions at proper and earlier lead time.  

Renewable energy has been increasingly developed and utilized in the recent years; however, the grid-connection and power dispatch scheduling of renewable energy is still a challenging issue because of the intermittent and unpredictable characteristics of renewable energy. High-quality meteorological forecasts play a critical role in renewable energy application and management, such as the unit scheduling for a power system. This study focuses on the evaluation of both forecast quality and economic benefit of the probabilistic forecasts of 100-meter wind speed from WRF ensemble prediction system (WEPS), and the purpose is to help users optimize their decision-making in strong wind situations, which influence the security and service life of wind turbines. The evaluation results indicate：(1) The probabilistic forecasts of 100-meter wind speed from WEPS display more obvious over-forecasting during northeast monsoon periods than southwest monsoon, and over northern Taiwan than the other areas; (2) Both reliability and potential usefulness of the probabilistic wind forecasts are satisfactory; (3) Almost all users can obtain economic values (EV) if they make decisions based on WEPS during the typhoon seasons for perennial time; and (4) Decision-makers can benefit more from the probabilistic forecasts than the ensemble mean forecasts derived from the same EPS because probabilistic forecasts take forecast uncertainty into account.

On 19 April 2019, a mature squall-line mesoscale convective system (MCS) with the characteristics of a leading convective line and trailing stratiform landed on Taiwan, resulting in strong gust wind and heavy rainfall. This squall-line MCS became asymmetric after landfall on Taiwan. There were two asymmetric features discussed in this study: one was the upwind-side asymmetry due to different ridge orientations, and the other was lee-side asymmetry associated with hydraulic jumps.    Two sets of idealized numerical simulations using the Weather Research and Forecasting (WRF) model were conducted to examine the impacts of realistic Taiwan topography on a squall-line MCS. Model results showed numerous similarities between the idealized simulations and real-case observations on 19 April 2019. The low-level Froude number which considered the terrain height (F_mt) was calculated to examine the blocking effect of the Taiwan terrain, and the cold pool (determined by the –2 K isotherm) was found to be completely blocked by the 500-m height contour. The northeast-southwest orientation of the Snow Mountain Range (SMR), and the north-south orientation of the Central Mountain Range (CMR) led to the upwind side asymmetry. On the other hand, the lee-side asymmetry was associated with the different intensities and occurrence locations of the hydraulic jump between the northern SMR and southern CMR, and the Froude number which considered cold pool depth (F_cp) was used to determine flow regimes. The low-level Froude number (F_mt) was determined by the terrain-height and low-level shear. A series of sensitivity experiments were conducted to better understand how the structure of the squall-line MCS varied along with the F_mt. Spatial correlation coefficient was utilized to quantify the degree of MCS symmetry in terrain-height experiments, and the low-level shear experiments were performed to better understand the characteristics of the hydraulic jump. 

System for Convection Analysis and Nowcasting (SCAN) is one of the operational forecast systems used by the Central Weather Bureau (CWB), which can provide the storm cells information by the function of Storm Cell Identification and Tracking algorithm (SCIT). This study presents an 8-year analysis of summer storm cells in Taiwan (from 2015 to 2018, May-Aug) based on the SCIT, discusses the characteristics of cells under different weather systems (synoptic or weak synoptic) and geographical environments (north or south of Taiwan). The results show that the movement of summer storm cells in Taiwan is mostly influenced by topography. Large amount of storm cells are triggered and distributed near the mountain areas. In addition, the lifetime is around one hour, and the maximum reflectivity is concentrated at 45~55dBZ. The statistical analysis also reveals that the one-hour forecast error in the southern region is higher than the error in the northern region, and the error in the synoptic day is higher than which in the weak synoptic day. For storm tracking nowcast, a method of Potential Track Area (PTA) is developed through the statistical results of storm track errors. It is able to define 0-1-hour severe storm warning area and quantitatively improve the warning capability of severe weather. By using the Probability of Detection (POD) to validate the performance, it is found that the PTA can capture the storm cells well in different types of weather systems.

The predictability of precipitation is very low because of its stochastic nature; therefore, the prediction of precipitation beyond 5 days is a big challenge for meteorologists. However, demand for medium- (3-to-10 days) and extended-range (10-to-30 days) precipitation forecasts by potential users in agriculture, forestry, livestock, and water resource management has grown significantly. In this study, a statistical post-processing technique combining similarity matching (SM) and probability-matched mean (PM), called SMPM, is developed to perform bias correction and downscaling of 1-to-14 day precipitation forecasts in Taiwan. The aim is to provide users with more accurate quantitative precipitation forecasts (QPFs) and reliable probabilistic quantitative precipitation forecasts (PQPFs). SMPM searches for the best analogs to the current forecast in a historical set of predictions. Similarity is defined based on large-scale circulation indices instead of precipitation patterns, as the former (1) is much more predictable, and (2) correlates well with local precipitation. The predicted SMPM precipitation forecast ensemble is the observed precipitation patterns corresponding to the historical forecasts that most resemble the current large-scale circulation forecast. For a single value forecast (QPF) with a realistic range of precipitation values, PM is applied on the mean of the forecast ensemble. Forecast evaluation shows that SMPM successfully corrects the precipitation pattern and amount of the raw forecasts, with some fine scale details added. Besides, SMPM significantly improves the accuracy of precipitation forecasts for reservoir water sheds, which helps water-resource management decision making. Also, the reliability of the PQPFs based on the SMPM ensemble is significantly higher than that from the raw precipitation forecast ensemble.  

The purpose of this study is to investigate the impact of assimilating additional polarimetric parameters with reflectivity (Z_H) and radial wind (V_r) in the high impact weather systems. A squall line case forced by the synoptic southwesterly wind and a local afternoon thunderstorm case are selected to conduct the assimilation experiments with WRF-LETKF Radar Assimilation System (WLRAS). In addition, different microphysics parameterization schemes, including GCE, MOR, WSM6 and WDM6, are examined in the experiments. The results in both cases show that assimilating Z_DR in addition to Z_H and V_r with single moment schemes (GCE and WSM6) can capture better raindrop mean size in the analysis field, yet it deteriorates simulated Z_H and K_DP compared with the observations. Differ from GCE and WSM6, assimilating Z_DR in double moment schemes (MOR and WDM6) would not lead to significant deterioration in the simulated Z_H and K_DP. This is because the prognostic hydrometeor variables include both mixing ratio and total number concentration, which provides more correction flexibility in the model. Additionally, the improvement of simulated K_DP in the analysis caused by assimilating K_DP is more obvious in WSM6 and WDM6 scheme than in GCE and MOR scheme. In conclusion, this study illustrates that double moment schemes can be adapted to the extra information from additional polarimetric parameters more flexibly than single moment schemes and might have more benefit in analysis and further improve the quantitative precipitation forecast for the heavy rainfall events.

Recently, South Korea have been suffering from an increasing number of short-duration extreme rainfall events (more than 20 mm/hr) and it is imperative to understand how these extreme events will change in the future. These meteorological disasters usually occur at small spatial (less than 5 km) and short time (hours) scales through the drastically developed local deep convection system. However, regional climate models (RCMs) have horizontal resolution of 12.5 ~ 50 km and utilize convective parameterization schemes, limited for simulating the short-duration extreme rainfall events. In this study, we run a convection permitting model (CPM) for historical and future periods using the CCLM (2.5 km resolution) and examine its performance and projection for the hourly extreme precipitation. Results from RCM runs (25 km resolution) which provide lateral boundary forcings to CPM runs are compared to identify CPM’s added values. The present-day simulations indicate that CPM can reproduce the observed hourly extreme precipitation intensity for given temperature ranges (from 8 to 28°C) and also capture the observed extreme precipitation-temperature (P-T) scaling, which overall follows the Clausius-Clapeyron relation. In contrast, the RCM run underestimates extreme precipitation intensity at temperatures above 18°C, more strongly at higher temperatures, resulting in a much weaker P-T scaling than the observed. Under the global warming scenario (RCP8.5), CPM projects an overall increase in extreme precipitation intensity over South Korea, which is stronger than the corresponding RCM simulation. CPM runs show that the P-T scaling will remain in a warmer world with an extension of the P-T slope into high temperatures, indicating the increased occurrence of short duration extreme precipitation events. The role of precipitation type (convective vs. large-scale) in the present-day and future P-T scaling will be discussed based on the evaluation of convective available potential energy (CAPE).

The urban heat island (UHI) is one of the most well-known urban climate phenomena. Recently, as opposed to UHI, the urban cool island (UCI) phenomena with lower daytime temperature in high-rise building areas than suburban have been reported. So far, not many studies have investigated the relationship between UCI phenomena and precipitation. Thus, in this study, we examined the effect of the urbanization factor (i.e., anthropogenic heat flux (AH), building height (BH)) that can affect UHI and UCI on urban precipitation by using WRF coupled to SLUCM. The AH experiments consisted of four types (CTRL, AH25, AH75, AH225) with hourly maximum AH prescribed 0, 25, 75, and 225 W·m^-2, respectively. The AH increased urban and downwind area precipitation, especially during the daytime. The AH is added directly to the sensible heat flux, which increases surface temperature, especially during the daytime, associated with the UHI. The surface heat is then transferred throughout the planetary boundary layer (PBL) by turbulence, which increases air instability. The BH experiments consisted of three types (CTRL, BH12, BH18) with BH of 7.5, 12, and 18 m, respectively. In the BH experiments, precipitation as well as surface temperature increased during the nighttime but increased during the daytime. The ground heat flux increased during the daytime and decreased during the nighttime as the BH increased and the sensible heat flux changed in the opposite direction to the ground heat flux for energy balance. The ground heat flux can also explain the phenomenon of UCI in areas where high-rise buildings are concentrated. Thus, increasing BH resulted in a decrease in the surface temperature during the daytime due to the thermodynamic factors and stabilizing the atmosphere throughout the PBL, causing reducing precipitation. The AH and the BH played a role in alleviating changes in precipitation during the daytime.

Diurnal cycle of precipitation over the Indochina Peninsula during the South East Asian summer monsoon season was examined by using non-hydrostatic (5km grid), and convection permitting (2k grid) regional climate models (NHRCM). Our results indicated that those fine grid models lead better performance to represent diurnal cycle of precipitation due to realistic representation of geographical distribution. The models successfully simulated the local circulation corresponding to the intensification of precipitation. They are consistent with the satellite-based observed diurnal cycle of the precipitation. The model simulation indicates that convergence area over the mountain on the south of Khorat Plateau occurred in the afternoon in association with the occurrence of precipitation. This migrated northward and contributed to the precipitation peak over the plateau at the night time. The models outperformed parent model MRI-AGCM in terms of the timing of diurnal peak of precipitation and total amount of precipitation, which is typical added value in the dynamical downscaling. Bias correction was also applied to the model results to study potential impact of climate change.

Current climate models often have significant wet biases in the Tibetan Plateau and encounter particular difficulties in representing the climatic effect of the Central Himalaya Mountain (CHM), where the gradient of elevation is extremely steep and the terrain is complex. Yet, there were few studies dealing with the issue in the high altitudes of this region. In order to improve climate modeling in this region, a network consisting of 14 rain gauges was set up at elevations > 2800 m above sea level along a CHM valley. Numerical experiments with Weather Research and Forecasting model were conducted to investigate the effects of meso- and micro-scale terrain on water vapor transport and precipitation. The results show that the simulations with high horizontal resolution, even without the Turbulent Orographic Form Drag (TOFD) scheme, can not only increase the spatial consistency (correlation coefficient 0.84–0.92) between the observed and simulated precipitation, but also considerably reduce the wet bias by more than 250%. Adding the TOFD scheme further reduces the precipitation bias by 50% or so at almost all stations in the CHM. The TOFD scheme reduces precipitation intensity, especially heavy precipitation (> 10 mm h⁻¹) over high altitudes of the CHM. Both high horizontal resolution and TOFD enhance the orographic drag to slow down wind; as a result, less water vapor is transported from lowland to the high altitudes of CHM, causing more precipitation at lowland area of the CHM and less at high altitudes of CHM. Therefore, in this highly terrain complex region, it is crucial to use a high horizontal resolution to depict mesoscale complex terrain and a TOFD scheme to parameterize the drag caused by microscale complex terrain.

This study provides an example of application of radar-based rainfall estimates for water hazard mapping and disaster risk communication. It is a prototype activity in the Borderless Radar Information Networking over South and Southeast Asia (BRAIN) project. This study investigates rainfall and its social impacts in the Kuma River basin flowing through Kumamoto Prefecture in Japan during the torrential rainfall event in July 2020. Using the Radar/Raingauge-Analyzed Precipitation product by the Japan Meteorological Agency (JMA), catchment-mean rainfall values are computed at all places (10,487 grid points) in the river basin and at the time scales from 1 hour to 48 hours for the 17-year period from 2004 to 2020. Then, probability density functions (PDFs) of catchment-mean rainfall are derived at each grid point and at each time scale; using the PDFs based on past records, catchment-mean rainfall values during the torrential rainfall event are converted into return periods as an index of rareness of rainfall. The return period values can be summarized as a geographical pattern of the maximum return period of catchment-mean rainfall during the event and in the 48 time scales. This geographical pattern clearly shows the locations of dangerous places with long return periods where embankment break, overtopping or inundation was confirmed at 35 points. The simple mapping of water hazard proposed in this study uses rainfall and flow direction data only. The latter flow direction data are available at any river basins over the globe. And thus, it is easy to apply this methodology in many countries and regions where long-term past records of rainfall are available. We call this methodology the Simple Water Hazard Mapping (SimpleMap). Outputs of the SimpleMap show simple information about rainfall. Therefore, they can be quite easily understood. It is an essential point in disaster risk communication. 

Raw outputs from Global Climate Models are hardly efficient in climate change impact assessment at catchment scale. In this study, departing from the conventional Top-Down model, a Bottom-Up approach is adopted to identify the changes in hydrological responses of the catchment due to changes in climate and land use. Climatological parameters (precipitation and temperature) have been analyzed for the period 1901-2013 and statistical measures like standard deviation, skewness, kurtosis, anomalies are calculated. The Mann-Kendall’s test and Sen slope analysis were used to check if any trend present in the datasets. Critical thresholds of climatic parameters and land use change are estimated to be classified as vulnerable for each hydrological indicator. The vulnerability index is chosen as the maximum acceptable change in climate and vegetation beyond which the hydrological responses turn out to be vulnerable. A correlation is established between these indices and watershed characteristics like topography, land use, and soil moisture. The RISE model setup was used for rainfall-runoff modelling and simulation of streamflow in daily timescale. Vulnerability of two year return period flood is quantified in response to various factors like wet spells, seasonal rainfall, vegetation condition, soil condition, proceeding drainage. Interaction between flood and its driving factors has been addressed with the help of comparative hydrology over a large range of watersheds. For generalization of the approach, fifty-six sub basins of the Brahmaputra river basin have been explored. Errors can be avoided up to great extent using this strategy because the impact assessment is independent of future projections of change drivers. Thus it sets a hope for water resources planners and managers to take more accurate decisions in ungauged basins in spite of large uncertainty.

Our previous research confirmed the high reliability of inundation pattern analysis using MODIS reflectance data, even for local inundation in Bangladesh. We found that waterlogging flood was associated with prolonged monsoon-inundation and local rainfall. Present research targeted to explain how waterlogging flood inside several polders on south-western parts of Ganges-Brahmaputra delta is related with local rainfall, monsoon inundation, water discharge rate and water level at distributaries related to targeted polders from the year 1970 to 2019. To detect temporal evolution of inundations and waterlogging flood, we will utilize Landsat reflectance data. Reliable seasonal rainfall data, water discharge ratio data and water level data for targeted rivers from 1970 to 2019 will be used to see the relationship of waterlogging flood. Several Tidal River Managements (TRM), a community-oriented solution for waterlogging flood, implemented by government agencies to solve this waterlogging flood on south western part of the Ganges Brahmaputra delta that are likely continued to further north.  But previously some protests had been arisen from the stakeholders (villagers, land owners, local elites) during and after implementation of TRM. Thus, our research includes community’s historical adaptation and perception to solve waterlogging flood other than TRM. Primary observation of the area and perception of the stakeholders confirmed the existence of waterlogging even in January 2020 and they are about to loss this year rice production. And they might accept the alternative of TRM only if the solution will not harm their crop and fish cultivation. The result may be helpful for policy makers.      

Bhubaneswar, the eastern recently developing city of India, has experienced rapid urbanization since 2000. The city has also witnessed heavy rainfall events that have been exhibiting spatiotemporal variability. The pre-monsoonal (March-April-May) rainfall events over Bhubaneswar city taking the India Meteorological Department (hourly station and daily gridded) datasets has been studied. The data is analyzed for the 1980-2018 period (39 years). Wavelet and trend analysis for temporal changes of rainfall reveals that the intensity of precipitation has increased over the study period; which is about 0.4 mm/season. The increase in the rainfall is preferentially high over the city and along the right side of the city compared to the left side of the city. To examine this type of variation in rainfall, a supervised classified land-use-land-cover (LULC) map of the Bhubaneswar region has been examined for 1980, 1990, 2000, 2010, and 2019 years. To produce these LULC map, cloud free Landsat imageries has been downloaded from United States Geological Survey. From the analysis, it is found that the urban built-up area is increasing over the region throughout the study period and the increment is about 100 km² from 1980 (22.5 km²) to 2019 (122 km²). There is a positive correlation with increase in urbanization and rainfall over Bhubaneswar and this correlation is also increasing with time.

Prediction of lightning is an area of interest and research for the decades. The study evaluates the relationship between lightning flash occurrences and the rate of change of stability indices for 17 years (1998-2015) over Bangladesh. In addition lightning climatology (1998-2015) and stability indices climatology for the same time period are studied to locate the hot spot over Bangladesh and its seasonal variation. A climatological study shows that lightning flash rate shows highest magnitude over North-East part of Bangladesh during pre-monsoon season. Lightning Imaging Sensors (LIS) on board Tropical Rainfall Measuring Mission (TRMM) derived lightning flashes (1998 to 2015) are analyzed and correlated with the JRA-55 reanalysis data derived Instability indices (Showalter Stability Index, Total Total Index, K Index, Precipitable Water, Cape, Mean Shear, Storm Relative Environmental Helicity, Vorticity Generation Parameter) to investigate the association among them. A brief spatial and temporal relationship between lightning flashes count and stability indices over Bangladesh are explored. A threshold analysis of stability indices based on K-means clustering technique is constructed to forecast the lightning flashes intensity.  

Lightning is a dangerous natural hazard and major concern of for the death of people and property loss. Due to heavy intensity lightning in Bihar region which kills around 120 people in less than 3 days in June month. The climatology of mean lightning flash rate for the global level is already produced but the detailed analysis of over India is not properly studied. In this study, Lightning climatology for the period of 17 years from 1998 to 2014, over India region by using Lightning Imaging Sensor (LIS) of 0.1° × 0.1° very high spatial resolution of Tropical Rainfall Measuring Mission (TRMM) satellite. The detailed analysis and comparison of these climatology has been studied with different resolution datasets of LIS-OTD sensor of satellite TRMM which are Low resolution (2.5), High resolution (0.5) respectively. Diurnal, monthly, seasonal and annual variations in the occurrence of lightning flashes rate density has also been analysed. This study has been focused on different aspect such as change in the lightning flash rate density with respect to current scenario with previous years to find out the hot spots over different regions of Indian land mass. The diurnal lightning event is mainly occurred in the afternoon/evening time duration. The highest lightning occurred in May Month and least in December. The distribution of lightning flash counts by season over India landmass is mainly in pre-monsoon (March-May) and monsoon (June-Sept) and decreased afterwards. Spatially, the distribution of lightning flashes at North Eastern region along Bangladesh, Jammu& Kashmir (South western region) border with Pakistan and Shivalik Himalaya (Himalayan foothills). Also, further work is to investigate trends in the lightning flash rate frequency, intensity and temporal shifting changes over the India, which will be helpful in better understanding of Lightning events. Keywords: LIS-OTD, TRMM, Lightning flash density, Trend analysis, Climatology.