Weather Prediction System Research Articles

Accurate Wind Observation and Robust Control for Drones in the Field of Micro-meteorology⋆

Urban logistics has been the primary target for recent drone applications. However, these drones face varying weather conditions in their flight horizon that can limit their performance. From another perspective, this flight horizon is the focus of micro-meteorology data gathering for a city-wide accurate weather prediction system. Hence, these drones can carry out logistics while gathering micro-meteorology data. This paper focuses on the latter and aims to realize the drone capabilities required for forecasting micro-meteorology, especially observing highly accurate wind data and robust control capability against strong wind disturbances. The generation of wind maps using a 2D wind sensor mounted on a drone and wind gust estimation using a nonlinear disturbance observer (NDOB) are proposed as the capability for accurate wind observation. Next, the position and velocity control of the drone using super twisting sliding mode control (STSMC) and NDOB are designed for robust control. Lastly, numerical simulation and real-world experiments verify the effectiveness of the whole system proposed in this paper.

A 3-D Cloud Detection Method for FY-4A GIIRS and Its Application in Operational Numerical Weather Prediction System

A 3-D Cloud Detection Method for FY-4A GIIRS and Its Application in Operational Numerical Weather Prediction System

Determinants of Climate Change Adaptation Strategies in Bundelkhand Region, India

The study attempted examining climate change adaptation strategies in the most backward regions of India, i.e., the Bundelkhand region. Data were collected from 200 farmers during April and May of 2017, and a binary logistic regression model was used to analyze the data. The farmers were most likely to improve irrigation facilities if they have access to institutional credit, information about the weather, and have taken crop insurance. Also, farmers perceived that an increase in temperature, a decline in rainfall, and a decline in water tables motivate and influence farmers to change their cropping pattern from wheat (a water-intensive crop) to chickpea (a less water-intensive crop) to cope with the changing climate. Hence, the present study recommends that since water table is declining continuously and creating water crisis even in the rainy season for agriculture and domestic consumption, community participation in the conservation of water-bodies is vital. Further, the adoption of crop varieties that require less water, are drought tolerant and mature early will help increase farm productivity and reduce cultivation costs. Lastly, an advance regional weather forecasting system capable of providing farmers with accurate information that will enable them to change cropping patterns and adjust farming practices are required.

Toward a Deep-Learning-Network-Based Convective Weather Initiation Algorithm From the Joint Observations of Fengyun-4A Geostationary Satellite and Radar for 0–1h Nowcasting

Nowcasting of convective weather is a challenging and significant task in operational weather forecasting system. In this study, a new convolution recurrent neural network (ConvRNN)-based regression network for convective weather prediction is proposed, which is named as the Convective Weather Nowcasting Net (CWNNet). The CWNNet adopts the joint observations of Fengyun-4A geostationary satellite and the ground-based Doppler weather radar data of the last 0–1-hour as the inputs of the model to predict the radar reflectivity factor maps of next 0-1 hour. The statistical validating results clearly demonstrate that the mean values of the probability of detection, false alarm ratio, threat score, root mean square error and mean absolute error evaluating the performance of CWNNet for 1-h nowcasting reach 0.87, 0.137, 0.71, 3.365 dBZ and 1.038 dBZ, respectively. Due to that the geostationary meteorological satellite is capable of capturing the features of convective initiation, the CWNNet shows a good performance in convective initiation nowcasting. Besides, several case studies also further indicate that the CWNNet can predict convective initiation more than 30 minutes in advance by monitoring the convective clouds. The CWNNet based on the joint satellite and radar data shows a better nowcasting performance than that only employing single data source. Thus, it can effectively produce more reliable nowcasting for convective weather events.

Comparison of Spring Forecasting Ability between GY-WRF and GRAPES in Danghe South Mountain

The Plateau Numerical Weather Forecast System (GY-WRF) was evaluated for its ability to forecast spring weather in Danghe South Mountain. The forecasted temperature, relative humidity, and wind were compared with GRAPES using data from four automatic weather stations in March and April 2022. The results showed that GY-WRF had the smallest mean absolute error (MAE) for each factor at station 4, and a small MAE for temperature, relative humidity, and wind speed at station 2, but a slightly larger MAE for wind direction. GRAPES had a small MAE for wind direction, and a similar trend in wind speed MAE to GY-WRF. GY-WRF was better at predicting the diurnal variation trend of temperature, while GRAPES had a better effect on forecasting the minimum temperature. Both GY-WRF and GRAPES were capable of pre-dicting the trend of wind direction conversion at each station, with GY-WRF being better at predicting the diurnal variation trend of wind speed.

Parameterisation of the GNSS troposphere tomography domain with optimisation of the nodes’ distribution

Water vapour is a highly variable constituent of the troposphere; thus, its high-resolution measurements are of great importance to weather prediction systems. The Global Navigation Satellite Systems (GNSS) are operationally used in the estimation of the tropospheric state and assimilation of the results into the weather models. One of the GNSS techniques of troposphere sensing is tomography which provides 3-D fields of wet refractivity. The tomographic results have been successfully assimilated into the numerical weather models, showing the great potential of this technique. The GNSS tomography can be based on two different approaches to the parameterisation of the model’s domain, i.e. block (voxel-based) or grid (node-based) approach. Regardless of the parameterisation approach, the tomographic domain should be discretised, which is usually performed in a regular manner, with a grid resolution depending on the mean distance between the GNSS receivers. In this work, we propose a new parameterisation approach based on the optimisation of the tomographic nodes’ location, taking into account the non-uniform distribution of the GNSS information in the troposphere. The experiment was performed using a dense network of 16 low-cost multi-GNSS receivers located in Wrocław and its suburbs, with a mean distance of 3 km. Cross-validation of four different parameterisation approaches is presented. The validation is performed based on the Weather Research and Forecasting model as well as radiosonde observations. The new approach improves the results of wet refractivity estimation by 0.5–2 ppm in terms of RMSE, especially for altitudes of 0.5–2.0 km.

Outdoor heat stress assessment using an integrated multi-scale numerical weather prediction system: A case study of a heatwave in Montreal

More frequent and severe extreme weather events such as heatwaves are among the most serious challenges to society in coping with the changing climate. To evaluate the impacts of the heatwave on large-scale urban areas, a multi-scale weather forecasting system is designed by integrating different resolutions of the Canadian urbanized version of the Global Environmental Multiscale (GEM) Numerical Weather Prediction (NWP) model, cascading from 10 km to 2.5 km, and 250 m. The multi-scale model is implemented in Montreal, Canada, for modeling the 2018 heatwave. Simulation results are well-validated against measurement data, including Moderate Resolution Imaging Spectroradiometer (MODIS) satellite imagery and ten weather stations in the city. The Universal Thermal Climate Index (UTCI) map was calculated to identify vulnerable regions in the city against the heatwave. Land-use types in hotspots and coldspots are analyzed to find dominant factors in the formation of hot and cold areas. It is found that natural landscapes such as vegetation, trees, and water bodies are the dominant features of most coldspots. On the other hand, roads, parking lots, less tree covers, and industrial activities are the common land use features in the hotspots. A weak correlation is found between heat-related death locations and the outdoor UTCI map, implying that the assessment of an outdoor heatwave may not address overheated buildings and communities. This paper shows the importance of built environments – their properties and occupants' socio-demographic factors in the study of heat-related mortalities in cities.

Forecast Sensitivity-based Observation Impact (FSOI) in an analysis–forecast system of the California Current Circulation

Forecast Sensitivity-based Observation Impacts (FSOI) in an analysis–forecast​ system of the California Current System (CCS) are quantified using an adjoint-based approach. The analysis–forecast system is based on the Regional Ocean Modeling System (ROMS) and a 4-dimensional variational (4D-Var) data assimilation approach. FSOI was applied to four different metrics of forecast skill that target important features of the CCS circulation along the central California coast. A particular focus of the FSOI analysis is the impact of assimilation of measurements of the radial component of surface currents from a network of high frequency (HF) radars since this is a new data stream in the near-real-time system considered here. On average, ∼50–60% of all observations assimilated into the model yielded improvements in the forecast skill. Conversely, the remaining ∼40–50% of data degrade the forecasts, in line with similar findings in numerical weather prediction systems. Much of the improvement in forecast skill arises from remotely sensed observations, including HF radar data; on average only ∼50% of in situ measurements contribute to a reduction in forecast error. This is partly due to the large volume of remote sensing observations compared to in situ observations. However, in situ observations are an order of magnitude more impactful than remotely sensed data when viewed in terms of the average impact per observation.

Testing Variational Bias Correction of Satellite Radiance Data in the ACCESS-C: Australian Convective-Scale NWP System

Radiance observations are typically affected by biases that come mainly from instrument error (scanning or calibration) and inaccuracies of the radiative transfer model. These biases need to be removed for successful assimilation, so a bias correction scheme is crucial in the Numerical Weather Prediction (NWP) system. Today, most NWP centres, including the Bureau of Meteorology (hereafter, "the Bureau"), correct the biases through variational bias correction (VarBC) schemes, which were originally developed for global models. However, there are difficulties in estimating the biases in a limited-area model (LAM) domain. As a result, the Bureau's regional NWP system, ACCESS-C (Australian Community Climate and Earth System Simulator-City), uses variational bias coefficients obtained directly from its global NWP system ACCESS-G (Global). This study investigates independent radiance bias correction in the data assimilation system for ACCESS-C. We assessed the impact of using independent bias correction for the LAM compared with the operational bias coefficients derived in ACCESS-G between February and April 2020. The results from our experiment show no significant difference between the control and test, suggesting a neutral impact on the forecast. Our findings point out that the VarBC-LAM strategy should be further explored with different settings of predictors and adaptivity for a more extended period and over additional domains.

Optimization of wind farm portfolios for minimizing overall power fluctuations at selective frequencies – a case study of the Faroe Islands

Abstract. Hourly modeled wind turbine power output time series – modeled based on outputs from the mesoscale numerical weather prediction system Weather Research and Forecasting Model (WRF) – are used to examine the spatial smoothing of various wind farm portfolios located on a complex isolated island group with a surface area of 1400 km2. Power spectral densities (PSDs), hourly step-change functions, and duration curves are generated, and the 5th and 95th percentiles and the standard deviations of the hourly step-change functions are calculated. The spatial smoothing is identified from smaller high-frequency PSD amplitudes, lower hourly fluctuations, and more flat duration curves per installed wind power capacity, compared with single wind turbine outputs. A discussion on the limitation of the spatial smoothing for the region is included, where a smoothing effect is observed for periods of up to 1–2 d, although it is most evident at higher frequencies. By maximizing the smoothing effect, optimal wind farm portfolios are presented with the intention of minimizing overall wind power fluctuations. The focus is mainly on the smoothing effect on the 1–3 h timescale, during which the coherency between wind farm power outputs is expected to be dependent on how the regional weather travels between local sites, thereby making optimizations of wind farm portfolios relevant – in contrast to a focus on either lower or higher frequencies on the scale of days or minutes, respectively, during which wind farm power output time series are expected to be either close to fully coherent due to the same weather conditions covering a small region or not coherent as the turbulences in separate wind farm locations are expected to be uncorrelated. Results show that an optimization of the wind farm capacities at 14 pre-defined wind farm site locations has a minimal improvement on the hourly fluctuations compared with a portfolio with equally weighted wind farm capacities. However, choosing optimized combinations of individual wind farm site locations decreases the 1–3 h fluctuations considerably. For example, selecting a portfolio with four wind farms (out of the 14 pre-defined wind farm site locations) results in 15 % lower 5th and 95th percentiles of the hourly step-change function when choosing optimal wind farm combinations compared with choosing the worst wind farm combinations. For an optimized wind farm portfolio of seven wind farms, this number is 13 %. Optimized wind farm portfolios consist of distant wind farms, while the worst portfolios consist of clustered wind farms.

An Optimal Interpolation–Based Snow Data Assimilation for NOAA’s Unified Forecast System (UFS)

Abstract Within the National Weather Service’s Unified Forecast System (UFS), snow depth and snow cover observations are assimilated once daily using a rule-based method designed to correct for gross errors. While this approach improved the forecasts over its predecessors, it is now quite outdated and is likely to result in suboptimal analysis. We have then implemented and evaluated a snow data assimilation using the 2D optimal interpolation (OI) method, which accounts for model and observation errors and their spatial correlations as a function of distances between the observations and model grid cells. The performance of the OI was evaluated by assimilating daily snow depth observations from the Global Historical Climatology Network (GHCN) and the Interactive Multisensor Snow and Ice Mapping System (IMS) snow cover data into the UFS, from October 2019 to March 2020. Compared to the control analysis, which is very similar to the method currently in operational use, the OI improves the forecast snow depth and snow cover. For instance, the unbiased snow depth root-mean-squared error (ubRMSE) was reduced by 45 mm and the snow cover hit rate increased by 4%. This leads to modest improvements to globally averaged near-surface temperature (an average reduction of 0.23 K in temperature bias), with significant local improvements in some regions (much of Asia, the central United States). The reduction in near-surface temperature error was primarily caused by improved snow cover fraction from the data assimilation. Based on these results, the OI DA is currently being transitioned into operational use for the UFS. Significance Statement Weather and climate forecasting systems rely on accurate modeling of the evolution of atmospheric, oceanic, and land processes. In addition, model forecasts are substantially improved by continuous incorporation of observations to models, through a process called data assimilation. In this work, we upgraded the snow data assimilation used in the U.S. National Weather Service (NWS) global weather prediction system. Compared to the method currently in operational use, the new snow data assimilation improves both the forecasted snow quantity and near-surface air temperatures over snowy regions. Based on the positive results obtained in the experiments presented here, the new snow data assimilation method is being implemented in the NWS operational forecast system.

Solar X-ray and EUV imager on board the FY-3E satellite

The solar X-ray and Extreme Ultraviolet Imager (X-EUVI), developed by the Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences (CIOMP), is the first space-based solar X-ray and Extreme ultraviolet (EUV) imager of China loaded on the Fengyun-3E (FY-3E) satellite supported by the China Meteorological Administration (CMA) for solar observation. Since started work on July 11, 2021, X-EUVI has obtained many solar images. The instrument employs an innovative dual-band design to monitor a much larger temperature range on the Sun, which covers 0.6–8.0 nm in the X-ray region with six channels and 19.5 nm in the EUV region. X-EUVI has a field of view of 42′, an angular resolution of 2.5″ per pixel in the EUV band and an angular resolution of 4.1″ per pixel in the X-ray band. The instrument also includes an X-ray and EUV irradiance sensor (X-EUVS) with the same bands as its imaging optics, which measures the solar irradiance and regularly calibrates the solar images. The radiometric calibration of X-EUVS on the ground has been completed, with a calibration accuracy of 12%. X-EUVI is loaded on the FY-3E satellite and rotates relative to the Sun at a uniform rate. Flat-field calibration is conducted by utilizing successive rotation solar images. The agreement between preliminarily processed X-EUVI images and SDO/AIA and Hinode/XRT images indicates that X-EUVI and the data processing algorithm operate properly and that the data from X-EUVI can be applied to the space weather forecast system of CMA and scientific investigations on solar activity.

Impact of urban parameterization and integration of WUDAPT on the severe convection

Amplified rates of urban convective systems pose a severe peril to the life and property of the inhabitants over urban regions, requiring a reliable urban weather forecasting system. However, the city scale's accurate rainfall forecast has constantly been a challenge, as they are significantly affected by land use/ land cover changes (LULCC). Therefore, an attempt has been made to improve the forecast of the severe convective event by employing the comprehensive urban LULC map using Local Climate Zone (LCZ) classification from the World Urban Database and Access Portal Tools (WUDAPT) over the tropical city of Bhubaneswar in the eastern coast of India. These LCZs denote specific land cover classes based on urban morphology characteristics. It can be used in the Advanced Research version of the Weather Research and Forecasting (ARW) model, which also encapsulates the Building Effect Parameterization (BEP) scheme. The BEP scheme considers the buildings' 3D structure and allows complex land–atmosphere interaction for an urban area. The temple city Bhubaneswar, the capital of eastern state Odisha, possesses significant rapid urbanization during the recent decade. The LCZs are generated at 500 m grids using supervised classification and are ingested into the ARW model. Two different LULC dataset, i.e., Moderate Resolution Imaging Spectroradiometer (MODIS) and WUDAPT derived LCZs and initial, and boundary conditions from NCEP GFS 6-h interval are used for two pre-monsoon severe convective events of the year 2016. The results from WUDAPT based LCZ have shown an improvement in spatial variability and reduction in overall BIAS over MODIS LULC experiments. The WUDAPT based LCZ map enhances high-resolution forecast from ARW by incorporating the details of building height, terrain roughness, and urban fraction.

Integrating the Spatial Autocorrelation Structure of Irregularly Distributed Wind Measurement Stations to Model Wind Patterns Using R

Wind behavior in Sri Lanka, mainly exposed to two major monsoon seasons, is complex to model spatially. The acute changes in the topography further make it difficult to map the wind patterns. Irregularly distributed wind measurements are the main source of data available to interpret and model this dynamic geographic phenomenon. Sri Lanka is in need of an accurate and timely wind map in order to support wind energy resource planning and to enable weather prediction systems to be more accurate. Strong and continuous winds in the southern part of the island encourage wind harvesting for energy. Wind energy is arguably the most affordable per megawatt-hour renewable energy source, and it is growing nearly as quickly as conventional energy generation techniques in the country. The objective of this paper is twofold. First is to use the spatial autocorrelation structure of 23 data points using diurnal data measured every 10 minutes on an irregularly distributed grid to construct continuous wind speed fields for the 4 monsoon seasons. This has been done for the year 2015 as a reference year, where a complete set of data was acquired. Secondly, it was intended to provide the reader how R opensource with advanced statistical modelling capabilities and graphics could be utilized conveniently for vague geographic phenomena analysis. The spatial autocorrelation of the wind speed and the direction was determined in the beginning by using the semi-variograms. Candidate semi-variogram models, especially spherical and exponential models, were tested to be more suitable for the observations. Ordinary Kriging, which uses the samples in the local neighborhood to avoid the mean estimation, was used to perform the spatial interpolations Kriging. For each month of the year 2015, four weekly wind maps were generated, both for the direction and speed. These 48 wind maps were further analyzed for the wind patterns. The changes in the morning and evening wind patterns, and the significant wind pattern changes along the coastal areas and the central highlands are extracted and presented in the study. The generated maps could be useful in identifying potential areas for wind farming. By feeding the data in a time series to the presented model, it would be possible to deduce wind pattern changes in Sri Lanka on a yearly basis. As a whole, the paper presents a reasonable model that could be used to model a diverse set of vague phenomena with an understanding of the uncertainties involved.

Weather Prediction for Singapore—Progress, Challenges, and Opportunities

Singapore is a tiny city-state located in maritime Southeast Asia. Weather systems such as localized thunderstorms, squalls, and monsoon surges bring extreme rainfall to Singapore, influencing the day-to-day conduct of stakeholders in many sectors. Numerical weather prediction models can provide forecast guidance, but existing global models struggle to capture the development and evolution of the small-scale and transient weather systems impacting the region. To address this, Singapore has collaborated with international partners and developed regional numerical weather prediction systems. Steady progress has been made, bringing added value to stakeholders. In recent years, complex earth system and ultra high-resolution urban models have also been developed to meet increasingly diverse stakeholder needs. However, further advancement of weather prediction for Singapore is often hindered by existing challenges, such as the lack of data, limited understanding of underlying processes, and geographical complexities. These may be viewed as opportunities, but are not trivial to address. There are also other opportunities that have remained relatively unexplored over Singapore and the region, such as the integration of earth system models, uncertainty estimation and machine learning methods. These are perhaps key research directions that Singapore should embark on to continue ensuring value for stakeholders.

Improving Arctic Weather and Seasonal Climate Prediction: Recommendations for Future Forecast Systems Evolution from the European Project APPLICATE

Abstract The Arctic environment is changing, increasing the vulnerability of local communities and ecosystems, and impacting its socio-economic landscape. In this context, weather and climate prediction systems can be powerful tools to support strategic planning and decision-making at different time horizons. This article presents several success stories from the H2020 project APPLICATE on how to advance Arctic weather and seasonal climate prediction, synthesizing the key lessons learned throughout the project and providing recommendations for future model and forecast system development.

Assessment of Retrieved GMI Emissivity over Land, Snow, and Sea Ice in the GEOS System

Abstract Directly assimilating microwave radiances over land, snow, and sea ice remains a significant challenge for data assimilation systems. These data assimilation systems are critical to the success of global numerical weather prediction systems including the Global Earth Observing System–Atmospheric Data Assimilation System (GEOS-ADAS). Extending more surface sensitive microwave channels over land, snow, and ice could provide a needed source of data for numerical weather prediction particularly in the planetary boundary layer (PBL). Unfortunately, the accuracy of emissivity models currently available within the GEOS-ADAS along with other data assimilation systems are insufficient to simulate and assimilate radiances. Recently, Munchak et al. published a 5-yr climatological database for retrieved microwave emissivity from the Global Precipitation Measurement (GPM) Microwave Imager (GMI) aboard the GPM mission. In this work the database is utilized by modifying the GEOS-ADAS to use this emissivity database in place of the default emissivity value available in the Community Radiative Transfer Model (CRTM), which is the fast radiative transfer model used by the GEOS-ADAS. As a first step, the GEOS-ADAS is run in a so-called stand-alone mode to simulate radiances from GMI using the default CRTM emissivity, and replacing the default CRTM emissivity models with values from Munchak et al. The simulated GMI observations using Munchak et al. agree more closely with observations from GMI. These results are presented along with a discussion of the implication for GMI observations within the GEOS-ADAS.

Deep Learning Provides Substantial Improvements to County‐Level Fire Weather Forecasting Over the Western United States

AbstractThe recent wildfires in the western United States during 2018 and 2020 caused record‐breaking fire damage and casualties. Despite remarkable advances in fire modeling and weather forecasting, it remains challenging to anticipate catastrophic wildfire events and associated damage. One key missing component is a fire weather prediction system with sufficiently long lead time capable of providing useful regional details. Here, we develop a hybrid prediction model of wildfire danger called CFS with super resolution (CFS‐SR) as a proof of concept to fill that void. The CFS‐SR model is constructed by integrating the Climate Forecast System version 2 with a deep learning (DL) technique from Single Image Super Resolution, a method widely used in enhancing image resolution. We show that for the 2018–2019 fire season, the CFS‐SR model significantly improves accuracy in forecasting fire weather at lead times of up to 7 days with an enhanced spatial resolution up to 4 km. This level of high resolution provides county‐level fire weather forecast, making it more practical for allocating resources to mitigate wildfire danger. Our study demonstrates that a proper combination of ensemble climate predictions with DL techniques can boost predictability at finer spatial scales, increasing the utility of fire weather forecasts for practical applications.

A satellite lightning observation operator for storm-scale numerical weather prediction

Abstract. This study aims at simulating satellite-measured lightning observations with numerical weather prediction (NWP) system variables. A total of eight parameters, calculated with the AROME-France NWP system variables, were selected from a literature review to be used as proxies for satellite lightning observations. Two different proxy types emerged from this literature review: microphysical and dynamical proxies. Here, we investigate which ones are best related to satellite lightning and calibrate an empirical relationship between the best parameters and lightning data. To obtain those relationships, we fit machine learning regression models to our data. In this study, pseudo flash extent accumulation (FEA) observations are used because no actual geostationary lightning observations are available yet over France, and non-geostationary satellite lightning data represent a sample that is too small for our study. The performances of each proxy and machine learning regression model are evaluated by computing fractions skill scores (FSSs) with respect to observed FEA and proxy-based FEA. The present study suggests that microphysical proxies are more suited than the dynamical ones to model satellite lightning observations with the AROME-France NWP system. The performances of multivariate regression models are also evaluated by combining several proxies after a feature selection based on a principal component analysis and a proxy correlation study, but no proxy combination yielded better results than microphysical proxies alone. Finally, different accumulation periods of the FEA had little influence, i.e. similar FSS, on the regression model's ability to reproduce the observed FEA. In future studies, the microphysical-based relationship will be used as an observation operator to perform satellite lightning data assimilation in storm-scale NWP systems and applied to NWP forecasts to simulate satellite lightning data.

Development of a Low-Cost Portable Cluster for Numerical Weather Prediction

Numerical weather prediction (NWP) systems are crucial tools in atmospheric science education and weather forecasting, and high-performance computing (HPC) is essential for achieving such science. The goals of NWP systems are to simulate different scales of weather systems for educational purposes or to provide future weather information for operational purposes. Supercomputers have traditionally been used for NWP systems; however, supercomputers are expensive, have high power consumption, and are difficult to maintain and operate. In this study, the Raspberry Pi platform was used to develop an easily maintained high-performance NWP system with low cost and power consumption—the Improved Raspberry Pi WRF (IRPW). With 316 cores, the IRPW had a power consumption of 466 W and a performance of 200 Gflops at full load. IRPW successfully simulated a 48-h forecast with a resolution of 1 km and a domain of 32,000 km2 in 1.6 h. Thus, IRPW could be used in atmospheric science education or for local weather forecasting applications. Moreover, due to its small volume and low power consumption, it could be mounted to a portable weather observation system.