For roughly 70 years, Toshiba has been in the weather radar systems business. We have established a track record and gained expertise by supplying systems to the national government such as the Japan Meteorological Agency; local governments; and infrastructure providers such as rail and power companies. These systems are customized to their individual usage needs. Accurately assessing what is currently happening in the skies, predicting what will happen in the future, and supplying this information so that it can be effectively utilized helps prevent and mitigate weather-related disasters, which are on the rise worldwide. It also helps reduce the anxiety people feel about weather changes and assists people in planning their activities. In this issue, we will look at Toshiba’s weather data services, which use our experience, know-how, and technologies to analyze raw data from weather radars to predict rain, snow, hail, and sudden wind that will soon occur. We will also present some examples of these services in action.

In recent years, the world has seen a high number of extreme weather phenomena. In Japan, heavy rain, such as localized torrential rain (sudden downpours), causes damage, but people’s lives are threatened by more than just rain. Snow, hail, and wind also can put people in danger and cause lots of losses. Both the frequency and the economic impact of the weather disasters occurring around the globe have risen five-fold over the past 40 years^*1, and the frequency of heavy rain within Japan has roughly doubled compared to four decade ago^*2.

*1 Source: “WMO Atlas of Mortality and Economic Losses from Weather, Climate and Water Extremes (1970-2019) (WMO-No. 1267)”
*2 Extreme weather phenomenon information on the Japan Meteorological Agency website (https://www.data.jma.go.jp/cpdinfo/extreme/extreme_p.html)

Through delivering weather radar systems which enable accurate observation of rainfall, sudden winds, lightning strikes, and the like, Toshiba has assisted operators who manage social infrastructure such as rivers and dams, railroads, and electrical grids and who must pay close attention to weather changes. These systems are made up of shared hardware, such as parabolic antennas, combined with software for analyzing raw radar data from antennas, customized to the uses of individual clients.

We are now offering the software functions of these weather radar systems as data services. The value generated by our advanced weather radar analysis technologies, developed based on our extensive experience, can now be leveraged by an even wider range of companies and organizations. This is a new initiative by Toshiba, based on our desire to mitigate danger, not only in the form of preventing and mitigating disasters, but also by reducing the dangers presented by sudden weather changes in a variety of areas, such as public transportation, farming, and event operation. It was made possible by fact that the Ministry of Land, Infrastructure, Transport and Tourism opened up weather radar data to the private sector so that raw data could be used commercially*.

* Announcement on the Ministry of Land, Infrastructure, Transport and Tourism website (https://www.mlit.go.jp/report/press/mizukokudo03_hh_001111.html)

Through the use of Toshiba technologies to analyze the raw data observed by weather radar, we are able, for example, to accurately predict localized torrential rain, from the formation to the development of rainclouds. Similar to the emergency earthquake alert system, this system can urge people to avoid danger and can help minimize damage by assessing current aerial weather phenomena in real time and providing notice before those dangers become manifest.

Currently, the Ministry of Land, Infrastructure, Transport and Tourism operates 65 weather radars  throughout Japan (as of July 2024)*. After importing the raw data from these weather radars into the cloud and processing it, Toshiba predicts when rain will fall, analyzes particle types of precipitation such as rain, snow, graupel, and hail, and searches for wind gusts by means of detecting and tracking wind vortexes. The technologies used to process data in the cloud leverage Toshiba’s unique know-how. For example, the technologies used to convert raw data from weather radars, which is in polar coordinate form, into mesh data with rectangular coordinates, and the technologies used to cleanse the raw data, eliminating noise, are the keys to performing higher quality analysis and producing more accurate predictions (Fig. 1).

* Radar precipitation gauge information on the Ministry of Land, Infrastructure, Transport and Tourism website (https://www.mlit.go.jp/tec/tec_fr_000040.html)

What kind of actual benefits can be expected from using weather data services? The easiest one to conceptualize is reducing the damage caused by rain.

As mentioned earlier, Japan has been seeing a rise in heavy rains. Sudden heavy rains can cause underpasses to flood and rivers to overflow their banks. These disasters can threaten people’s very lives. Toshiba technologies can accurately forecast changes in rainclouds over the following thirty minutes. For example, they can predict if rain will fall (and, if so, whether it will be heavy rain) or if the skies will clear. Knowing about heavy rains thirty minutes in advance enables people and organizations to make operating decisions such as closing down roads or shutting down operations. It makes it possible to issue instructions to take shelter indoors. This forecast information can help people avoid danger (Fig. 2).

There are also technologies that predict hailfall dozens of minutes ahead through analyzing the particle types of precipitation such as graupel and hail in growing rainclouds. Hail often starts falling suddenly, so detecting the signs of hail in advance has the potential to help reduce the amount of damage to assets such as homes and cars. Hail can be dangerous, growing to diameters of 5 cm or more. Large pieces of hail can cause extensive damage, breaking home windows and damaging car bodies. With advance notice of this danger, people could take steps to reduce hail damage, such as by closing their storm shutters and moving their cars to safe areas or protecting them with hail-proofing covers.

The service is also effective at combatting snow damage. Heavy snow can paralyze public transportation such as roads and railroads. It can have a significant impact on people’s ability to commute or go out. There are many different types of snow, from powdery snow to very wet snow. The ease with which snow can pile up depends on the kind of snow. Toshiba’s technologies can accurately assess snow types from weather radar data. These technologies can be used to make decisions regarding public transportation operation, whether to use snow-melting equipment, and the like. They can also detect danger by assessing how much snow will build up on power wires.

Furthermore, the system can detect wind gusts, which are difficult to forecast using anemometer measurements. These predictions could be used, for example, to prevent major accidents by warning people working at high elevations, exposed to the dangers of wind gusts, or by warning public transportation operators.

Let’s briefly go over these Toshiba technologies, which have the potential to protect people from danger in various situations.

First, let’s look at our technologies for accurately predicting rainfall up to 30 minutes in advance. These involve not only technologies for primary processing of raw data from weather radars, but also technologies for synthesizing this data from multiple weather radars to generate 3D data and technologies for analyzing this synthesized 3D data to determine the total amount of rain in the clouds. The entire territory of Japan is divided into a mesh of segments measuring 250 m on a side, and a trend of the vertically integrated liquid (VIL) water content of each segment, which is the total amount of rain in the clouds, is analyzed to forecast rainfall. For example, if the VIL is increasing, this indicates that a large mass of water in the skies, which can turn into torrential rain, is growing up. Once a large mass of water forms in the skies, it will inevitably fall to the ground. Through the use of VIL, which is focused on the skies instead of the ground, the system can accurately forecast rainfall 30 minutes into the future.

The next technology is the one used to determine what kinds of particles are present in rainclouds. The horizontally polarized and vertically polarized radio waves emitted by the weather radars are reflected by rainclouds and back to the weather radars. Toshiba uses information such as the ratios of intensities of these radio waves and their mutual correlations to determine if the particles that will fall to the ground will be rain, snow, or hail, all in real time. The distribution of the classified particles in cross section of raincloud helps us to understand the structure of rainclouds. indicates the type of rain that will fall. For example, the stratiform clouds, which are the clouds responsible for the drizzle that is often seen during the rainy season, have a typical striated structure. At higher altitudes, temperatures are low, and snow and ice crystals are present. The lower the altitude, the warmer the air is, causing the snow to melt and become sleet. Near the ground, it becomes rain. In the case of the cumulonimbus clouds responsible for hail, the distribution shows that hail present at high altitudes falls to ground level (Fig. 3).

Next, let’s look at the technologies we have developed for detecting wind gusts. Wind gusts are detected by determining the direction of wind in rain clouds by observing the Doppler effects of the radio waves emitted by weather radars. Due to the Doppler effect, when the wind is blowing toward the measurer (the weather radar), the wavelengths of the received radio signals will be higher. When the wind is blowing away, the frequencies will be lower. This is the similar reason that the sound of the siren on an approaching ambulance is higher when the ambulance is approaching (the frequencies become higher and the wavelengths shorter) and then lower when the ambulance is growing more distant (the frequencies become lower and the wavelengths longer). When the wind is circling, such as in a tornado, the direction of the wind will be opposite on each side of the center of the vortex. Detecting this kind of vortex pattern makes it possible to track the tornado.

Toshiba is also carrying out verification projects that use these weather data services. The first is a project to verify the effectiveness of rainfall and flooding predictions used in localized torrential rain countermeasures. We worked with a telecommunications company and a construction consultant to create an integrated service that links to each company’s systems. We then verified the effectiveness of the flooding prediction service offered to local governments.

Toshiba analyzes raw data from weather radars to predict 30 minutes in advance where, when, and how much rain will fall, or the signs of heavy rain and its amount. Based on this predictive information, the construction consultant used their flooding predictive simulation technologies to forecast the overflowing of rivers and other inland waters and the flooding of roads and buildings. When a major disaster was predicted, the telecommunications company’s information delivery platform would provide people in high-risk areas with flood area prediction information and other flooding information in the form of a dynamic hazard map. The project evaluated the effectiveness of the service offered by these systems, working together, from the perspectives of how well they assisted with rapidly and accurately making and implementing decisions based on high-precision rainfall forecasts, and how well they assisted with mitigating flood damage.

* Check here to read our news article.

The next example is our joint verification project with Mitsui Sumitomo Insurance Company, Limited and venture company Spectee Inc. We aimed to help reduce the amount of damage caused by hail by using Toshiba’s technologies for classifying particles. The background to this initiative is the extensive localized hail damage that was caused by heavy hail in 2022. We tested whether alerts to insurance policy-holders of Mitsui Sumitomo Insurance and others of signs of hail or localized torrential rain could lead to their avoidance actions, such as moving or protecting their automobiles, and reduce the damage caused by disaster.

First, Toshiba analyzed weather radar data to gather information on particles and rainfall. Based on this information, we predicted the sign of hail and localized torrential rain and issued email alerts to project participants in areas expected to suffer some degree of rain or hail damage. When participants received these alerts, they could see hail forecasts for several dozen minutes and rain forecasts for the following 30 minutes on a map on their smartphones. The project also provided alert recipients with posts and comments from social media in areas in which hail was actually falling. This information came from Spectee, which excels at analyzing disaster information posted to social media. Based on these predictions and actual conditions, Mitsui Sumitomo Insurance verified the effectiveness of the hail alerts and the need for system operation (Fig. 4). The project met with a high level of approval, and a full-fledged service has since been launched.

* Check here to read our news article.

In Japan, the Meteorological Service Act requires the approval of the Director-General of the Japan Meteorological Agency to engage in operations involving making weather-related predictions, such as weather forecasting. Toshiba Digital Solutions received approval (Approval No. 238) in October 2023, so we are now able to issue short-term predictions regarding weather-related phenomena nationwide*.

* List of operators authorized by the Japan Meteorological Agency to engage in weather prediction operations (https://www.jma.go.jp/jma/kishou/minkan/minkan.html)

Toshiba has advanced analysis capabilities and prediction technologies for assessing what is currently happening in the skies, and what will happen in the future, over a wide area and with a high degree of area granularity. The weather data services which are made possible by these technologies contribute to disaster prevention and mitigation in society. Their use in a variety of fields, such as social infrastructure and private sector companies, helps people live more peaceful and richer lives. Weather is something that affects the activities of everyone and every company. We hope that the ability to visualize the skies, made possible by Toshiba’s weather data services, will be effectively used by numerous people, organizations, and companies, helping generate new value.

• Weather Data Service