HUAWEI CLOUD MVP Zhou Zheng: Weather forecasting is a technical activity. Big data, supercomputing, and AI are all indispensable

Abstract: In such an industry related to people's livelihood, the potential of artificial intelligence, big data, and supercomputing technologies is unlimited. Huawei Cloud MVP Zhou Zheng is one of the technology practitioners, and he is leading the team Bring a gentle but powerful spring breeze to the domestic meteorological industry.

This article is shared from the HUAWEI CLOUD community " [One hundred million yards first, cloud gathering Jinling] HUAWEI CLOUD MVP Zhou Zheng: Weather forecasting is a technical activity, big data, supercomputing, AI, all indispensable ", the original author: We are all clouds expert.

The weather is related to the production and life of all people on the earth. In ancient times, people watched the sky to eat. Now, we watch the weather forecast to make preparations for work and life in advance, such as:

  • Forecast typhoons and floods, give early warnings, and make emergency plans;
  • Combine pollution emission data and air flow to forecast air quality;
  • Marine aquaculture revises the cultivation strategy in time according to climate change;
  • ……

In addition, fields such as national defense, transportation, insurance, and smart cities all rely on weather data.

In such an industry related to people’s livelihood, the potential of artificial intelligence, big data, and supercomputing technologies is unlimited. HUAWEI CLOUD MVP Zhou Zheng is one of the technology practitioners. He is leading the team for the domestic meteorological industry. Brings a gentle but powerful spring breeze.

Weather market dividend: 20 times in 5 years

2015 was a turning point in the field of meteorology. Since then, the era of monopoly has come to an end, and weather services have gradually opened up to commercial companies.

According to statistics, the meteorological market will have a 20-fold growth space in the next few years, and it will reach a market size of 300 billion yuan in 2025.

Comparing the data that has doubled in the previous 7 years, the dividends brought by the market opening are obvious, and a large wave of emerging startups has poured in. In 2018, Zhou Zheng also resolutely stepped into this wave.

Before founding Jiufang Technology, Zhou Zheng had already accumulated a certain amount of technology in the field of meteorology. As a Ph.D. jointly trained by Wuhan University and Ohio State University, he is a senior engineer in the Department of Earth System Science of Tsinghua University, and also served as a department director and senior engineer of the National Supercomputing (Wuxi) Center.

Zhou Zheng is good at system structure design, core algorithm design, etc. He has participated in the research and development of a number of major national scientific research projects, published papers, software copyrights, and various national awards.

It was his research work at the National Supercomputing Wuxi Center that led to his meteorological big data entrepreneurship. Zhou Zheng led nearly 20 colleagues of the former department "Earth System Numerical Simulation Department" to explore the road to the commercialization of supercomputer industrialization and commercial meteorological services.

The combination of technical talents has allowed this startup company, which has emerged from the national scientific research team, to successfully achieve corporate profits for three consecutive years.

It is not easy to make the forecast accurately

As the founder & CEO of the company, Zhou Zheng leads the core R&D personnel. Based on the "Shenwei·Light of Taihu Lake" supercomputer system, combined with supercomputing and artificial intelligence technologies, he has developed an independent intelligent grid forecasting platform made in China. Technical breakthroughs have been made in the field of forecasting.

Taking marine numerical forecasting as an example, Zhou Zheng introduced that in terms of marine aquaculture, if the seawater flows too fast, the kelp will break off. If the temperature is too high, the sea cucumbers will directly turn into water, and if the temperature is suddenly too low, it will be difficult for the crabs to survive.

Therefore, they used the "three-dimensional ocean element intelligent forecasting system" to analyze the conditions of the seafood farming area by scientific means. It can not only forecast the information of ocean temperature, salinity, and ocean currents, but also forecast the nutrients, marine biochemistry and other elements. .

"What we do is to report the weather forecast." It seems that this sentence is simple, but it is also the most challenging. Zhou Zheng said that their position is the producer of data, not the porter of data.

Therefore, Zhou Zheng and his team coupled multiple weather forecast models such as the atmosphere, ocean, waves, sea ice, land surface, etc., gathered multiple types of data such as satellites, radars, and stations, and combined numerical forecasting, artificial intelligence, and supercomputing technologies. Form a complete industrial capability and technical barriers from weather forecasting to application of weather information.

And unlike most meteorological companies, Zhou Zhenghe's team has self-developed weather forecast numerical models, as well as the ability to directly process source data and applications. This is also their killer.

They combined data with AI algorithms, and used powerful computing power to produce weather forecast data with higher accuracy and precision. It is understood that they can further improve the time and space accuracy of atmospheric and ocean forecasts to minute-level updates and updates. 100-meter grid level. For example, in air pollution monitoring, Zhou Zheng and his team have achieved an accuracy of nearly 90%; in the El Niño phenomenon, it can be predicted 6 months in advance and so on.

The cooperation between Zhou Zheng and HUAWEI CLOUD has also brought their weather forecasting capabilities to a higher level. For example, through the HUAWEI CLOUD high-performance computing HPC cluster to create an air pollution early warning system. Among them, the HPC management node delivers computing tasks through the SGE resource scheduling system. The computing nodes built on the elastic cloud server in the cluster undertake the main computing tasks. At the same time, all processing results are stored in the elastic file service and the storage results are shared.

It is understood that the HUAWEI CLOUD HPC solution is optimized for HPC Workload to realize the on-demand rental of HPC resources, which can help users reduce costs, improve the efficiency and accuracy of weather forecasts, and help the meteorological business to quickly transform into intelligence, cloudification, and digitalization. .

at last

After years of cultivating the meteorological industry, Zhou Zhenghe's team has also obtained a series of honors and qualifications through technical conquests over the years. It has won numerous awards in various national, provincial and municipal competitions, and has accumulated 26 intellectual property rights such as patents and software works. They have also completed more than 10 national-level scientific research projects and published a large number of scientific research results. And joined the China Meteorological Service Association, the National Marine Information Industry Development Alliance, the China Ports Association, and Huawei's solution partner strategic cooperation plan.

In the future, Zhou Zheng will also provide refined weather data and comprehensive solutions for G-end and B-end customers, including cooperation with Huawei Cloud, and provide industry customers with more refined and targeted solutions based on self-developed numerical models and artificial intelligence forecasting technology. The products and services of the company cover the fields of intelligent weather forecasting, atmospheric environmental protection, shipping navigation, finance and insurance, etc.

Click to follow, and learn about Huawei Cloud's fresh technology for the first time~

阅读 857



1.3k 声望
1.7k 粉丝
0 条评论


1.3k 声望
1.7k 粉丝