As DeepSeek is leading the wave of change in many fields, Professor Chen Bin is using artificial intelligence (AI) to decode the mystery of dust storms in his laboratory at Guanyun Building of Lanzhou University.
From sketching the first sandstorm transport path map by hand to solving aerosol research bottlenecks with AI algorithms; from collecting samples in northern deserts during his student days under mentorship to now leading the development of an AI-driven atmospheric aerosol inversion system, over two decades, Chen Bin has gradually unveiled the complex veil of atmospheric aerosols in the Loess Plateau's deserts. Standing before his self-authored book "Artificial Intelligence Technologies in Atmospheric Science," Chen firmly believes northwest China can equally produce outstanding scientific achievements.
The reason for taking root —— is to turn regional disadvantages into scientific research advantages
In the autumn of 2003,18-year-old Chen Bin from Hubei Province boarded a northbound train carrying a woven bag, entering Lanzhou University. The scenery outside the window transformed from lush greenery to desolate landscapes – from the verdant fields of his hometown to the barren Loess Plateau of Northwest China. Little did he know that this barren land would become the "laboratory" for his future research to tackle global environmental challenges.
Years later, when asked about his decision to settle in Lanzhou, Chen Bin would always recall a remark by his mentor, Academician Huang Jianping: "The geographical disadvantages of Northwest China are precisely its strengths in scientific research. " Not only did Academician Huang contribute groundbreaking research on semi-arid climate changes in the region, but he also demonstrated through his achievements that world-class scientific accomplishments can be achieved in western China.
In 2008, Chen Bin followed his mentor to a weather observation station at Jingtai Farm, where he truly grasped what his mentor meant by "scientific research advantages." The site, built with crude sand walls, was worn and patchy; the dirt roads muddy and uneven. The only water source was a tributary of the Yellow River located two kilometers away. "We carried iron buckets back and forth every day just to fetch water—using only half a cup to wash our faces. Everyone's face was covered in grime," he recalled, mimicking the way he used to wash up. "But every grain of sand here holds precious meteorological data, making it a prime spot for studying dust transmission and aerosols."
Harsh environments give rise to groundbreaking discoveries. It was here that researchers, through aerosol trajectory inversion, confirmed that the dust affecting Japan primarily originates from Mongolia rather than western China. This finding not only resolved an international debate but also revealed the unique value of northwest dust storms in climate research: each grain of sand acts like a natural cipher, and the northwest stands as the ideal "feng shui" ground for decoding these environmental secrets.
In arid and semi-arid regions, dust aerosols now account for 50% of total atmospheric pollutants. These seemingly tiny particles interact with human-made pollutants such as industrial emissions and vehicle exhaust, collectively reshaping climate patterns from regional to global scales. Chen Bin’s research centers on this critical issue, developing a three-dimensional intelligent inversion system for atmospheric aerosols that enables precise identification of aerosols and clouds under non-clear-sky conditions. By integrating data from multiple satellites, his team has built a comprehensive, high-resolution dataset on pollutants, offering robust theoretical support for uncovering the key factors influencing dust and its transport mechanisms.
Behind these research achievements lies Chen Bin's profound understanding of the scientific value of the northwest region: "In the northwest, we are not merely enduring sandstorms—we are coming to understand them."
Fruits of Deep Roots – AI Empowers Breakthroughs in Scientific Research
In 2015, Chen Bin hit a bottleneck in his research. "Traditional studies of dust aerosols typically rely heavily on vast amounts of satellite and ground-based observational data. However, due to the diverse sources and inconsistent formats of these datasets, effective integration remains a challenge, often leading to critical information being overlooked. This limitation not only restricts the depth of research but also undermines the accuracy of predictions."
Following his advisor’s suggestion, Chen turned his attention to AI technology. At the time, AI was rarely applied in atmospheric sciences, so the path forward required feeling his way forward step by step. He gradually realized that the core constraint of traditional methods lay in insufficient data processing capability, and that the theoretical framework of meteorology urgently needed the power of data-driven approaches to break free from these constraints. With this understanding, he decided to transcend the limitations of traditional meteorological analysis and tackle the challenge through data-driven approaches.
He began focusing on learning AI, officially embarking on an interdisciplinary academic journey. "At that time, I treated myself as an undergraduate student, actively seeking out any available resources on AI," he recalled. By repeatedly watching online AI tutorials and studying the book *Introduction to Artificial Intelligence*, he gradually built a comprehensive knowledge system in AI. Building upon his foundation in meteorological theory, Chen Bin forged a new research pathway powered by data-driven methods.
As he clicked through his presentation with a remote pointer, Chen explained: "The most revolutionary breakthrough of AI lies in its ability to autonomously discover knowledge.""By pointing at the satellite cloud imagery, he explained: 'Through the attention mechanism module in AI, the model can automatically identify key aerosol targets under non-clear-sky conditions and accurately classify aerosol types. This technological breakthrough overcomes the limitations of traditional methods in complex weather scenarios."
Building on such technical expertise, Chen Bin's team has set its sights on a more challenging goal: developing an AI-powered intelligent forecasting system for extreme weather across the Belt and Road region. Given the sparse distribution of monitoring stations in this area, predicting extreme weather events such as sandstorms and strong winds remains highly difficult. To address this, Chen Bin and his team plan to leverage AI technologies.Enhance current operational capabilities with a goal of improving forecast accuracy for sandstorms and high-wind events by 10%. This objective serves not only as their five-year research roadmap but also as a response to President Xi Jinping’s directive for meteorological work: “precise monitoring, accurate forecasting, and refined services.” The aim is to provide more accurate and detailed extreme weather forecasts for countries along the Belt and Road Initiative, helping them better address the challenges posed by climate change and strengthen disaster prevention and resilience.
Rooted in Practice—Bridging Theory and Application
How can theory and practice be truly integrated in teaching? Chen Bin offers his answer through a freshly printed book, still smelling of ink, and a unique "pilot's license."
While teaching at the intersection of meteorology and artificial intelligence, Chen encountered a critical gap: no existing textbook effectively combined AI with meteorological science. Moreover, current AI textbooks were overly theoretical, making them difficult for atmospheric science students to grasp. To bridge this divide, Chen took the initiative to write a textbook that not only clarifies AI concepts but also applies them to real-world meteorological scenarios.In 2021, Chen Bin began planning this book, systematically organizing relevant knowledge based on his interdisciplinary experience in learning AI.
Over the next three years, he revised and refined the manuscript multiple times, eventually publishing *Artificial Intelligence in Atmospheric Sciences* in September 2024 through the China Meteorological Press. The book not only covers foundational AI theories but also integrates concrete applications in meteorology. Since its release, it has undergone three print runs, reaching over 2,000 readers, and has been selected as part of Gansu Province's "14th Five-Year Plan" for undergraduate-level higher education teaching materials.
Chen Bin's Artificial Intelligence Technologies in Atmospheric Sciences
Stepping into Chen Bin's office, one’s eyes are immediately drawn to the slightly worn copy of *Artificial Intelligence Technologies in Atmospheric Sciences* on his desk—its pages gently curled from frequent use. With a smile, he says, “Whenever I get a new idea, I flip through this book. No wonder it’s looking so old.” He then pulls a document from his drawer and adds, “Not long ago, a chief expert from the China Meteorological Administration told me this book was selected as one of the seven recommended references for the National Weather Forecasting Vocational Skills Competition in the meteorological sector. I was deeply encouraged, yet I also know there are still many shortcomings that need constant improvement.”
Yet Chen Bin understands that textbook knowledge can only convey theory, while truly helping students grasp atmospheric science requires more.Understanding must be deepened through hands-on experience. To this end, he led his team to obtain drone pilot licenses designed for instructors, integrating drone-based sensing into meteorological education. Chen Bin explained, "Drone malfunctions are rarely due to hardware failure—more often, they stem from weather conditions, especially in complex terrains where turbulent airflows can severely impact flight stability." By operating drones firsthand, Chen not only helped students grasp how weather influences aircraft performance but also allowed them to experience how meteorology and technology converge to solve real-world challenges.
With support from the university’s research office and his College, Chen also established the Lanzhou University Institute of Meteorological Artificial Intelligence.Chen Bin has not only transformed AI research achievements in meteorology into practical teaching, but also actively aligned with the university's strategy of promoting interdisciplinary integration and innovative education. Despite a full teaching schedule, he has launched multiple elective and general education courses that integrate artificial intelligence with meteorology, and established a micro-specialization in "Big Data Applications – Digital and Smart Earth," accumulating over 200 undergraduate teaching hours annually. These new offerings have enriched the curriculum and provided students with diverse learning opportunities, equipping them with cutting-edge meteorological technologies and applications in the era of big data and AI.
Drone Training for the Micro-Program Course "Digital Earth Drone Detection Technology"
Rooted Aspiration—Carrying the Torch in the Wilderness
To Professor Chen Bin, his mentor Academician Huang Jianping was not only a guide in academia but also a role model in life. "At the end of 2006, I was recommended for graduate studies under Professor Huang. Back then, I was just a naive rural youth. He personally took me into the field, conducting observations and research, gradually helping me discover my own strengths," Chen recalled. In 2010, when he was selected for government-sponsored overseas study, Chen faced financial pressure from repaying national loans taken during his undergraduate years. Without hesitation, Professor Huang not only provided academic guidance but also used his own salary to pay off the debt.He was freed from all worries. At that time, Chen Bin was like a humble tumbleweed in the northwest—unnoticed and unassuming. It was his mentor's selfless care and unwavering support that transformed this rural youth into a national-level young talent.
Chen Bin’s thoughts drift back to those vibrant years spent with Professor Huang’s team—seniors, classmates, and mentors—conducting field observations together: facing howling sandstorms in vast deserts, enduring bone-chilling winds atop snowcapped mountains, trudging across grasslands, measuring the earth step by step with their own feet. "We’ve been to the most challenging places," he recalls. Today, those unforgettable experiences have become precise meteorological curves on his computer screen and data points marked on satellite cloud images.
Academician Huang Jianping once said, "Doing research is like planting trees in the desert—each generation has its own responsibility." This statement has become an enduring belief in Chen Bin's heart, as well as his heartfelt expectation for his students.
Group photo of all students from the Meteorological Artificial Intelligence Research Team
Hao Mingxiang, a Ph.D. student of Chen Bin, recalled: "When I had my first online interview with Professor Chen, he emphasized academic integrity, the regional characteristics of Northwest China, and the principles of being a good person and doing solid work. This deeply impressed upon me his steadfast commitment to education and his strong emphasis on students' moral development." It was precisely this sense of responsibility and deep concern for character building that inspired Hao Mingxiang to switch from computer science to meteorology and join Chen Bin's "Artificial Intelligence in Meteorology" research team.
Similarly, Song Zhihao began engaging in scientific research under Chen Bin’s guidance as an undergraduate. With Chen’s mentorship, Song has worked tirelessly in his studies, publishing seven SCI papers ranked in the second quartile or higher, embodying Lanzhou University’s motto.
This May, Chen Bin will join Academician Huang Jianping and team members on the "Retracing the Taklamakan Scientific Expedition." This journey is not merely a revisit to former research sites, but also an on-the-ground assessment of ecological restoration efforts in western China. From Lanzhou to the vast expanse of the Taklamakan Desert, it’s a long and arduous trek—much like the work of ecological governance, a marathon passed from one generation of scientists to the next. Despite the difficulties, Chen bin believes that, just like the sand grass, although the road is long and difficult, it is full of hope!
