In April of 1950, a team of scientists made a series of successful 24-hour forecasts over North America. Today, cutting-edge compute platforms are predicting the weather up to several weeks in advance. Supercomputers have vastly improved weather forecasting since the first systems were introduced in the 1960s, and they will become increasingly important to fuel tomorrow’s scientific breakthroughs.
Research institutions are investing heavily in high performance computing (HPC) technologies to streamline data analytics and deliver groundbreaking levels of IT performance. These machines are designed with optimal speed, efficiency, and durability in mind, allowing scientists to monitor temperature, barometric pressure, windspeed, and a host of other data points to make weather predictions in real time.
ENHANCING WEATHER MONITORING AND RESEARCH
Across a variety of markets—agriculture, aviation, public transportation, government, and more—weather can have a major impact on the way business owners and their workforces operate. People rely on accurate, timely weather forecasts to effectively protect and run their homes and businesses. But with skyrocketing volumes of geospatial data to collect, analyze, and store, traditional IT is no longer enough to support these increasingly demanding workloads.
Supercomputing delivers the compute capabilities to solve extremely complex and data-intensive problems. Investing in an HPC architecture enables users to overcome the barriers of legacy operating systems, empowering innovation at any scale, supporting a variety of latency-sensitive applications, and accelerating insights like never before. Supercomputers are furthering weather forecasting in key ways:
- Bolstering data management capabilities to utilize enormous datasets
- Enabling higher-resolution models
- Addressing performance bottlenecks such as memory and bandwidth latencies
The National Center for Atmospheric Research (NCAR) deployed a supercomputer named Cheyenne to dramatically enhance weather research. NCAR scientists are tackling a number of ambitious projects that will extend weather predictions from months to years in advance. Cheyenne can execute 5.34 quadrillion calculations per second, allowing scientists to rapidly harness data insights and use simulations as real-time forecasts.
TRANSFORMING THE FUTURE OF WEATHER FORECASTING
Technological advancements are driving a major paradigm shift to weather research. To further this progress, Hewlett Packard Enterprise (HPE) is excited to announce a partnership with the Thai Meteorological Department (TMD). Together, our goal is to implement an HPC solution that will significantly improve our ability to forecast weather changes and predict natural disasters. This industry-leading technology will facilitate atmospheric research and natural disaster planning by accelerating data analytics and increasing the accuracy of results, while leveraging the flexibility, scalability, and cost-efficiency required to support immense and unpredictable workloads.
To achieve this, TMD selected the HPE Apollo 2000 to power the department’s Weather Research and Forecast (WRF) model. The WRF harnesses data from a variety of sources to help TMD forecast rainfall within a four-square kilometer (roughly one-and-one-half-square mile) spatial resolution and forecast weather more than seven days in advance. Featuring HPE 3PAR StoreServ Storage and HPE 3PAR Gen5 ASIC, HPE Apollo servers offer the efficiency, portability, flexibility, and performance to transform weather monitoring and research in Thailand—helping citizens better prepare for potentially life-threatening storms.
To learn more about how supercomputing is transforming the weather sector and delivering more reliable, up-to-the-minute information, follow me on Twitter at @Bill_Mannel. I also invite you to follow @HPE_HPC for the latest news and updates on HPC innovation.