AI in meteorology and climate prediction

0

The Malaysian Meteorological Department (MMD) works tirelessly to save lives from weather disasters. Located just above the equator, both parts of the country, the Malay Peninsula and the island of Borneo, are surrounded by ocean waters. Weather events in Malaysia can quickly accelerate and cause widespread havoc. Although weather data is plentiful and free, the country’s existing numerical weather prediction computer system could not produce sufficiently detailed weather forecasts to give adequate warning of any growing threat. A new system was needed to efficiently assimilate weather data and produce detailed weather forecasts more quickly to provide early warnings to the public of severe weather.

Malaysia’s old computer system generated weather forecasts at an approximate resolution of tens of kilometres. That meant there was plenty of room for error. An emerging weather threat should become large enough for the system to detect at this distance, meaning significant storms could cause massive damage long before detection.

In order to get a more finished and accurate reading of real-time weather conditions in Malaysia’s tropical rainforests, beaches, high-rise commercial and rural areas, as well as air routes, the country needed to upgrade its system for performing very complex numerical weather forecasts. (NWP) which require a huge amount of computing power. This is a huge challenge because to double the resolution of the same forecast area (from 10 to 5 kilometers), a numerical weather prediction model would require about 16 times the computing power to provide the forecast in the same period.

Across the South China Sea, about a three-hour jet flight from Malaysia, lie the more than 7,600 islands that make up the Philippines. While weather is also a constant concern for the Philippine Atmospheric, Geophysical and Astronomical Services Administration (PAGASA), that country’s agency has also set itself a dual goal of economic development through climate change predictions. Being able to accurately identify optimal locations for everything from cropland to huge wind farms, offshore oil rigs and other expensive economic engines is a fundamental capability for countries trying to build an economy. of the 21st century.

To successfully simulate regional climate change, PAGASA data scientists adopted a regional numerical climate model, producing a set of simultaneous climate predictions. Each forecast member is initialized with slightly different meteorological data in order to statistically represent the uncertainty inherent in errors in the collection and assimilation of meteorological data. The outputs from all ensemble passes are then averaged to produce a probabilistic local climate outlook over a three-month period, rather than just providing a one-week weather forecast.

For example, a perfect location for most crops is one where flooding is not a routine problem. A seven-day weather forecast cannot reveal the expected number of flooding incidents in a specific geographic area over months or years, but a climate simulation can. Given the changes caused by influences ranging from global climate change to regional climate cycles such as El Nino, the demands of the PAGASA model on the computing infrastructure are extreme and complicated.

Another example is wind farms which are ideally located where the wind blows most often. Being able to predict wind speed and frequency over longer periods allows planners to pinpoint these locations. Climate simulations also play an important role in weather forecasting by finding and accurately predicting air humidity and pools of warm ocean water that are a strong indicator of the development of a cyclone or storm. ‘a typhoon.

Malaysia and the Philippines were looking for complete systems to handle these huge IT workloads. Neither country was interested in buying hardware and software piecemeal and then fighting to make it all work together.

Lenovo responded to the challenge by creating two complete supercomputer system solutions ready to run these workloads immediately.

For Malaysia, Lenovo, together with its partner, Numerical Weather Prediction Consultants (NWPC LLC) of the United States, installed a NextScale water-cooled supercomputer system at the MMD data center and a weather forecasting system based on the Weather Research and Forecasting (WRF) weather model from the National Center for Atmospheric Research (NCAR). It operates today in Malaysia with a resolution of up to 1 km for all of Malaysia (and 333 meters for Kuala Lumpur), to issue overlapping 7-day forecasts and not to miss any weather event without being detected. This system is the first of its kind and the most powerful supercomputer in Malaysia.

“To run the latest WRF software, we needed the HPC system to benchmark the performance of at least 100 TFLOPs. We looked at many proposals from different vendors, and Lenovo Scalable Infrastructure (LeSI) came out on top in terms of performance and cost-effectiveness – it met all of our compute and storage requirements at the best price. This system has allowed us to at least double the accuracy of weather forecasts in Malaysia. Dr. Wan Azli B Wan Hassan, Deputy Managing Director (Strategic and Technical) at MMD.

For the Philippines, Lenovo has built a powerful system specifically suited to the daunting task of producing climate simulations as well as weather forecasts. This supercomputer system is based on the Lenovo ThinkSystem SD530 and uses Intel’s XEON Scalable Systems processors. Lenovo has partnered with NCAR and NWPC to provide a climate solution also based on the WRF model at climate resolutions up to 3 kilometers.

“As the national meteorological and hydrological center of the country, PAGASA has acquired from Lenovo a high-performance computing system for operational climate prediction, the first in the country. This system will be able to produce short- and long-term local climate outlooks to assist various decision-making bodies, while planning activities for socio-economic sectors in the Philippines. Ana Liza S. Solis,. Chief, Climate Monitoring and Prediction Section, PAGASA.

Share.

Comments are closed.