SMART 2022 (Sciences of Meteorology with Artificial-intelligence in Research and Technology for the Beijing 2022 Winter Olympics), is a research, development and demonstration project on high-impact weather observations and forecasts in support of the Beijing 2022 Winter Olympics, funded by the Ministry of Science and Technology of China. The objectives of the project is to provide the games with meteorological services and support that is “Wonderful, Extraordinary, Outstanding” – the overall goal of these great sporting games.
(1) Enhanced meteorological observations. The programme performs a series of high-resolution observation experiments already underway in Chongli and Haituoshan mountain areas using multi-source sensors (e.g., automated weather stations, Doppler radars, microwave radiometers, wind profilers, wind lidars, etc.). The programme also develops high-resolution reanalyses for small-scale weather attributes in the two mountain areas based on multi-sensor observations, high-resolution data assimilation/integration and special numerical simulations.
(2) Very short-term forecasting and nowcasting. The goal is to develop high-resolution forecast techniques for the 0–24 hour period based on rapid local data assimilation, improvement of key physics in high-resolution models, integration and blending of multi-source data, large-eddy simulation, and downscaling of numerical weather prediction (NWP) over complex terrain. The ultimate objective is to achieve up to 500-metre resolution covering the greater Beijing–Tianjin–Hebei region and 100-metre resolution in Olympic mountain domains covering the two key skiing and sliding areas with 10-min update cycling.
(3) Short and medium range prediction within 10 days. This includes development of prediction techniques for 1–10-day period based on framework of China’s Global and Regional Analysis and Prediction System, data assimilation, high-resolution ensemble forecasting and bias correction of NWP over complex terrain. The goal is to provide improved 3-hour update cycling for deterministic forecasts, 3 km resolution for 1–3-day ensemble forecasts and 6-hour update cycling for deterministic forecasts of 3–10 days.
(4) Seamless forecasting and early risk warning for key points and events. This programme is to provide seamless forecasting of key weather forecasting parameters out to 240 hours for different skiing events, venues and key sites based on conventional statistics, Analog Ensembles (AnEn), and machine learning methods using observations and grid forecast data. Also develops early warning techniques for risks of disruptive weather at individual sporting venues and other key sites.
(5) Intelligent meteorological support services. This is to develop meteorological service techniques for a variety of end users, including skiing and ice sport events, key traffic channels, helicopter rescue operations, public viewing, broadcast media, and other applications. These services will employ a variety of innovative approaches including information technology, artificial intelligence, data mining and cloud computing.