Artificial intelligence takes on Shenzhen forecast
A groundbreaking artificial intelligence (AI) weather forecaster has been released by the team behind the PanguWeather prediction model.
The newest iteration of the AI model, Zhiji, which focuses on regional weather, gives a five-day forecast with a precision sharpened from 25km to 3km.
Its launch came less than a month after Pangu-Weather, developed by Huawei Technologies, was named China’s best scientific innovation of 2023.
Since its release last August, Pangu has revolutionised forecasting, offering quicker and more accurate predictions than traditional meteorological methods.
Pangu-Weather burst onto the scene in July 2023, when a paper detailing the AI model was published in Nature journal. A month later, it was launched on the European Centre for MediumRange Weather Forecasts (ECMWF) website.
It hit a major milestone when it was able to complete a seven-day weather forecast in just 10 seconds – more than 10,000 times faster than traditional methods. Then on February 29, PanguWeather was ranked first among China’s top 10 scientific advances in 2023 by the National Natural Science Foundation of China.
According to a Huawei report in late February, Pangu delivered more accurate forecasts for crucial weather elements, such as temperature, pressure, humidity and wind speed, than numerical simulations. Plus its error margin for predicting the paths of tropical cyclones was 25 per cent lower than the ECMWF.
By leveraging AI to predict weather patterns, scientists can bypass complexities associated with traditional methods of forecasting. Now, researchers have used Pangu as a foundation to develop the regional model, Zhiji.
Created in collaboration with the Shenzhen Meteorological Bureau, Zhiji has been trained with high-resolution data from southern China.
According to the Huawei team, Zhiji can provide a five-day forecast with a precision of 3km for Shenzhen and surrounding areas. While the Central Meteorological Bureau provides hourly forecasts with street-level precision, these are generally only available for the following 24 hours.
“Zhiji is capable of forecasting core meteorological elements such as wind speed, temperature, humidity and precipitation. Since its trial operation began in February, it has provided valuable insights to the Shenzhen Meteorological Bureau on multiple occasions,” Huawei reported.
AI and manual predictions have their strengths and weaknesses. AI has the edge in predicting typhoon paths; numerical simulations are more accurate in determining wind strengths.
“Scientists can now integrate results from numerical simulations with forecasts provided by Zhiji to make the most advantageous judgments,” a Huawei spokesman said.
According to researchers, this year’s flood season will be the true test for Zhiji 1.0. They expect to see the model further optimised with improvements to algorithms.
Ongoing work on the technology aims to enhance its rainfall forecasting capabilities, including providing specialised forecasts like heatstroke indexes and comfort levels, and improving the resolution of heavy rain forecasts to 1km.