Intelligent Meteorological Monitoring Assistant
With the advent of the Big Data era, we can all experience the pressure of being bombarded by huge amount of information every day. Being an active user on the internet, you may receive hundreds of emails, WhatsApp messages, Twitter tweets, photos on the Instagram, in addition to countless interesting video clips that you want to view on YouTube, and the many Facebook posts waiting for you to comment or "like" each day. The weather forecasters in the Central Forecasting Office of the Hong Kong Observatory (HKO) are also facing with similar challenges of information overload, for example, high-density automatic weather station data, high-resolution Numerical Weather Prediction model products, frequent scans from meteorological radars and satellites, etc. Currently, the data processed by HKO amount to over 150 million pages of text each day!
Figure 1 Dozens of computer panels filled with various types of weather information on the workbench of the weather forecaster at HKO.
In 2013, the HKO initiated an in-house project to develop the Intelligent Meteorological Monitoring Assistant (IMMA) for alerting forecasters of emerging weather conditions. The system has since been put into operational. Being a knowledge-based expert system, IMMA is a computer program designed to simulate the problem-solving behaviour of an expert in a specific domain or discipline, in our case, weather forecasting. The system applies expert knowledge to real world problems based on given facts and well-defined rules. Facts are pieces of information but on their own are of limited use (for example, a lightning signal detected over the waters to the south of Hong Kong). It is necessary to apply rules to reveal the usefulness and value of all the facts as a whole (for instance, prompt the forecaster to consider issuing the thunderstorm warning if the lightning is close enough to Hong Kong).
Figure 1 Workflow of the Intelligent Meteorological Monitoring Assistant.
At present, IMMA processes more than 530,000 data points every minute. It does not merely gather data and display them as necessary, but also turns them into actionable advice to aid the forecasters in performing weather monitoring and forecasting duties. It serves as a tireless assistant to the forecasters working in a busy and information overwhelming environment, such that the amount of time spent on weather watch and the associated decision-making processes can be reduced considerably.
Figure 3 Alert from IMMA highlighting the potential for a rainstorm situation, considering the rainfall over inland Guangdong and the prediction from HKO's nowcasting system.
With the help of IMMA, the knowledge and experience in weather forecasting can be retained and managed in a more effective manner. In the future, when combining with location-based and social networking services on mobile devices, the technology may be utilized to develop new services for people on the go by providing them with personalized advice at different locations under different weather conditions.