Track Similarity-based Typhoon Search Engine for Disaster Preparedness

Plenty of variance lies in the development stage of a typhoon, including its location, direction, and surrounding meteorological conditions. Affected by such numerous contributing factors, it is not always easy for nowaday technology to make accurate typhoon forecasts. The experiences from the past typhoons may provide decision-makers with sufficient information to anticipate potential damage and thus develop appropriate strategies. The typhoon track plays a vital role in selecting appropriate historical cases based on the need to grasp the situation during the disaster preparedness phase. This paper summarizes our published journal article, which proposed an algorithm to compare the similarity between the current typhoon forecast track and all the typhoon tracks of the western North Pacific in the past. Based on the forecast track points of the current typhoon, the algorithm suggests a list of historical typhoons with the highest similarity in tracks. Hence, the tracks and the disastrous area of those historical typhoons can be used as crucial alerts and hints for disaster preparedness. Inside the algorithm, the mechanism follows the Recentness Dominance Principle, which is elaborated in our published journal article. The principle states that the more recent the forecast track point is, the higher the weight that it possesses and thus should be emphasized. For implementation, the study develops a user-friendly front-end interface that synchronizes the latest forecast typhoon tracks from six major meteorological institutions automatically, including CWB, HKO, JMA, JTWC, KMA, and NMC. The result comes with a convenient and concise search engine that assists decision-makers in finding similar historical typhoon records.