The number of foreign residents and tourists in Japan has been dramatically increasing in recent years. Despite the fact that Japan is prone to natural disasters, with each climate-related event turning into an emergency such as with record rainfalls, floods and mudslides almost every year, non-Japanese communication infrastructure and everyday disaster drills for foreigners have received little attention. This study aims to understand how a resilient communication ecosystem forms in various disaster contexts involving foreigners. Within a framework of information ecology we try to get an overview of the communication ecosystem in literature and outline its structure and trends in social media use. Our empirical case study uses Twitter API and R programming software to extract and analyze tweets in English during Typhoon 19 (Hagibis) in October 2019. It reveals that many information sources transmit warnings and evacuation orders through social media but do not convey a sense of locality and precise instructions on how to act. For future disaster preparedness, we argue that the municipal government, as a responsible agent, should (1) make available instructional information in foreign languages on social media, (2) transfer such information through collaboration with transmitters, and (3) examine the use of local hashtags in social media to strengthen non-Japanese speaker's capacity to adapt.