Floods are classified into four types according to the intensity and duration of heavy rainfall and the characteristics of the affected area: river floods, urban floods, coastal floods, and flash floods. A flash flood refers to flooding that occurs when heavy rain falls on a narrow area over a short period of time, typically triggered by rainfall exceeding 50 to 100 mm per hour or more. The increasing frequency of flash floods in recent years has raised concern that they represent a new rainfall pattern driven by climate change.
Because flash floods can occur within six hours of heavy rainfall — and sometimes within as little as one hour — they are nearly impossible to predict and allow only limited response time, unlike conventional floods or typhoons. In response, the UN World Meteorological Organization (WMO) operates a Flash Flood Guidance system in more than 60 countries, combining real-time meteorological data with hydrological modeling to forecast risk. However, significant limitations remain in terms of prediction accuracy and securing sufficient response time.
[Flash Flood ⓒESG.ONL]
Flash floods occur primarily in urban areas with poor drainage, landslide-prone zones, and areas near rivers — where water levels can rise in an instant before warnings are even heard, leading to a high likelihood of casualties from incidents such as flooding of subway stations and semi-basement residences, and the entrapment of vehicles. According to data released by the WMO on July 9, 2025, flash floods cause more than 5,000 deaths and USD 50 billion in economic losses annually — recording overwhelmingly greater damage than other forms of flooding.
Since advance response to flash floods is nearly impossible, companies must also carefully assess the exposure risks of their business sites and supply chain partners, and establish emergency plans that can be executed rapidly. Transparently disclosing in ESG reports the status of response measures to extreme weather events — including flash floods — is becoming a new standard, and what is needed is an effort to build a system capable of rapid response rather than perfect prediction.
by Editor O