Global Near Real-Time MODIS-based Flood Mapping - Upcoming Improvements and Transition to LANCE.
Flooding is the most destructive, frequent, and costly natural disaster faced by modern society, and is expected to increase in frequency and damage with climate change, sea-level rise, and population growth. Every year new record floods are widely reported; the unusual is becoming the normal. The toll of these events, in financial costs, displacement of individuals, and deaths, is substantial and continues to rise as climate change generates more extreme weather events.
The disaster management community requires frequently updated and easily accessible information to better understand the extent of flooding and better coordinate response efforts. With funding from various NASA programs over the past decade, we have developed and operated the Near Real-Time Global Flood Mapping System (https://floodmap.modaps.eosdis.nasa.gov
). The system applies a water detection algorithm to low latency MODIS imagery provided by NASA LANCE and outputs critical flood extent information for global events. Using imagery from both the Terra and Aqua platforms allows an initial daily assessment of flooding extent by late afternoon, and more robust assessments after accumulating cloud-free imagery over 2 or 3 days. Cloud cover is the primary limitation in detecting surface water from MODIS imagery, by both obscuring the surface, and false positives due to cloud shadows . Other issues include the relatively coarse scale of the MODIS imagery, the difficulty of detecting flood waters in areas with continuous canopy cover, confusion of terrain shadows with water, and accurately identifying detected water as flood as opposed to normal water extents.
The Near Real-Time Global Flood Mapping System is currently transitioning from a PI-managed system at Goddard to LANCE-MODIS and includes optimization and improvements to several elements of the system. Anticipated results include lower-latency products, improved removal of false-positives, and an assessment of whether reported flood is "regular" (e.g., recurring due to normal annual hydrological cycles), or unusual (e.g., likely of most concern as a disaster). The LANCE integration will also leverage the ESDIS Worldview interface to visualize Global Flood Mapping System products for current and previous flood events in the context of other NASA EOS science data.