In our new publication published last week in the AGU journal, Earth’s Future, co-authors and I find that the choice of dataset used in index-based flood insurance programs directly controls the years in which payouts are made and how timely those payouts are, with potentially great implications for communities living in flood-prone areas.

The University of Arizona is developing a global inundation model in collaboration with the NASA Goddard Space Flight Center. The model uses VIIRS satellite imagery to map fractional inundation globally every day at 375-meters resolution. If deployed, the model will be NASA’s first operational flood mapping service which uses machine learning.

While Earth Observations (EO) are now widely used to understand where and when floods happen, rarely are they used to monitor what happens in communities after floods, and how that might differ amongst different parts of society. In our new paper Earth Observation to Address Inequities in Post-Flood Recovery, led by my colleague Hannah Friedrich along with co-authors from the University of Arizona and beyond, we argue that EO is well positioned to monitor post-flood recoveries. Furthermore, EO-based monitoring is necessary to identify inequitable recovery outcomes that result from unjust practices of disaster mitigation and assistance.

In June 2023, along with Hannah Friedrich (University of Arizona) and Emmalina Glinskis (University of California Berkeley), I co-hosted a panel session at the Columbia Climate School’s Managed Retreat Conference, “At What Point Managed Retreat?: Habitability and Mobility in an Era of Climate Change”.

As part of the NASA New Investigator Program “Understanding flood risk in human altered landscapes from cities to farms: inferences from satellites and machine learning” led by P.I. Dr Beth Tellman, in May 2023, I traveled to Bangladesh with Beth and two other colleagues from the Social Pixel Lab.

Mapping floods is important for response, relief, preparedness, planning and mitigation. Satellites offer a more direct approach to mapping floods (than other approaches such as physical inundation models), and through advancements in satellite sensors and algorithms, satellite estimates of flooding are increasing in temporal frequency, spatial resolution and accuracy.