Near real-time (NRT) flood observation using Synthetic Aperture Radar (SAR) Satellites:

Are you ready to provide flood inundation extent for EVERY flood event?

We developed the only operational and national-scale inundation mapping system in the world using SAR data and are now deploying this system to NOAA with further improvements. SAR data has the advantages of high spatial resolution (from <1 m to 30 m), and close to 100% weather penetration, which makes it the most reliable sensor for flood-inundation mapping. Recently, public agency-owned and commercial satellites can provide sub-daily to 6-day revisiting intervals, which is a gamer changer of flood observatory. This unique product has been used for validating and calibrating dynamic models and rapid disaster response. Based on existing radar physics and statistics, we keep on advancing satellite-based inundation mapping methods by incorporating emergent big datasets, and deep learning approaches (learn more).

Flood risk analysis and prediction by dynamic modeling: we developed fully distributed hydrological models for long-term water cycle simulation, flood risk analysis, and seasonal and short-term flood forecasts. Our hydrological models have been applied by users over all continents. Currently, we attempt to integrate more AI techniques and big data from autonomous systems and satellites into these models, to improve the applicability to ungauged locations.

Analysis of flood impact and drivers:

Remote sensing, Meteorology, demography, economics- cook ALL available data to predict flood impact

Floods threaten human society in many aspects, including residential and food security while flood severity is primarily determined by weather, topography, and the built environment (including infrastructure). Through building AI models and generating some big datasets, we attempt to answer the following questions: 1) To what extent the damages could future climate causes through floods?

 

 

 

 

2)  how do different backgrounds (e.g., socioeconomic status, education level, etc.) and law enforcement affect a homeowner’s perspective on purchasing flood insurance?

And 3) is social equity the key to reducing flood vulnerability?

Climate change and anthropogenic activities on biodiversity:
It’s time to enjoy deep learning in ecology, what are you waiting for?
This is a highly collaborative area being contributed by my group and other ecological and biogeographic groups. Based on the abundance of remote sensing data, AI techniques, and quantitative techniques, we try to first identify the distribution of species’ population, and/or estimate their physiological traits, then model their vulnerability to climate extremes and to human activities (Learn more from our AGU poster).