Earth-observing Radar Satellite Data for Semi-Real-Time Flood Exposure Analysis: Case Study of Iran 2019 flood

Abstract:

Extreme precipitation led flooding is one of the major causes of global economic losses, and pose tremendous pressure on human society. Data provided by Earth-observing Synthetic Aperture Radar (SAR) satellites enables mapping of the disaster extent in semi-real-time and is of great importance for rescue and mitigation efforts as well as future hazards assessments. During January through April 2019, Iran experienced one of its worse flash floodings, affecting a large part of the country, including Golestan, Khuzestan, Fars, and Lorestan provinces. Here, we analyzed the full archive of Sentinel-1 C-Band SAR satellite, including 777 images for the period January 1^st through April 15^th, to map the extent of the 2019 flood of Iran. We applied a Bayesian framework to SAR intensity images to calculate the probability of a SAR pixel being flooded, for which the likelihood probability density function was estimated using a fast marching algorithm. We calculated the percentile area of each province and county that is flooded per month. We found that 25 out of 31 provinces are severely affected by the flood event. Next, using these results in combination with Iran’s recent census data, we calculate the land and population flood exposures. We estimated that more than ~10 million people were exposed to flooding, in addition to the severe damages of the historical monuments. Statistical analysis of the meteorological data indicated that more intense precipitation in the southern part of the country correlates with higher flooding exposure. However, other factors such as topographic slope and weak flood defense mechanism also play a role in the severity of 2019 flooding. The probabilistic flood maps generated here are highly useful in data assimilation into numerical models for future flood forecasting. This study also highlights the importance of Earth-observing SAR data sets with open-access policy, such as Sentinel and future NISAR, for rapid response to disasters, where field observations are not available due to severe weather condition and the broad extent of the affected area.

Keywords:

Synthetic Aperture Radar (SAR), Fast Marching Algorithm, Probabilistic flood map, 2019 Iran Flood, Precipitation Analysis