IN020-05
The COVID-19 mitigation effects and the incoming risk in the flood season

Thursday, 10 December 2020: 16:16
Virtual
Xinyi Shen1, Chenkai Cai2, Qing Yang3, Emmanouil N Anagnostou1 and Cory Merow4, (1)University of Connecticut, Civil and Environmental Engineering, Storrs, CT, United States, (2)Hohai University, Nanjing, China, (3)Guangxi University, College of Civil Engineering and Architecture, Nanning, China, (4)University of Connecticut, Department of Ecology and Evolutionary Biology, Storrs, CT, United States
Recorded Presentation
Abstract:
The COVID-19 has infected more than 4 million people in the US and 16 million in the world by July. As a respiratory infectious disease, the transmissivity of COVID-19 is greatly affected by the frequency of close-range interaction. To slow down the spread, most polities adopt mitigation measures at the cost of reducing socioeconomic activities (SA). Flooding can displace a large population, which in turn, can intensify the pandemic1 (Fig. 1). We build a mitigation model2 and a flood damage model3 to evaluate the impacts of restricting SA and flood damage on the transmissivity of COVID-19.

In the mitigation model, the SA is proxied by the daily Nitrogen Dioxide (NO2) product from Sentinel-5P. We find that restricting SA has a leading contribution to lowering the reproductive number of COVID-19 (18.3±3.5%), more significantly than the weather. The reduction effects by restricting SA become more pronounced (23±3.0%), and the weather becomes less impactful in more developed contries, where the indoor climate is mostly controlled. We estimate the spared infectees by restricting SA in all polities then find most predicted high-risky areas became epicenters later by either loosening the SA restriction or not having sufficient facilities.

To characterize the flood damage, we first prove that the flood displacement is strongly correlated to house claims (r2=0.84), then build a machine learning model to predict flood claims for the CONUS. Using river stage, precipitation, surge, synthetic aperture radar (SAR) derived flood maps4,5, topographic, geomorphological, and house location information as predictors, the claim model yields high accuracy at county level (r2=0.91,relative bias=-13%). Finally, we find the high-risk areas of both flooding and COVID-19 are concentrated along the southern and eastern coasts and along some part of the Mississippi River.

1 Shen, X., Cai, C. et al. “The US COVID-19 Pandemic in the Flood Season” Sci Total Environ (2020b), (submitted)

2 Shen, X., Cai, C. & Li, H. Sci Total Environ (2020a), (minor revision)

3 Yang, Q., Shen, X. et al. Prediction of flood claims over the CONUS by building a classification/regression-hybrid machine learning scheme, (2020), (submitted)

4 Yang, Q. Shen, X. et. al. Bulletin. Amer. Met. Soc. (2020), (under review)

5 Shen, X., Anagnostou, et al. Remote Sens. Environ (2019)

<< Previous Abstract | Next Abstract >>