Spatially Explicit Modeling of 2019-nCoV Epidemic Trend based on Mobile Phone Data in Mainland China

Xiaolin Zhu; Aiyin Zhang; Shuai Xu; Pengfei Jia; Xiaoyue Tan; Jiaqi Tian; Tao Wei; Zhenxian Quan; Jiali Yu

We predicted daily infected cases of each city up to March 12, 2020 under three different scenarios (scenario 1 -the current trend maintained; scenario 2 -control efforts expanded; and scenario 3 -person-to-person contacts increased due to work resuming). Our prediction shows that the whole mainland China will have 72172, 54348, and 149774 people infected up to March 12, 2020 under the above three scenarios respectively (Supplementary data Table 4 ). To provide an intuitive picture about epidemic dynamics in different scenarios, we showed in Figure 7 the number of cumulative infections in each city on March 12. The infected people will mainly distribute in the central and eastern provinces, the number of western cities at a relatively low level under all scenarios. In scenario 3 (Figure 7 .c), many cites have much larger number of infections than scenario 1 (Figure 7 .a) and 2 (Figure 7 .b), suggesting that work resuming will bring large challenge to control the disease timely. This difference is significant for those cities with largest number of infections by March 12 (Figure 8) , such as Wuhan and other cities in Hubei province. To understand the specific attributes of epidemic dynamics under different scenarios, we investigated the temporal changes of daily new infections across all cities in mainland China. In Figure 9 , we show the results in Wuhan, Hubei province excluding Wuhan, other provinces, and four first-tier cities. Compared with the scenario 1 where current trend is maintained, the daily new infections in scenario 2 reduces quickly in the second half of February. In scenario 3 where transmissibility rebounds after the public holiday in all cities, the peak of new infections will postpone ten days and the magnitude will be twice of that in scenarios 1 and 2. Our simulation suggests that strict quarantine of inner-and inter-city population movement during February would have a significant effect on the suppression of virus spreading. Table 4 ) suggests that the exponential growing of disease will stop and the spreading would be controlled gradually. Under current trend, our model estimated that the number of new infections in 79.7% cities already reached the peak point before February 11 and in all other cities, it will reach the peak point by February 21. With the control effort expanded, all cities will have the peak point of new infections by February 14, one week earlier than current trend. However, the peak point of new infections will be greatly delayed under scenario 3 that a few cities in Hubei province will show the peak point of new infections by February 26.