Predicting COVID-19 Cases in South Korea Using Stringency and Niño Sea Surface Temperature Indices.

Most coronavirus disease 2019 (COVID-19) models use a combination of agent-based and equation-based models with only a few incorporating environmental factors in their prediction models. Many studies have shown that human and environmental factors play huge roles in disease transmission and spread, but few have combined the use of both factors, especially for SARS-CoV-2. In this study, both man-made policies (Stringency Index) and environment variables (Niño SST Index) were combined to predict the number of COVID-19 cases in South Korea. The performance indicators showed satisfactory results in modeling COVID-19 cases using the Non-linear Autoregressive Exogenous Model (NARX) as the modeling method, and Stringency Index (SI) and Niño Sea Surface Temperature (SST) as model variables. In this study, we showed that the accuracy of SARS-CoV-2 transmission forecasts may be further improved by incorporating both the Niño SST and SI variables and combining these variables with NARX may outperform other models. Future forecasting work by modelers should consider including climate or environmental variables (i.e., Niño SST) to enhance the prediction of transmission and spread of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2).

Understanding chaos in COVID-19 and its relationship to stringency index: Applications to large-scale and granular level prediction models.

Understanding the underlying and unpredictable dynamics of the COVID-19 pandemic is important. We supplemented the findings of Jones and Strigul (2020) and described the chaotic behavior of COVID-19 using state space plots which depicted the changes in asymptotic behavior and trajectory brought about by the increase or decrease in the number of cases which resulted from the easing or tightening of restrictions and other non-pharmaceutical interventions instituted by governments as represented by the country's stringency index (SI). We used COVID-19 country-wide case count data and analyzed it using convergent cross-mapping (CCM) and found that the SI influence on COVID-19 case counts is high in almost all the countries considered. When we utilized finer granular geographical data ('barangay' or village level COVID-19 case counts in the Philippines), the effects of SI were reduced as the population density increased. The authors believe that the knowledge of the chaotic behavior of COVID-19 and the effects of population density as applied to finer granular geographical data has the potential to generate more accurate COVID-19 non-linear prediction models. This could be used at the local government level to guide strategic and highly targeted COVID-19 policies which are favorable to public health systems but with limited impact to the economy.