Impact of the COVID-19 pandemic on inbound air travel to Canada–Supplemental

The figure shows output from the null hypothesis (H₀) model residuals as a time series, autocorrelation function (ACF) and a histogram. Residuals should appear random (i.e., white noise) in models that adequately capture trends, including temporal trends. Output from the residual time series appears mostly random except for a steep trough shortly after 2020. In the ACF plot, there is no statistically significant autocorrelation in residuals at the assessed time lags. The histogram suggests the residuals are mostly random by having a close to normal distribution shape.

The figure shows output from the interrupted time series (ITS) model residuals as a time series, autocorrelation function (ACF) and a histogram. Residuals should appear random (i.e., white noise) in models that adequately capture trends, including temporal trends. Output from the residual time series appears mostly random for the pandemic and the pre-pandemic periods, separately. However, differences in the range of residuals between both periods suggest that the model does not adequately account for differences in the magnitude of seasonal patterns between both periods. In the ACF plot, there is no statistically significant autocorrelation in residuals at the assessed time lags. The histogram suggests the residuals are mostly random by having a close to normal distribution shape.