**5. Time Series Comparisons of the WARMF and Regression Models**

The preceding analysis focused on comparisons of mean EC values and differences between model-predicted EC and observations for the Regression and WARMF models for various forecast lead times. In this section, time series comparisons of the EC predicted by each model compared to EC observations for the same time period were made for selected lead times. As shown in Figure 11, both models have relatively small mean EC differences at forecast lead times of less Δ Day + 4. From Δ Day + 5 to Δ Day + 8, mean differences increased. After Δ Day + 9, the EC predictions of both models reached a relatively constant plateau. Figure 12 also shows a comparison of Regression model EC forecasts and observations at Δ Day + 4, Δ Day + 8 and Δ Day + 12. As illustrated, there was a good match between observations and forecasts. However, as the forecast lead time increased the differences between model forecast of EC and observations also increased. This relationship between model EC forecasts and observations can be quantified using the root mean square error (RMSE) statistic which increases from 69.4 at Δ Day + 4 to 103 at Δ Day + 8 to 154 at Δ Day + 12. Figure 12 shows a similar relationship between model EC forecasts and observations for the WARMF model. In this case, the RMSE increases from 99.8 at Δ Day + 4 to 123 at Δ Day + 8 to 166 at Δ Day + 12. As illustrated by the figures and RMSE values, the Regression model performed somewhat better than the WARMF model in predicting EC for similar lead times.

**Figure 11.** Comparison of Regression model forecasts and observations of EC at various lead times for the period between 22 February 2018 and 22 May 2020.

**Figure 12.** Comparison of WARMF model forecasts and observations of EC at various lead times for the period between 22 February 2018 and 22 May 2020.
