Fig. 8.

A comparison of the impacts of forecast missingness in the applications to the U.S. (panel (a)) and Europe (panel (b)). Within each panel, the histogram on the left shows the number of locations forecasted by each contributing forecaster in the week of October 11, 2021, colored by whether or not the forecaster was among the top 10 forecasters eligible for inclusion in the relative WIS weighted ensemble selected for prospective evaluation. The plot on the right shows the estimated weights that would be used if all of the top 10 models (each represented by a different color) were available for a given location (on the left side), and the effective weights used in each location after setting the weights for models that did not provide location-specific forecasts to 0 and rescaling the other weights proportionally to sum to 1. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)