The quotation above is one of my favorites! It’s so simple, but speaks to our reliance on models…and really all predictive data. Predictive modeling can be a useful tool to forecast possible outcomes and better understand systems. But as a former modeler, I know that models are only as good as the data used for input and as robust as the system used to run it. That is, if you don’t have a lot of data to input into the model, it’s essentially useless.
However, the “wrongness” of the model determines its usefulness. Not all models are useless, and almost all good models tell us something. We saw this lesson reemerge last week, as we all prepared for Hurricane Joaquin and tried to predict its trajectory and potential impact. There were several models, most notably:
After Hurricane Joaquin dumped almost two feet of rain on South Carolina and then turned to sea, there were many questions about the accuracy of forecasting data. Several articles explored the accuracy of the ECMWF model, and asked why the GFS model did not seem to hold up. Some went as far as to question the ability of the GFS to predict any extreme weather events accurately. What should not be lost in this argument is that both the ECMWF and the GFS models are world-class models…but still models, which means that they will make incorrect predictions from time to time. However, the more robust and enriched a model is, the more accurately it will perform.
These models proved valuable in helping states (and countries, like the Grand Bahamas Islands) prepare for the potential impact of the storm. And as we know, these natural events have a public health consequence. It may be time to start thinking about the value of predictive weather models as a public health tool, since these events have the potential for serious impacts on public health. Fortunately, we have a few options of top-notch global weather models to assist in predictions, and can explore inputs and other measures of rigor to determine which tool to rely on for a particular event. As public health workers, spending more time understanding the accuracy of particular tools will help as we prepare to offset or minimize public health impacts of disasters. It’s just another example of the interrelatedness of preparedness!