re: Weather forecasts and statistics

quote:

---

I've got something on Saturday and the local weather has been showing 100% chance of thunderstorms all week. Then today, it's suddenly going down. I can get that the models are going to change up as you get better info, but you left yourself no wiggle room when you said that there will be 100% chance earlier in the week.

---

While my computer modeling experience does not involve weather forecasting, I can tell you that all modeling is based around "Garbage in, Garbage out." Essentially, your model's prediction is only as good as the data you use to run it. Looking at a 10-day forecast is taking current weather conditions and atmospheric data, and extrapolating that information with assumptions that the current weather patterns will persist. As you get closer to the target day, current conditions are used to update the prediction. Because the weather is constantly evolving based on a wide range of variables, it is impossible to create 100% accurate predictions.

Scientists use statistical methods to estimate the probability of an event. These types of mathematical principles drive lay people nuts, as they aren't used to dealing so freely with such high levels of uncertainty. Also, as posted above, the rain percentage covers the entire area of the prediction. Essentially it will likely rain somewhere in the predicted area, but that doesn't mean that everyone will get precipitation.