Increasing forecast length equates to increasing uncertainty. For example, I can tell you with 100% certainty that it will not rain in the next 10 mins at my house (notice there is a time and space qualifier). Time and space are linked. Also, some regimes are more stable then others and this leads to more certainty in the forecast. Fortunately, we can tell which regimes, to a pretty good degree, are inherently less predictable. Forecasting is based on models that solve nonlinear PDEs as an initial value problem, via numerical methods that approximate the continuous equations, where the initial state has uncertainties and the system is inherently chaotic. Plus some of the physics are not well understood and are therefore difficult to accurately model.
All forecasts have either an implicit or explicit probability associated with them. On average, there is little deterministic skill beyond 4-5 days for precip forecasts. It is better in the cold season than the warm season. Seasonal (3 month) outlooks show some skill for regional (time and space again) above or below average precip and temperatures. There are some biases in the forecast the public hears. Forecasts you hear on TV (via the interpretation of model results) tend to be "wetter" than they should be. Progress in modeling tends to be slow and incremental but the increase in the fidelity of forecasts compared to the 70s is nothing short of amazing.
Quote:
Originally Posted by rubicon
My working for nephew is a meteorologist NOAA when he was in college and I asked him to give me a weather report for my interstate travel home he was precise and absolute. Once he graduated and began to work in his field these "qualifiers"popped up mostly , probable 50% chance. I asked him why the change when he was working for Accu weather. He replied because he had rich clients who wanted weather forecast before they sailed their yachts.
Please hire one handed meteorologists  |