Perry Staltic
Member
- Joined
- Dec 14, 2020
- Messages
- 8,186
When hurricane forecasters attempt to predict a hurricane's path, they run numerous models that predict similar, but different paths based on each models' programming, input data and assumptions. Each run of a set of models is called an ensemble. As the hurricane moves, each model starting point is reset to the hurricane's current position and another ensemble is run. This image shows how the starting position for each model is reset to the hurricane's current position (feedback). All of those lines represent 51 different model runs. Without the realtime feedback the models would be garbage. Continuous feedback anchors a model in reality and gives it predictive ability.
Climate forecasters do the same thing with one exception: they never reset their model run starting points to current temperatures. They just let the models rip and then average all of them together, which becomes the forecast. In the image below the thick black line represents the average of the 73 model runs. But notice how far it is from reality, which is represented by blue and green circles and squares (satellite and balloon measurement data). Since it doesn't incorporate feedback, it's not anchored in reality and is just junk science with no forecasting ability.
Climate forecasters do the same thing with one exception: they never reset their model run starting points to current temperatures. They just let the models rip and then average all of them together, which becomes the forecast. In the image below the thick black line represents the average of the 73 model runs. But notice how far it is from reality, which is represented by blue and green circles and squares (satellite and balloon measurement data). Since it doesn't incorporate feedback, it's not anchored in reality and is just junk science with no forecasting ability.
Last edited: