See here in this thread:

Difference between "Forecasting" and "Predicting" a hydrological variable? (researchgate.net/post/Difference-between-Forecasting-and-Predicting-a-hydrological-variable#:~:text=Forecasting%20pertains%20to%20out-of,used%20to%20estimate%20the%20regression.)

Weather is a good illustration of this difference. We forecast the weather by running a detailed computer model based on physics and meteorology starting at the present and extending out into the future with some assumptions about the boundary conditions. We predict the price of gas (or Twitter stock) next week based on the previous six weeks and a few assumptions (shape of the curve). I have done a lot of work with hydrological parameters, working as an applied mathematician with geologists and hydrologists. In that case I had to infer these parameters (which we can't directly measure) given data that we were able to measure. That was an exercise in nonlinear constrained optimization.