Solar Forecast over a day

Below is the solar forecast over a day for a PREVIOUS day.

Why does it change during the day? Does it continuously take weather into account?

Can I then assume that the last datapoint of a day, in this case 4.24kWh is the most a curate prediction? In other words, no unknowns should be left except the constants that define the installation (efficiencies, azimuth, angles etc etc and assuming complete clear view on the sky)