It’s proverbially true that an observed correlation between two variables does not imply there is some causal relation between the two…but what about the other way round? What if there is a known relation between two variables (for example from physics, laboratory experiments etc) does this imply there will be a correlation between the variables?
The answer is no, sort of. The point about correlation is that it is an observation of apparent relations between data sets. Two phenomenon can have an actual connection but the outcome might not be observable. Why not? Well, just because something exists doesn’t mean that your data collection methods can observe it: they might not be sensitive enough (e.g. your telescope might not be big enough) and closely related to that lots of other things will also have causal connections.
For example, the physics connecting CO2 and atmospheric temperature is very old, much older than the consensus among climate scientists that CO2 emissions are causing global warming. Why? Well in part because for much of the Twentieth-Century the connection was not easily observable — greenhouse gases are not the only drivers of surface temperature. Postwar data collection improved and the impact of rising CO2 (and other greenhouse gases) also became more prominent in surface temperatures and the impact of other factors (such as particles causing cooling) became better understood.
Put another way causation implies correlation but only if all other factors can be controlled (or accounted for) and your data collection is good enough.