Anomaly relative to the climate without the extra human forcings, like how you could call the dead forest from acid rains an "anomaly". A simple average of a reference period is used as reference for simplicity but is by the fact that climate change is happening much faster than typical climate variations so it's fine for now: the difference is (still) essentially entirely out fault.
I think the point parent is making is that it is possible to cherrypick a reference period.
You can say that the melting of glaciers is an anomaly but when an ancient forest appears below the melted glacier how can we agree on what an anomaly is?
I don't want to deny global warming but I agree with the parent that it's difficult to state what an anomaly is.
It is true that you don’t want anomalies to be anomalous only with respect to some arbitrarily chosen (or sought) feature of the reference periods. The solution to this is to do sensitivity analysis of the anomaly to these ‘arbitrary’ features (length, starting/ending points). With most climate things, anomalies will persist. This is a standard thing to do.
