To explain further:
For electrical grade enamelled copper wire, the CSA (cross section area D2𝝅/4) only determines the resistance per unit length. The type of coating on the wire determines its max. service temperature. For wires in free space or in small bundles the temperature rise can be calculated from the RMS current through the wire(s), so, given the ambient temperature, it is possible to determine the max. continuous RMS current a wire may safely carry.
However in transformers the temperature rise is largely determined by the total losses in the transformer, its shape and surface area, as the thermal resistance transformer surface to ambient typically dominates (except for actively cooled transformers) though the RMS current through the wire does contribute to localised heating due to the thermal resistance winding to core or other windings. Transformer manufacturers generally try to balance the I2R losses in the primary and secondaries to avoid hotspots, but are limited by the minimum wire gauge their winding equipment can handle which may result in the primary or high voltage secondaries being heavier gauge than ideal, and by the wire gauges they actually stock. Also the length of wire for N turns is determined by the distance from the core, so a winding at the surface will have higher losses. Therefore a particular wire gauge secondary can have different ratings in different transformer and can even have different current ratings in the same multi-secondary transformer depending on where it is in the winding stackup.