A few years back, I was wondering what sort of temperatures were realistic for welding bladesmithing-type steels and nobody seemed to be able to give a firm answer. I took a type S thermocouple and pyrometer along to a hammerin where a number of good smiths made quite a lot of Damascus over a couple of days, using a vertical gas forge that one of them had brought along. He fired it up and adjusted it how he wanted it, then everyone used it. Nobody said it was too hot and nobody said it was too cold. All the comments that I heard were to the effect that it was good and hot. Late on the Sunday, I stuck the thermocouple in and took some measurements. Temperature took a bit of a dive when a workpiece was first put in, but quickly recovered to around 1285 degC/ 2345 degF. Typical temperature when the workpiece was removed was between 1295 and 1310 deg C (2363 to 2390 degF). With no workpiece in, the temperature reached 1330 degC (2426 degF).
As is the way of such things, curiosity got the better of us and it got turned up to see what it would do (pressure set to maximum and choke adjusted to give peak temperature). We got 1440 degC (2624 degF) where the workpiece normally sits, with a spot reading of 1470 degC (2678 degF) an inch or so below. The experiment melted the lining.
Whilst it is clearly possible to weld at lower temperatures, the consensus seemed to be that 1300 degC (2340 degF) was about right and that if welds were not good at that temperature it was not because of the temperature.
Identified steels being welded included 15N20, 20C and EN42, along with a pretty varied assortment of other stuff.