All of our sources suggest that Roman armies on the march constructed the castra whenever it was deemed necessary, and especially in hostile environments. Having said that, the vast majority of our sources (for any Roman period) are frustratingly vague on the details of campaigns, deployment, etc. Caesar gives the most detail, and even so, he often leaves the details vague. In Caesar, building the camp seems to be the default preference on a "normal" day of march. Both of his major accounts, however, are spent almost exclusively in hostile territory. Would a legion build a castra while traveling between Brundisium and Capua? I think probably, yes, just because soldiers are slaves to routine, but I'm not sure where or how to find evidence for that. Livy, maybe.

The castra could take many forms, from a quick enclosure surrounded by a simple ditch, to a complex series of fortifications with watchtowers and reinforced gates. We hear about all sorts in the sources, and see many variations in the archaeological record. Some were built carefully, some in great haste and under dire circumstances. Despite the differences, their basic function was always the same: to provide temporary security to the army and its supplies; to act as point of departure (ie deployment to the battlefield) or return (in event of a withdrawal); to create a strong point or fixed feature on the battlefield. One of the major (and often unmentioned) functions was to offer some basic security for getting rest. Tired armies don't fight well, and the castra allowed (at least in theory) for at least some of the soldiers to get some much-needed shut-eye after a long march.

Given the great variability in construction, scarcity of resources was not a problem. All you needed for a basic castra was a trusty shovel, some sharpened stakes, and a lot of hard labor. Each soldier knew his part of the fortifications, and once the anchor points were set down by the army engineers, the individual soldiers worked in concert to build the thing very rapidly. When a couple of thousand people are digging a ditch at the same time, it happens pretty fast. The earth from the ditch was piled behind (inside) the ditch, creating a modest rampart. If timber and time were available, you could elaborate from there, but the basic ditch and rampart were sufficient in an emergency. The only real geographical obstacle to the basic construction would be frozen ground, or bedrock. We have the remains of Roman castra all over the place, from the northern reaches of Scotland to the desert under Masada. There are some really elaborate, "permanent" examples on the Danube frontier

Catherine Gilliver, in her excellent coverage in Cambridge History of Greek and Roman Warfare V2, does a good job at quickly collecting some evidence for Roman camp-building mentality.