The study of prehistoric sites and the ecological and cultural information it provides is often punctual, residual, and related to local context (local sedimentation, site specificities, choices).

Such parameters are complex to extrapolate at the regional scale because they rarely bear global signal. On the other hand, global parameters such as climatic ones (palaeo-temperatures, topography, precipitations) are difficult to downscale, and make high-resolution models complex to handle.

The recent advances in quaternary geomorphology are currently allowing researchers to reach new understandings of landscape evolution and the associated dynamic processes at the scale of the valleys, on middle-term evolution (last glacial period, Holocene) as well as for long-term evolution (multiple glacial-interglacial stages). Processes such as the specific evolution of past hydrosystems and past erosion dynamics may have a noticable impact at the scale of the sites (evolution, preservation, generation of highly visible landscape features), and also at the scale of the valleys (behaviour of the hydrosystems, vegetation repartition, faunal preferential traffic), and questions de ways ecosystems can differ from one valley to another, according to local factors.

We present some examples of valley landscape evolution in Europe during glacial / interglacial periods, especially focusing around the Ardèche valley (France), and discuss the potential impact of some processes in the fauna dynamics, and in the end, past human choices.

We will try do discuss how taking into account the specificities at the scale of site's surroundings and at the scale of the valley can play a role – or not – in the modeling of past dynamics, and be considered - or not - in the definition of eco-cultural niches.