Late Cenozoic burial history and dynamics of the Northern Caucasus molassebasin: implications for foreland basin modelling

The collisional history of the Caucasus segment of the Alpine-Himalayan fol d belt started at the end of the Eocene. The associated development of the Northern Caucasus foreland basin occurred during two syn-collisional stages , each displaying different subsidence patterns. The 34-16 Ma (Maikopian or pre-foreland stage) displays a long-wavelength subsidence of a broad area whereas the 16-0 Ma (foreland stage) displays asymmetrical foreland subside nce. The first is correlated with the termination of subduction in the sout hern area and can be associated with mantle flow induced by the re-equilibr ation of the subducted slab. The along-strike configuration of the second, molasse basin, stage contradicts the hypothesis in which topographical load ing is considered to be the main control on foreland subsidence. There is a clear anti-correlation between basin depths and orogen heights: the deepes t parts of the basin are at the tips of the orogen, where mountain heights are negligible while the area adjacent to the highest mountains (Central Ca ucasus) is uplifted. The influence of other types of loading in foreland ba sin development has therefore been investigated and it has been possible to find a good fit with both gravity and basin architecture by including only two additional model parameters into the model. These are crustal and lith ospheric thickening/thinning The results demonstrate that crustal thickenin g and removal of Lithospheric roots are responsible for supporting the high Central Caucasus Mountains and uplift of adjacent areas. The subsidence of the basins at the orogen tips is explained by loading of Lithospheric root s. Both effects are important in the geographically intermediate areas. in general, it is concluded that topography should not be considered as the ma in control on foreland subsidence, but only as one of several counterbalanc ing mechanisms. The existing flexural model therefore needs improvements su ch that it can be related directly to 'subsurface loading' by real structur es and processes during collision. (C) 1999 Elsevier Science B.V. All right s reserved.