A Dynamical Classification of the Cosmic Web

A dynamical classification of the cosmic web is proposed. The large scale environment is classified into four web types: voids, sheets, filaments and knots. The classification is based on the evaluation of the deformation tensor, i.e. the Hessian of the gravitational potential, on a grid. The classification is based on counting the number of eigenvalues above a certain threshold, lambda_th at each grid point, where the case of zero, one, two or three such eigenvalues corresponds to void, sheet, filament or a knot grid point. The collection of neighboring grid points, friends-of-friends, of the same web attribute constitutes voids, sheets, filaments and knots as web objects. A simple dynamical consideration suggests that lambda_th should be approximately unity, upon an appropriate scaling of the deformation tensor. The algorithm has been applied and tested against a suite of (dark matter only) cosmological N-body simulations. In particular, the dependence of the volume and mass filling fractions on lambda_th and on the resolution has been calculated for the four web types. Also, the percolation properties of voids and filaments have been studied. Our main findings are: (a) Already at lambda_th = 0.1 the resulting web classification reproduces the visual impression of the cosmic web. (b) Between 0.2 < lambda_th < 0.4, a system of percolated voids coexists with a net of interconected filaments. This suggests a reasonable choice for lambda_th as the parameter that defines the cosmic web. (c) The dynamical nature of the suggested classification provides a robust framework for incorporating environmental information into galaxy formation models, and in particular the semi-analytical ones.