Semita: a novel cosmological filament finding algorithm

We introduce an algorithm aimed to identify large-scale filaments founded on the conception that these structures are bridges of matter that connect high density peaks. Our method is based on two standard tools, the Minimal Spanning Tree (MST) and the Friends of Friends (FoF) algorithm. Briefly, the complete process consists of five stages. Initially, we use the FoF algorithm to select intermediate density regions to stave off underdense zones (voids). Next, we build the MST by restricting it only to the regions defined in the previous step. Then the tree is pruned according to the length of its branches keeping the most dominant and discarding the more tenuous ones. Finally, the filaments are individualised according to the mass of the ends and smoothed using a B-spline fitting routine. To assess the results of the filament finder we apply it to a cosmological simulation. By focusing our analysis on those filaments whose halos at the ends have large masses, we found that the radial density profile, at scales around $1\, h^{-1} \mathrm{Mpc}$, follow a power law function with index -2. Even though the method only relies on halo positions, it is capable to recover the expected velocity field in filamentary structures. Large infall velocities coming from low density environment approach perpendicularly to the filaments and diverges toward to the ends. By studying the transverse velocity dispersion, we estimate the dynamical linear density following Eisenstein et al. (1997), finding a good correspondence between that and the actual mass per unit length of the filaments.