I've seen data like this before. My research focused on acquiring/processing both vertical and horizontal component data in subsurface environments characterized mainly by saturated and/or unconsolidated facies. So, it's entirely possible even for a that kind of surface-wave to display such low apparent velocities.
What I think you are seeing at that particular offset and traveltime is the Rayleigh-wave beginning to 'disperse' - AKA seismic dispersion beginning to manifests itself as a results of the wavefield convolving with layering and near-surface heterogeneity.
As for a refraction arrival, if it's indeed present, it should be extremely weak and probably only discernible via some kind of gain (e.g. AGC).