Volumetric Lightmaps are per-pixel sampling of probes of various sizes, based on the density and number of static meshes. It only provides indirect lighting and sky occlusion(if you bake the Skylight). So you won’t get any contact shadows, baked AO, or multi-bounce light that you’d otherwise get in surface lightmaps with Lightmass. It’s a good tradeoff for large meshes that would require multiple parts and/or large lightmaps, or like Tilmann said, foliage that shouldn’t be baked, but provided that you have enough sample density, the indirect lighting is almost indistinguishable between surface and volumetric lightmap objects.