I think you are looking at it in a linear way. But maybe the relationship between scale and causal information is nonlinear. Causal information might oscillate as the scale increases. So that physics, chemistry, and bioligy are concentrated around scales that are associated with peaks in causal information.