Digging through some literature this afternoon, I ran into a 2007 paper by Denise Su and Terry Harrison , who mounted several explanations for why Laetoli, Tanzania, has a relatively low abundance of Au. afarensis fossils compared to other sites. They suggest a size-sorting bias due to predation, which would disproportionately have affected postcranial bones:
The evidence supports the inference that, with increasing body size of the prey species, there is a greater chance that skeletal elements will survive complete destruction by carnivore scavenging. This results in a higher representation of postcranial elements relative to craniodental remains as body size increases. However, the converse means that species in the lower weight categories are increasingly susceptible to being entirely destroyed by carnivores. Hominins, which occur in the lowest end of the range for WC II, would be among those large mammals expected to be most heavily affected by the greatest number of species with the ability to completely remove skeletal elements from the skeletal assemblage at Laetoli. If this model is correct, then it is not unexpected that so few postcranial bones of hominins have been recovered. In fact, it is precisely what would be predicted at an open-air site with subaerial deposition in which the skeletal assemblage was readily accessible to carnivore scavengers.
They also point to the variation in the abundance of living chimpanzees and other primates in different ecologies.
Extant chimpanzees occur at a wide range of population densities across equatorial Africa according to habitat. Densities range from 0.08–0.09/km2 in open woodland (Ugalla, Tanzania; Mount Assirik, Senegal) to 3.1–4.7/km2 in closed woodland and forest (Gombe, Tanzania) (Plumptre and Cox, 2006), an almost sixty-fold difference between marginal and optimal habitats. This observation gives us a better appreciation of how different types of habitat can influence the biomass of African hominoids. It can be assumed that similar levels of population-density variation would have characterized A. afarensis across its geographic range. Given this fact, if A. afarensis at Hadar is indeed more common than at Laetoli, then it would suggest that Hadar had habitats that were more optimal for sustaining higher population densities of A. afarensis compared to Laetoli. The magnitude of the difference between the specimen counts between Hadar and Laetoli are equivalent to the difference in population density between modern-day chimpanzees living in closed and open woodland habitats, respectively.
Based on the fauna -- which interestingly include some rat genera now absent in East Africa but present in India and Southeast Asia -- along with other indicators, they infer that Hadar had more woody cover than Laetoli during the period represented by the hominins.