- journal_title：PALAIOS
- Contributor：Rebecca C. Terry
- Publisher：SEPM Society for Sedimentary Geology
- Date：2008-
- Format：text/html
- Language：en
- Identifier：10.2110/palo.2007.p07-071r
- journal_abbrev：PALAIOS
- issn：0883-1351
- volume：23
- issue：6
- firstpage：402
- section：ARTICLE
Raptors concentrate the remains of their small mammal prey in pellets rich in skeletal material. Stratified pellet deposits beneath long-term roost sites should, therefore, represent valuable archives of Holocene faunal change. Accurate paleoecological reconstruction from such deposits, however, requires a complete assessment of factors that may bias the ecological information that such records preserve. Three factors that could bias or obscure the community structure of a small mammal death assemblage relative to the living community include: (1) short-term transient dynamics of prey populations; (2) feeding activity of the raptors; and (3) extent of time averaging represented in individual stratigraphic horizons. Here I model (1) how much summed time is necessary for a raptor-derived small mammal death assemblage to capture a long-term (centennial to millennial) signal of relative abundance; and (2) the accuracy of the relative abundance information preserved in such death assemblages given short-term (decadal) cycling of small mammal prey populations. Results generated from an empirically parameterized model of prey dynamics assuming a multi-species type III functional response of raptors to fluctuations in density of two prey species suggest that the maximum extent of time averaging necessary to capture a stable relative abundance signal in a death assemblage is ∼140 years. This estimate is highly conservative, yet still remains fine enough to analyze phenomena operating at the centennial to millennial time scales critical for addressing long-term community response to habitat transitions through the Holocene. Results also suggest that the mismatch between relative abundance information in the living community and the death assemblage is generally low (<1%), except for a few specific parameter combinations that result in the population dynamics of the prey species being extremely similar to one another.