Upcoming seminar on species distribution models

Monday, November 17, 2014

Yellow Bellied Flycatcher. Photo: Bill DeLuca

On December 4th, Jim Nichols from the USGS Pautuxent Wildlife Research Center will visit the NE CSC to present, "Static Species Distribution Models: Adequate for Description, But Not For Prediction."

  The talk will be held in 134 Morrill Science Center Conference Room on Thursday, December 4 at 1:00pm.

Bio:  Jim Nichols has spent his career as a research scientist at Patuxent Wildlife Research Center, Laurel, Maryland, where he worked first for the U.S. Fish and Wildlife Service and now as a Senior Scientist with the U.S. Geological Survey (Biology). He has broad interests that generally focus on the dynamics and conservation/management of animal populations and communities. Nichols has worked on a variety of taxa, with methodological emphases on population modeling, biostatistics and decision theory.

Abstract:  Ecologists frequently try to predict the future geographic distributions of species.  Most studies assume that the current distribution of a species reflects its environmental requirements (i.e., the species’ niche). However, the current distributions of many species are unlikely to be at equilibrium with the current distribution of environmental conditions, both because of ongoing invasions and because the distribution of suitable environmental conditions is always changing. This mismatch between the equilibrium assumptions inherent in many analyses and the disequilibrium conditions in the real world leads to inaccurate predictions of species’ geographic distributions and suggests the need for theory and analytical tools that avoid equilibrium assumptions. Here, we develop a general theory of environmental associations during periods of transient dynamics. We show that time-invariant relationships between environmental conditions and rates of local colonization and extinction can produce substantial temporal variation in occupancy-environment relationships. We then estimate occupancy-environment relationships during three avian invasions. Changes in occupancy-environment relationships over time differ among species but are predicted by dynamic occupancy models. Since estimates of the occupancy-environment relationships themselves are frequently poor predictors of future occupancy patterns, research should increasingly focus on characterizing how rates of local colonization and extinction vary with environmental conditions.