Ecological neighborhoods of small mammal communities at flint hills national wildlife refuge, Kansas :a multi-scale comparison of community and habitat structure.

Abstract

A multi-scale investigation was conducted to examine patterns of similarity among small mammal communities at the Flint Hills National Wildlife Refuge, Kansas, and determine at what scales these patterns of similarity correlated with habitat and landscape characteristics. Application of the ecological neighborhood concept requires scaling an investigation,l to a particular ecological process over an appropriate time period, and to the response or influence of the organism during that time. Although multiple processes determine small mammal community structure it is possible to infer the extent of the ecological neighborhoods by evaluating patterns of similarity of small mammal community structure and their environmental correlates. Monthly small mammal and vegetation surveys were conducted from June 2004 to August 2005 at nine sites representing three series of successional grasslands. A total of 5302 individuals consisting of 11 small mammal species was captured over 22,860 potential trapnights. Multi-scale spatial analysis was conducted at station, site, and landscape levels and temporal analysis by season and duration of the study (15 months). Ordination (nonmetric multidimensional scaling) and subsequent correlation analyses demonstrated a relationship between small mammal community structure and vegetation structure. Although small mammal community structure correlated with vegetation at site level, these correlations also existed at the station level, with more corresponding axes between the two ordinations. Thus, station-level vegetation was a better predictor of small mammal community structure than successional type, suggesting that the processes relating small mammal communities to vegetation operate at a scale finer than the patch. The general lack of association between landscape characteristics and small mammal community structure suggests differential access was not a determining factor in this system, except during the April-June season. During that time, it appears that larger scale processes influence small mammal communities than are acting during other times of the year.