Temporal variation in habitat structure and shrubsteppe bird dynamics.

Abstract

We investigate the role of temporal variation in habitat physiognomy in influencing the dynamics of shrubsteppe bird populations and communities. During a 3-y (1977-1979) study of 14 sites in the northern Great Basin of North America, annual precipitation varied substantially, with one of the driest years on record followed by one of the wettest. This resulted in significant physiognomic variation (increasing height and coverage of vegetation, decreasing horizontal patchiness), mediated largely by changes in the annual elements of the flora, particularly forbs and grasses. Shrub species coverage values, on the other hand, demonstrated no statistically detectable year-to-year changes, nor were they correlated with any physiognomic variation. Despite large scale physiognomic changes, no bird species' abundance varied in a statistically significant fashion; neither could variation in bird abundances be correlated with variation in either physiognomy or shrub species coverages.Multivariate analyses revealed essentially the same patterns as the univariate analyses: substantial changes in physiognomy, few changes in shrub species coverages or bird species abundances, and little correlation of temporal variation among the three data sets. Calculation of the Euclidean distances that sites "moved" in multivariate physiognomic, bird species, or shrub species hyperspaces yields synthetic gradients of annual "turnover" of sites with respect to those data sets. Sites identified as demonstrating high physiognomic turnover were characterized by high coverage of grass and forbs, while low turnover sites had greater coverage of shrubs and higher shrub species diversity. Relatively high bird turnover sites had greater numbers of Western Meadowlarks and Black-throated Sparrows, while more stable sites had high numbers of Brewer's Sparrows. Physiognomically, high bird turnover sites were grassier and had greater total vegetation coverage, while low bird turnover sites had more bare ground and higher horizontal patchiness. A site's position on the avifauna turnover axis, however, was uncorrelated with its position on the physiognomic turnover axis. Shrub species showed virtually no annual turnover.Reanalysis of a previous Principal Components Analysis (PCA) of these same data sets that was applied without regard to year of sampling revealed that the first physiognomic component (41% of the total physiognomic variability) did in fact have a strong temporal element, and that this element was consistent with the changes in univariate characters noted above. No other physiognomic component could be associated with annual variation, nor could any components of parallel bird abundance or shrub species coverage PCA's.Regional patterns indicate that sites tended to be very consistent from year to year in their relationships to one another as defined by their relative locations in either physiognomic or shrub species hyperspaces, but varied independently of one another with respect to their bird species abundances and composition. Examination of the temporal consistency of site relationships between "bird space" and "vegetation space" reveals that bird communities are to a large degree independent of a site's physiognomic position, but instead are strongly associated with its position in shrub species hyperspace.The overall patterns that emerge from these analyses are consistent with the so-called "checkerboard effect" that results from the apparently random annual redistribution of individual birds, and leads to the conclusion that populations of shrubsteppe birds are not existing at maximum density or "carrying capacity." Such observations are consistent with contentions that these populations lack close biological coupling with coexisting species and that interactions among these species (e.g. competition) are likely to play little if any role in the organization of their communities.