Regression
quantiles is an especially useful technique for describing the variation
in heterogeneous data distributions, a commonly observed pattern in ecological
data when all important processes are not measured. Furthermore, the utility
of statistical models increases if their implications can be visualized
geographically. To identify appropriate spatial applications of regression
quantile models for grassland nesting birds, we used a moving window technique
that applied model coefficients across the Boulder, Colorado landscape.
The moving window allows quantification of the neighborhood of each cell,
calculation of regression quantile changes in slope, and assignment of
predicted abundance levels for each map unit (e.g., grid cell). Our statistical
models include independent variables that describe the landscape in and
around the Boulder Open Space.
U.S. Department of the Interior | U.S. Geological Survey
URL: http://www.fort.usgs.gov/resources/spotlight/blossom/blossom_grasslandbird.asp
Page Contact Information: AskFORT@usgs.gov
Page Last Modified: 11:42:37 AM
