Main Title |
Smoothing of Breeding Bird Survey Data to Produce National Biodiversity Estimates. |
Author |
Yang, K. S. ;
Carr, D. D. ;
O'Connor, R. J. ;
|
CORP Author |
George Mason Univ., Fairfax, VA. ;Maine Univ. at Orono.;Corvallis Environmental Research Lab., OR. |
Publisher |
1996 |
Year Published |
1996 |
Report Number |
EPA/600/A-96/102; |
Stock Number |
PB97-123145 |
Additional Subjects |
Birds ;
Biological indicators ;
Surveys ;
Bird populations ;
Habitats ;
Environmental monitoring ;
Statistical analysis ;
Species diversity ;
Statistical analyses ;
Reprints ;
Biodiversity
|
Internet Access |
|
Holdings |
Library |
Call Number |
Additional Info |
Location |
Last Modified |
Checkout Status |
NTIS |
PB97-123145 |
Some EPA libraries have a fiche copy filed under the call number shown. |
|
07/26/2022 |
|
Collation |
8p |
Abstract |
The Breeding Bird Survey (BBS) is an annual survey Conducted since 1966 at approximately 3000 25-mile long routes within the continental U.S. The purpose of the present analysis is to derive from the BBS a biodiversity measure based on incidence of 615 species. This analysis purposely avoids use of concomitant information concerning habitat because a later analysis will relate the biodiversity to habitat. A preliminary analysis of abundance and prevalence data focused attention on 342 species at 1870 routes in the 1990 data. The first smoothing step addressed a problem in census efficiency; occasionally observers fail to detect species present on route. The smoother used a distance-limited median-based near-neighbor smoother to impute bird occurance where birds were not observed. To provide a basis for later comparison against habitat indicators, the second smoothing step used near-neighbor logistic regression to estimate probabilities on the US Environmental Protection Agency's Environmental Monitoring and Assessment Program hexigon grid. The biodiversity measure was the number of species of birds exceeding a specific threshold of probability at each EMAP grid point. Expert review of single year smoothing identified a second issue. Species could be present on a BBS route in some years, but not others. Review of long-running BBS routes suggested a ten-year period provides a stable estimate of species' presence at a route. Almost all species appearing at a route are detected sometime during ten years. This motivated a new analysis based on a ten-year window from 1981-1990. The reanalysis used a logistic regression on 615 species of birds at 1909 routes. The paper presents selected maps from the first and second smoothing efforts. This paper provides some background on interpretation of results but primarily emphasizes computation details. Later papers will delve into interpretation and model criticism. |