|
|
Research
Seasonal Forecast of St. Louis
Encephalitis Virus Transmission, Florida
Jeffrey Shaman,*
Jonathan F. Day,† Marc Stieglitz,* Stephen Zebiak,‡ and Mark Cane*
*Columbia University, New York, New York, USA; †University of Florida,
Gainesville, Florida, USA; and ‡International Research Institute for Climate
Prediction, Palisades, New York, USA
Appendix B (Online Only)
Dependence in the Transmission
Number Category Data
Unfortunately, the sentinel chicken data was recorded in a form that
leaves it unclear as to which chicken in which cage becomes seropositive
(only the number of chickens is known). Additionally, the distribution
of the seroconversion data is not normal. As a result, reconstruction
of individual chicken time series and demonstration of independence or
even the level of dependence is difficult.
Several types of analysis, however, reveal that some level of dependence
may exist. Correlation of both the hard-clipped values of the weekly
times series of the number of chickens with HI antibodies at each sentinel
site, and the raw values, produces correlation coefficients that range
from r=0-0.6, suggesting there is dependence among some of the sentinel
sites. Furthermore, to examine within site dependence we compared the
distributions of the weekly number of seropositive chickens within each
flock (representing within site clustering of seroconversion events) and
the weekly number of flocks with one or more seropositive chickens (representing
between site clustering of seroconversion events) to Poisson distributions,
which would represent random clustering of such events in time. Both
empirical distributions were more clustered than the Poisson distributions;
however, the amount of clustering within and between flocks appeared similar.
These indications of dependence are accounted for in our empirical analysis
of the transmission number category.
|
