I was wondering whether it's appropriate to analyze multiple years data with multiple repeated obs by using GLIMMIX procedure. I have a weekly based mosquito database for 5 years and a weekly-based weather station record. The mosquito was collected from 20 traps every week for 5 years (multiple observations for each trap/week). The weather information will be the same for each trap for specific week. I set those traps (specified by IDs) as repeating measurements and put them in the mixed model analysis. I use GLIMMIX because I wanna adjust for the spatial autocorrelations of traps by using sp(exp) covariance matrix structure.
I understand that it is not normal for repeated measurement analysis but I could not think about other more appropriate way to analyze it. Is there any statistical problem about this model? Thanks
proc glimmix data=MODEL;
model mosquito= temp prec rh / cl solution stdcoef;
random intercept /subject=id solution
type=sp(exp) (longitude latitude);