Is overdispersion unidentifiable if each survey site is only visited once?

Question

I am creating a binomial logistic regression model with two variables: average tree height within a survey site and tree density within that same site. For each site, I only have one measurement for each variable. In this case, is it correct that I cannot measure overdispersion because I cannot group the data into subsets?