Simplified version of my model:

glm(cbind(young, adults) ~ month + effort, family = "binomial")

i.e., I study proportion of young as a dependent variable on month (or season), taking into account observer effort. However, the observer effort is dependent on the month:

How to solve this problem? I want to take into account both variables.

I was looking into literature but haven't found any solution. My naive idea is to compute mean effort for each month and instead of taking effort, take difference between effort and this mean. But this is just a naive idea, I would like to hear your advice. Thanks!