calculating power

I am currently setting up a prospective cohort study. In wave 1, I will be recruiting 2000 participants and average response rate is 70% so I expect 1400 participants at wave 2. I have a continuous predictor variable, which I will be splitting into 4 quartiles and 3 outcome variables. I have attempted to calculate power in G*Power using the logistic regression function, however, I fear that this can only give me a power range (power at wave 1 and power at wave 2) and Im missing the power for the interaction between time 1 and time 2. Has anyone calculated power for a cohort study before and can tell me what I am doing wrong? Sorry I can't give any more information about IVs and DVs, I need this to be confidential.

Hope you can help!