I am analyzing my experiment and I have encountered a doubt. It is a split plot design, agriculture experiment where main plot is received four different treatments (incl. control) and four subplots with four varieties. Area under disease development curve as a response, calculated from percentage of diseased tissue based on six weekly assessments.
I am not clear how to treat my replicates, as replicates or as blocks? Does it make sense to treat those as blocks, since those were in the same field (four strips and within each 16 subplots). I will have this experiment for the next two years and those should be treated as blocks. Now if I take these as replicates, should I calculate the mean of each disease status at each measurement and calculate the area, or take the average of four areas calculated per replicate?
Going through literature I have realized that I could try to work on repeated measures model and also test the influence of weather variables on disease development. Any suggestions or advice on this? How to deal with correlation here?