Modified Mann-Kendall using hourly dataset?

Hi have a ten year dataset of hourly readings of water quality parameters (pH, temperature, dissolved oxygen, turbidity and specific conductivity). Since the data is on an hourly basis, it probably has a high autocorrelation. Therefore, I would like to run the modified (and seasonal) Mann Kendall. I have read the advantage and disadvantage of the MKD, MKDD and MKRD (Blain 2013). QUESTION; can I run the MKD, MKRD or MKDD with an hourly data set or should I transform it somehow (daily weekly or monthly mean, median, min or max)? Thanks!