Each of the 12 colonies is sampled on its own staggered cadence: population every inspection interval, Varroa load every sampling interval — matching real practice, where mite counts run on their own schedule. A colony is flagged red the first time a check finds its mite load M above the threshold T, and amber when population falls in two consecutive checks (ΔP<0 twice running). The real mite level rises continuously underneath, but the record only "knows" about it at the next scheduled check — the gap between the two is the whole point of a surveillance programme: detection lag = week_recorded − week_true_level_crossed(T) Shorter sampling intervals and a lower threshold shrink that lag; a flagged colony is treated at its next check, and the mite count resets.