On correcting the time-lag bias in aerial-surveyed bark beetle infestation data