The detectability of Arctic leads using thermal imagery under varying atmospheric conditions

The way in which thermal contrast between leads of varying widths and thickness can be distinguished from the background multiyear ice surface under varying atmospheric conditions is examined. The normalized brightness temperature difference between image pixels that include lead fractions and of the background ice is employed to determine thresholds of detection accounting for sensor FOV and various atmospheric phenomena that influence the Arctic radiation balance during winter. Surface temperatures are prescribed as a function of ice thickness, and the effects of the intervening atmosphere are simulated by varying the optical depths of hypothetic cloud or haze layers varying in microphysical characteristics. It is found that the limits of lead detection may be determined as a function of pixel lead fraction and atmospheric optical depth if suitable values of normalized contrast are used as detection criteria.