Silvia Conti

The researchers employed a global probability sample of 600 blocks, each measuring 5 × 5 kilometres, to map tree cover loss using satellite data with high resolution (ranging from 3 to 10 metres) and identify individual disturbances and subsequent land use changes within three years of the initial tree cover loss. The use of high-resolution imagery enabled the identification of pre-disturbance tree cover types, disturbance types, and post-disturbance land uses.

At the core of the problem is how sensors process and store calibration data. “Even if the raw analog-to-digital measurement were assumed to be noiseless, the compensated sensor output could still carry uncertainty,” explains Kaparounakis.

The study applies an advanced continuous change detection algorithm to NASA’s daily Black Marble NTL data. Over 1.16 million daily images were analysed to quantify the timing (day-of-year and year), intensity (area-averaged radiance change), type (abrupt or gradual), and direction (brightening or dimming) of ALAN changes for every 15-arc-second pixel (about 500 m at the equator) across the primary inhabited landmasses of Earth (70° N–60° S) from 2014 to 2022.