GLAD-S2 (Sentinel-2 Deforestation Alerts)

GLAD-S2 (Sentinel-2 Deforestation Alerts) is an event-based near-real-time forest disturbance alert product derived from Sentinel-2 multispectral imagery at 10 m native spatial resolution. Produced by the GLAD laboratory at the University of Maryland and distributed via Global Forest Watch (GFW), the product extends the Landsat-based GLAD-L methodology described in Hansen et al. 2016^[1] to the higher spatial resolution of Sentinel-2.

Confirmed coverage as of the research date encompasses Amazon Basin primary humid tropical forest^[2]. GFW marketing materials state a pan-tropical footprint was planned for expansion by end of 2021^[3], though that broader coverage extent has not been independently verified from a tier-1 source as of 2026-05-18.

Alerts are available through the GLAD Earth Engine Applications portal^[4] and Global Forest Watch^[3] under CC BY 4.0. The product is event-based and designed for detection of individual clearing events; it is not a suitable input for forest area estimation. Cloud cover during Sentinel-2's nominal 5-day revisit extends effective detection latency for cloud-persistent scenes. Alert confidence tiers distinguish probable detections (single observation) from confirmed detections (repeat observations)^[2].

An integrated multi-system analysis published in 2024 demonstrated improved detection timeliness and confidence when GLAD-S2 outputs are combined with complementary alert systems^[5]. No dedicated peer-reviewed validation paper for GLAD-S2 specifically has been identified; the product inherits the GLAD-L validation framework.