GLAD forest loss alerts
We operate two alert systems that detect loss of tree cover. GLAD-L alerts map tree cover loss in near real-time pantropically (30°N-30°S) and are based on Landsat data (formerly just GLAD alerts). GLAD-S2 alerts map tree cover loss in the primary humid tropical forest within the Amazon basin region.
GLAD-L
Since the opening of the Landsat archive in 2008, medium spatial resolution data have been available for use in alert-based applications. Since 2013, two Landsat sensors, the Enhanced Thematic Mapper Plus (ETM+) onboard Landsat 7, and the Operational Land Imager (OLI) onboard Landsat 8, have been systematically acquiring global multi-spectral observations at a 30m spatial resolution. The orbits of the two spacecraft are coordinated to enable potential 8-day repeat coverage globally. Given this cadence, the use of Landsat as a near-real time source of land change information is possible. The data displayed and made available here quantify forest disturbance events for the tropics using Landsat ETM+ and OLI data as an input. Daily updates are made for areas where quality land observations are acquired. We define forest cover as 5m tall trees with a canopy closure exceeding 30%. An alert is defined as any Landsat pixel that experiences a canopy loss in excess of 50% cover. For the full description of forest alert methodology please refer to the following publication: Humid tropical forest disturbance alerts using Landsat data (Hansen et al. 2016)
GLAD-S2
Loss of primary forest is mapped in near-real time at 10m resolution using Sentinel-2 multispectral data. Cloud, shadow, and water are detected in each new Sentinel-2 image and a forest loss algorithm is applied to all remaining clear land observations. The algorithm relies on the spectral data in each new image in combination with spectral metrics from a baseline period of two prior years. Confidence is built through repeated loss observations in the consequent images. Results are mapped only within the primary forest mask of Turbanova et al. (2018) with 2001-2018 forest loss from Hansen et al. (2013) removed. This system extends the methods of GLAD-L to Sentinel-2 data.
Use of the data
These alerting systems are designed to be an early indicator of forest loss to aid forest management activities and enforcement. These systems are meant to complement GLAD’s current annual global forest cover loss product, implemented in collaboration with Google and the World Resources Institute as part of Global Forest Watch. The annual product is based on a calendar year update, first prototyped using Landsat 7 data from 2000 through 2012 and annually updated thereafter (www.globalforestwatch.org and https://glad.earthengine.app/view/global-forest-change). The interim alert systems provide forest disturbance updates as observed, with the objective of providing conservative change detection with minimal errors of commission. In other words, the alert system should not be used to provide area estimates, but event-based information for areas experiencing forest cover loss. Such information is useful for forest management activities and enforcement.
Dataset access
Alerts from both systems can be accessed on Global Forest Watch. The GLAD-L and GLAD-S2 alerts are also available for visualization and download at GLAD maintained pages: GLAD-L and GLAD-S2.
To access the data within Google Earth Engine here are example scripts: GLAD-L, GLAD-S2.
If you are experiencing technical difficulties viewing or downloading the data, please contact the UMD GLAD team at glad@umd.edu.
Dataset reference
Hansen, M.C., Krylov, A., Tyukavina, A., Potapov, P.V., Turubanova, S., Zutta, B., Suspense, I., Margono, B., Stolle, F., Moore, R. (2016) Humid tropical forest disturbance alerts using Landsat data. Environmental Research Letters, 11, 034008. https://doi.org/10.1088/1748-9326/11/3/034008
