RAdar for Detecting Deforestation (RADD) is a deforestation alert product that uses data from the European Space Agency’s Sentinel-1 satellites to detect forest disturbances in near-real-time . The RADD alerts use a detection methodology produced by Wageningen University and Research (WUR), Laboratory of Geo-information Science and Remote Sensing. These alerts are particularly advantageous in monitoring tropical forests, as Sentinel-1’s cloud-penetrating radar and frequent revisit times (6-12 days) allow for more consistent monitoring than alert products based on optical satellite images. Alerts are available for the primary humid tropical forest areas of South America, sub-Saharan Africa and insular Southeast Asia at a 10m spatial resolution, with coverage from January 2019 to the present for Africa and January 2020 to the present for South America and Southeast Asia. Pre-processed Sentinel-1 images are collected from Google Earth Engine, then quality controlled and normalized using historical time-series metrics. Forest disturbance alerts are then detected using a probabilistic algorithm. Each disturbance alert is detected from a single observation in the latest image if the forest disturbance probability is above 85%. If the forest disturbance probability reaches 97.5% in subsequent imagery within a maximum 90-day period, alerts are then marked as "high confidence". The product has a minimum mapping unit of 0.1 ha (equivalent to 10 Sentinel-1 pixels) to minimize false detections. Alerts are detected within areas of primary humid tropical forest, defined by Turubanova et al. (2018) and with 2001-2018 forest loss (Hansen et al. 2013) and mangroves (Bunting et al. 2018) removed. For more information on methodology and validation, please refer to Reiche et. al. (2021). The version presented here (v1) has been updated from that described in the paper (v0), with changes to the forest mask and a reduction of the minimum mapping unit. The RADD alerts were made possible thanks to the support of a coalition of ten major palm oil producers and buyers. Under the project, Wageningen University and Research (WUR) developed the detection method and Satelligence first scaled the system in Indonesia and Malaysia and provided additional prioritization of alerts for on-the-ground follow up. Additional support was provided by the US Forest Service and Norway’s International Climate and Forest Initiative. The alerts are currently generated by WUR using Google Earth Engine.*This data product utilizes a special encoding*Each pixel (alert) encodes the date of disturbance and confidence level in one integer value. The leading integer of the decimal representation is 2 for a low-confidence alert and 3 for a high-confidence alert, followed by the number of days since December 31, 2014. 0 is the no-data value. For example:20001 is a low confidence alert on January 1st, 201530055 is a high confidence alert on February 24, 201521847 is a low confidence alert on January 21, 20200 represents no alert

This data set, created by the GLAD (Global Land Analysis & Discovery) lab at the University of Maryland and supported by Global Forest Watch, is the first Landsat-based alert system for tree cover loss. While most existing loss alert products use 250-meter resolution MODIS imagery, these alerts have a 30-meter resolution and thus can detect loss at a much finer spatial scale. The alerts are currently operational for select countries in the Amazon, Congo Basin, and Southeast Asia, and will eventually be expanded to the rest of the humid tropics.New Landsat 7 and 8 images are downloaded as they are posted online at USGS EROS, assessed for cloud cover or poor data quality, and compared to the three previous years of Landsat-derived metrics (including ranks, means, and regressions of red, infrared and shortwave bands, and ranks of NDVI, NBR, and NDWI). The metrics and the latest Landsat image are run through seven decision trees to calculate a median probability of forest disturbance. Pixels with probability >50% are reported as tree cover loss alerts. For more information on methodology, see the paper in Environmental Research Letters.Alerts remain unconfirmed until two or more out of four consecutive observations are labelled as tree cover loss. Alerts that remain unconfirmed for four consecutive observations or more than 180 days are removed from the data set. You can choose to view only confirmed alerts in the menu, though keep in mind that using only confirmed alerts misses the newest detections of tree cover loss.