RP_LAC_2019_S2: reference fire perimeters obtained from Sentinel-2 imagery over Latin America and Caribbean for the year 2019.

Resumen

The reference dataset RP_LAC_2019_S2 was obtained from S2 images over a set of 56 mosaics (sampling units) sampled following a custom design for the Sentinel mosaic grid system and to represent both major fire regimes and regimes with lower burned area of Latin America and Caribbean in the different ecoregions. For each mosaic, S2 time series were defined based on a set of conditions to minimize cloudiness and ensure series length and minimum time lag between image pairs. The S2 image pairs were classified with a Random Forest (RF) algorithm to provide burned perimeters representing burned areas between the two dates that were combined into a synthetic burned area reference dataset. This dataset represents for each unit burned and unburned polygons and masked areas. The RP_LAC_2019_S2 dataset is part of the Burned Area Reference Database (BARD), a database that compiles multitemporal global and regional burned area reference datasets for Earth Observation burned area products validation. Description of the project: This dataset has been developed for the validation of the S2BA (Sentinel-2 Burned Area) product in Latin America and Caribbean for the year 2019. S2BA is an automatic global burned area mapping algorithm (S2BA) based on Sentinel-2 Level-2A imagery in combination with Visible Infrared Imaging Radiometer Suite (VIIRS) and Moderate Resolution Imaging Spectrometer (MODIS) active fire data. The algorithm and product was developed by the University of the Basque Country UPV/EHU under the "Proyecto Estratégico Análisis y explotación de información geoespacial (GeoInf) PES20/54” Estratégico Análisis y explotación de información geoespacial (GeoInf) PES20/54”.