Landsat and Sentinel-2 Based Burned Area Mapping Tools in Google Earth Engine

Abstract

Four burned area tools were implemented in Google Earth Engine (GEE), to obtain regular processes related to burned area (BA) mapping, using medium spatial resolution sensors (Landsat and Sentinel-2). The four tools are (i) the BA Cartography tool for supervised burned area over the user-selected extent and period, (ii) two tools implementing a BA stratified random sampling to select the scenes and dates for validation, and (iii) the BA Reference Perimeter tool to obtain highly accurate BA maps that focus on validating coarser BA products. Burned Area Mapping Tools (BAMTs) go beyond the previously implemented Burned Area Mapping Software (BAMS) because of GEE parallel processing capabilities and preloaded geospatial datasets. BAMT also allows temporal image composites to be exploited in order to obtain BA maps over a larger extent and longer temporal periods. The tools consist of four scripts executable from the GEE Code Editor. The tools’ performance was discussed in two case studies: in the 2019/2020 fire season in Southeast Australia, where the BA cartography detected more than 50,000 km2, using Landsat data with commission and omission errors below 12% when compared to Sentinel-2 imagery; and in the 2018 summer wildfires in Canada, where it was found that around 16,000 km2 had burned.

A Preliminary Global Automatic Burned-Area Algorithm at Medium Resolution in Google Earth Engine

“…An alternative methodology was applied instead in the form of the Simple Non-Iterative Clustering (SNIC), a non-iterative superpixel segmentation algorithm [74] already implemented in GEE that has been shown to be efficient in terms of computation and memory The composite with the maximum P dy may also contain pixels with low burn probability values and some areas burned on dates outside the processing month that should be removed (Figure 11a). The best option for removing these areas would be a two-phased strategy involving first identifying burned seeds (pixels with a strong-burned signal) and then extending the burned region around these seeds up to a threshold [66]; this strategy has already been used extensively for BA detection [12,13,16,18,22,25,39,41,61]. Several approaches were tested to implement this strategy, such as spatial dilation from burned seeds, cumulative cost maps, and grouping of connected pixels, although they all proved to be too time-and memory-consuming and always ended up exceeding GEE's user memory limit.…”

“…The Committee on Earth Observing Satellites' Land Product Validation Subgroup (CEOS-LPVS) defined validation as the process of assessing by independent means the accuracy of the data products from the system outputs [76]. Out of four validation stages defined by CEOS-LPVS, Stage 3 requirements are usually followed by BA product validations [16,18,20,21,41,77,78], which consist of an assessment characterized in a statistically robust way over multiple locations [79]. BA products are validated by comparison with reference data (RD) created in multiple image pairs and located in validation sites typically selected by stratified random sampling [20,77,80,81].…”

“…This QA exercise was limited to 50 areas-with several image pairs analyzed at each one-and is deemed to be sufficient since it covers several biomes. This sampling methodology has already been implemented in GEE as the VA tool from Burned-Area Mapping Tools (BAMT) [41] and, even though it could not be used globally because it exceeds the GEE user memory limit in large areas, part of its code was used in order to access the S2 catalog and analyze its availability in GEE. The whole analysis was carried out for data from 2019, which is the first year when S2 L2A data have already been consistently processed (Figure 1).…”

“…This has been made possible by Google Earth Engine (GEE), a free cloud-computing platform with several satellite data catalogs at different spatial resolutions (mainly MODIS, Landsat and Sentinel-2) and global-scale analysis capabilities [35]. Even though the first significant work on the topic was published almost a decade ago [36] and a significant number of studies has used GEE since then [37], there are still few published works on the field of burned areas [38][39][40][41][42], especially at a global scale [22].…”

Multi-Sensor, Active Fire-Supervised, One-Class Burned Area Mapping in the Brazilian Savanna

“…Another notable limitation of most BA mapping algorithms is the need for human intervention in training sample acquisition [28,[40][41][42][43], making it a challenge to maintain operational products over large areas and long periods. This limitation may be circumvented by exploring two specific aspects of the problem: the relationship between active fires and burned areas and the possibility of envisaging BA mapping under the framework of one-class classification [44].…”