The regrowth of vegetation after fire is important for generating healthy plant communities in forests but also scrubland and other regions. Monitoring this progress can be tricky, particularly when conditions are cloudy or data solely depend on optical measurements.
Radar data, however, are now being used as a means to better understand how plant communities are bouncing back from fires. Such measurements will be important in our forecasts for climate change, due to the ability of vegetation to absorb carbon, while also enabling the prediction of future fires.
Synthetic-aperture radar (SAR) instruments are carried on a variety of satellites, including the European Space Agency’s Sentinel-1 satellite as well as other airborne instruments. NASA has, in fact, been deploying its Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR) using repeated aircraft flyovers.
In fact, a new SAR satellite using similar technology will be launched by NASA  and India’s Space Research Organization, with the launch planned for 2022 and data likely to be used to monitor vegetation growth.
UAVSAR aircraft have been flying since 2009, with data about wildfires in Southern California collected in 2010, 2017, and 2020 from regions burned by different brush or forest fires.
Radar signals can be differentiated based on barren land, unburnt terrain, brush-covered hillsides, and area with new or fresh vegetation growth. Radar signal data can be color-coded based on vegetation, with data collected stacked over time to determine how vegetation has changed.
Results show that the Angeles National Forest, near Los Angeles, has a varying patchwork of regrowth and fire-affected areas over the last decade.[1]Â A recent paper examined 27 flights by UAVSAR aircraft with imagery at 50 m scale using radar data. Overall, results show that, comparable to the southern California data, a patchwork of recovery is evident in different regions in the United States, showing unevenness in growth in regions. Interestingly, the authors also see this regrowth as creating potential vulnerability for new fires, particularly if drought conditions return again to the Western United States.[2]
Other studies have shown comparable utility for SAR imagery. In addition to vegetation data obtained from aircraft, using SAR data from European Remote Sensing-2 and RADARSAT-2 satellites, data ranges spanning C-band (VV/HV polarization) and L-band (HH polarization) used to created images were acquired both before and after fires between 2002-2016 in Alaska’s Arctic tundra regions. In this case, data from the C-band images demonstrated that recovery was slower for vegetation relative to data derived from Normalized Difference Vegetation Index (NDVI) using optical data.
The use of NDVI is the traditional method to observe vegetation growth. Data derived from the C-band proved more accurate and shows that recovery to pre-fire levels takes somewhat longer than what was evident in NDVI. The results demonstrate that SAR has great potential for post-fire mapping in tundra regions.[3]
While the results have generally been encouraging for SAR data used for mapping post-fire vegetation growth, there are results showing some potential limitations. For instance, using the Advanced Land Observating Satellite-2 (ALOS-2), L-band SAR data showed that local incidence angle and topography could have a greater impact on results, whereby shadowing and miss-identification could occur. This may further support the benefit of using C-band data. The results, overall, could be due to radar signals bouncing in the uneven terrain that creates interference and more noisy data, creating somewhat false interpretation.[4]
Despite some evident limitations, overall SAR data and imagery have proven highly useful for measuring vegetation recovery after fires. This will be important as the ability for vegetation, particularly forests, to recover from fires are crucial to keeping a healthy ecosystem and in efforts to combat climate change, given that new forests and vegetation growth areas can absorb a lot of atmospheric carbon. Our ability to map regrowth will also be important for anticipating future fires and data can be used to model fire progression, which is an increasing threat due to climate change.
References
[1] For more on a recent article about SAR data collected for post-fire vegetation recovery, see: https://earthobservatory.nasa.gov/images/147872/a-mosaic-of-fire-data.
[2] For more on the patchy nature of vegetation recovery from fires, see: Parker, J., Donnellan, A., Glasscoe, M.T., 2021. Survey of Transverse Range Fire Scars in Ten Years of UAVSAR Polarimetry (preprint). Climatology (Global Change). https://doi.org/10.1002/essoar.10505784.1.
[3] For more on using SAR data for tundra fire monitoring, see: Zhou, Z., Liu, L., Jiang, L., Feng, W., Samsonov, S.V., 2019. Using Long-Term SAR Backscatter Data to Monitor Post-Fire Vegetation Recovery in Tundra Environment. Remote Sensing 11, 2230. https://doi.org/10.3390/rs11192230.
[4] For more on potential limitations with SAR data in elevated regions, see: Mutai, S., Chang, L., 2020. POST-FIRE HAZARD DETECTION USING ALOS-2 RADAR AND LANDSAT-8 OPTICAL IMAGERY. ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci. VI-3/W1-2020, 75–82. https://doi.org/10.5194/isprs-annals-VI-3-W1-2020-75-2020.
