Bridging the biomass gap: Advances, challenges and prospects in unlocking Radar's potential in terrestrial ecosystems

Abstract

Radar (radio detection and ranging) technology operates in all weather and lighting conditions, making it a valuable tool for several purposes. It offers advantages such as the ability to penetrate vegetation canopies and ground surfaces. This study examines the progress, challenges, and prospects of using radar technology for improved biomass estimation in terrestrial ecosystems. Various sensors have been utilized for biomass estimation, including synthetic aperture radar (SAR), interferometric SAR (InSAR), and ground penetrating radar (GPR). SAR applications in tropical forests have employed C-band SAR sensors, such as Sentinal-1 A/B, while RADARSAT-2 has been used in temperate and subtropical forests. Advanced SAR (ASAR) has been effective in both tropical and temperate forests. L-band SAR sensors, such as phase array L-band SAR (PALSAR), have been used in tropical, temperate, boreal, and mixed forests, while unmanned aerial vehicle (UAV)-SAR and polarimetric L-band imaging SAR (PLIS) have been employed in boreal, mixed, and temperate forests, respectively. L-band InSAR has been used in boreal forests, and X-band InSAR, including Tandem-X and Shuttle Radar Topography Mission (SRTM) digital elevation model (DEM), has been implemented in boreal, tropical, and temperate forests. P-band SAR and polarimetric interferometry SAR (PolInSAR) have been used with ONERA aerial system SETHI in tropical forests, and GPR with field portable GEOTECH in subtropical forests. SAR sensors face challenges including signal attenuation due to incidence angle, polarization sensitivity, heterogeneous landscapes, data availability, and temporal decorrelation. InSAR sensors encounter phase decorrelation, baseline issues, and various errors. GPR sensors face signal attenuation, ground penetration, target detection, and ground truth validity challenges. To improve biomass estimation, this article suggests considering forest type-specific models, using multi-frequency SAR, employing polarimetric SAR (PolSAR) for scattering analysis, integrating radar with light emission and detection (LiDAR) or optical data, implementing advanced data processing techniques, and utilizing artificial intelligence (AI) and stochastic modeling.