Characteristics of Vertical Forest Backscatter Profiles Measured using Radar Tomography

Licentiate thesis, 2023

Forests are an integral part of the Earth’s climate system, having shaped the conditions of life throughout millions of years, but the coverage and conditions of forests are under rapid change. To monitor this change and predict the impact on climate, yearly global mapping of forest Above-Ground Biomass (AGB) is needed. Synthetic Aperture Radar (SAR) is a well-suited technique for this purpose, able to sense through dense canopies with high spatial resolution.

Tomographic SAR (TomoSAR) resolves the vertical distribution of backscatter by constructing 3D tomograms of forested areas. Here, results obtained in the TomoSense project are presented, where sensitivity of P- and L-band TomoSAR to AGB of a temperate forest is analysed. For the first time, differences between the AGB dependence of spruce and beech vertical backscatter profiles are identified. Moreover, unique observations of ground slope influence on TomoSAR AGB retrieval are presented. The effect for P-band is significant, but not for L-band. For spruce, AGB was estimated for P-/L-band (ground slopes below 10◦), with R2 = 0.86/0.75 and RMSE = 15.6/12.5%. Without separating forest types, R2 = 0.77/0.54 and RMSE = 11.4/12.0%.

Finally, unique L-band radar tomography observations using the BorealScat tower-based radar are presented. Time series of the vertical backscatter profile of a boreal forest during the summer of 2018 is analysed. Weekly diurnal cycles up to 1.3 dB are observed, showing clear differences depending on both canopy height layer and polarization. These differences are new results which are likely related to tree transpiration phenomena but needs further study.