Description:

Remotely sensed estimates of aboveground biomass (AGB) are important inputs to ecological and climate modeling. They have evolved over the last 30 years. AGB measurements used to center around the correlation between remote-sensing signal strength (radar or optical) and AGB. With the advent of interferometric SAR (InSAR) and lidar, remote measurements of tree height have been combined with average-power measurements to improve accuracy and resolution of AGB estimates. In this project, radar profiles from multi-baseline InSAR, called "tomoSAR", and/or lidar profiles from waveforms will be used to estimate AGB. Properties of the vertical profiles of radar (tomograms) and lidar (waveforms) bearing on AGB include InSAR phase and amplitude, averages of InSAR phase or waveform height, and/or Fourier transforms of both tomograms and waveforms. These properties are principal observables for NASA’s Surface Topography and Vegetation (STV) observation program.

The postdoc on this project will first learn or review characteristics of measurements with InSAR and tomoSAR, as well as lidar waveforms. She/he will use data in hand from airborne-UAVSAR tomoSAR and airborne lidar and develop algorithms for estimating AGB from their profiles. The data from UAVSAR were taken in boreal Saskatchewan, Canada in 2017 and 2022, while the lidar data were taken over the same area in 2017. Lidar data were also taken in 2020 over tropical forests near the Atlantic coast of Brazil. Fieldwork was done at the boreal and tropical sites. 

The postdoc will study diverse approaches to analysis and combining of data types to arrive at AGB. “Mechanistic” analyses involve postulating relationships between remote sensing observations and characteristics of forest structure, derived from biophysical descriptions of forest vegetation. “Statistical analyses” involve realizing remote sensing estimates by choosing them so as to produce optimal agreement with field “truth”estimates, as in “parameter estimation” and “supervised learning”. The postdoc will also investigate "unsupervised learning", a form of statistical analysis that is based on derived, internal properties of measurements. It will be applied to tomoSAR and lidar data. 

It is envisioned that the postdoc will get exposure to the above methods of AGB estimation from remotely sensed forest structure, and focus on one or two approaches for a publication concerning the technology and/or the science of the remotely-sensed region(s). 

Field of Science:

• Technology Development

Advisors: