Speaker
Description
The Climate Change Initiative (CCI) Biomass project foresees the generation of forest aboveground biomass (AGB) estimates for several years between 2005 and 2022 at high resolution and globally. This is a challenging task, as accurately measuring the organic mass stored in forests relies on mathematical models applied to observations, either near or far-field. The difficulty lies in achieving both spatial detail and temporal consistency in AGB estimates, primarily because no single satellite mission has consistently collected data with the same level of geometric and radiometric quality over more than a decade. The only viable solution to overcome such limitations is to combine observations and models to reduce both random noise and individual systematic biases.
The CCI Biomass CORE retrieval algorithm, presently in its fifth version, combines estimates of AGB from C- and L-band SAR missions. These estimates are generated using physically-based models that incorporate allometric functions trained on spaceborne LiDAR observations. By combining AGB estimates from multiple SAR observations with varying frequencies and polarizations, the algorithm reduces uncertainty compared to individual AGB estimates. Furthermore, a temporal stability measure is applied to minimize unnatural fluctuations in AGB estimates, given the moderate-to-weak sensitivity of SAR observables to AGB.
Although the integration of models based on multiple satellite observations has matured, AGB estimates still carry a significant degree of uncertainty. This is due to the limited sensitivity of SAR observations to diverse forest structures and the generalizations made within the modeling framework. Indeed, a workaround solution was implemented to train the retrieval model because of the lack of a densely populated dataset of AGB measurements from ground surveys. Consequently, the algorithm may not fully adapt to local forest structural conditions.
In this talk, we will present the evolution of the CCI CORE algorithm and discuss its current limitations. Additionally, we will explore potential avenues for reducing these limitations by considering various scenarios involving different satellite missions. This discussion will be situated within the context of quantifying AGB changes, highlighting the current disparity between user requirements and the capabilities available to AGB map producers.
