Estimating aboveground biomass density using hybrid statistical inference with GEDI lidar data and Paraguay's national forest inventory

Paraguay's forests are known to be poorly represented in GEDI's current calibration dataset, and here we demonstrate that local models calibrated opportunistically with on-orbit GEDI data and field surveys from Paraguay's national forest inventory can be used with GEDI's statistical estimators of aboveground biomass density. We specify a protocol for opportunistically matching GEDI observations with field plots to calibrate a field-to-GEDI biomass model for use in GEDI's hybrid statistical framework. Country-specific calibration using on-orbit data resulted in relatively accurate and unbiased predictions of footprint-level biomass, and importantly, supported the assumption underlying model-based inference that the model must "apply" to the area of interest. [...] On average, the standard errors are 47% lower for estimates based on GEDI than the forest inventory, representing a significant gain in precision. Our research demonstrates that GEDI can be used by national forest inventories in countries that seek reliable estimates of aboveground biomass density, and that local calibration using existing field plots may be more appropriate in some applications than using GEDI global models, especially in regions where those models are sparsely calibrated.

BULLOCK Eric;  HEALEY Sean;  YANG Zhiqiang;  ACOSTA Regino;  VILLALBA Hermelinda;  INSFRÁN Katherin Patricia;  MELO Joana;  WILSON Sylvia;  DUNCANSON Laura;  NAESSET Erik;  ARMSTON John;  SAARELA Svetlana;  STÅHL Göran;  PATTERSON Paul;  DUBAYAH Ralph; 

2023-07-12

IOP PUBLISHING LTD

JRC132945

1748-9326 (online),   

https://dx.doi.org/10.1088/1748-9326/acdf03,    https://publications.jrc.ec.europa.eu/repository/handle/JRC132945,   

10.1088/1748-9326/acdf03 (online),