A key component in the monitoring of aboveground biomass (AGB) is the use of remotely sensed (RS) data, which provides the only truly synoptic view of a landscape. However, the quantification of AGB in high-biomass ecosystems, most notably forests, has been found to be problematic for medium and coarse scale active RADAR and passive sensors due primarily to the saturation problem. This has led to large uncertainties in our understanding of carbon and biomass stocks for approximately 30% of the terrestrial surface of the planet. Active research in the utility of alternative sensors and approaches to monitoring AGB has progressed using analyses applied to aerial LiDAR, hyperspatial optical, and time series analyses. However, an ongoing challenge is properly quantifying the uncertainties in AGB estimates from RS, as well as the sources of these errors. In this analysis, we seek to estimate AGB at 30 meter resolution using small-footprint, airborne LiDAR data for over 50 sites in California, and estimate the uncertainty and the sources of uncertainty in these estimates.
Results/Conclusions
Our initial results confirm the ability of individual tree crown recognition (ITCR) techniques applied to airborne LiDAR can produce relatively accurate estimates of AGB for a diverse set of forested ecosystems in California. However, we found significant sources of uncertainty stemming from the following elements of the analysis: 1) errors in the structural estimates from the LiDAR data, 2) omission and commission errors from the LiDAR analysis, 3) a lack of knowledge of the species composition at a given site, and 4) errors in the allometric equations used in the analysis. Many of these errors are often under- or unreported in published analyses, potentially leading to an overly optimistic view of the accuracy of these products. Our analysis, while providing a concrete quantification of errors for sites in California, also provides a framework with which to perform future propagation of error analyses when using remote sensing to estimate AGB.