THE UNCERTAINTY OF SPACEBORNE OBSERVATION OF VEGETATION STRUCTURE IN THE TAIGA-TUNDRA ECOTONE: A CASE STUDY IN NORTHERN SIBERIA

Abstract

The ability to characterize vegetation structure in the taiga-tundra ecotone (TTE) at fine spatial scales is critical given its heterogeneity and the central role of its patterns on ecological processes in the high northern latitudes and global change scenarios. This research focuses on quantifying the uncertainty of TTE forest structure observations from remote sensing at fine spatial scales. I first quantify the uncertainty of forest biomass estimates from current airborne and spaceborne active remote sensing systems and a planned spaceborne LiDAR (ICESat-2) across sparse forest gradients. At plot-scales, current spaceborne models of biomass either explain less than a third of model variation or have biomass estimate uncertainties ranging from 50-100%. Simulations of returns from the planned ICESat-2 for a similar gradient show the uncertainty of near-term estimates vary according to the ground length along which returns are collected. The 50m length optimized the resolution of forest structure, for which there is a trade-off between horizontal precision of the measurement and vertical structure detail. At this scale biomass error ranges from 20-50%, which precludes identifying actual differences in aboveground live biomass density at 10 Mg•ha-1 intervals. These broad plot-scale uncertainties in structure from current and planned sensors provided the basis for examining a data integration technique with multiple sensors to measure the structure of sparse TTE forests. Spaceborne estimates of canopy height used complementary surface elevation measurements from passive optical and LiDAR to provide a means for directly measuring TTE forest height from spaceborne sensors. This spaceborne approach to estimating forest height was deployed to assess the spaceborne potential for examining the patterns of TTE forest structure explained with a conceptual biogeographic model linking TTE patterns and its dynamics. A patch-based analysis was used to scale estimates of TTE forest structure from multiple sensors and provided a means to simultaneously examine the horizontal and vertical structure of groups of TTE trees. The uncertainty of forest patch height estimates provides focus for improving spaceborne depictions of TTE structure patterns associated with recent change that may explain the variability of this change and the vulnerability of TTE forest structure.