Filling Gaps in Global Data Sets: The Role of New Vegetation Remote Sensing Data Products

Abstract:

Terrestrial ecosystem and carbon cycle feedbacks contribute substantial uncertainty to model projections of future climate, but the real controls and sensitivities are poorly constrained by observations, especially at large scales. Local studies provide the most direct information about cause and effect in the terresrial carbon system, but are difficult to scale up. Atmospheric analyses provide the best large-scale information, but inferences from atmospherioc CO₂ are so coarse they are difficult to relate to specific mechanisms. Remote sensing is at the appropriate scale but vegetation index-based measures do not correspond directly to specific ecosystem processes and have been of limited success in parameterizing or falsifying models. We analyze extant global data sources for three key terrestrial carbon cycle quantities: ecosystem carbon fluxes, aboveground biomass and plant growth parameters (traits), revealing a sampling pattern far from optimum. Flux measurements are sparse in regions where fluxes are highest and most variable, biomass measurements are limited where biomass is highest, and plant trait data are scarcest where plant diversity is highest. The likely carbon cycle tipping point regions in high and low latitudes are scarcely observed by the standards of the mid-latitudes. Strategic and synergistic use of available in situ and emerging space-based observations can provide critical information needed to detect and forecast change. New remote sensing data products can contribute significantly to filling sampling gaps in the global observing system, as well as potentially addressing sensitivities at appropriate scales for testing global models. Environment and logistics limit extensive deployment of additional in situ observations in challenging tropical and Arctic-Boreal environments, so increasing use of space-based techniques is important for increasing sampling density and reducing uncertainty. Much greater integration between the remote sensing,atmospheric, biodiversity and terrestrial modeling communities is needed to reduce uncertainty and ensure that the critical satellite missions are selected and executed.