Developing a methodology for consistent plant biomass and NPP estimates across NEON domains

Background/Question/Methods

The NEON Terrestrial Observation System is designed to measure key taxa that are not readily observed via instruments, and will enable a continental-scale understanding of both the drivers of and the responses to environmental change. With respect to plant biomass and productivity, one major goal for the terrestrial field sampling is to partition NPP into functional vegetation components (e.g. trees, shrubs, herbaceous vegetation, belowground biomass, etc.) within the NEON tower footprint at the same spatial scale from which tower eddy flux data are collected. To rigorously measure plant biomass and production within the NEON tower footprint, there are several key sampling design issues that need to be addressed: 1) How can measurements made across a diverse set of ecosystem types be standardized, given that different methods and plot sizes are required for sampling different vegetation components?; and 2) Is there a general approach capable of determining the appropriate number, size, shape, and spacing of plots for sampling biomass production and loss across multiple ecosystem types? To address these sampling design questions, we measured vegetation in and around the proposed NEON tower footprint in a short-grass steppe ecosystem (D10 core site), and in an open mixed evergreen/deciduous forest in northern Florida (D3 core site). In both ecosystems, we measured leaf area index (LAI) and vegetation structure, and at the D10 site we measured NPP via clip harvests.

Results/Conclusions

We have adopted uncertainty as the metric by which biomass and NPP measurements will be standardized across NEON sites. That is, NEON will adopt methods to estimate biomass that are appropriate to the vegetation components present at each site, and sampling effort will be calibrated to achieve a common level of uncertainty across vegetation components. To illustrate this concept, we assumed LAI is a coarse proxy for biomass at the landscape scale, and we show how LAI data can be used to adjust sampling effort and sample placement in very different ecosystems in order to meet pre-defined levels of sampling uncertainty. We also performed semivariogram analyses of these data to determine optimal sample spacing at each site. In the future, we will compare empirical results, such as those presented here, with airborne remote sensing data and sampling simulations in order to optimize plant biomass and NPP sampling.