MSB-ENSA: Foliar traits and ecosystem variability across NEON domains

University Of Wisconsin-Madison, Madison WI

Investigators

Abstract

Accurately predicting plant species response to global change is difficult. Responses vary within and among species, with variation in the local and regional environment, and in response to external pressures such as insect infestations making extrapolation from individual plant to species-specific studies difficult. Plant leaf functional traits (e.g. leaf chemistry, plant pigments) have recently emerged as a potentially useful way to characterize and understand the variability in plant function (e.g. growth, stress, nutrient uptake), and plant responses to environmental change. Plant leaf traits or "foliar traits" have been shown to be strongly correlated with global variation in plant function and can be detected from remotely sensed imagery. Remote sensing offers the possibility to characterize and map the spatial variation in foliar functional traits to gain a better understanding of biological responses to global change. Remotely sensed and ground based measurements from the National Ecological Observatory will be used to develop a suite of foliar traits from 81 locations that encompass the range of ecosystems found in the United States. One component of the National Ecological Observatory infrastructure is its Aerial Observation Platform, which collects remote imagery annually over the locations using the latest generation imaging technologies. These imaging sensors have an unprecedented ability to map biological function, including plant chemistry and physiology, as well as the biomass and structure of plants. This award will result in the first comprehensive data set and methods for mapping plant biochemistry and physiology across the range of ecosystem types in the US and will enable characterization of how plant traits vary across space and time. The leaf traits, supporting data, maps and enabling equations and software will be made publically available via existing databases. Graduate students and post doctoral candidates will be engaged in the research. One of the great promises of imaging spectroscopy ? also known as hyperspectral remote sensing ? is the ability to map the spatial variation in foliar functional traits, such as nitrogen concentration, pigments, leaf structure, photosynthetic capacity and secondary biochemistry, that drive terrestrial ecosystem processes. Such foliar trait characterization offers an organizing principle that can be used to understand the occurrence and evolutionary differentiation of function across different taxonomic or phylogenetic levels and to detect functional differences across ecosystems. The National Ecological Observatory provides one of the first opportunities to characterize and compare these different ecosystem types, in terms of their biodiversity and ecosystem services such as maintaining air and water quality and sequestering and storing carbon. The imaging spectrometer on the National Ecological Observatory Airborne Observation Platform (AOP) will be used to regularly estimate trait variation across the major biomes of North America, providing high-resolution data suitable for scaling these traits continentally, as well as ecosystem modeling across domains. The award will enable hyperspectral mapping algorithms for a large number of plant foliar traits (such as nitrogen concentration, pigments, LMA, photosynthetic capacity and secondary metabolites) to be developed (or modified from existing algorithms), validated, applied and made publicly available. To accomplish this will entail fundamental research in remote sensing to understand how optical properties - detectable using imaging spectroscopy - permit mapping traits across biomes. Additionally the data will permit the evaluation of how trait retrieval algorithms differ across biomes and how they are affected by vegetation structure, physiognomy, or other ecosystem properties. Lidar data will be used to test and control for the influence of canopy vertical structure on trait mapping. The resulting maps of vegetation trait variation will be analyzed to determine the requisite geographic area that must be sampled to fully characterize trait variation across ecosystem types and, subsequently, compare to global variation. The ultimate objective is the synthesis, testing and validation of cross-biome trait retrieval models for hyperspectral imagery. By characterizing how multi-dimensional trait space is filled across landscapes and comparing the resulting trait maps with global data, this research will identify the extent to which hyperspectral imagery can be used to both extrapolate trait variation on Earth and fill geographical gaps in existing knowledge of biome- to continental- scale trait variation. Comprehensive trait information for the biomes represented in the Observatory is largely absent except for localized studies, and this work will enable better understanding and prediction of the response of global terrestrial ecosystems to disturbance, stress and change. This work will provide the scientific community with data products necessary to better understand local-, regional-, continental- and global-scale variation in ecosystem function, in part through linkage to a range of other data (e.g. flux tower estimates of GPP). The project will train students and postdoctoral scholars in cross-disciplinary research that merges ecology, remote sensing, and the quantitative analyses of dense data sets, creating a new generation of researchers that can address cross-cutting questions in global ecology.