|
Characterizing Arctic plant traits with near-surface and unmanned aerial system (UAS) remote sensing
Shawn
Paul
Serbin, Brookhaven National Laboratory, sserbin@bnl.gov
(Presenter)
The inadequate representation of plant traits and trait variation across space and time in terrestrial biosphere models (TBMs), including many that underlie the land-surface component of Earth System Models (ESMs), is a key driver of uncertainty in model hindcasts and forecasts of terrestrial carbon, water, and energy cycling and storage. This is particularly relevant for biomes with only sparse observational data availability such as the Arctic and tropics. In the Arctic, uncertainty in the modeling of carbon uptake and associated processes and fluxes has been tied to the lack of key data on plant properties that regulate these processes. What is needed is an approach to bridge the scales between detailed ongoing in-situ observations of Arctic vegetation in remote locations and the larger, landscape context needed to inform models on parameter variation in relation to climate, soils, topography, perturbations and other edaphic conditions. Remote sensing approaches, particularly spectroscopy, imaging spectroscopy, high resolution imaging, and thermal infrared (TIR) thermography, represent powerful observational datasets capable of scaling plant properties and capturing broad-scale spatial and temporal dynamics in many important vegetation properties related to terrestrial ecosystem functioning, offering a an important and direct data constraint on model projections or as critical benchmarks against prognostic model predictions. In temperate ecosystems we have shown how leaf and imaging spectroscopy (IS) can be used to map a broad range of plant traits across large areas of the continental U.S. and through time. Here we extend these approaches to the high Arctic to evaluate the capacity to scale and map vegetation properties, including biochemical, morphological and physiological leaf traits from the leaf to landscape scales. We focus on the development of linkages between a range of plant species and remote sensing within our two core study areas within the Barrow Environmental Observatory (BEO), Barrow, and Nome Alaska regions. We coupled measurements of leaf chemistry and physiology, including leaf-level gas exchange, with measurements of full range (i.e. 0.3 to 2.5 microns) leaf optical properties (reflectance and transmittance), TIR, and optical imagery from near-surface (leaf, tram) to unmanned aerial system (UAS) platforms. We show how leaf-level spectra-trait models for Arctic vegetation, developed using data collected in the BEO during the 2014-2016 growing seasons, are comparable with existing models from other biomes. In addition, tram and UAS platforms show a strong capacity to scale leaf-level traits to the larger landscape and capture patterns through time. Importantly, despite strong variation in leaf morphology and physiology, we are finding a good potential for spectral models to capture trait variation and highlights the possibility to map traits in the high Arctic. Our next steps include the use of the existing and future NASA ABoVE airborne campaign data to scale from tram/UAS to the larger regions to develop algorithms for mapping key traits across broad regions in the Arctic. Presentation: ASTM4_Poster_Serbin_46_26.pdf (46794k) Associated Project(s): Poster Location ID: 46 Session Assigned: Vegetation Dynamics and Distribution
|