For Researchers
For Students
For Professionals
Outreach & Events
About the Lab

Menu

For Researchers
For Students
For Professionals
Outreach & Events
About the Lab

Search

Close

Investigating the patterns of vegetation-permafrost dynamics as determined by landscape-scale disturbance

Sponsor: Department of Energy

Daniel Hayes and Wouter Hantson - University of Maine
Shawn Serbin, Andrew McMahon, Dedi (Daryl) Yang, and Ran Meng - Brookhaven National Laboratory
Peter Thornton and Stan Wullshleger - Oak Ridge National Laboratory

multi-stage sampling of the arctic tundra

Project Title

Project Participants

Project Duration

Project Sponsor

Project Contact Name

Project Contact Email

Project Description

The University of Maine conducts research in collaboration with Brookhaven National Laboratory and Oak Ridge National Laboratory toward a subset of science objectives outlined in the Phase 3 proposal of the Next-generation Ecosystem Experiments (NGEE) in the Arctic. The <a href="https://ngee-arctic.ornl.gov/">NGEE-Arctic project</a> is using observations, experiments and modeling to improve our predictive understanding of Arctic ecosystem processes and their feedbacks to global climate. The research objectives defined for NGEE-Arctic are designed to address a series of overarching <a href="https://ngee-arctic.ornl.gov/science-questions">science questions</a> focused on vegetation, hydrology, biogeochemistry, and permafrost dynamics. UMaine brings expertise in remote sensing and modeling to conduct research aimed at generating model-data and process understanding of vegetation-permafrost dynamics as determined by landscape-scale disturbances processes.

The challenges associated with collecting ground-based measurements across large and remote Arctic environments limit the availability of data required for model development and testing. To address this issue, UMaine has partnered with the NGEE project since 2015 in collecting field information concurrent with spectral measurements of individual plants and vegetation canopies using both ground-based and small unmanned aerial systems (sUAS). The goal of the current work is to build a scaling framework for synthesizing and extrapolating these existing and new field data across increasingly broader spatial and temporal resolution and extent. The framework is based on a scaling concept whereby current spatio-temporal patterns of Arctic-Boreal vegetation and associated ecological function are an outcome of the disturbance history and trajectory of permafrost thaw of the landscape.

The research activities being conducted in this project are organized within three objectives designed to build such a scaling framework:
<ul>
 	<li>Measure the local-scale heterogeneity of vegetation composition and structure along gradients of disturbance associated with permafrost thaw features on the Arctic landscape.</li>
 	<li>Map permafrost thaw features by association with the spatial patterns of vegetation composition and structure at the landscape scale.</li>
 	<li>Characterize the trajectories of permafrost thaw at the regional scale as the patterns of vegetation composition and structure change over time in response to landscape disturbances.</li>
</ul>
For the first objective, we are building on existing “spectral libraries” relating remote sensing measurements to vegetation composition and functional traits for individual plants, plot-level canopies and transects across thaw gradients characterizing thermokarst features. The field-based work is being conducted at and around the three NGEE-Arctic research sites on the Seward Peninsula, Alaska. Hyperspectral scans of vegetation reflectance are captured at the ground level with a hand-held ASD spectroradiometer and at the canopy scale with a VNIR spectrometer mounted on a UAS. Canopy height models of the plots are developed using high-resolution RGB and NIR imagery from a sUAS processed as photogrammetric point clouds.

For the second objective, the plot-scale data will be used as the link to develop landscape-scale maps of plant functional traits across larger footprints acquired by airborne imagery. In particular, we will generate 5m resolution maps of key species functional groups related to thaw gradients at the landscape scale (e.g., shrub-tussock tundra, sedges and graminoids, sphagnum, etc.). These maps will be developed based on AVIRIS-NG imagery acquired by NASA over the NGEE-Arctic sites from 2017 to 2020. The airborne hyperspectral imagery will be classified as maps of composition and other plant traits using data mining algorithms trained on the plot-scale ground- and sUAS- based spectral libraries.

For the third objective, we will scale-up the composition-thaw patterns from the landscape maps to the broader region by characterizing trends in vegetation dynamics using time-series imagery form the historical Landsat record. Here, we will relate an ensemble of satellite-derived spectral vegetation indices to the species functional groups mapped over the AVIRIS-NG footprints. These indices will then be analyzed by time-series segmentation algorithms designed to characterize trends in the disturbance and recovery of vegetation as they relate to different permafrost thaw trajectories at the regional scale.

The work conducted by UMaine complements the modeling objectives of the broader NGEE-Arctic project by contributing to the science questions guiding its Phase 3 activities. Specifically, we will investigate disturbance processes (i.e. Question 6) as the spatio-temporal control on ecological structure and function, including the trajectories of permafrost thaw (Q1), the patterns of shrubs and other plant functional groups and their traits (Q3, Q4), and the impacts on landscape wetting and drying (Q5).

Project Type - select all applicable

Applied Research Project
Faculty Research Project
Industry Outreach Project
Public Service Project
Student Project

Project Applications - select all applicable

Arctic Ecosystems and Permafrost Landscapes
Carbon Cycle
Climate Change
Conservation
Forestry
Project-Based Learning
Remote Sensing Applications
Terrestrial Biosphere Modeling
Terrestrial Ecosystem
Tourism

Project Technologies - select all applicable

GIS
Hyperspectral
Lidar
Machine Learning
Multispectral
Piloted Aircraft
Remote Sensing
Statistical Modeling
UAV

Featured Image

http://wheatlandlab.org/wp-content/uploads/2020/01/scaling.jpg

Additional Project Images

http://wheatlandlab.org/wp-content/uploads/2020/01/field-spec.jpg

http://wheatlandlab.org/wp-content/uploads/2020/01/spectral-plots.jpg

http://wheatlandlab.org/wp-content/uploads/2020/01/image-cube.jpg

http://wheatlandlab.org/wp-content/uploads/2020/01/Wouter2.jpg

http://wheatlandlab.org/wp-content/uploads/2020/01/Wouter5.jpg

http://wheatlandlab.org/wp-content/uploads/2020/01/Wouter4.jpg

http://wheatlandlab.org/wp-content/uploads/2020/01/Wouter3.jpg

Project Video URL - if any

Project URL - if any

Applications:

Arctic Ecosystems and Permafrost Landscapes, Carbon Cycle, Climate Change, Remote Sensing Applications

Technologies:

Hyperspectral, Machine Learning, Multispectral, Remote Sensing, UAV

Contact:

Daniel Hayes

daniel.j.hayes@maine.edu

taking field spectrometer measurements on the arctic tundra

spectral signatures along vegetation transects

hyperspectral image cube

taking field spectrometer measurements on the arctic tundra

taking sUAS-based photos of the arctic tundra

Arctic tundra

Arctic tundra

Previous

Next

Recent News

UMaine & NASA to map carbon in forests!

Planes, UAVs, and AI to count seabirds in coastal Maine

New lidR book now available online!

WPI Institute on Project-Based Learning – Summer 2020

Upcoming Events

Getting Started with QGIS

Spectroscopy of plants – What light spectra can tell us about functional, taxonomic, and phylogenetic plant biodiversity

Leveraging High Resolution Forest Inventory and Satellites to Predict Fire Damage Severity at the Micro-Stand Level

Moving Beyond Enhanced Forest Inventory (EFI): Accelerating the Digitalization of the Forest Value-Chain

Current Projects

Forest Disturbance Detection and Aboveground Biomass Modeling Using Moderate-Resolution, Time-Series Satellite Imagery

Synthesizing Disparate LiDAR and Satellite Datasets through Deep Learning to Generate Wall-to-Wall Forest Inventories of New England

Airborne Hyperspectral Data Application in Health Stress Detection of Blueberry Fields and Ash Trees

The Effects of Climate Change on Forest Ecosystem Phenology and its Impacts on Tourism in Maine

About

The Lab
Our People
Calendar
Contact us

Follow Us

Twitter Youtube Instagram

Supporting the geospatial education, research, outreach, and innovation needs of students, forest industry, and natural resource partners in Maine since 2013

Copyright 2021 All Rights Reserved

Insert/edit link

Enter the destination URL

URL

Link Text

Open link in a new tab

Or link to existing content

Search

No search term specified. Showing recent items. Search or use up and down arrow keys to select an item.