3826 Rate this article:

Student Spotlight Spring 2023

Erin Eckles Tuesday, April 11, 2023

Exploring Forest Ecosystem Dynamics: Spring Student Spotlight Winner Paige Williams

Remote sensing technologies have revolutionized our ability to study and understand the complex and interconnected systems of Earth. One researcher making strides in this field is the Spring 2023 student spotlight winner, Paige Williams, a Ph.D. candidate in the Department of Forest Resources and Environmental Conservation (FREC) at Virginia Tech.

Along with her Ph.D. research, Williams has been working with the Structure and Function of Forested Ecosystems (SAFE) science team at NASA Goddard Space Flight Center since 2020. The team is working towards developing an Earth observing system to simultaneously quantify structure and function, assessing global productivity changes in forests. Williams’ studies and her collaboration with NASA underscore the critical role of remote sensing to understand forest ecosystems and climate change. “I was trained to use ENVI® Software in an undergraduate remote sensing course and a graduate level hyperspectral course at Virginia Tech. Since then, ENVI‘s ease of use and reliability has enabled me to produce significant scientific results,” according to Williams.

Ecosystem structure and function provide insight on the dynamics of the carbon cycle which contribute to the pools and fluxes of carbon in the system. Ecosystem structure for forests can be defined as the canopy height, crown shapes, leaf area and orientation, etc., while the function for forests can be described by the rates of photosynthesis and respiration which inform CO2 fluxes of the system.

To investigate the dynamics of Earth's interconnected systems, Williams integrates remote sensing data captured by drones, planes, and satellites with field observations from flux towers. By combining high-resolution observations of forest structure from LiDAR and function from hyperspectral, she seeks to improve our understanding of ecosystem dynamics and the response to changing climate conditions. “ENVI’s ability to incorporate multiple different datasets has provided an imperative feature for producing, analyzing, and visualizing results for my dissertation research,” said Williams.

Williams uses ENVI to process and analyze the collected hyperspectral and LiDAR data to improve the quantification of forest productivity. This combination provides a rich dataset for environmental analysis, producing high-resolution 3D images of forests. “As a geospatial scientist I am familiar with using ArcGIS Pro, but due to its incompatibility with hyperspectral data, ENVI has been a crucial tool for my hyperspectral image processing,” said Williams.

The capabilities in ENVI have allowed her to visualize the diurnal and vertical distributions of vegetation indices that relate to plant photosynthesis to reveal structural mediation on the canopy light environment. Williams also uses ENVI to calculate vegetation indices related to plant photosynthesis, classify different levels of illumination using a panchromatic band simulation, and incorporate digital surface models (DSM) to visualize vertical distributions of panchromatic reflectance and the photochemical reflectance index (PRI).

Williams holds a B.S. degree in Environmental Informatics with a GIS minor, and an M.S. degree in Forestry Remote Sensing from Virginia Tech. She is currently pursuing her doctoral degree in Forestry in the Remote Sensing Interdisciplinary Graduate Education Program at Virginia Tech. She is a licensed Part 107 Remote Pilot and has Geospatial Information Technology and Remote Sensing Certificates. Williams has also been part of several international and interdisciplinary research projects in Panama, India, and Burkina Faso, Africa, working as a GIS and remote sensing team member.

Williams’ passion for the natural beauty of forests, coupled with her academic interests in GIS and remote sensing, drives her to contribute to ecological modeling and sustainable natural resource management. She plans to graduate in December 2023 or May 2024 and is considering a professor position at an academic institution or pursuing a postdoctoral position at NASA Goddard Space Flight Center.

Williams' work exemplifies the importance of integrating remote sensing technologies to understand Earth's complex ecosystems. Her collaboration with NASA Goddard Space Flight Center and her commitment to advancing our knowledge of forest ecosystems and climate change highlight the critical role remote sensing plays in addressing global environmental challenges.

*This research has been presented at the American Geophysical Union Fall Meeting in 2020 (https://ui.adsabs.harvard.edu/abs/2020AGUFMB081.0010W/abstract) and 2021 (https://ui.adsabs.harvard.edu/abs/2021AGUFM.B25I1586W/abstract) and also SilviLaser in 2021 (https://doi.org/10.34726/wim.2026). The manuscript is currently in prep for review and is expected to be published by the end of the year.

**All images are published with the permission of Paige Williams and her fellow co-authors.

Figure 1. Study area sites in Parker Tract, NC shown with the 420 m x 420 m bounding boxes in the 20-year-old loblolly pine stand centered on the US-NC2 flux tower (plot 1). A more mature loblolly pine stand is shown in plot 2, and mixed deciduous broadleaf stand (plot 3) is overlain on top of (a) high-resolution NAIP imagery from October 2014 (a). The imagery in (b) is G-LiHT’s 2 m lidar canopy height model from the October 26, 2013 midday flight (blue to red corresponds to increasing canopy height).

Figure 2. Plot 1 image output results at 2 m resolution from the canopy hyperspectral processing in ENVI shown in order of time across the two days. Images scaled to 1:2,000. First row: simulated panchromatic band scaled to 0-600. Second Row: PRI images scaled to -0.1 - 0.1. Third row: Illumination classification with sunlit as yellow, mixed as green, and shadow as purple.

Figure 3. Plot 2 image output results at 2 m resolution from the canopy hyperspectral processing in ENVI shown in order of time across the two days. Images scaled to 1:2,000. First row: simulated panchromatic band scaled to 0-600. Second Row: PRI images scaled to -0.1 - 0.1. Third row: Illumination classification with sunlit as yellow, mixed as green, and shadow as purple.

Figure 4. Plot 3 image output results at 2 m resolution from the canopy hyperspectral processing in ENVI shown in order of time across the two days. Images scaled to 1:2,000. First row: simulated panchromatic band scaled to 0-600. Second Row: PRI images scaled to -0.1 - 0.1. Third row: Illumination classification with sunlit as yellow, mixed as green, and shadow as purple.

Figure 5. Scatterplots from ENVI for Plot 1 of 2m DSM elevations versus panchromatic reflectance % (first column), shadowed PRI (second column), mixed PRI (third column), and sunlit PRI (fourth column) for the North-South flights on October 26th for the morning (first row), midday (second row), and evening (third row).

Figure 6. These are the 3D structural results for plot 1 (US-NC2) of the CHM (first column, blue to red indicates increased height), illumination class (middle column, sunlit as yellow, mixed as green, and shadow as purple), and PRI (last column, low values in red to high values in blue) for the morning (first row), midday (second row), and evening (third row) NS flights. Time of flights (EST), solar zenith angle (SZA, in degrees), and the solar azimuth angle (SAA, in degrees) are shown for each row.