Project Funding: 2010 - 2013

NRA: 2009 NASA: New Investigator Program   

Funded by NASA

Abstract:
Tropical forests hold unique opportunities and difficulties for ecological study and inspire fascination for the general scientific audience. These forests are made up of heterogeneous canopies and forest communities with unique assemblages of tree species, complex vegetation dynamics and history, and high biodiversity. Forest structural components include canopy geometry and tree architecture, size distributions of trees, and species diversity. Structural properties of forests are closely linked with ecosystem functioning and are the result of the dynamics of growth and disturbance along both temporal and spatial scales. Ground-based field measurements allow for accurate mapping of vegetation structure, but on a very limited spatial and temporal scale with high cost and unknown biases. In order to extract information related to changes in forest structure, we must devise ways to relate measurements on the ground with those inferred from space-based remote sensing data. To better understand tropical forest structure and dynamics, a unique set of strategies tuned to observations of ecosystems dynamics, such as gap formation and blow-downs, specific to tropical forests is vital. Observable changes in structure are related to forest dynamics and hence the rates and sign of carbon flux in forest systems. Because tropical forests contains a large stock of carbon, understanding tropical forest dynamics and associated forest structure are important for estimating regional and global carbon and biogeochemical cycles. We will examine the forest structure of a tropical forest at La Selva Biological Reserve, Costa Rica. Using both optical and lidar high resolution remotely sensed data we will estimate forest structure using a crown delineation algorithm and textural methods. As part of an extensive validation exercise, we will compare our results with field based data. We will use a 3D canopy model, that uses trunk or crown diameter distributions and allometric equations, to generate a synthetic canopy. Using geometric series of tree size distributions, we will generate thousands of synthetic forest vegetation profiles. These synthesized forest canopy profiles can be rapidly and efficiently compared with vegetation profiles and matches can easily be identified. In this way, unique vegetation profiles are associated with forest structure. Our challenge is to plumb the information in the high resolution optical and lidar data and the full waveform of larger footprint lidar (GLAS ICESat) to extract forest structure information, particularly the tree size distributions on the landscape level. Tropical forests hold a sense of awe and wonder, inspiring exploration amongst researchers, students, and the general population. In addition, remote sensing and use of satellite imagery have become common in popular culture. However, the connections between these ubiquitous technologies and ecological research are less clear to the general population. High resolution image data provide a wonderful opportunity to create a bridge because there are obvious parallels between what one sees in an image, and how that image is represented statistically, allowing for opportunities for education and public outreach. We will develop a scaffolded approach to improve understanding of the underlying concepts and potential of ecosystem imaging, from remote sensed platforms and their application to understand forests, specifically tropical forests, which hold a sense of interest based on their complexity, diversity, and unique flora and fauna. We will rely on existing infrastructure in the University of New Hampshire's Joan and James Leitzel Center for Mathematics, Science, and Engineering Education and their network of middle and high school teachers to develop and implement our education plan. Development of learning tools and educational components will inform how scientists use imaging as a research tool in Earth System Science.

2015 NASA Carbon Cycle & Ecosystems Joint Science Workshop Poster(s)

• Estimating forest structure in tropical forested sites using lidar point cloud data   --   (Franklin B. Sullivan, Michael Palace, Michael Keller, Ekena Rangel Pinage, Maiza Nara dos Santos)   [abstract]

2011 NASA Carbon Cycle & Ecosystems Joint Science Workshop Poster(s)

• Fake it till you make it: What can we learn about simulating forest structure?   --   (Michael Palace, Bobby H. Braswell, Stephen Hagen)   [abstract]