Harvard Forest Symposium Abstract 2013

Title: Evaluating Forest Inventory in a New England Mixed Hardwood Forest Using Quickbird Data and Image Segmentation
Primary Author: Noam Raffel (Clark University)
Additional Authors: Noam Raffel (Clark University)
Abstract:
Evaluating Forest Inventory in a New England Mixed Hardwood Forest Using Quickbird Data and Image Segmentation

Authors: Noam Raffel1+, John Rogan1, and James Toledano2

1Clark University Graduate School of Geography, Clark University, Worcester MA 01610

2Clark Labs, Clark University, Worcester MA 01610

+Corresponding author: nraffel@clarku.edu

Abstract

Forests are complex ecosystems that are vulnerable to natural threats from invasive species and anthropogenic disturbance, requiring routine monitoring. Forest inventory is implemented to monitor how forests respond to various stressors and distribution of tree type for timber harvesting. This study evaluates fine spatial resolution satellite data to perform stand inventory in northeastern mixed hardwood forest using a new method of supervised segmentation classification using 0.6 m Quickbird imagery and LIDAR data in central Massachusetts. Growth form maps representing hardwood versus conifer stands yielded 78% - 82% accuracy. Conifer species accuracy was 78%, with individual species accuracy ranging 75% - 85%. Unlike conifer species delineation, hardwood species could not be classified to acceptable standards for inventory purposes.
Research Category: Conservation and Management

You are here