ABSTRACT FINAL ID: B53D-05;

TITLE: An Early Warning System for Identification and Monitoring of Disturbances to Forest Ecosystems

SESSION TYPE: Oral

SESSION TITLE: B53D. Remote Sensing of Long-Term Ecological Trends I

AUTHORS: Aaron A. Marshall^1,2, Forrest M. Hoffman², Jitendra Kumar², William W. Hargrove³, Joseph Spruce⁴, Richard T. Mills²

INSTITUTIONS:
¹Thomas Nelson Community College, Yorktown, VA, United States.
²Oak Ridge National Laboratory, Oak Ridge, TN, United States.
³Eastern Forest Environmental Threat Assessment Center (EFETAC), USDA Forest Service, Asheville, NC, United States.
⁴NASA Stennis Space Center, Bay St. Louis, MS, United States.

ABSTRACT BODY: Forest ecosystems are susceptible to damage due to threat events like wildfires, insect and disease attacks, extreme weather events, land use change, and long-term climate change. Early identification of such events is desired to devise and implement a protective response. The mission of the USDA Forest Service is to sustain the health, diversity, and productivity of the nation‘s forests. However, limited resources for aerial surveys and ground-based inspections are insufficient for monitoring the large areas covered by the U.S. forests. The USDA Forest Service, Oak Ridge National Laboratory, and NASA Stennis Space Center are developing an early warning system for the continuous tracking and long-term monitoring of disturbances and responses in forest ecosystems using high resolution satellite remote sensing data. Geospatiotemporal data mining techniques were developed and applied to normalized difference vegetation index (NDVI) products derived from the Moderate Resolution Imaging Spectroradiometer (MODIS) MOD 13 data at 250 m resolution on eight day intervals. Representative phenologically similar regions, or phenoregions, were developed for the conterminous United States (CONUS) by applying a k-means clustering algorithm to the NDVI data spanning the full eight years of the MODIS record. Annual changes in the phenoregions were quantitatively analyzed to identify the significant changes in phenological behavior. This methodology was successfully applied for identification of various forest disturbance events, including wildfire, tree mortality due to Mountain Pine Beetle, and other insect infestation and diseases, as well as extreme events like storms and hurricanes in the United States. Where possible, the results were validated and quantitatively compared with aerial and ground-based survey data available from different agencies. This system was able to identify most of the disturbances reported by aerial and ground-based surveys, and it also identified affected areas that were not covered by any of the surveys. Analysis results and validation data will be presented.

KEYWORDS:
[0430] BIOGEOSCIENCES / Computational methods and data processing,
[1914] INFORMATICS / Data mining,
[0480] BIOGEOSCIENCES / Remote sensing,
[4341] NATURAL HAZARDS / Early warning systems.

SPONSOR NAME: Aaron Marshall

CONTACT: Aaron Marshall <amarshall at email dot wm dot edu>