Estimate of annual vegetation productivity for Kenya’s Amboseli National Park from 2002-2012, based on 8-day Moderate Resolution Imaging Spectroradiometer (MODIS) Normalized Difference Vegetation Index (NDVI) imagery. Areas with significant (90 percent) trends are highlighted in color.
Team Location: Goddard Space and Flight Center (GSFC), Greenbelt, Maryland
Authors:
Oluwakonyinsola Adesoye (Eleanor Roosevelt High School, Greenbelt, Maryland)
Anne Clark Baker (Clark University)
Michael Z. Gao (Johns Hopkins University)
Emily Voelker (University of Maryland, College Park)
Ryan T. Williams (Clark University)
Quinten Geddes (DEVELOP Goddard Center Lead)
Mentors/Advisers:
Jeffrey Masek, Ph.D. (NASA, GSFC)
Frederick Policelli (NASA, GSFC)
Gerasimos Michalitsianos (Science Systems and Applications Inc., GSFC)
Past/Other Contributors:
John David (Science Systems and Applications Inc., GSFC)
DEVELOP West Africa Eco Forecasting Team Spring 2013 (GSFC)
Abstract:
In southern Kenya, the Amboseli National Park is home to hundreds of plant and animal species. Over the years, vegetation landcovers have changed drastically, greatly affecting other organisms in the Amboseli. The woodland landcovers have contracted from about 30 percent of the area to only 10 percent, while swamps and grasslands have been increasing. The Amboseli Conservation Centre (ACC) is an organization dedicated to monitoring the park. However, they have had difficulties in distinguishing the different types of vegetation using remote sensing, causing discrepancies between their ground-truthed data and remotely sensed results. These limitations prevent the ACC from producing comprehensive solutions for restoring and maintaining the Amboseli. By processing Moderate Resolution Imaging Spectroradiometer (MODIS) vegetation index imagery in TIMESAT, an open source time series analysis software, this project observed the baseline phenological characteristics of different vegetation types, which allowed for a derived temporal classification to distinguish each vegetation type using a decision tree methodology. By providing the ACC with a refined vegetation classification method that can be coupled with their ground-truthed data, this project increases capability for monitoring of habitat degradation issues in the Amboseli.