Allowing Existing Vegetation to Empirically Define Its Own “Phenological Year” to Quantify the Degree and Timing of Seasonality Across the United States

Authors

William W. Hargrove
USDA Forest Service Southern Research Station
Danny C. Lee
USDA Forest Service Southern Research Station
Steven P. Norman
USDA Forest Service Southern Research Station
Forrest M. Hoffman (forrest at climatemodeling dot org)
Oak Ridge National Laboratory
Jitendra Kumar
Oak Ridge National Laboratory

Session

Phenology
Tuesdy, April 11, 2017 10:20 am–10:04 am
Constellation E

Abstract

Polar analysis of the annual distribution of MODIS NDVI allows for temporally seamless determination of vegetation seasonality, irrespective of the human calendar. Evergreen vegetation appears as a perfect circle in such a polar graph presentation. Circular statistics can be used to calculate a mean resultant vector for the annual NDVI distribution over a number of years. The magnitude of this mean NDVI resultant vector describes how far “off-center” the center of mass of the NDVI distribution is, and therefore characterizes the degree of seasonality of the mixture of existing vegetation present within each MODIS cell. The angle of the mean NDVI resultant vector indicates the Day-of-Year of the center of mass of the entire multi-year NDVI distribution, showing the most-active date for NDVI greenness. Conversely, the anti-vector date indicates the least-active or most dormant greenness date, dividing the annual cycle at the most sensible beginning and ending date, and creating a unique, vegetationally defined “phenological year” in every MODIS cell. Once the strength and mean date of seasonality are quantitatively defined, national maps can be drawn showing the nature of dominant vegetation seasonality.


Forrest M. Hoffman (forrest at climatemodeling dot org)