Abstract:
|
Plant phenology is the study of life cycles in plants, and how they are affected by variations in climate. Phenological events, such as first flowering in spring, are very sensitive to changes in climate and are therefore important indicators of the impacts of climate change. Early studies of landscape phenology were constrained to small areas due to the difficulty of conducting surveys over large regions. However, in recent decades, remote sensing data on satellite derived variables (such as chlorophyll content) have become widely available, which has provided an impetus to studies of phenology on a global scale. To study phenology with remote sensing data, variables representing phenological events must be identified. The current methodology employed by scientists involves fitting models separately to each year to allow for changes over time. We present an approach using dynamic linear models on weekly MTCI data from 2003 to 2007 to model phenological patterns over southern India. We estimate the variables of onset of greenness, peak of greenness, and end of senescence using an iterative search, and propose a method for describing uncertainty using Monte Carlo simulations.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.