Analysis of remote sensing data to monitor changes in phenology

Faculty Mentors:

Erin Schliep (Statistics, NCSU)
Josh Gray (Center for Geospatial Analytics, NCSU)

Prerequisites: Basic statistics, computing

Outline: As the climate changes, so does phenology. Slight changes in phenological variables such as the start (SOS) and end of season (EOS) can have major ecological and agricultural ramifications [1]. Remote sensing measures of normalized difference vegetation index (NDVI) are now available at fine temporal (daily) and spatial (30 m) scales. These data for the past 30 years will be used to study phenological evolution in agricultural regions of the US Midwest [2].

Objectives: Students will analyze daily NDVI measurements using nonlinear regression models to estimate the SOS in each year and each pixel, and regression models to understand the climate drivers of this phenomenon and to map changes in SOS.

Outputs: Local estimates of the change in phenology and deepen our understanding.

References:

[1] North, J., E. Schliep, and C. Wikle, On the spatial and temporal shift in the archetypal seasonal temperature cycle as driven by annual and semi-annual harmonics. Environmetrics, 2021. 32: p. e2665.

[2] Gao, X., J. Gray, and B. Reich, Long-term, medium spatial resolution annual land surface phenology with a Bayesian hierarchical model. Rem Sens Environ, 2021. 261: p. 112484.