Year of Award

2015

Document Type

Thesis

Degree Type

Master of Science (MS)

Degree Name

Resource Conservation

Department or School/College

College of Forestry and Conservation

Committee Chair

Steve Seibert

Commitee Members

Brady Allred, Matt Reeves

Keywords

rangeland phenology, random forest model, spatial model, rangeland, climate model

Subject Categories

Natural Resources and Conservation | Other Forestry and Forest Sciences | Statistical Models

Abstract

Plant phenology has long been used as an indicator of climate. Recent changes in plant phenology are evidence of the influence of climate change. Modeling plant phenology has become an effective tool to understand the impacts of climate change. Using machine learning techniques I developed a modeling process for accurately predicting phenology across a diverse landscape. This model uses individual site data to set site specific climate thresholds for plant phenology. This model also identifies the limiting factors to vegetation phenology for rangelands in the western United States. NDVI remotely sensed data was used to quantify land surface phenology and DAYMET data was used to quantify climate variables. I found that random forest modeling can predict observed plant phenological dates across western rangelands to within a single day for start of season, end of season and day of max NDVI. The model can also identify the most highly correlated variables for phenological events in the study area and highlight which variables limit growth in different vegetative communities. These results confirm previous work on drivers of temperate phenology. This study’s results show that random forest modeling can accurately identify the most important climate variables for phenological events and use those variables to predict phenological events on a large spatial scale.

Download

Share

COinS
 

© Copyright 2015 Joseph R. St. Peter