| Abstract Detail
Recent Topics Posters Burke, Antigone [1], Murray, Bryan [2], Adams, Henry D. [3], Gholizadeh, Hamed [2]. Selecting vegetation indices to estimate live fuel moisture in Oklahoma grasses. Climate models predict an increase in the number and severity of droughts in Oklahoma, thereby increasing wildfire risk. In wildfires, live fuel moisture (LFM), the water content in living plants, plays an important role. High LFM decreases available fuels and slows rate of consumption. Currently, in Oklahoma, LFM is estimated using relative greenness calculated from Normalized Difference Vegetation Index (NDVI). There have been several studies suggesting the use of other remote sensing vegetation indices’ to better predict LFM. The goal of this research is to assess the ability of different vegetation indices to predict LFM in dominant Oklahoma grasses (Andropogon gerardii, Bouteloua gracilis, Panicum virgatum, and Schizachyrium scoparium).
A. gerardii, B. gracilis, P. virgatum, and S. scoparium were grown and then subjected to water stress in a greenhouse. Leaf-level spectra and LFM data were measured daily at midday. Spectra were recorded using a hyperspectral spectroradiometer (ASD FieldSpec® 3, 2151 bands, 350-2500 nm) with plant probe attachment. The relationships between vegetation indices and LFM were analyzed using linear, polynomial, and piecewise models. Models were ranked using AIC.
The relationship between LFM and leaf reflectance, as well as the best-supported model varied among species. These preliminary findings support the need for unique LFM models for different Poaceae species/prairie communities. My continuing research will use the relationship between LFM and vegetation indices, along with soil moisture and weather data, to improve wildfire risk models in Oklahoma.
1 - Oklahoma State University, Natural Resource Ecology and Management, 008C Ag Hall, Stillwater, Oklahoma, 74078, United States
2 - Oklahoma State University
3 - Washington State University
Keywords:
fire
Oklahoma
live fuel moisture
remote sensing.
Presentation Type: Recent Topics Poster
Number: PRT002
Abstract ID:1308
Candidate for Awards:None |
