Estimating Grassland Biophysical and Biochemical Properties Using Remote Sensing and Modelling

Grasslands comprise about 42% of the world's plant cover and provide critical ecosystem services to the planet. In recent decades, worldwide grasslands have degraded greatly in terms of area and condition, mainly due to human activities and environmental changes. It is therefore critical to investigate grassland heath and adopt corresponding conservation strategies. Remote sensing is an efficient tool for investigating grasslands, owing to its broad spatial coverage and capability of acquiring data repeatedly. Previous studies have utilized remote sensing to investigate various grassland features, such as disturbances (e.g., wildfire, grazing), species composition, and vegetation properties (e.g., leaf area index, chlorophyll). However, limited by the attributes of remote sensing data (e.g., spatial or spectral resolution) or image analysis methods, there are still research gaps in studying grasslands using remote sensing that warrant further exploration. In this research, multi-source remote sensing data with different spatial, temporal, and spectral features were applied to investigate different grassland properties. Specifically, multi-temporal medium-spatial resolution multispectral imagery acquired by satellite was utilized to investigate a wildfire disturbance on a mixed grassland. Multi-temporal ultra-high spatial resolution multispectral imagery acquired by Unmanned Aerial Vehicle was processed to identify species composition and estimate vegetation properties in a tall grassland using an empirical method. Lastly, multi-temporal high-spatial resolution hyperspectral imagery acquired by manned helicopter was utilized to estimate vegetation properties of a tall grassland using radiation transfer modelling. Results show that grassland features can be accurately retrieved from the chosen imagery, but one should keep in mind that different imagery types have differences in spectral, spatial, and temporal characteristics, and thus are suitable for different research purposes. These research findings provide insights on selecting appropriate remote sensing data and processing techniques for investigating different grassland features, and further contribute to the understanding of grassland ecosystems and grassland conservation.Ph.D.2018-12-19 00:00:0