Investigation of Remote Sensing Derived Surface Temperature and Normalized Difference Vegetation Index and Implications for Land Cover Classification
No Thumbnail Available
Date
2013-08
Authors
Advisors
Journal Title
Series/Report No.
Master of Natural Resources Professional Papers (North Carolina State University. College of Natural Resources)
Journal ISSN
Volume Title
Publisher
North Carolina State University. College of Natural Resources
Abstract
ABSTRACT
Davis, Francis George. Investigation of Remote Sensing Derived Surface Temperature and Normalized Difference Vegetation Index and Implications for Land Cover Classification. (Under the direction of Dr. Gary Blank, Dr. Ryan Emanuel, and Dr. Heater Cheshire).
This research investigates the relationship between surface temperature (Ts) and the Normalized Difference Vegetation Index (NDVI) as a proxy for leafy green biomass across a landscape in Southern Kenya. The two parameters, derived from Landsat 7 data are compared between and across land cover types in 2000 and 2005 to see how the relationship varies among different ecosystems. This work serves to investigate how plant community composition affects energy transformation. It is well documented that NDVI and Ts are inversely proportional and correlated. However, based on the assumption that different ecosystems process solar energy differently, it is hypothesized that the strength and direction of this relationship will vary significantly between different land cover types. Due to this variable relationship Ts is a useful ancillary data source for land cover classification, especially in the early stages of a hierarchical classification protocol. Ts is demonstrated to be more robust and resilient to inter-annual climatic variation compare to NDVI. Furthermore, Tukey’s Honest Significant Difference based on mean rank found that the number of land classes that were significantly different increased when both Ts and NDVI were used rather than either alone.
