Skip to main content
ARS Home » Research » Publications at this Location » Publication #146538


item Cosh, Michael

Submitted to: Journal of Hydrology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/15/2003
Publication Date: 9/1/2003
Citation: Cosh, M.H., Brutsaert, W. 2003. Microscale structural aspects of vegetation density variability. Journal of Hydrology. 276:128-136.

Interpretive Summary: Vegetation is a critical element of the land surface for understanding land surface atmosphere interactions. To properly model the surface, the spatial structure of the vegetation must be determined. The field of fractal analysis has opened up a new area of research for investigating spatial variables, such as vegetation, and this work establishes the validity of characterizing vegetation density as a fractal at a scale of concern for many atmospheric and land surface modelers. Three types of analyses are used to reveal the fractal nature of the variable, and previous studies are supported for showing the relationship between these methods from a statistical point of view. Further conclusions are drawn about how the land use/land cover affects the fractal dimension of vegetation, which a common calculation when dealing with this area of research. This work will have an impact on modelers who are attempting to solve the problem of spatial variability at scales smaller than models currently resolve, by allowing the variable's distribution to be characterized by a fractal, which has the distinction of being self-similar.

Technical Abstract: Current radiometric remote sensing technology enables high resolution data of vegetation biomass to be obtained in the form of the Normalized Difference Vegetation Index (NDVI). The present study uses such data to examine the spatial structure of vegetation density at the land surface by wavelet, semivariogram, and spectral analyses. Within the range of 30 m to 800 m, the results show that vegetation density of arable cropland is persistent and can be characterized as a fractal with a dimension of 1.59. Within the same range of distances, vegetation density across a more diverse landscape (including besides cropland, also pasture and savanna) exhibits a fractal dimension of 1.69. In the Southern Great Plains, this range is of the order of the typical size of agricultural fields.