Repository logo

Research Data

Permanent URI for this collection

https://hdl.handle.net/10217/171906

Browse

Browsing Research Data by Author "Boone, Randall B"

Now showing 1 - 1 of 1
  • No Thumbnail Available
    ItemOpen Access
    Data associated with Boone (2020) "Hierarchical global plant biophysical regions as potential analysis units"
    (Colorado State University. Libraries, 2020) Boone, Randall B
    Regional and global vegetation simulations can be problematic when analysis units to which parameters are assigned do not align with plant productivity and phenology. Having a suite of pre-defined biophysical regions at a variety of scales that correspond to differences in plant productivity and phenology would allow analysts to select a set of analysis units at the scale needed. In other cases, environmental or social responses may be hypothesized to be related to differences in plant dynamics. One may compare the discrimination in such data that biophysical regions at difference scales provide to determine which best distinguishes the responses in question, such that like responses fall within the same regions to the degree possible. If those relationships are significant, the responses may then be extrapolated based on the biophysical regions. I defined hierarchical biophysical regions based on plant productivity and phenology by clustering global 0.083° Normalized Difference Vegetation Indices over a 10-year period. Agglomerative average-linkage distances based on squared error between clusters was conducted using an iterative sampling approach to merge more than 2 million clusters into fewer and fewer clusters based on NDVI greenness profiles comprised of 240 values over 10 yrs, until all cells were in a single cluster. Greater and greater differences in greenness profiles were ignored at higher levels of the hierarchy. Using a difference increment of 0.1, 253 non-duplicative sets of clusters were created, and 107 of those were included in animations that may be used to explore differences in global plant dynamics. Differences in clusters were quantified based on comparing the focal set of cluster results with 10 other cluster sets. Analysts may use the hierarchical clusters to improve the alignment of their parameter sets that inform plant growth and other dynamics with real-world plant dynamics.