GNN attributes of interest

The big picture of how these will be used: Categorical data must be an exact match (i.e. they pre-filter the potential selected plots). Numeric data can be matched with a kdtree approach (i.e. finding the nearest plot in n-dimensional attribute space)

Categories

There seem to be lots of categorical variables that overlap in purpose. The best appear to be:

Structure

• VEGCLASS (vegetation class e.g. "Conifer, large, mod/closed")
• SIZECL (size class)
• COVCL (coverage class i.e. open vs closed)

Species

• IMAP_DOMSPP (dominant species)
• HDW_PLIV (hardwood species of greatest importance)
• CON_PLIV (conifer species of greatest importance)
• FORTYPPIV (forest type; if mixed, names are hyphenated)

Numeric

• TREER (count of tree species)
• SDI_REINEKE (stand density index)
• BAA_GE_3 (basal area)
• THP_GE_3 (trees per hectare)
• STNDHGT (stand height)
• CANCOV (canopy cover; also has a _CON, _HDW, _DOM for conifer, hardwood and dominant)
• QMDA_DOM (quadratic mean diameter of dominant species; QMDC and QMDH to separate conifer vs hardwood)
• BAC_PROP (Proportion of basal area that is conifer)
• IMAP_QMD (mean diameter (inches) of top 25% of the highest trees)

Reference

See http://www.fsl.orst.edu/lemma/main.php?project=common&id=mr&model_region=200&ref=nwfp15