This repository centers around (1) extracting, processing, and geo-locating AOP data with field-collected Foliar Nitrogen data and (2) wrangling and visualizing the extracted data for the D03 (2014) and D17 (2013) prototypes.

The analysis is currently divided into two main pieces: (1) Analyzing plot-scale (i.e. all bright pixels averaged) NDNI vs. Foliar Nitrogen across sites, and (2) Analyzing single-stem-scale spectra (i.e. a single pixel, VIS - NIR wavelengths) vs. Foliar N across sites (this is the PLS analysis).

Below is an abstraction of the workflow for my 2015 AGU poster. Airborne data extraction scripting file workflow (where "[]" indicates a folder/file and "{}" indicates an input/output):

(Step 1 - Extract AOP data)

{inputs: all HDF5 files for a single site; plot boundaries shapefile; flightlineID per plotID} >> [canopyN/ExtractPlotSpectra_siteID.py] >> {outputs: NDNI pixels for each unique plotID, as separate HDF5 files}

(Step 2 - Regress NDNI ~ Foliar N)

{inputs: NDNI calculated per pixel per plotID; foliar nitrogen data per plot} >> [canopyN/processFieldData_siteID] >> {outputs: plot averaged NDNI; plot averaged foliar nitrogen}

(Step 3 - Extract Single Pixel per tree stem coordinate)

{inputs: HDF5 file per plotID; field coordinates of tree stems} >> [pls/pull_spectra_from_h5_siteID.R] >> {outputs: a 3D array of reflectance per unique stemID per plotID}

(Step 4 - Regress Single Pixel against foliar N per stemID)

{inputs: stemID reflectance; stemID foliar_N} >> [cflagg/pls_foliarN.Rmd] >> {outputs: PLS regression results and figures}