A open-access VIS-NIR soil spectroscopy dataset of the Grayson Riverbend Preserve for exploratory data analysis. Potential modeling outputs with this dataset include soil carbon, nitrogen, and salinity modeling.
Forthcoming...
Forthcoming...
All spectral pretreatment steps were performed using R 4.2.2. using the prospectr package. The details are provided in the RMarkdown document named Grayson_spectral_preprocessing.Rmd. The baseline spectra, which have been smoothed using a Savitzky-Golay (SG) filter with polynomial order of 3 and a window size of 25, can be found under Grayson_savistzky_golay_spectra.csv. The derivative of the SG spectra is named Grayson_savitzky_golay_derivative.csv. Continuum removal was also applied to the SG baseline spectra and the data can be found under Grayson_continuum_removal.csv. Finally, a PCA of the baseline SG spectra was also added.
To aid in analysis and modeling, I have provided a series of Pearson's correlation plots for each of the 4 datasets and % total C.
Fig. 1: Pearson's Correlation between %C and Savitzky-Golay smoothed spectra.
Fig. 2: Pearson's Correlation between %C and the first derivative of Savitzky-Golay smoothed spectra.
Fig. 3: Pearson's Correlation between %C and the Savitzky-Golay smoothed spectra after continuum removal.