Abstract This dataset includes all of the data needed to validate and/or reproduce the manganese calibration model described in the manuscript. The reference database contains metadata for the new Mn-bearing standards, minerals, and mixtures that are > 2.9 wt.% MnO. In addition, the files include the MnO composition data for all standards used, non-normalized spectral data, mean peak area spectrum, results of outlier determination, RMSECV data, regression vectors, and Test Set predictions. Description of files Supplement_MnO_Cal_Input_outliers_wvl.csv contains MnO composition (Column B), the name and filename of the sample (Column C and D), metadata (Columns E through P), and spectral files for each standard (Columns Q through IBX). The standard filename is in the format MonthDay_Clock_Year_CCS_SampleName_SampleNumber, where CCS is defined as “clean calibrated spectrum.” The first column of the metadata includes whether the sample is a mixture or not for samples with >2.9 wt.% MnO, and if so, what standards are mixed to make the mixture. Column F states the sample type, which is whether the sample is a standard, mixture, or a mineral, and if so, what mineral it is for samples with >2.9 wt.% MnO. Columns G and H state whether the sample spectrum was used in the Test Set or Training Set. Columns I through N show the outcome of the outlier identification using different methods described in the manuscript. Each standard is composed of five spectra, 50 individual shots averaged together. The second line has the wavelength channel while the subsequent lines contain the spectral intensity data for each sample spectrum. Each of the CV*.csv files contain all of the Root Mean Squared Error of Cross Validation (RMSECV) data used for each technique as described in the manuscript. The minima determined from each cross validation run is listed in the Supplement_CV_summary_results.xlsx file. The technique with the lowest RMSECV in the Supplement_CV_summary_results.xlsx file is highlighted in green. This summary information was used to construct Table 1 in the manuscript. Supplement_Model_Means.csv contains the average peak area spectrum from the model, which was used to create the peak area spectra from the normalized spectra for the models tested as described in the manuscript. Supplement_Test_Set_Predictions.csv contains the Test Set predictions for each of the full models, submodels, and blended models that were used by the double blended model described in the manuscript. Supplement_Model_Regression_Vectors.csv contains the regression vectors for the submodels used in the double blended model described in the manuscript.