I. File list
------------
PPM1_code.sas		SAS script file that contains code to run PPM_1.0 (compatible with SAS 9.2 or greater)
PPM1_code_EG.egp	SAS script file that contains code to run PPM_1.0 (compatible with SAS EG)
ppm1_data.sas7bdat	PPM_1.0 training data required for running PPM_1.0
README.txt		This file.  Contains file list, design, installation and operating instructions
User_data.xls		Example xls spreadsheet that is formatted for user's paleosol data.  The file contains data from the Ngira paleosol (Driese et al., 2016)


II. Design
----------
The PPM_1.0 was designed in SAS version 9.2 and SAS EG.  This program calculates the Partial Least Squares (PLS) regressor scores for paleosol geochemical oxides using the 685 mineral soil B horizons as a training set.  The resulting regressor scores are modeled using the PPM_1.0 thin-plate spline (TPSPLINE) model.

The PPM_1.0 works in the following SAS versions:

SAS EG 4.3
SAS 9.2 (at least)


III. Installation and operating instructions
--------------------------------------------
1. After downloading zip file, extract all files, "PPM1_data" folder, to computer.  Do not alter the name of the folder.
2. Open SAS or click on the PPM1_code.sas file, which will open in SAS.
3. In the SAS code, define the file paths for PPM1 SAS dataset and user Excel data set.  These include the LIBNAME and FILENAME.  This step shows SAS where to find the training and user data.  This section of the code is near the top of the file and looks like this:

LIBNAME Location "C:\...\PPM1_data";
FILENAME user "C:\...\User_data.xls";
/*\*/*\*/*\*/*\*/*\*/*\*/*\*/*\*/*\*/*\*/*\*/*\*/*\*/*\*/*\*/*\*/*\*/*\*

4. Determine your file paths and modify the above section of code.  Do not alter any other portion of this code.
5. Enter User data into User_data.xls spreadsheet.  ***Follow the format in the example record provided in the spreadsheet.  Pedon ID cannot exceed 25 characters. Geochemical data should be entered in oxide weight percent, wt. %. SAS will remove records if information is missing.  You must have a Pedon ID and value for each oxide and correction factors (see 6 below) for SAS to recognize the record.
6. In the User_data.xls spreadsheet, Indicate whether your data weathered in an Glacial Till (1=yes, 0=no) or Subglacial Till (1=yes, 0=no).
7. Save the User_data.xls.
8. Run SAS script.  The output will show in a new window/tab within SAS.

*If you require further assistance please watch the online tutorial video.


IV. Example Output from two example records provided
----------------------------------------------------

Predictions


Obs Pedon_ID low_MAP best_MAP high_MAP low_MAT best_MAT high_MAT
1   Ngira20  1299    1769      2238	16.8	20.9	25.0




The output includes the best MAP (best_MAP) and MAT (best_MAT) predictions along with the low and high Root Mean Squared Prediction Errors (RMSPE), low_MAP, low_MAT, high_MAP and high_MAT, in mm yr-1 and °C. Note that the RMSPE are often higher than the Root Mean Squared Error (RMSE) for MAP (228 mm) and MAT 2.46 degrees C). As of the year 2015, most paleosol-based paleoclimate reconstructions report RMSE.