Proxy Development

The Paleosol-Paleoclimate Model version 1.0 (PPM1.0)

The Random Forest model for Mean Annual Precipitation (RF-MAP)

Soil and Paleosol pH


Proxy Information

Inputs: 11 major and minor elemental oxides from soil/paleosol uppermost B horizons

Response: Mean annual precipitation and temperature (joint response)

Method: Thin-plate spline fitted to partial least-squares regression scores

Software used: SAS

For more information, go visit Gary Stinchcomb

Inputs: 10 major and minor elemental oxides from soil/paleosol uppermost B horizons

Response: Mean annual precipitation 

Method: Random forest regression

Software used: JMP, RStudio


Code to run RF-MAP in RStudio can be found here.

Inputs: Bulk oxides (in moles), including Al2O3, CaO, Fe2O3, and TiO2, from soil/paleosol uppermost B horizons

Response: Soil pH (1:1 soil:H2O method) 

Method: Sigmoidal regression

Software used: JMP, RStudio

Related Publications