Machine learning bedrock property prediction

Knowledge graph centered on Machine learning bedrock property prediction with 15 nodes and 36 connections. Top connected: Crested Butte, Engelmann Spruce, Picea/Abies spp., Surface parameters and bedrock properties covary a, Snodgrass Mountain.

Method synopsis

Extremely randomized trees regression model to predict subsurface electrical resistivity from surface geomorphological and vegetation parameters, enabling watershed-scale bedrock characterization.

Synthesized from method descriptions across 1 paper using this protocol.

Procedure as described in the canonical source

Steps below were extracted from the paper that introduces this protocol — Surface parameters and bedrock properties covary across a mountainous watershed: Insights from machine learning and geophysics (2022), Science Advances. Implementations in other papers (listed below) may differ.

1. Machine learning model training
   Extremely randomized trees regression (ERTr) model trained using scikit-learn package in Python. Model used topographic elevation, slope, aspect, and vegetation type as estimators to predict shallow subsurface electrical resistivity. Training performed on 13,524 data points with 70% used for training and 30% for validation.
   Quantities: 13,524 total data points, 70% training/30% validation split, 2705 validation pointsDuration: Not specifiedConditions: Computational analysis on scaled data from northern and southern Mancos Shale domains
   Equipment: Python, scikit-learn package, ERTr algorithm

Typical Equipment

• Python
• scikit-learn package

Output Measurements

• predicted electrical resistivity maps
• fracture density estimates
• model accuracy metrics

Papers Using This Protocol (1)

Species Studied With This Protocol

Places Where Used