Table 1 The list of selected input features for Random Forest forecast.

From: Forecasting induced seismicity in Oklahoma using machine learning methods

Pore pressure rate*

Pore pressure rate (mean), pore pressure rate (max), pore pressure rate (variance), pore pressure rate (20th percentile), pore pressure rate (gradient between minimum and maximum values in space)

Pore pressure*

Pore pressure (mean), pore pressure (max)

Poroelastic stress rate (PES rate)*

PES rate (mean), PES rate (max), PES rate (variance), PES rate (20th percentile), PES rate (gradient between minimum and maximum values in space)

Poroelastic stress (PES)*

PES (mean)

Injection rate*

Monthly injection volume (sum)

Injection PSI*

Monthly injection pressure (skewness)

Cumulative volume*

Cumulative injection volume (20th percentile)

1 year injection volume*

1 year injection volume (max), 1 year injection volume (40th percentile)

Injection depth

Well depth (mean), well depth (max), well depth (min), well depth (variance)

Basement depth

Basement depth (mean), basement depth (max), basement depth (min), basement depth (variance)

Distance to basement

Distance (mean), distance (max), distance (min), distance (variance)

Region parameters

Grid number 0–31 to represent different locations

  1. The complete input features also include the history data of the features denoted by asterisk (*).
Back to article page