This repository contains exploratory spatial data analysis (ESDA) functions and scripts that were used in Smith et al. "Exploratory Spatial Data Analysis for Geothermal Resource Assessments: An Appalachian Basin Case Study" paper submitted to Geothermics.
These functions are developed for ESDA on geothermal spatial datasets and are applicable to other spatial datasets.
This repository depends on functions located in the following repositories:
calvinwhealton -> geothermal_pfa -> outliers
jds485 -> Geothermal_DataAnalysis_CrossSections
The methods used in ESDA_Main.R include:
-
Identification and processing of data in the same spatial location i) Nugget semi-variance calculation for data in same exact spatial location ii) Plot of similarity for data in the same locations based on a covariate (depth of measurement)
-
Local Median/Mean Deviation (LocalDevition.R)
-
Local Spatial Outlier Analysis (see geothermal_pfa repository for R script containing the functions) i) Plots by depth slices
-
Q-Q plots
-
Nonparametric Local Outlier Analysis i) KS test on depth rank distribution of outliers in local neighborhoods ii) Chi-squared test for depth bins
Additional Contributions:
-
Discrete Color Function for Map Making (ColorFunctions.R)
-
Sensitivity analysis of outlier algorithm for this dataset (OutAlgoSensitivity.R)
-
Jackknife confidence intervals for semi-variograms (JackknifeSemivariogramConfInts.R)
-
Diagnostic plots to discover operators that had systematically rogue data (bias and/or variance) as a result of their data recording behavior. (OperatorDiagnostics.R)
Note on running script: Input data used to run the script for Smith et al. is provided in the repository. Output data is too large to host on Github. Please write to Jared Smith (jds485@cornell.edu) if you would like to have the output Rdata file.