A Practical Guide to Geostatistical Mapping

1.1 Basic concepts......Page 22
1.1.1 Environmental variables......Page 25
1.1.2 Aspects and sources of spatial variability......Page 26
1.1.3 Spatial prediction models......Page 30
1.2.1 Inverse distance interpolation......Page 33
1.2.2 Regression on coordinates......Page 34
1.3 Statistical spatial prediction models......Page 35
1.3.1 Kriging......Page 36
1.3.2 Environmental correlation......Page 41
1.3.4 Hybrid models......Page 45
1.4 Validation of spatial prediction models......Page 46
2.1 The Best Linear Unbiased Predictor of spatial data......Page 48
2.1.1 Mathematical derivation of BLUP......Page 51
2.1.2 Selecting the right spatial prediction technique......Page 53
2.1.3 The Best Combined Spatial Predictor......Page 56
2.1.4 Universal kriging, kriging with external drift......Page 57
2.1.5 A simple example of regression-kriging......Page 60
2.2 Local versus localized models......Page 62
2.3 Spatial prediction of categorical variables......Page 63
2.5 Spatio-temporal regression-kriging......Page 65
2.6 Species Distribution Modeling using regression-kriging......Page 67
2.7.1 Some theoretical considerations......Page 72
2.7.2 Choice of auxiliary maps......Page 74
2.8 Regression-kriging and sampling optimization algorithms......Page 75
2.9.1 Soil mapping applications......Page 76
2.9.3 Species distribution modeling......Page 77
2.10.1 Alternatives to RK......Page 79
2.10.2 Limitations of RK......Page 80
2.10.3 Beyond RK......Page 81
3.1.1 ILWIS......Page 84
3.1.2 SAGA......Page 87
3.1.3 GRASS GIS......Page 92
3.2 Statistical computing: R......Page 93
3.2.1 gstat......Page 95
3.2.2 The stand-alone version of gstat......Page 96
3.2.4 Isatis......Page 98
3.3 Geographical visualization: Google Earth (GE)......Page 99
3.3.1 Exporting vector maps to KML......Page 101
3.3.2 Exporting raster maps (images) to KML......Page 103
3.3.3 Reading KML files to R......Page 108
3.4.1 Strengths and limitations of geostatistical software......Page 109
3.4.2 Getting addicted to R......Page 111
3.4.4 Towards a system for automated mapping......Page 117
4.1 Global data sets......Page 120
4.1.1 Obtaining data via a geo-service......Page 123
4.1.2 Google Earth/Maps images......Page 124
4.1.3 Remotely sensed images......Page 127
4.2 Download and preparation of MODIS images......Page 129
4.3 Summary points......Page 135
5.2 Data import and exploration......Page 138
5.2.1 Exploratory data analysis: sampling design......Page 144
5.3.1 Regression modeling......Page 148
5.3.2 Variogram modeling......Page 151
5.3.3 Spatial prediction of Zinc......Page 152
5.4.1 Fitting a GLM......Page 154
5.4.2 Variogram fitting and final predictions......Page 155
5.5.1 Geostatistical simulations......Page 157
5.5.2 Spatial prediction using SAGA GIS......Page 158
5.5.3 Geostatistical analysis in geoR......Page 161
5.6.1 Visualization of uncertainty......Page 166
5.6.2 Export of maps to Google Earth......Page 169
6.1 Introduction......Page 174
6.2.1 Point-sampled values of HMCs......Page 175
6.2.2 Gridded predictors......Page 178
6.3.2 Regression modeling using GLM......Page 181
6.3.3 Variogram modeling and kriging......Page 185
6.4 Automated generation of HMC maps......Page 187
6.5 Comparison of ordinary and regression-kriging......Page 189
7.2 Loading the data......Page 194
7.2.1 Download of the world maps......Page 195
7.2.2 Reading the ISRIC WISE into R......Page 196
7.3 Regression modeling......Page 200
7.5 Adjusting final predictions using empirical maps......Page 205
7.6 Summary points......Page 207
8.1.1 Preparation of maps......Page 210
8.1.2 Auxiliary maps......Page 211
8.2.1 Kernel density estimation......Page 213
8.2.3 Simulation of pseudo-absences......Page 217
8.2.4 Regression analysis and variogram modeling......Page 218
8.3 Final predictions: regression-kriging......Page 221
8.4 Niche analysis using MaxEnt......Page 223
9.1 Introduction......Page 228
9.2 Data download and exploration......Page 229
9.3.1 Variogram modeling......Page 230
9.3.2 DEM filtering......Page 231
9.3.3 DEM generation from contour data......Page 232
9.4 Extraction of Land Surface Parameters......Page 233
9.5.1 Fuzzy k-means clustering......Page 234
9.5.2 Fitting variograms for different landform classes......Page 235
9.6.2 Selection of training pixels......Page 236
9.7 Extraction of memberships......Page 239
10.2 Data download and import......Page 242
10.3.1 Variogram modelling......Page 244
10.3.2 Geostatistical simulations......Page 246
10.4 Generation of stream networks......Page 248
10.5 Evaluation of the propagated uncertainty......Page 250
10.6.1 Objective selection of the grid cell size......Page 252
10.6.2 Stream extraction in GRASS......Page 254
10.6.3 Export of maps to GE......Page 257
11.1 Introduction......Page 262
11.2 Data download and preprocessing......Page 263
11.3 Regression modeling......Page 265
11.4 Space-time variogram estimation......Page 269
11.5.1 A single 3D location......Page 270
11.5.2 Time-slices......Page 272
11.5.3 Export to KML: dynamic maps......Page 273
11.6 Summary points......Page 276