Advanced Mapping of Environmental Data (Geographical Information Systems)

Advanced Mapping of Environmental Data: Geostatistics, Machine Learning and Bayesian Maximum Entropy......Page 5
Table of Contents......Page 7
Preface......Page 13
1.1. Introduction......Page 17
1.2.1. Spatial data analysis: typical problems......Page 19
1.2.2. Spatial data analysis: methodology......Page 21
1.2.3. Model assessment and model selection......Page 24
1.3.2. Software......Page 28
1.4. Conclusion......Page 30
1.5. References......Page 31
2.1. Introduction......Page 35
2.2. Spatial clustering and its consequences......Page 36
2.2.1. Global parameters......Page 37
2.2.2. Spatial predictions......Page 38
2.3.2.1. Topological indices......Page 39
2.3.2.2. Statistical indices......Page 40
2.3.3. Dimensional resolution: fractal measures of clustering......Page 42
2.3.3.1. Sandbox method......Page 43
2.3.3.2. Box-counting method......Page 46
2.3.3.3. Lacunarity......Page 49
2.4. Validity domains......Page 50
2.5. Indoor radon in Switzerland: an example of a real monitoring network......Page 52
2.5.2. Topological index......Page 53
2.5.3.1. Morisita index......Page 54
2.5.3.2. K-function......Page 55
2.5.4.1. Sandbox and box-counting fractal dimension......Page 56
2.5.4.2. Lacunarity......Page 58
2.6. Conclusion......Page 59
2.7. References......Page 60
3.1. Assumptions of geostatistics......Page 63
3.2. Family of kriging models......Page 65
3.2.2. Ordinary kriging......Page 66
3.2.3. Basic features of kriging estimation......Page 67
3.2.5. Lognormal kriging......Page 72
3.3.2. Kriging with external drift......Page 74
3.3.3. Co-kriging......Page 75
3.3.4. Collocated co-kriging......Page 76
3.3.5. Co-kriging application example......Page 77
3.4.1. Indicator coding......Page 80
3.4.2. Indicator kriging......Page 82
3.4.3.1. Indicator kriging for 241Am analysis......Page 85
3.4.3.2. Indicator kriging for aquifer layer zonation......Page 87
3.4.3.3. Indicator kriging for localization of crab crowds......Page 90
3.5.1. Simulation vs. estimation......Page 92
3.5.2. Stochastic simulation algorithms......Page 93
3.5.3. Sequential Gaussian simulation......Page 97
3.5.4. Sequential indicator simulations......Page 100
3.5.5. Co-simulations of correlated variables......Page 104
3.6. References......Page 108
4.1. Introduction......Page 111
4.2.1. The three learning problems......Page 112
4.2.2. Approaches to learning from data......Page 116
4.2.3. Feature selection......Page 117
4.2.4. Model selection......Page 119
4.2.5. Dealing with uncertainties......Page 123
4.3. Nearest neighbor methods......Page 124
4.4.1. Multi-layer perceptron neural network......Page 125
4.4.2. General Regression Neural Networks......Page 135
4.4.3. Probabilistic Neural Networks......Page 138
4.4.4. Self-organizing (Kohonen) maps......Page 140
4.5.1. VC dimension and structural risk minimization......Page 147
4.5.2. Kernels......Page 148
4.5.3. Support vector machines......Page 149
4.5.4. Support vector regression......Page 153
4.5.5. Unsupervised techniques......Page 157
4.5.5.1. Clustering......Page 158
4.5.5.2. Nonlinear dimensionality reduction......Page 160
4.7. References......Page 162
5.1. Introduction......Page 165
5.2. Air temperature modeling with machine learning algorithms and geostatistics......Page 166
5.2.1.1. Data description......Page 167
5.2.1.2. Variography......Page 168
5.2.1.3. Step-by-step modeling using a neural network......Page 169
5.2.1.4. Overfitting and undertraining......Page 170
5.2.1.5. Mean monthly air temperature prediction mapping......Page 172
5.2.2.1. The Föhn phenomenon......Page 175
5.2.2.2. Modeling of instant air temperature influenced by Föhn......Page 176
5.2.3.1. Temperature inversion phenomenon......Page 179
5.2.3.2. Terrain feature extraction using Support Vector Machines......Page 180
5.2.3.3. Temperature inversion modeling with MLP......Page 181
5.3. Modeling of precipitation with machine learning and geostatistics......Page 184
5.3.1.1. Data description......Page 185
5.3.1.2. Precipitation modeling with MLP......Page 187
5.3.2.1. Data description......Page 189
5.3.2.2. Practical issues of MLP modeling......Page 190
5.3.2.3. The use of elevation and analysis of the results......Page 193
5.3.3.1. Neural network residual kriging......Page 195
5.3.3.2. Neural network residual simulations......Page 198
5.3.4. Conclusions......Page 200
5.4.1. k-nearest neighbor algorithm......Page 201
5.4.2. Automatic mapping of spatial data......Page 203
5.4.2.1. KNN modeling......Page 204
5.4.2.2. GRNN modeling......Page 206
5.4.3. Automatic classification of spatial data......Page 208
5.4.3.1. KNN classification......Page 209
5.4.3.2. PNN classification......Page 210
5.4.3.3. Indicator kriging classification......Page 213
5.4.4. Automatic mapping – conclusions......Page 215
5.5.1. SOM analysis of sediment contamination......Page 216
5.5.2. Mapping of socio-economic data with SOM......Page 220
5.6.1. Indoor radon measurements......Page 225
5.6.2. Probability mapping......Page 227
5.6.3. Exploratory data analysis......Page 228
5.6.4.1. Variogram for indicators......Page 232
5.6.4.2. Variogram for Nscores......Page 233
5.6.5. Neighborhood parameters......Page 234
5.6.6.1. Probability maps with IK......Page 235
5.6.6.2. Probability maps with SGS......Page 236
5.6.7.1. Influence of the simulation net and the number of neighbors......Page 237
5.6.7.2. Decision maps and validation of results......Page 238
5.7. Natural hazards forecasting with support vector machines – case study: snow avalanches......Page 241
5.7.1. Decision support systems for natural hazards......Page 243
5.7.2. Reminder on support vector machines......Page 244
5.7.2.1. Probabilistic interpretation of SVM......Page 245
5.7.4. Temporal forecasts......Page 246
5.7.4.1. Feature selection......Page 247
5.7.4.2. Training the SVM classifier......Page 248
5.7.4.3. Adapting SVM forecasts for decision support......Page 249
5.7.5.1. Data preparation......Page 253
5.7.5.2. Spatial avalanche forecasting......Page 255
5.8. Conclusion......Page 257
5.9. References......Page 258
6.1. Conceptual framework......Page 263
6.2.1. The spatiotemporal continuum......Page 267
6.2.2. Separable metric structures......Page 269
6.2.3. Composite metric structures......Page 271
6.2.4. Fractal metric structures......Page 272
6.3. Spatiotemporal random field theory......Page 273
6.3.1. Pragmatic S/TRF tools......Page 274
6.3.2. Space-time lag dependence: ordinary S/TRF......Page 276
6.3.3. Fractal S/TRF......Page 278
6.3.4. Space-time heterogenous dependence: generalized S/TRF......Page 280
6.4.1. The fundamental equations......Page 283
6.4.2. A methodological outline......Page 289
6.4.3. Implementation of BME: the SEKS-GUI......Page 291
6.5. A brief review of applications......Page 297
6.5.1. Earth and atmospheric sciences......Page 298
6.5.2. Health, human exposure and epidemiology......Page 307
6.6. References......Page 315
List of Authors......Page 323
Index......Page 325