1 Introduction ............................................... 1
1.1 The need of innovative modelling methods ................... 3
1.1.1 Inverse modelling ................................... 4
1.1.2 Modelling biodiversity and neutral theory ........... 5
1.1.3 The state of global deforestation ................... 6
1.2 Main objectives of the thesis .............................. 7
1.2.1 Chapter 2 ........................................... 8
1.2.2 Chapter 3 ........................................... 8
1.2.3 Chapter 4 ........................................... 8
2 Fast calibration of a dynamic vegetation model ............. 9
2.1 Introduction ............................................... 9
2.2 Material and methods ...................................... 10
2.2.1 Nonlinear regression model ......................... 10
2.2.2 Parameter estimation ............................... 11
2.2.3 General calibration procedure ...................... 12
2.2.4 Modelling the dynamics of tropical forest .......... 13
2.2.5 Data generation .................................... 14
2.3 Results ................................................... 16
2.3.1 Virtual data test case ............................. 16
2.3.2 Real case - single species group ................... 18
2.3.3 Real data case - multiple species group ............ 21
2.4 Discussion ................................................ 22
3 Size-structured neutral theory ............................ 25
3.1 Introduction .............................................. 25
3.2 Material and methods ...................................... 27
3.2.1 Study site ......................................... 27
3.2.2 The size-structured spatially explicit neutral
model .............................................. 27
3.2.3 Linking model and data ............................. 30
3.3 Results ................................................... 34
3.3.1 Prediction of ecological patterns .................. 34
3.3.2 Global parameter screening ......................... 37
3.4 Discussion ................................................ 40
4 The global state of forest fragmentation .................. 43
4.1 Introduction .............................................. 43
4.2 Material and methods ...................................... 44
4.2.1 Database: Hansen's forest cover map ................ 44
4.2.2 The advanced cluster analysis algorithm ............ 44
4.2.3 Compression of labelled forest fragment maps ....... 45
4.2.4 Forest fragment characteristics .................... 45
4.2.5 Structural fragment analysis for different
ecoregions ......................................... 47
4.2.6 Mapping forest conditions in relation to
fragmentation ...................................... 48
4.3 Results ................................................... 49
4.3.1 Global analysis of forest fragment structure for
the year 2000 ...................................... 49
4.3.2 Global analysis of forest fragment condition for
the year 2000 ...................................... 53
4.4 Discussion ................................................ 56
4.5 Conclusion ................................................ 57
5 Discussion and perspectives ............................... 59
5.1 Model calibration and uncertainty assessment .............. 59
5.2 Understanding the mechanisms that shape biodiversity ...... 61
5.3 Analysis of high-resolution forest cover maps ............. 61
5.4 Outlook ................................................... 63
A Appendix for Chapter 2 .................................... 65
A.l Uncertainty assessment .................................... 65
A.2 Identifiability ........................................... 66
A.3 Model selection ........................................... 66
В Appendix for Chapter 3 .................................... 69
B.l Parameter screening ....................................... 69
B.l.l The morris method for global parameter screening ... 69
B.1.2 Individual screening results ....................... 69
B.2 Calibration ............................................... 71
С Appendix for Chapter 4 .................................... 73
C.l Data structures of the cluster analysis algorithm ......... 73
C.2 Compression method used on labelled images ................ 73
C.3 Summary of results for a tree cover threshold of 25% ...... 74
C.4 Edge detection method ..................................... 75
List of Figures ................................................ 79
List of Tables ................................................. 82
Bibliography ................................................... 83
| 
