The book aims to help students, researchers and interested practitioners monitor and model changes in land use and land cover, such as deforestation, cropland expansion, ecosystem degradation and urban sprawl, using the popular, free-and-open-source R environment. It bridges the wide pedagogical gap between basic GIS and R texts and advanced R analyses published in scientific journals.
Using real-world worked examples, the book demonstrates the complete operational chain for land use change analysis and modelling of land use futures. Clearly structured chapters take the reader from data acquisition through change map comparison, estimation of accuracy, analysis and quantification of change sequences, to statistical modelling of drivers of change, model development and estimation of model goodness-of-fit.
Key Features:
¿ Includes exercises in each chapter that make use of a wide range of supplementary materials and third-party datasets available online.
¿ All the code necessary to follow the exercises is provided in the book. Datasets are available on Zenodo at: https://zenodo.org/communities/hewitt_mas_mmluccr. The SIMLANDER and APoLUS models, which are used extensively, are also freely available online at https://simlander.wordpress.com/.
Inhaltsverzeichnis
Part I Introduction to Monitoring and Modelling of LUCC in R 1 Why R? 1.1 Introduction: The shifting landscape of geographical data analysis 1.2 The limitations of the traditional GIS paradigm analysis 1.4 Bridging GIS and specialized modelling 1.5 Conclusion 2 Preliminary steps (installing R and packages) 2.1 Installation of R and RStudio 2.1.1 Windows 2.1.2 Linux 2.1.3 Mac 2.1.4 System requirements for spatial packages 2.2 Introduction to R and RStudio 2.3 Installing R packages 2.3.1 From the official CRAN repository 2.3.2 From R-Forge 2.3.3 From GitHub and GitLab 2.3.4 From Bioconductor 2.4 Managing packages 2.5 Data availability notice Part II Visualization and analysis of LUCC 3 Spatial data in R 3.1 Introduction 3.2 Point coverages 3.3 Line coverages 3.4 Polygon coverages 3.5 Raster data 3.5.1 Raster data organisation in the sp package 3.5.2 SpatialPixels and SpatialPixelsDataFrame 3.5.3 SpatialGrid and SpatialGridDataFrame 3.5.4 The raster package and class RasterLayer 3.5.5 The terra package and class SpatRaster 3.6 Chapter summary and next steps 3.7 Answers to exercises 3.7.1 Exercise 1 3.7.2 Exercise 2 3.7.3 Exercise 3 3.7.4 Exercise 4 3.7.5 Exercise 5 4 Accuracy assessment of change maps 4.1 Introduction 4.2 Materials 4.3 Creation of colour image composites 4.4 Preparation of the LUCC map 4.5 Calculation of change area statistics 4.6 Calculating sample size and number of samples per category 4.7 Carrying out the georeferenced sampling 4.8 Photointerpreting the sample units 4.9 Error matrix preparation and accuracy indices computation 4.10 A brief interpretation of the results 4.11 Chapter summary and next steps 4.12 Answers to exercises 4.12.1 Exercise 1 4.12.2 Exercise 2 4.12.3 Exercise 3 5 Visualisation and analysis of change dynamics 5.1 Areas and rate of change 5.1.1 Area computing 5.1.2 Rates of change 5.2 Change matrix and change map 5.2.1 Raster data 5.2.2 Vector data 5.3 LUCC budget 5.4 Barplot and Sankey diagram 5.5 Change intensity 5.5.1 Interval level 5.5.2 Category level 5.5.3 Transition level 5.6 Markov matrix 5.7 Detection of hot spots with the Getis-Ord method 5.7.1 Getis-Ord Gi* index 5.7.2 Detection of areas intensively converted to orchards 5.8 Summary and next steps 5.9 Answers to exercises 5.9.1 Exercise 1 5.9.2 Exercise 2 5.9.2.1 Rate of change 5.9.2.2 LUCC budget 5.9.2.3 Intensity analysis 6 Analysis of land use/cover sequences 6.1 Introduction 6.2 Preparation of the datasets 6.3 Sequence analysis 6.3.1 Creation of state sequence objects 6.3.2 Sequence plots 6.3.3 Sequence frequency plots 6.3.4 State distribution plots (chronograms) 6.3.5 Transversal and longitudinal entropy index 6.3.6 Transition rates 6.3.7 Subsequence analysis 6.3.8 Identifying discriminating subsequences 6.4 Summary and next steps 6.4.1 Response to Exercise 1 Part III LUCC Modelling 7 Preparing the data for modelling 7.1 The explicitly spatial nature of land use data 7.2 Download data for this chapter 7.3 Working directories, file paths and directory management 7.4 Pattern searching within the R environment 7.5 Importing and visualising vector data with sf 7.6 Importing and visualising raster data with terra 7.7 Querying and modifying raster data 7.8 Plotting simple feature vector objects 7.9 Cropping a raster map to a polygon boundary 7.10 Writing out raster files with writeRaster 7.11 Reclassifying raster data 7.12 Simple scripting 7.13 Working with colours 7.14 Chapter summary and next steps 7.15 Answers to exercises 7.15.1 Exercise 1 7.15.2 Exercise 2 7.15.3 Exercise 3 7.15.4 Exercise 4 8 Linking land use drivers and patterns of change 8.1 Understanding land use change - visual inspection of land use maps. 8.2 Quantifying land use change across dates using cross-tabulation analysis. 8.3 Calculating and visualising land use change for urban land 8.4 Computing distance to urban areas 8.5 The influence of transport networks 8.6 Importing and rasterizing road networks 8.7 Creating a Euclidean distance map 8.8 The influence of physical landscape characteristics 8.9 Creating a slope map with the terrain() command 8.10 Testing the explanatory power of the variables 8.10.1 What test is suitable for land use data and explanatory drivers? 8.10.2 So what test is suitable for land use data and explanatory drivers then? 8.11 Building a simple linear regression model with our land use drivers 8.11.1 Preparing the data 8.11.2 Building a simple linear model 8.11.3 Examining residuals to check against model assumptions 8.12 Spatial autocorrelation and geographically weighted regression 8.12.1 Do the relationships detected by the linear model vary across the study area? 8.12.2 Setting up the GWR analysis 8.12.3 Running the GWR analysis using gwr.basic() 8.12.4 Reading the GWR output 8.12.5 Visualising the results of the GWR analysis 8.12.6 Interpreting the results of the GWR analysis 8.13 Chapter summary and next steps 8.14 Answers to exercises 8.14.1 Exercise 1 8.14.2 Exercise 2 8.14.3 Exercise 3 8.14.4 Exercise 4 9 From association to simulation - a simple cellular automata land use model 9.1 Building the model part 1: land use drivers. 9.1.1 What have we discovered about land use change in our study area? 9.1.2 From distance to urban areas to the neighbourhood effect 9.1.3 From distance to roads to accessibility 9.1.4 From terrain slope to suitability for urban development 9.1.5 Is urban development always completely predictable? Introducing the stochastic factor 9.2 Building the model part 2: transition potential 9.2.1 The transition potential computation 9.2.2 Zoning: masking out areas we do not want to change 9.3 Building the model part 3: How much land to allocate? 9.3.1 Land use demand 9.4 Putting it all together 9.5 Running a simulation to 2050 9.6 Chapter summary and next steps 9.7 Answers to exercises. 9.7.1 Exercise 1 9.7.2 Exercise 2 9.7.3 Exercise 3 9.7.4 Exercise 4 9.7.5 Exercise 5 9.7.6 Exercise 6 9.7.7 Exercise 7 9.7.8 Exercise 8 9.7.9 Exercise 9 9.7.10 Exercise 10 10 Is my model a good model? Assessing model quality using multiple met-rics 10.1 Introduction to the chapter 10.1.1 Visual inspection 10.1.2 Making visual inspection easier 10.1.3 Simple quantification of agreement/disagreement between maps 10.1.4 Creating a simple bar plot to visualise comparison statistics 10.1.5 Evaluating trade-offs 1: hits versus misses (true positive rate). 10.1.6 Evaluating trade-offs 2: false alarms versus correct rejects (false pos- itive rate) 10.1.7 Evaluating trade-offs 3: plotting results in receiver operating charac-teristic (ROC) space 10.1.8 Summary: cell-by-cell measures 10.1.9 Pattern measures 10.1.10 Chapter summary and next steps 10.2 Answers to exercises 10.2.1 Exercise 1 10.2.2 Exercise 2 10.2.3 Exercise 3 10.2.4 Exercise 4 10.2.5 Exercise 5 10.2.6 Exercise 6 10.2.7 Exercise 7 10.2.8 Exercise 8 10.2.9 Exercise 9 10.2.10 Exercise 10 11 Modelling competition between multiple land uses. Introduction to APo-LUS 11.1 Introduction to the chapter 11.2 History of the APoLUS model, documentation and scientific publication. 11.3 Downloading APoLUS and preparing the workspace 11.4 Plotting the underlying land use maps 11.5 Getting basic information on the model 11.6 Running a simulation 11.7 Model calibration 11.7.1 Modifying the neighbourhood rules 11.7.1.1 Solar energy attractiveness to solar energy 11.7.2 Modifying the accessibility effect 11.7.3 Modifying suitability 11.7.4 Zoning 11.8 Chapter summary and next steps 11.9 Answers to exercises 11.9.1 Exercise 1 11.9.2 Exercise 2 11.9.3 Exercise 3 11.9.4 Exercise 4 11.9.5 Exercise 5 11.9.6 Exercise 6 11.9.7 Exercise 7 11.9.8 Exercise 8 11.9.9 Exercise 9 11.9.10 Exercise 10