Introduction to Environmental Data Science

Description

Introduction to Environmental Data Science focuses on data science methods in the R language applied to environmental research, with sections on exploratory data analysis in R including data abstraction, transformation, and visualization; spatial data analysis in vector and raster models; statistics and modelling ranging from exploratory to modelling, considering confirmatory statistics and extending to machine learning models; time series analysis, focusing especially on carbon and micrometeorological flux; and communication. Introduction to Environmental Data Science is an ideal textbook to teach undergraduate to graduate level students in environmental science, environmental studies, geography, earth science, and biology, but can also serve as a reference for environmental professionals working in consulting, NGOs, and government agencies at the local, state, federal, and international levels. Features • Gives thorough consideration of the needs for environmental research in both spatial and temporal domains. • Features examples of applications involving field-collected data ranging from individual observations to data logging. • Includes examples also of applications involving government and NGO sources, ranging from satellite imagery to environmental data collected by regulators such as EPA. • Contains class-tested exercises in all chapters other than case studies. Solutions manual available for instructors. • All examples and exercises make use of a GitHub package for functions and especially data. Cover Half Title Title Page Copyright Page Contents Author/editor biographies List of Figures 1. Background, Goals and Data 1.1. Environmental Data Science 1.2. Environmental Data and Methods 1.3. Goals 1.3.1. Some definitions: 1.4. Exploratory Data Analysis 1.5. Software and Data 1.5.1. Data 1.6. Acknowledgements I. Exploratory Data Analysis 2. Introduction to R 2.1. Data Objects 2.1.1. Scalars and assignment 2.2. Functions 2.3. Expressions and Statements 2.4. Data Classes 2.4.1. Integers 2.5. Rectangular Data 2.6. Data Structures in R 2.6.1. Vectors 2.6.2. Lists 2.6.3. Matrices 2.6.4. Data frames 2.6.5. Factors 2.7. Accessors and Subsetting 2.7.1. [] Subsetting 2.7.2. [[]] The mysterious double bracket 2.7.3. $ Accessing a vector from a data frame 2.8. Programming scripts in RStudio 2.8.1. function : creating your own 2.8.2. if : conditional operations 2.8.3. for loops 2.8.4. Subsetting with logic 2.8.5. Apply functions 2.9. RStudio projects 2.9.1. R Markdown 2.10. Exercises: Introduction to R 3. Data Abstraction 3.1. The Tidyverse 3.2. Tibbles 3.2.1. Building a tibble from vectors 3.2.2. tribble 3.2.3. read_csv 3.3. Summarizing variable distributions 3.3.1. Stratifying variables by site using a Tukey box plot 3.4. Database operations with dplyr 3.4.1. Select, mutate, and the pipe 3.4.2. filter 3.4.3. Writing a data frame to a csv 3.4.4. Summarize by group 3.4.5. Count 3.4.6. Sorting after summarizing 3.4.7. The dot operator 3.5. String abstraction 3.5.1. Detecting matches 3.5.2. Subsetting strings 3.5.3. String length 3.5.4. Replacing substrings with other text (“mutating” strings) 3.5.5. Concatenating and splitting 3.6. Dates and times with lubridate 3.7. Calling functions explicitly with :: 3.8. Exercises: Data Abstraction 4. Visualization 4.1. plot in base R 4.2. ggplot2 4.3. Plotting one variable 4.3.1. Histogram 4.3.2. Density plot 4.3.3. Boxplot 4.4. Plotting Two Variables 4.4.1. Two continuous variables 4.4.2. Two variables, one discrete 4.4.3. Color systems 4.4.4. Trend line 4.5. General Symbology 4.5.1. Categorical symbology 4.5.2. Log scales instead of transform 4.6. Graphs from Grouped Data 4.6.1. Faceted graphs 4.7. Titles and Subtitles 4.8. Pairs Plot 4.9. Exercises: Visualization 5. Data Transformation 5.1. Data joins 5.2. Set operations 5.3. Binding rows and columns 5.4. Pivoting data frames 5.4.1. pivot_longer 5.4.2. pivot_wider 5.4.3. A free_y faceted graph using a pivot 5.5. Exercise: Transformation II. Spatial 6. Spatial Data and Maps 6.1. Spatial Data 6.1.1. Simple geometry building in sf 6.1.2. Building points from a data frame 6.1.3. SpatVectors in terra 6.1.4. Creating features from shapefiles 6.2. Coordinate Referencing Systems 6.3. Creating sf Data from Data Frames 6.3.1. Removing geometry 6.4. Base R’s plot() with terra 6.4.1. Using maptiles to create a basemap 6.5. Raster data 6.5.1. Building rasters 6.5.2. Vector to raster conversion 6.6. ggplot2 for Maps 6.6.1. Rasters in ggplot2 6.7. tmap 6.8. Interactive Maps 6.8.1. Leaflet 6.8.2. Mapview 6.8.3. tmap (view mode) 6.8.4. Interactive mapping of individual penguins abstracted from a big dataset 6.9. Exercises: Spatial Data and Maps 6.9.1. Project preparation 7. Spatial Analysis 7.1. Data Frame Operations 7.1.1. Using grouped summaries, and filtering by a selection 7.2. Spatial Analysis Operations 7.2.1. Using topology to subset 7.2.2. Centroid 7.2.3. Distance 7.2.4. Buffers 7.2.5. Spatial overlay: union and intersection 7.2.6. Clip with st_crop 7.2.7. Spatial join with st_join 7.2.8. Further exploration of spatial analysis 7.3. Exercises: Spatial Analysis 8. Raster Spatial Analysis 8.1. Terrain functions 8.2. Map Algebra in terra 8.3. Distance 8.4. Extracting Values 8.5. Focal Statistics 8.6. Zonal Statistics 8.7. Exercises: Raster Spatial Analysis 9. Spatial Interpolation 9.1. Null Model of the Original Data 9.2. Voronoi Polygon 9.2.1. Cross-validation and relative performance 9.3. Nearest Neighbor Interpolation 9.3.1. Cross-validation and relative performance of the nearest neighbor model 9.4. Inverse Distance Weighted (IDW) 9.4.1. Using cross-validation and relative performance to guide inverse-distance weight choice 9.4.2. IDW: trying other inverse distance powers 9.5. Polynomials and Trend Surfaces 9.6. Kriging 9.6.1. Create a variogram. 9.6.2. Fit the variogram based on visual interpretation 9.6.3. Ordinary Kriging 9.7. Exercises: Spatial Interpolation III. Statistics and Modeling 10. Statistical Summaries and Tests 10.1. Goals of Statistical Analysis 10.2. Summary Statistics 10.2.1. Summarize by group: stratifying a summary 10.2.2. Boxplot for visualizing distributions by group 10.2.3. Generating pseudorandom numbers 10.3. Correlation r and Coefficient of Determination r2 10.3.1. Displaying correlation in a pairs plot 10.4. Statistical Tests 10.4.1. Comparing samples and groupings with a t test and a non-parametric Kruskal-Wallis Rank Sum test 10.4.2. Analysis of variance 10.4.3. Testing a correlation 10.5. Exercises: Statistics 11. Modeling 11.1. Some Common Statistical Models 11.2. Linear Model (lm) 11.3. Spatial Influences on Statistical Analysis 11.3.1. Mapping residuals 11.4. Analysis of Covariance 11.5. Generalized linear model (GLM) 11.5.1. Binomial family: logistic GLM with streams 11.5.2. Logistic landslide model 11.5.3. Poisson regression 11.5.4. Models employing machine learning 11.6. Exercises: Modeling 12. Imagery and Classification Models 12.1. Reading and Displaying Sentinel-2 Imagery 12.1.1. Individual bands 12.1.2. Spectral subsets to create three-band R-G-B and NIR-R-G for visualization 12.1.3. Crop to study area extent 12.1.4. Saving results 12.1.5. Band scatter plots 12.2. Spectral Profiles 12.3. Map Algebra and Vegetation Indices 12.3.1. Vegetation indices 12.3.2. Histogram 12.3.3. Other vegetation indices 12.4. Unsupervised Classification with k-means 12.5. Machine Learning Classification of Imagery 12.5.1. Read imagery and training data and extract sample values for training 12.5.2. Training the CART model 12.5.3. Prediction using the CART model 12.5.4. Validating the model 12.6. Classifying with 10 m Sentinel-2 Imagery 12.6.1. Subset bands (10 m) 12.6.2. Crop to RCV extent and extract pixel values 12.6.3. Training the CART model (10 m) and plot the tree 12.6.4. Prediction using the CART model (10 m) 12.7. Classification Using Multiple Images Capturing Phenology 12.7.1. Create a 10-band stack from both images 12.7.2. Extract the training data (10 m spring + summer) 12.7.3. CART model and prediction (10 m spring + summer) 12.8. Conclusions and Next Steps for Imagery Classification 12.9. Exercises: Imagery Analysis and Classification Models IV. Time Series 13. Time Series Visualization and Analysis 13.1. Structure, Seasonality, and Decomposition of Time Series 13.2. Creation of Time Series (ts) Data 13.2.1. Frequency, start, and end parameters for ts() 13.2.2. Associating times with time series 13.2.3. Subsetting time series by times 13.2.4. Changing the frequency to use a different period 13.2.5. Time stamps and extensible time series 13.3. Data smoothing: moving average (ma) 13.4. Decomposition of data logger data: Marble Mountains 13.5. Facet Graphs for Comparing Variables over Time 13.6. Lag Regression 13.6.1. The lag regression, using a lag function in a linear model 13.7. Ensemble Summary Statistics 13.8. Learning more about Time Series in R 13.9. Exercises: Time Series V. Communication and References 14. Communication with Shiny 14.1. Shiny Document 14.1.1. Input and output objects in the Old Faithful Eruptions document 14.1.2. Input widgets 14.1.3. Other input widgets 14.2. A Shiny App 14.2.1. A brief note on reactivity 14.3. Shiny App I/O Methods 14.3.1. Data tables 14.3.2. Text as character: renderPrint() and verbatimTextOutput() 14.3.3. Formatted text 14.3.4. Plots 14.4. Shiny App in a Package 14.5. Components of a Shiny App (sierra) 14.5.1. Initial data setup 14.5.2. The ui section, with a tabsetPanel structure 14.5.3. The server section, including reactive elements 14.5.4. Calling shinyApp with the ui and server function results 14.6. A MODIS Fire App with Web Scraping and observe with leafletProxy 14.6.1. Setup code 14.6.2. ui 14.6.3. Using observe and leafletProxy to allow changing the date while retaining the map zoom 14.7. Learn More about Shiny Apps 14.8. Exercises: Shiny References Index