Chapter 13 Handy Hints

This chapter contains a few handy hints which may be useful if you’re having any issues loading R packages or running Whitebox tools. We covered most of these during our session in Week 9 but these are included here for reference.

If there are any questions regarding Assessment 2 or the practical instructions, please do get in touch at my @manchester email.

Dealing with errors

Errors are an inevitable aspect of coding, irrespective of how experienced you are! As we gain programming experience, these errors increase in difficulty from initially simple problems, such as incorrect file paths or errors in code syntax, to the more complex.

Being able to deal with errors is therefore an essential programming skill. My top tips for fixing errors are:

  1. Read the error message (red text in the R console). If this appears, it usually means that something has gone wrong! Error messages can vary in quality, but at the very least they should tell you:

    • where the error was encountered i.e. which line of code failed.
    • some indication as to why the error occurred e.g. No such file or directory

  2. If the error message doesn’t provide a simple solution to the problem, then the next step should be to work through your code step-by-step, testing each line or code block to ensure it is working correctly (usually using print() statements).

To illustrate this, here is some code which has failed:

# Function to check and install packages
check.packages <- function(pkg){
  new.pkg <- pkg[!(pkg %in% installed.packages()[, "Package"])]
  if (length(new.pkg)) 
    install.packages(new.pkg, dependencies = TRUE)
  sapply(pkg, require, character.only = TRUE)
}

# Checks and installs packages
packages <- c("ggplot2", "here", "raster", "whitebox")
check.packages(packages)

# Fill depressions in a DEM
wbt_fill_depressions(dem = here("data", "practical_2", "dem_10m.tif"),
                     output = here("output", "practical_2", "dem_10m_fill.tif"))

# Loads raster
filled_dem <- raster(here("data", "practical_2", "dem_10m_fill.tif"))

# Plots using ggplot
g <- ggplot() +
  # Adds raster layer using layer_spatial
  layer_spatial(filed_dem, aes(fill = after_stat(band1))) +
  # Sets ggplot theme
  theme_classic() + 
  # Axis and legend labeling
  labs(fill = "Elevation (m)", x = "Easting", y = "Northing" +
  theme(legend.position = "top") +
  # Sets fill symbology
  scale_fill_continuous(type = "viridis"  na.value = NA)

# Shows plot
g

I would first inspect the error message (No such file or directory) but if I didn’t understand it, I would then troubleshoot by:

  1. First running the check.packages(packages) function. Were all the packages installed correctly i.e. TRUE?

  2. Then checking the wbt_fill_depressions() function. Does this produce an error? Does it correctly produce an output raster in the intended location? Does the output look sensible? Plotting the raster might be useful here to check.

  3. Then running the raster() function to load the raster into R. Here the code would fail and I could then start to investigate further:

    • Are there any syntax errors? e.g. use of commas, correct number of brackets, spelling errors?

    • Are the inputs defined correctly?

In the above example, the code fails because the raster package is trying to load dem_10m_fill.tif from data/practical_1 when it was actually saved in output/practical_1.

Running the code again might produce another error, in which case I would repeat the above process, checking the code step-by-step until the error can be located and fixed.

Can you spot any errors in the code below? There are three to find…

# Plots using ggplot
g <- ggplot() +
  # Adds raster layer using layer_spatial
  layer_spatial(filed_dem, aes(fill = after_stat(band1))) +
  # Sets ggplot theme
  theme_classic() + 
  # Axis and legend labeling
  labs(fill = "Elevation (m)", x = "Easting", y = "Northing" +
  theme(legend.position = "top") +
  # Sets fill symbology
  scale_fill_continuous(type = "viridis"  na.value = NA)

TLDR: the best way to fix errors is to be systematic in your approach.

Work your way through the code step-by-step, checking it is producing the intended outputs at each stage - just because the code doesn’t produce an error, it doesn’t mean it is producing the output that you want!

Whitebox functions vs. the raster package

So far, there has been some confusion regarding the role of whitebox functions, the raster package and ggplot2. Each of these has a unique role, as visualised below, while analysis needs to be conducted in the correct order to avoid errors.

Analysis structure. [1] Create output files using whitebox [2] Load outputs into R using raster [3] Plot them using ggplot

To illustrate this, below we have an example of the wbt_d8_pointer function, which creates a D8 pointer file based on an input digital elevation model (dem):

# Creates a D8 pointer file (.tif) at the output location, based on the input dem
wbt_d8_pointer(dem = here("data", "practical_2", "mersey_dem_fill.tif"),
               output = here("output", "practical_2", "mersey_dem_D8_pointer.tif"))

This uses the here package for locating the input file (mersey_dem_fill.tif), which is stored in data/practical_2, and for saving the output, which is stored as mersey_dem_D8_pointer.tif in output/practical_2.

If this runs successfully (make sure to carefully check file paths, file names, use of commas), then an output raster will be created at the specified location.

Crucially, this does not load the raster into the R environment, so any attempt to plot the raster using ggplot2 would fail.

To achieve this, a raster must be first loaded into the R environment as follows:

# Loads D8 pointer raster using the raster and here packages
mersey_pointer <- raster(here("output", "practical_2", "mersey_dem_D8_pointer.tif"))

Again, we have determined the input file location using here package (output/practical_2/mersey_dem_D8_pointer.tif) and stored the raster as an object in the R environment called mersey_pointer.

With this stored, it can then be plotted using ggplot, setting layer_spatial to the mersey_pointer object name:

# Plots using ggplot
p <- ggplot() +
  layer_spatial(mersey_pointer, aes(fill = stat(band1))) +
  theme_classic() + 
  labs(fill = "Pointer value", x = "Easting", y = "Northing") +
  scale_fill_continuous(type = "viridis",  na.value = NA)
p

TLDR: for subsequent analysis, make sure to adhere to the following structure:

  1. Create output files using whitebox functions.

  2. Load those output files into R using the raster package (stored as objects)

  3. Plot the R objects using ggplot2

Any deviation from this order (e.g. plotting without loading using raster, loading files that have not yet been created using whitebox) will cause errors!

R Projects

Make sure to initialise an R project for your work (Section 5.4) and ensure this is created for your GEOG70581 directory. This should resemble the following, with the R project file and a separate R script for the Eskdale and Mersey Basin practicals:

File paths

Remove spaces in directories or file paths. As a rule, avoid using spaces when naming files or folders as this can cause issues, particularly when coding:

  • “P:/Environmental Monitoring Concepts/GEOG70581/…” Bad
  • “P:/Environmental_Monitoring_Concepts/GEOG70581/…” Good

Output files

Make sure to inspect output files. When you run a command (e.g. a Whitebox function), check that it produces the intended output (e.g. a raster file (.tif) in the output directory). If there is an error or the output hasn’t been created, subsequent code will fail.

Packages

If you’re having difficulty loading the raster or sf packages on a University managed computer, this is due to multiple package installs. Go to the packages window, identify the older version of each package, and tick the check box to load it into your R library.

Code structure

While you will not be assessed on the quality of your code (you do not have to submit your code for the assessment), remember that well-formatted code (with comments and good structure) is easier to read and understand and will be less prone to error.

Inspect the code below:

# Function to check and install packages
check.packages <- function(pkg){
  new.pkg <- pkg[!(pkg %in% installed.packages()[, "Package"])]
  if (length(new.pkg))
    install.packages(new.pkg, dependencies = TRUE)
  sapply(pkg, require, character.only = TRUE)
}

# Checks and installs packages
packages <- c("ggplot2", "sf", "here", "raster", "whitebox", "ggspatial", "patchwork")
check.packages(packages)

# Sets file path for DEM
dem <- here("data", "practical_1", "dem_10m.tif")

# Breach and fills depressions
wbt_fill_depressions(dem, here("output", "practical_1", "dem_10m_fill.tif"))
wbt_breach_depressions(dem, here("output", "practical_1", "dem_10m_breach.tif"))

# Calculates D8 pointer
wbt_d8_pointer(here("output", "practical_1", "dem_10m_fill.tif"),
               here("output", "practical_1", "dem_10m_D8_pointer.tif"))

# Calculates D8 accumulation file (SCA), with log-transformed values
wbt_d8_flow_accumulation(here("output", "practical_1", "dem_10m_fill.tif"),
                         here("output", "practical_1", "dem_10m_flow_accumulation.tif"),
                         out_type = "specific contributing area",
                         log = "TRUE")

This code includes comments for each main code block, line spaces to distinguish different parts of the code, and is written in a logical order (e.g. first loading packages, then loading/selecting data, running tools). Any non-essential code has been removed e.g. additional calls to install.packages() or library().

If there are any other questions, please do get in touch.

Alcalde, J. et al. (2017) ‘Framing bias: The effect of figure presentation on seismic interpretation’, Interpretation, 5(4), pp. T591–T605. doi:10.1190/INT-2017-0083.1.

Amrhein, V. et al. (2019) ‘Scientists rise up against statistical significance’, Nature, 567(7748), pp. 305–307. doi:10.1038/d41586-019-00857-9.

Andrade, C. (2019) ‘The P Value and Statistical Significance: Misunderstandings, Explanations, Challenges, and Alternatives’, Indian Journal of Psychological Medicine, 41(3), pp. 210–215. doi:10.4103/IJPSYM.IJPSYM_193_19.

Crameri, F. (2018) ‘Geodynamic diagnostics, scientific visualisation and StagLab 3.0’, Geoscientific Model Development, 11(6), pp. 2541–2562. doi:https://doi.org/10.5194/gmd-11-2541-2018.

Fairfield, J. and Leymarie, P. (1991) ‘Drainage networks from grid digital elevation models’, Water Resources Research, 27(5), pp. 709–717. doi:https://doi.org/10.1029/90WR02658.

Freeman, T.G. (1991) ‘Calculating catchment area with divergent flow based on a regular grid’, Computers & Geosciences, 17(3), pp. 413–422. doi:10.1016/0098-3004(91)90048-I.

Geissbuehler, M. and Lasser, T. (2013) ‘How to display data by color schemes compatible with red-green color perception deficiencies’, Optics Express, 21(8), pp. 9862–9874. doi:10.1364/OE.21.009862.

Goodman, S. (2008) ‘A Dirty Dozen: Twelve P-Value Misconceptions’, Seminars in Hematology, 45(3), pp. 135–140. doi:10.1053/j.seminhematol.2008.04.003.

Grabs, T.J. et al. (2010) ‘Calculating terrain indices along streams: A new method for separating stream sides’, Water Resources Research, 46(12). doi:https://doi.org/10.1029/2010WR009296.

Hawkins, E. (2015) ‘Scrap rainbow colour scales’, Nature, 519(7543), pp. 291–291. doi:10.1038/519291d.

Jenson, S.K. (1991) ‘Applications of hydrologic information automatically extracted from digital elevation models’, Hydrological Processes, 5(1), pp. 31–44. doi:https://doi.org/10.1002/hyp.3360050104.

Jenson, S.K. and Domingue, J.O. (1988) ‘Extracting topographic structure from digital elevation data for geographic information-system analysis’, Photogrammetric Engineering and Remote Sensing, 54(11), p. 15931600. Available at: http://pubs.er.usgs.gov/publication/70142175 (Accessed: 2 December 2020).

Levine, T. (2009) ‘Using colour in figures: Let’s agree to differ’, Traffic, 10(3), pp. 344–347. doi:https://doi.org/10.1111/j.1600-0854.2008.00863.x.

Lindsay, J.B. et al. (2008) ‘Mapping outlet points used for watershed delineation onto DEM-derived stream networks’, Water Resources Research, 44(8). doi:https://doi.org/10.1029/2007WR006507.

O’Callaghan, J.F. and Mark, D.M. (1984) ‘The extraction of drainage networks from digital elevation data’, Computer Vision, Graphics, and Image Processing, 28(3), pp. 323–344. doi:10.1016/S0734-189X(84)80011-0.

Quinn, P. et al. (1991) ‘The prediction of hillslope flow paths for distributed hydrological modelling using digital terrain models’, Hydrological Processes, 5(1), pp. 59–79. doi:https://doi.org/10.1002/hyp.3360050106.

Quinn, P.F. et al. (1995) ‘The in(a/tan/β) index: How to calculate it and how to use it within the topmodel framework’, Hydrological Processes, 9(2), pp. 161–182. doi:https://doi.org/10.1002/hyp.3360090204.

Seibert, J. and McGlynn, B.L. (2007) ‘A new triangular multiple flow direction algorithm for computing upslope areas from gridded digital elevation models’, Water Resources Research, 43(4). doi:https://doi.org/10.1029/2006WR005128.

Tarboton, D.G. (1997) ‘A new method for the determination of flow directions and upslope areas in grid digital elevation models’, Water Resources Research, 33(2), pp. 309–319. doi:https://doi.org/10.1029/96WR03137.